1

Empathy in healthcare settings

Helen Scott

Supervisor: Professor Jo Silvester

Thesis submitted for PhD

Department of Psychology

Goldsmiths, University of London

March 2011
Declaration:

I declare that the work presented in this thesis is my own

Signed................................................

Helen Scott

25/03/2011
 1

Table of contents

List of tables ............................................................................................................ 7

List of figures .......................................................................................................... 8

Acknowledgements.................................................................................................. 9

Abstract ................................................................................................................. 10

Chapter 1 - Introduction ......................................................................................... 12

1.1. Patient satisfaction ................................................................................... 13

1.2. Patient adherence to treatment ................................................................. 14

1.3. Improved health outcomes ....................................................................... 15

1.4. Professional satisfaction........................................................................... 16

1.5. Sustaining and developing empathy ......................................................... 17

Chapter 2 - Literature Review ................................................................................ 20

2.1 Empathy: An Overview ........................................................................... 20

2.2 Defining empathy .................................................................................... 21

2.2.1 An emotional dimension of empathy................................................. 21

2.2.2 A cognitive dimension of empathy.................................................... 24

2.2.3 A moral dimension of empathy ......................................................... 26

2.2.4 A behavioural dimension of empathy ................................................ 28

2.3 Multidimensional models of empathy in healthcare.................................. 29

2.3.1 The multidimensional model of medical empathy (Squier, 1990) ...... 30

 2.3.2 The Multidimensional Model of Empathy (Davis, 1983)................... 32

2.3.3 Process model of clinical empathy (Larson & Yao, 2005) ................. 38

2.4 Models of empathy: limitations and next steps ......................................... 40

Chapter 3. Measures of empathy ............................................................................ 42

3.1. Reliability and validity............................................................................. 42

3.2. An overview of empathy scales................................................................ 44

3.2.1. Empathy Construct Rating Scale (ECRS).......................................... 45

3.2.2. The Barrett-Lennard Relation Inventory (BLRI) ............................... 47

3.2.3. Jefferson Scale of Physician Empathy (JSPE) ................................... 49

3.2.4. The Reynolds Empathy Scale ........................................................... 51

3.2.5. The CARE Measure.......................................................................... 53

3.2.6. Hogan’s Empathy Scale .................................................................... 55

3.2.7. The Questionnaire Measure of Emotional Empathy .......................... 56

3.2.8. The Interpersonal Reactivity Index ................................................... 57

3.3. Summary of empathy measures ............................................................... 59

3.4. Overview of Thesis and Research Questions ............................................ 64

Chapter 4 – Individual differences and empathy: Personality and the IRI ............... 68

4.1. The Five Factor Model (FFM) of Personality ........................................... 68

4.2. Construct validity .................................................................................... 73

4.3. Method .................................................................................................... 74

4.3.1 Participants and Procedure ................................................................ 74

 4.3.2 Measures .......................................................................................... 74

4.3.3 Analysis............................................................................................ 77

4.4. Results ..................................................................................................... 77

4.4.1 Correlations ...................................................................................... 79

4.4.2 Joint factor analysis .......................................................................... 80

4.5. Discussion ............................................................................................... 86

4.5.1 Perspective Taking and Empathic Concern ....................................... 88

4.5.2 Personal Distress .............................................................................. 90

4.5.3 Fantasy ............................................................................................. 90

4.5.4 Limitations and next steps................................................................. 92

Chapter 5 – Emotional intelligence and empathy.................................................... 93

5.1 Emotional Intelligence ............................................................................. 94

5.2 Empathy and Emotional Intelligence........................................................ 96

5.3 Method .................................................................................................. 101

5.3.1 Participants ..................................................................................... 101

5.3.2 Procedure and Measures ................................................................. 102

5.4 Results ................................................................................................... 108

5.4.1 Correlations .................................................................................... 109

5.4.2 Joint factor analysis ........................................................................ 111

5.5 Discussion ............................................................................................. 113

5.5.1 Perspective Taking and Empathic Concern ..................................... 113

 5.5.2 Personal Distress ............................................................................ 114

5.5.3 Fantasy ........................................................................................... 115

5.5.4 Limitations ..................................................................................... 116

5.5.5 Summary and next steps ................................................................. 116

Chapter 6 – Self-report empathy and other-rated empathic behaviour ................... 118

6.1 Empathy and behaviour ......................................................................... 118

6.2 Criterion-related validation .................................................................... 121

6.3 Method .................................................................................................. 125

6.3.1 Context and participants ................................................................. 125

6.3.2 Measures ........................................................................................ 127

6.4 Results ................................................................................................... 132

6.4.1 Correlations .................................................................................... 134

6.4.2 Cultural differences: comparing groups .......................................... 137

6.4.3 Correlations by group ..................................................................... 141

6.5 Discussion ............................................................................................. 146

6.5.1 Empathy and behaviour .................................................................. 148

6.5.2 Empathy and culture ....................................................................... 149

6.5.3 Limitations ..................................................................................... 151

6.5.4 Summary and next steps ................................................................. 153

Chapter 7 - Empathy from the patient perspective ................................................ 154

7.1 The patient perspective: judgements of empathy .................................... 155

 7.2 Empathy and behaviour ......................................................................... 157

7.3 Antecedents: situational characteristics .................................................. 158

7.4 Quantitative and qualitative methods ..................................................... 158

7.5 Method .................................................................................................. 161

7.5.1 Participants ..................................................................................... 161

7.5.2 Procedure ....................................................................................... 161

7.5.3 Analysis.......................................................................................... 162

7.6 Findings and Discussion ........................................................................ 169

7.6.1 Interpersonal Processes ................................................................... 170

7.6.2 Antecedents of empathy.................................................................. 175

7.6.3 Other areas of the template ............................................................. 179

7.6.4 Reflections and limitations.............................................................. 182

7.6.5 Summary ........................................................................................ 183

Chapter 8: Concluding discussion ........................................................................ 184

8.1 Summary of the findings ........................................................................ 184

8.2 Implications for theory........................................................................... 187

8.3 Implications for practice ........................................................................ 189

8.4 Future research ...................................................................................... 191

8.5 Closing points ........................................................................................ 191

References ........................................................................................................... 193

Appendix 1: Development of Research ................................................................ 219

Appendix 2: The Interpersonal Reactivity Index .................................................. 221

Appendix 3: Information Sheet (doctors) ............................................................. 223

Appendix 4: Consent Form (doctors) ................................................................... 224

Appendix 5: Factor loadings for the Bar On EQ-i items ....................................... 225

Appendix 6: Information sheet (patients) ............................................................. 228

Appendix 7: Consent form (patients).................................................................... 230

Appendix 8: Study four interview schedule .......................................................... 231

Appendix 9: Coded Interview .............................................................................. 232

List of tables

Table 3.1: Descriptions of existing measures of empathy................................ 61

Table 3.2: Reliability and validity information for existing empathy measures. 63

Table 4.1: Description of NEO PI-R five domains and facets............................ 77

Table 4.2: Descriptive statistics for the NEO PI-R and Empathy (IRI) scales.... 79

Table 4.3: Correlations between NEO PI-R factors and Empathy (IRI) scales... 81

Table 4.4: Factor loadings for joint factor analysis of IRI and NEO PI-R.......... 85

Table 5.1: Description of EQ-i five factors and fifteen subscales....................... 98

Table 5.2: Factor names, descriptions, means, S.D.s and reliabilities for eight

subscales resulting from EQ-i factor analysis..................................................... 109

Table 5.3: Comparison of the medical and non-medical samples 111

Table 5.4: Correlations between EQ-i and Empathy (IRI) scales....................... 112

Table 5.5: Factor loadings for IRI and EQ-i subscales........................................ 114

Table 6.1: Definitions and sample behavioural indicators of competencies

related to empathy (from Patterson et al 2000)................................................... 133

Table 6.2: Descriptives all variables (n=192)...................................................... 135

Table 6.3: Bivariate correlations for all variables (n = 192)................................ 137

Table 6.4: Descriptive statistics for UK/Ireland and Overseas trained doctors... 142

Table 6.5: Partial Correlations controlling for age and intelligence between

IRI, EQ-I and assessor-rated competency scores................................................ 145

Table 6.6: Partial correlations controlling for age and intelligence between

IRI, EQ-i and ‘patient’-rated competency scores................................................ 147

 8

List of figures

Figure 2.1: The ‘cycle’ of empathy (Barrett-Lennard, 1993)......................... 27

Figure 2.2: Model of empathic understanding and adherence to treatment

regimens (adapted from Squier, 1990)............................................................ 31

Figure 2.3: A multidimensional model of empathy (from Davis, 1996)........ 34

Figure 2.4: Process model of clinical empathy (from Larson & Yao, 2005).. 39

Figure 3.1: Framework for investigating empathy addressed by the studies

within this thesis ............................................................................................ 67

Figure 6.1: Process model of clinical empathy (from Larson & Yao, 2005).. 119

Figure 7.1: Process model of clinical empathy (from Larson & Yao, 2005).. 156

Figure 7.2: Initial Template for analysing patient interviews......................... 164

Figure 7.3: Final template from analysis of patient interviews...................... 168

Figure 8.1: A process model of empathy in the healthcare setting from this

thesis............................................................................................................... 188
 9

Acknowledgements

There are so many people to whom I am grateful for their support in the

completion of this PhD. First and foremost, to my supervisor Professor Jo

Silvester. Jo, throughout this journey you have provided me with wisdom and

inspiration, as well as so much time and practical support and, in the end, the

much needed space to get it finished. Thank you so much for your never-ending

patience and encouragement at every stage.

Thanks also to my colleagues who have given me support along the way. Anna,

Lara, Paul, Maddy and Sharon have all given me advice, motivation and great

friendship. I would like to thank Professor Fiona Patterson for giving me an

opportunity to collect data in the GP assessment centre for my second and third

studies. I would also like to acknowledge the participants who were involved in

this research; in particular, those from the James Cook University Hospital. It

was a real privilege to be able to interview patients at a time of illness.

As anyone who completes a PhD knows, this process can take you away from

your family both physically and mentally. Andrew and Beatrice, I apologise for

my absence and thank you for your love and patience with me. I am grateful to

my sister Louise for her absolute certainty that I would get there as well as a

demon eye for detail in proof reading. To my parents, whose unfailing pride in

me is a constant source of motivation, I thank you both. Dad, at times you were

the only one who kept bringing me back on track, which I have appreciated so

much. This thesis is dedicated to all of my family.

 10

Abstract

Empathy is an important concept associated with positive outcomes for

healthcare practitioners and their patients. In order to identify the best methods to

develop and sustain empathy in healthcare professionals there is a need for

greater understanding of the antecedents and behaviours involved in empathic

responding towards patients.

This thesis used a multidimensional model of empathy as a guide for research

aimed at understanding the antecedents and behaviours involved in empathic

interactions between patients and healthcare professionals. Studies one to three

were cross sectional and quantitative in design. Studies one and two investigated

relationships between self-reported empathy, personality and emotional

intelligence. Findings suggested that (1) perspective taking and empathic concern

were closely associated with agreeableness and extraversion, and also loaded on

to the single factor of emotional intelligence (2) fantasy was associated with

openness to experience but not emotional intelligence, and (3) personal distress

was positively related to neuroticism and negatively related to emotional

intelligence. Study three went on to investigate the relationships between

emotional intelligence, propensity to empathise and empathic behaviour among

doctors. Propensity to empathise was positively related to observer ratings of

empathic behaviour, but not when doctors had qualified in a different country.

Finally, study four qualitatively examined empathy in the healthcare context,

from patients’ perspectives. Situational and patient characteristics were also

identified as antecedents to empathy, further relating to employee engagement

 11

and work design. The specific behaviours associated with empathy as judged by

patients included helping and prosocial behaviours. Implications for the

development of empathy are discussed in terms of possible training, development

and work design interventions. Finally, areas for future research are identified.
 12

Chapter 1 - Introduction

“Some patients, though conscious that their condition is perilous, recover their

health simply through their contentment with the goodness of the physician”.

Hippocrates 460-380 BC

The importance of the doctor-patient relationship has been recognised throughout

the history of medicine. In a systematic review of the literature, Di Blasi,

Harkness, Ernst, Georgiou and Kleijnen (2001) found that those physicians who

adopt a warm, friendly and reassuring manner are perceived as more effective

than those who keep consultations formal. Empathy has been identified as

facilitating improved outcomes for both patients and doctors (e.g., Carmel &

Glick, 1996; Hardee, 2003; Hojat, Mangione, Nasca, Cohen, Gonnella, Erdmann,

Veloski & Magee, 2001), and an important trait for other healthcare

professionals including pharmacists, nurses and care staff (Lilja, Larsson,

Hamilton & Issakainen, 2000; Reynolds, Scott & Jessiman, 1999). Yet despite

the fact that empathy is recognised as a crucial component of helping

relationships, it remains a difficult concept for researchers to study and

practitioners to develop, not least due to continuing debates about its definition

and measurement. This chapter presents a rationale for conducting research on

empathy in healthcare professions. It begins by examining why empathy is

important by exploring potential outcomes and consequences for the patients and

healthcare professionals.
 13

1.1. Patient satisfaction

The extent to which a healthcare professional demonstrates empathy is assumed

to have an impact on the patient experience, and patient satisfaction is one of the

most frequently used outcome measures to evaluate empathy and communication

(e.g. Kaplan, Ware & Greenfield, 1998; Mercer & Reynolds, 2002). In the US,

patient appraisals are used as part of performance related pay schemes, and

account for as much as 20% of a doctor’s pay (Kolata, 2005). In Scotland the

CARE Measure (Consultation and Relational Empathy Measure: Mercer &

Reynolds, 2002; Mercer, Maxwell, Heaney & Watt, 2004) of patient satisfaction

has been accredited as a method for appraising General Practitioners (GPs), and

in the U.K. there has been a strategic move within the NHS to acknowledge the

patient perspective and develop doctor-patient partnerships (Department of

Health, 1996). The rationale for such approaches derives from the large number

of studies that have linked empathy and communication skills to patient

satisfaction across a wide range of healthcare settings including: paediatrics

(Wasserman, Inui, Barriatua, Carter & Lippincott, 1984); stroke units (Pound,

Gompertz & Ebrahim, 1995); diabetic clinics (Hornsten, Lundman, Selstam &

Sandstrom, 2005); rehabilitation facilities (McGilton, Irwin-Robinson, Boscart &

Spanjevic, 2006), and eating disorder clinics (Ramjan, 2004).

If empathy can have a positive effect on patient satisfaction, then it follows

sensibly that a lack of empathy should have a negative effect on patient

satisfaction with treatment. For some time now, legal experts have argued that

medical malpractice litigation is commonly due to problems arising from

 14

interpersonal skills of doctors (Avery, 1985), with investigations by medical

professionals in the 1990s providing support for this assertion. For example,

Frankel (1995) found that communication breakdowns played a key part in

litigation across a range of specialties including general surgery, gynaecology

and radiology. In a qualitative study of malpractice cases in the US, Beckman,

Markakis, Suchman and Frankel (1994) found evidence of relationship problems

between the doctor and patient in 32 of the 45 cases reviewed, with the most

common themes including devaluing patient or family views, delivering

information poorly, and failing to understand the patient’s or family’s

perspective. Although empathy was not the subject of investigation in these

studies per se, it was identified as a key issue, with the authors concluding that

“Doctors who can’t communicate are more likely to end up in court” (p.1365).

1.2. Patient adherence to treatment

While satisfaction with treatment is regarded as a valuable outcome of empathy,

perhaps more important are those studies that find actual improvements in patient

health. As Kaplan, Greenfield and Ware (1989) point out, patients may be

satisfied but this may not necessarily lead to the best healthcare outcomes or the

behaviour change required for treatment success. One possible reason for

improvement in patient health (or the lack of this) is the patient adhering to the

treatment prescribed by a doctor. Squier (1990) noted that “patients’ adherence to

treatment advice is generally necessary for maximum benefit to occur” (p.329),

with benefits resulting from greater sharing of concerns and motivation to get

better. In fact, several major reviews over the last three decades (e.g., Becker &
 15

Maiman, 1975; Becker & Rosenstock, 1984; Garrity, 1981) have concluded that

empathy is related to treatment adherence. Evidence has been found using a

range of methodologies, for example, in a qualitative study of the doctor-patient

relationship Frankel and Beckman (1989) identified that doctors who do not

attempt to identify with patients, or understand their reasons for non-compliance,

are more likely to appear frustrated and judgmental, rather than empathic and

supportive. The study also showed that feedback to the doctor on this issue can

have a positive impact on communication style as well as improved health

outcomes for patients. This finding is supported by quantitative research from

Winefield, Murrell and Clifford (1995). They analysed transcripts from 210 GP-

patient consultations and found that patient satisfaction after the consultation

predicted patients’ subsequent compliance with treatment. Patients who admitted

non-compliance with treatment were also more likely to complain that the doctor

has not listened to their perspective or treated them as an equal during the

consultation.

1.3. Improved health outcomes

A step on from adherence to treatment, other studies have directly investigated

the link between empathy and health outcomes. Empathy from a healthcare

professional can be a facilitator of motivation and support for patients (Kaplan,

Greenfield & Ware, 1989). By communicating this empathy to build patients’

perceptions of personal care, professionals have the power to influence patient

behaviour and also change their patients’ perceptions of their health status

(Tarrant, Windridge, Boulton, Baker & Freeman, 2003). Of 21 studies conducted

 16

between 1983 and 1993, Stewart (1995) found that 16 reported a positive

relationship between improved physician-patient communication and health

outcomes, including emotional health, symptom resolution, improved day to day

functioning and physiological measures such as blood pressure and pain control.

In nursing, La Monica, Madea and Oberst (1987) found less anxiety, depression

and hostility among clients being cared for by highly empathic nurses. For

patients with a variety of illnesses, physician empathy has also been linked to

decreases in patient anxiety, positive physiological effects, and improved health

outcomes (Frasure Smith, Lesperance & Talajic, 1995; Rietveld & Prins, 1998).

In their detailed analysis of physician communication behaviours associated with

improved health outcomes in patients with chronic disease, Kaplan, Greenfield

and Ware (1989) concluded that increased patient control and greater expression

of emotions by physicians and patients were consistently related to positive

health outcomes. This was found to be the case whether the outcome measure

was a subjective perception of health or an objective physiological measure such

as blood pressure or blood sugar levels. Such physiological outcomes have also

been reported in investigations of positive nurse-patient relationships (Reynolds

& Scott, 2000).

1.4. Professional satisfaction

In addition to recognising the importance of empathy for patients, researchers

have pointed out that professionals also need to recognise the benefits of

adopting empathic approach for themselves, in order for an empathic style to be

sustained. Moscrop (2001) argues that empathy with patients provides

 17

professionals with greater understanding of the meaning and importance of their

work. Indeed, such positive relationships between empathy and professional

satisfaction have been established. For example, in a study of continuing care

nurses in Canada, McGilton, Irwin-Robinson, Boscart and Spanjevic (2006)

found that those nurses who reported having developed closer relationships with

their patients also reported higher levels of job satisfaction. Conversely, an

absence of a therapeutic relationship has been linked to a struggle for

understanding and control (Ramjan, 2004).

1.5. Sustaining and developing empathy

Yet despite mounting evidence that empathy in healthcare professionals can have

an important impact on patient care and professional satisfaction, surprisingly

little progress has been made in efforts to develop interventions capable of

sustaining or significantly improving the levels of empathy demonstrated by

professionals. In fact, somewhat worryingly, it has been found that whilst

empathy usually increases with maturity, it typically declines over the period of

medical education and early stages of the medical career (Bellini, Baime & Shea,

2005; Bellini & Shea, 2005; Woloschuk, Harasym & Temple, 2004). Hojat,

Mangione, Nasca, Rattner, Erdmann, Gonnella and Magee (2004) found a

particular decrease in compassionate care during these early years.

Although there have been efforts to focus more on developing empathic

behaviour, Squier (1990) points out that compared to increased investment in

technological and pharmaceutical methods of treatment in the U.K., the

 18

development of relationship building skills is usually given lower priority due to

pressure on finances within the NHS. Similarly in the US the Association of

American Medical Colleges states that a key learning objective for medical

education should be that “physicians must be compassionate and empathetic in

caring for patients”. Yet according to Carmel and Glick (1996), whilst

compassionate-empathic physicians who focus on the welfare of patients by

‘curing and caring’ are desired by patients they are rarely found in medical

settings. Again, Sparr, Gordon, Hickam and Girard (1988) found that medical

students became more negative in their attitudes to the doctor-patient relationship

over the course of training, with bureaucratic pressures and experiences with

difficult patients cited as reasons for attitude changes. There is also

comparatively little evidence to suggest that changes in nursing education and

practice have had an impact on reported low levels of empathy in nursing

(MacKay, Hughes & Carver, 1990; Reynolds & Scott, 2000).

In attempts to “stop the rot” (Spencer, 2004), training interventions focused on

developing empathy have become more common and medical education has

begun to re-emphasise the doctor-patient relationship, with models of

communication and empathy included as a key component (e.g., Makoul, 2001;

Suchman, Markakis, Beckman & Frankel, 1997). Even with this renewed focus,

however, evaluations of training interventions designed to improve empathy and

communication skills have produced mixed findings. Stepien and Baernstein

(2006) found that only seven out of 13 articles reporting evaluations of empathy

interventions reported a significant increase in student empathy post-

intervention, although only three of the 13 studies reviewed gave a clear

 19

definition of empathy. It seems that after more than a century of research on

empathy, most articles still begin with a discussion of the meaning of the term.

At best, discussions take a broad approach and encompass many elements, at

worst they are conflicting. Interestingly, very few researchers in the medical field

have sought to develop an integrated theory of empathy based on empirical

research, but have relied instead on reviews of literature conducted in a range of

contexts. According to Squier, “the essence of good practitioner-patient

relationships lies in the presence of empathy” but there has been “little attempt to

integrate diverse perspectives into a systematic theory” (1990, p.326).

A basic premise of this thesis is therefore that greater understanding is required

of how empathy can be developed in patient-healthcare professional interactions

in order to foster the development of more effective training and development.

This chapter has considered why empathy is important in the helping professions.

Chapter two continues with a review of the existing literature concerned with

theories and models of empathy and empathic behaviour.

 20

Chapter 2 - Literature Review

Chapter one identified a need for greater understanding of empathic processes in

healthcare roles. Most discussions of empathy in healthcare open with a debate

about the nature and definition of the concept (Hardee, 2003). Chapter two

therefore provides a review of definitions and models of empathy, drawing from

both healthcare and psychology literature. Early approaches to defining and

understanding empathy will be discussed before presenting more recent

multidimensional models of empathic processes.

2.1 Empathy: An Overview

The term “empathy” first appeared in the psychology literature in Titchener’s

translation of the German word “einfühlung” literally meaning ‘feeling into’ or

the ability to perceive the subjective experience of another (1909). Since then,

empathy has been investigated in many fields of psychology including

developmental (e.g., Baron Cohen, 2003), social (e.g., Eisenberg, 1987), and

forensic (e.g., Blair, 2005), with research focusing on diverse topics such as:

relationships and marriage; violence and sexual offending; childhood

development, and autism. Within occupational psychology there is an emerging

interest in empathy relating to leadership and sales (e.g., Kellett, Humphrey &

Sleeth, 2002; Plank, Minton & Reid, 1996), and empathic processes in

relationships with clients and colleagues (e.g., Silvester, Patterson, Koczwara &

Ferguson, 2007). Studies of empathy in healthcare also span 50 years, with early

research undertaken by Aring (1958) in the medical profession, Peplau (1952) in

 21

nursing and Rogers (1957) in counselling. The focus of this chapter is

predominantly on theory and research from healthcare and psychology.

2.2 Defining empathy

Bachrach (1976) argues that “we know what we mean” (p.35) when we think of

empathy in a medical context, but as with many other psychological concepts,

there has been much debate over its definition, with articles often beginning with

a discussion of the exact meaning of the term (e.g., Ohmdahl, 1995). Early

researchers tended to differentiate between cognitive and emotional definitions,

although more recently efforts towards a more integrated approach have resulted

in the development of multidimensional models (e.g., Davis, 1996; Larson &

Yao, 2005). In broad terms, however, four dimensions of empathy have been

identified in healthcare research: an emotional (or affective) dimension; a

cognitive dimension; a moral dimension, and a behavioural dimension of

empathy (Morse, Anderson, Bottorff, Yonge, O'Brien, Solberg & McIlveen,

1992; Stepien & Baernstein, 2006). Each of these is reviewed briefly below.

2.2.1 An emotional dimension of empathy

Many psychology researchers have used an emotional definition of empathy in

their studies. For example, Feshbach’s early studies of conditions facilitating

empathy in young children defined empathy as a vicarious affective response to

another’s emotion (Feshbach & Roe, 1968). Strayer (1987) adopted a definition

of empathy as “the self’s feeling into the affect of another person” (p.236), and
 22

from a psychoanalytic perspective, Greenson (1960) suggested that “to

empathize means to share, to experience the feeling of another person” (p.418).

These definitions refer to empathy as shared affect, a temporary state of

identifying emotionally with another person. Proponents of the emotional

perspective define empathy as the process of vicariously feeling another person’s

emotional experience, which in the medical context has been identified as similar

to sympathy, defined as “feeling with the patient or feeling similar emotions that

the patient feels” (Mehrabian & Epstein, 1972, p. 1563). However, it has been

argued that sympathy is not appropriate in clinical contexts as it can “interfere

with objectivity in diagnosis and treatment” (Hojat et al., 2002, p.1563).

An alternative view of emotional empathy that has emerged more recently

suggests that merely to experience similar emotions to those of a target will not

lead to an observer being considered fully empathic (Batson, Fultz &

Schoenrade, 1987; Zahn-Waxler & Robinson, 1995). Rather than the sharing of

affect, these researchers define empathy as experiencing emotion that is

congruent with, but not identical to, the emotion of another person (e.g., Batson

& Shaw, 1991). Vreeke and van der Mark (2003) note that genuine empathy that

is effective in comforting the target is likely to involve the experience of

compassion and caring. In support of this argument, Davis (1996) also

emphasises the notion of congruent affect, as not all emotional reactions may be

perceived as empathic. Taking the example of frustration experienced by a target,

an observer who is perceived to share that frustration may be perceived as less

helpful by the target than an observer who expresses sympathy or compassion.

 23

Shared affect and sympathy have been distinguished experimentally, by asking

participants to watch video clips of two different scenarios. Gruen and

Mendelsohn (1986) found that while there are situations in which both empathy

and sympathy are aroused, sympathy can be present without shared affect. They

found that emotional empathy was more consistently linked to personality traits

associated with affiliation and putting others’ needs first, while sympathy was

more situation-specific.

Although reviews by Morse et al. (1992) and Stepien and Baernstein (2006) both

emphasise the importance of emotional engagement between the healthcare

practitioner and patient, the exact nature of the emotional response has not been

specified. Also, there are conflicting views such as those suggested by Hojat et

al. (2002) that an emotional response interferes with the objectivity of the

practitioner and reduces the effectiveness of diagnosis and care. However, no test

of these assertions has been made, and very few studies involve patients in their

methodologies. The emotional labour literature raises another issue with respect

to emotional empathy. Whereas surface acting emotional empathy can refer to

the appearance of care and concern that is not necessarily experienced, deep

acting empathy involves the actual experience and subsequent expression of

these emotions (Grandey, 2003). These are important issues to consider in efforts

to develop and sustain clinical empathy due to their potential impact on both the

practitioner and the patient. Whereas surface acting empathy might protect the

practitioner from becoming over-involved and emotionally exhausted (Maslach,

1978), it may not be perceived by patients as genuine. Conversely, while deep

acting emotional empathy might be perceived as more genuine by the patient,

 24

potential negative consequences might include over-involvement, decreased

objectivity and emotional exhaustion (Larson & Yao, 2005). These issues may

also be important in understanding factors that influence the maintenance of

clinical empathy over time (e.g. Hojat et al., 2004; Spencer, 2004).

2.2.2 A cognitive dimension of empathy

Although early researchers conceptualised empathy as a predominantly affective

process, it was suggested as early as 1929 that empathising involved

understanding of another person, by viewing and interpreting their actions,

movements and physical cues (Kohler, 1929). One of the most widely known

researchers to adopt this perspective was Hogan (1969) who defined empathy as

‘the intellectual or imaginative apprehension of another’s condition or state of

mind without actually experiencing that person’s feelings’ (p.308, italics added).

Cognitive definitions therefore conceptualised empathy as adopting another’s

perspective in order to understand that person’s thoughts, feelings and actions.

However, it has been suggested that healthcare professionals need more than the

objective understanding or ‘detached concern’ advocated by some medical

educators (Halpern, 2001). Morse et al. (1992) and Stepien and Baernstein

(2006) argue that without some emotional engagement, the healthcare

practitioner will not be perceived by the patient as genuinely empathic. The

importance of both cognitive and affective components of empathy is recognised

in more recent research. For example, occupational psychologists have identified

the importance of perspective taking in empathising with clients and providing

 25

helpful customer service. Axtell, Parker, Holman and Totterdell (2007) asked

347 agents from two UK call centres to complete self ratings of perspective

taking and emotional empathy. Both of these significantly predicted manager

ratings of helping behaviour. A similar result was found in a study of

manufacturing employees in helping internal customers, suggesting that the

results of the study may be generalisable (Parker & Axtell, 2001). These two

studies provide evidence that perspective taking is an important part of empathy

in an applied, emotional labour role. But as Hynes, Baird and Grafton (2006)

argue cognitive processes can still have an emotional focus. They identified the

separate aspects of emotional and cognitive perspective taking: emotional

perspective taking (EPT) is concerned with making an emotional attribution with

regards to another person’s behaviour or experience, whereas cognitive

perspective taking (CPT) involves making an attribution that requires no

emotional understanding. EPT but not CPT is deemed to be empathy. This is in

line with the studies previously discussed in which ratings of perspective taking

involved agents rating the extent to which they imagined their customers’

perspectives and thought about how they would feel in a customer’s situation

(Axtell et al, 2007; Parker & Axtell, 2001).

To demonstrate this distinction, Hynes, Baird and Grafton (2006) asked

participants to read scenarios which required them to imagine what a target was

feeling (EPT), thinking (CPT), or neither (control). Use of functional magnetic

resonance imaging allowed the researchers to show that only emotional

perspective taking activated specific areas of the medial orbital frontal lobes

known to be damaged in people with empathy dysfunction (Eslinger, 1998;

 26

Grattan, Bloomer, Archambault & Eslinger, 1994). Physiological evidence that

empathy specifically involves emotional perspective taking represents a

refinement of the definition of ‘perspective taking’ generally used to refer to

cognitive empathy. In the healthcare context, a cognitive dimension of empathy

can therefore be described as the process of perspective (or role) taking to

identify and understand patients’ emotions and perspectives (Morse et al., 1992;

Stepien & Baernstein, 2006).

2.2.3 A moral dimension of empathy

The third dimension of empathy identified in medical and nursing reviews

involves a moral component, defined as “an attitude of receptiveness, availability

and presence” (Zderad, 1970, p.30) that drives someone to help their patient. The

importance of this attitude or motivation is highlighted in the first stage of

Barrett-Lennard’s cyclical model of the phases of empathy (1993; see figure 2.1).

Based on Carl Rogers’ (1957) work on empathy in therapeutic relationships,

Barrett-Lennard intended his model to be applicable to explaining empathy in the

general population. PA represents the observer, who is empathising with a target,

PB, and in the ‘pre-empathy’ condition, PA is required to actively attend to the

subjective feelings and experiences of PB. If PA is not in an ‘empathic set’ the

process will not begin. By attending to PB in this way, PA connects by

recognising and interpreting the expressions of PB and only if the motivation to

attend to the needs of the patient is present in the ‘pre-empathy condition’ will

the further stages of the model take place. By including this ‘moral’ motivation
 27

as Morse et al. (1992) refer to it, empathy is seen as an altruistic rather than an

egoistic concept.

Pre-empathy Phase I Phase II Phase III Phase I in

conditions Empathy Empathy Empathy repeat form

(a) (b) (c) (d) (e)

PA PB PA PB PA PB PA PB PA PB

i) ’Empathic set’ Empathic Expressed Received Feedback;

expressed by PA resonation empathy empathy Fresh
expression and
ii) PB expressing resonation
self

Figure 2.1: The ‘cycle’ of empathy (Barrett-Lennard, 1993)

Evidence for altruistic empathy has been provided by Batson and colleagues

(e.g., Batson & Coke, 1981), who argue that only if the motivation to help is

selfless and altruistic and can it be considered as genuine empathic helping. To

demonstrate this, Coke, Batson and McDavis (1978) played a radio broadcast of

a person in distress to 144 undergraduate students. Results indicated that

participants who offered more help to the target in the broadcast were those who

reported greater empathic concern for others on a self report measure but not

higher personal distress. This led Coke et al. to conclude that empathy is an

altruistically motivated concept, including an emotional component. If empathy

were egoistic, those who reported greater personal distress would have also

offered help in order to reduce their own distress. However, there are concerns

with the ecological validity of this scenario as it is likely that real incidents that
 28

would produce empathic concern may also produce personal distress. For

example, witnessing the distress of someone with whom one shares a close

relationship would produce not only concern for that person but also personal

distress due to the feeling of ‘oneness’ with the target (Cialdini et al., 1997).

Also, the study does not consider individual differences in motivation, such that

some people may be motivated to act by personal distress whereas others may be

more motivated by empathic concern. Despite some limitations, findings suggest

that an altruistic motivation may lead to people being more likely engage in

helping behaviour. Indeed, altruistic concern to help relieve patient distress is

seen as a defining attribute of nurses (Odom-Forren, 2007).

2.2.4 A behavioural dimension of empathy

The final dimension of empathy identified as important for healthcare

practitioners relates to the behavioural or communication aspect of empathy.

Barrett-Lennard’s (1993) cyclical model presented above conceptualises

empathy as an active interpersonal process that involves communication and

behavioural interaction between the parties (see Figure 2.1). Observers’

perceptions of the target are given prominence in the model. Provided the

altruistic empathic set is present, in Phase II PA then expresses an empathic

response. According to Barrett-Lennard, this ‘expression’ could be intentional or

automatic, verbal or non-verbal, but for the expression to constitute empathy it

needs to convey that PA understands PB. The Phase II expressed empathy then

makes it possible for PB to receive this empathy in Phase III and the cycle repeats

following further expression from PB. If all conditions are satisfied ongoing and
 29

meaningful communication between the two parties will result. This experience

of being both heard and understood is proposed to bring about feelings of relief,

of being helped, connected or less alone. In research terms it would be of value

to know which methods or styles of expression are more impactful in bringing

about an awareness of being understood in PB and in an early exploration of this

phase, Mansfield (1973) videotaped initial interactions between psychiatric

nurses and patients to investigate empathic communication. Open

communication as well as non-verbal behaviours which demonstrated

compassion were identified as important. However, the study focused on single

initial interactions between nurse and patient when relationships were not

established, consequently information gained may not be generalisable across

other roles or more extended relationships. In addition, a study by Silvester,

Patterson, Koczwara & Ferguson (2007) found that doctors’ discussion of

personal topics and sensitive responses to patient cues predicted higher

judgments of physician empathy from observers. Other than this, very little has

been done to understand how empathy is communicated to patients and Morse et

al. (1992) concluded that there is an urgent need for more empirical research to

identify the behaviours perceived by patients as empathic.

2.3 Multidimensional models of empathy in healthcare

Following research evidence to support the existence of the four dimensions of

empathy identified as important in nursing and medical reviews, more recent

work concerned with understanding empathy in healthcare settings has focused

on developing multidimensional models of empathic processes. These

 30

encompass factors (patient and healthcare professional) that can lead to empathic

experience, demonstration of empathic behaviour and ultimate outcomes of

empathy for the different parties involved. Several multidimensional models of

empathy in healthcare roles have been developed. The three that have had most

influence are Squier’s (1990) multidimensional model of medical empathy,

Davis’ (1983, 1996) multidimensional model of empathy, and the process model

of clinical empathy (Larson & Yao, 2005). Each of these is reviewed in more

detail below.

2.3.1 The multidimensional model of medical empathy (Squier, 1990)

Squier (1990) developed his multidimensional model of medical empathy from a

review of existing literature (see Figure 2.2 for an adapted version). The central

theme of the model is that the healthcare practitioner needs to engage in both

emotional and cognitive empathic processes in order to maximise health benefits

for patients. Cognitive empathy facilitates the practitioner’s full understanding of

the patient’s health problems, but communication of this understanding allows

the patient to understand their illness and proposed treatment. Additionally,

emotional empathy on the part of the healthcare professional is seen as the main

predictor of stress reduction and increased patient satisfaction, which impacts on

the patient’s motivation to get well. The model provides a useful link between

the psychological process of empathy and patient outcomes, although these

pathways have yet to be fully tested. Indeed, no research has been located which

has directly tested Squier’s model and sixteen years after the model was
 31

PRACTITIONER INTERPERSONAL PATIENT

Accurate Information
Perspective
provision
Taking
Shared
understanding of
health problems
Open communication
Practitioner Greater knowledge Stronger adherence to
Empathic
relationship Greater satisfaction treatment regimes and
Understanding
skills Reduced stress improved health
Sharing of Increased motivation
emotional
concerns about
Emotional illness
Emotional
concern and
reactivity
motivation

Figure 2.2: Model of empathic understanding and adherence to treatment regimens (adapted from Squier, 1990)
 32

published, Stepien and Baernstein (2006) called for more empirical investigation

to address this. Potential limitations of the model include an absence of the

moral dimension of empathy identified within the healthcare literature as well as

a vague behavioural dimension, described only as sharing of concerns and

information. As no antecedents of empathy are specified, the model gives little

guidance as to how empathic processes can be developed or sustained among

healthcare practitioners.

2.3.2 The Multidimensional Model of Empathy (Davis, 1983)

Davis’ multidimensional model (1983; 1996) is perhaps the most widely

recognised psychological model of empathy (see Figure 2.3). Developed to

explain empathy in the general population, this model identifies antecedents of

empathy as including individual characteristics of the observer as well as

situational characteristics. Thus this model gives greater understanding of

potential ways to develop empathy than Squier’s model, which does not specify

antecedents of the empathic processes. With regard to individual characteristics,

Davis’ model suggests individual differences in empathy, with dispositional (or

trait) empathy relatively stable across time (Davis, 1983; Gladstein, 1987;

Strayer & Eisenberg, 1987). According to Davis an empathic disposition

develops during childhood as cognitive ability develops, and social and family

experiences take their influence (Hoffman, 1984). Changes during adulthood

would depend on unusual events such as brain injury or illness which may affect

cognitive capacity or personality (Damasio, 1994). Individual differences are

proposed in terms of both abilities and traits, as antecedents are said to include
 33

both the intellectual ability to engage in perspective taking and the dispositional

tendency to engage in empathic processes and experience emotional outcomes.

As well as individual characteristics, situational antecedents are identified as the

strength of the observer’s previous experiences and the degree of similarity

between observer and target. Both of these would enable more accurate

perspective taking and the activation of congruent emotional responses.

These antecedents are then proposed to provoke empathic processes in the

observer. These include the non-cognitive ‘primary circular reaction’ (Hoffman,

1984) by which even young infants appear to experience shared affect. Simple

cognitive processes are said to occur via classical conditioning as a result of the

learning history of the observer. Certain emotions that one may have

experienced, and at the same time perceived in others, can be activated simply by

observing that emotion in a target (Eisenberg, Fabes, Schaller, Miller, Carlo,

Poulin, Shea & Shell, 1991). This stage also includes the advanced cognitive

process of role or perspective taking, although Davis does not explicitly

recognise the distinction between emotional and cognitive role taking (Hynes,

Baird & Grafton, 2006). Following these processes, a range of outcomes within

both the observer (intrapersonal outcomes) and the target (interpersonal

outcomes) may result. Affective outcomes within the observer are divided into

parallel and reactive emotions. Parallel emotions are Davis’ term for the shared

affective response. The observer feels emotions the same as those experienced by

the target, which may result directly from the individual characteristics of the

observer, from the primary circular reaction or from simple cognitive processes
 34

ANTECEDENTS INTRAPERSONAL INTRAPERSONAL INTERPERSONAL

PROCESSES OUTCOMES OUTCOMES

THE PERSON AFFECTIVE OUTCOMES

Biological Capacities Parallel (shared affect)
Individual Differences Reactive (empathic concern, anger,
Learning History personal distress)

THE SITUATION NON- AFFECTIVE

Strength of Situation OUTCOMES
Observer/Target similarity Interpersonal Accuracy
Attributional Judgments

NON-COGNITIVE
Primary Circular Reaction HELPING +
Motor Mimicry

SIMPLE COGNITIVE
Classical Conditioning SOCIAL BEHAVIOUR +
Direct Association
Labelling

ADVANCED COGNITIVE AGGRESSION -

Role Taking

Figure 2.3: A multidimensional model of empathy (adapted from Davis, 1996)

 35

such as classical conditioning as mentioned previously. Reactive emotions on the

other hand are those which are different from, but congruent with, those of the target,

such as compassion and sympathy. These are purported to result from more

advanced cognitive processes such as perspective taking. Evidence for this

relationship was found in a study by Axtell et al. (2007), in which emotional

empathy self ratings were found to partially mediate the relationship between self

ratings of perspective taking and manager ratings of helping behaviour. This study

demonstrates the intertwined nature of perspective taking, emotional empathy and

helping behaviour.

Personal distress, defined as “the tendency to feel discomfort and anxiety in response

to needy targets” (Davis, 1996, p. 18) is highlighted as an interesting case in terms of

an emotional reaction. Personal distress may be a parallel emotion in some

circumstances. However it is not easily classified because it may not be the

reproduction of the target’s affective state, but more of a response to it. It is also

difficult to classify it as a reactive emotion because it may not necessarily be a

congruent response which the target would perceive as helpful. Davis’ inclusion of

personal distress in the model supports the view of Cialdini, Baumann and Kenrick

(1981) who see helping behaviour in terms of egoistic motivation, resulting from the

desire to relieve one’s own negative state, which may include tension, stress or guilt.

However as discussed earlier, it has been argued that true empathic helping should

be altruistically motivated (e.g., Batson & Coke, 1981). Personal distress may

generate an egoistic motivation to reduce one’s own negative state but this would not

constitute empathy. Only if empathic concern is experienced instead then the

motivation to help is selfless and altruistic and thus can be considered as empathic
 36

helping. Findings from studies such as that by Coke, Batson and McDavis (1978) in

which empathic concern and not personal distress predicted help for a target on a

radio broadcast have lead Batson, O’Quin, Fultz, Vanderplas and Isen (1983) to

conclude that personal distress is distinct from empathy, which ought to be purely

altruistic in nature. Therefore although personal distress may be an affective outcome

of observing the emotions of another, an egoistic motivation to help should exclude

it from a model of empathy. This potential criticism of Davis’ model seems

particularly relevant for healthcare practitioners who are seen as being defined by an

altruistic nature (Odom-Forren, 2007).

Other non-affective outcomes identified in Davis’ model of empathy include

interpersonal accuracy and attributions. Accuracy is the result of successfully

interpreting the other person’s thoughts and feelings. This relates to work by Ickes

(1993) on empathic accuracy, which conceptualises empathy as an ability rather than

a predisposition to engage in empathic processes. It should be noted that although

Davis acknowledges the need for accuracy, his measure of empathy (the

Interpersonal Reactivity Index, 1983; see Chapter 3) is a self-report measure that

does not attempt to measure ability from the perspective of a target.

The second non-affective outcome, attributions are judgements made by the observer

to explain the behaviour of the target. Several researchers have reported links

between empathy and attributions. Regan and Totten (1975) asked female students to

watch a conversation between two other female students and then rate one of them

on certain communication styles. Participants were asked to indicate whether they

thought each style resulted from the student’s personality or the situation.
 37

Instructions were manipulated so that participants were either asked to empathise

(emotional perspective taking) with the student or simply observe them. They found

that encouraging people to empathise resulted in observers attributing the student’s

behaviour more to the situation rather than personal disposition. This finding has

been replicated several times (e.g., Archer, Foushee, Davis & Aderman, 1979;

Betancourt, 1990).

It has also been suggested that Weiner’s (1980; 1985; 1986) attributional theory may

be useful in understanding helping responses (Fenwick, 1995). Weiner (1986)

suggests that it is the underlying structure of peoples’ explanations for causes of

events, rather than specific content, which determines the emotional and behavioural

consequences for an observer. According to Weiner’s attributional model of helping

behaviour, an observer will attempt to determine why help is needed before acting. If

the observer judges a negative outcome to be within a target’s control, this may lead

to negative emotions such as anger and annoyance, blaming the individual for the

outcome, and consequently, to help being withheld (Meyer & Mulherin, 1980;

Russell & McAuley, 1986). However, if the target is believed not to have control

over the situation, observers are more likely to experience sympathy, increasing the

likelihood of help and support being offered.

In a test of this in an applied medical setting, Silvester, Patterson, Koczwara &

Ferguson (2007) explored a socio-cognitive model of empathic ability incorporating

social-cognitive (distal) and behavioural (proximal) predictors of empathic

judgments. Doctors’ explanations for patient outcomes and their subsequent

interactions with patients were examined. Explanations were assessed via a modified
 38

Attibutional Style Questionnaire (ASQ: Peterson, Semmel, Baeyer, Abramson,

Metalsky & Seligman, 1982; Peterson & Villanova, 1988). Behaviours were

observed in simulated patient interactions as part of an assessment centre. Doctors’

causal attributions for patient outcomes predicted empathy judgments made by

observers, with doctors who attributed positive patient outcomes to causes that were

more internal and controllable to him or herself being judged more empathic and

more motivated to engage in help-giving behaviour. This suggests the need for the

observer to perceive they have control over the situation.

Finally, Davis’ model identifies the interpersonal outcomes associated with empathy

as helping, aggression and social relationships. Historically, empathy has most often

been linked with helping behaviour (Eisenberg, 1986). However, empathy has also

been found to have both a negative relationship with aggressive behaviours and a

positive relationship with prosocial behaviour (Eisenberg et al., 1991). Davis’

multidimensional model is by far the most complex and comprehensive model of

empathy reviewed here and it could be argued that aspects within the model relating

to the development of empathy during childhood are less relevant for the study of

adults working in the healthcare professions.

2.3.3 Process model of clinical empathy (Larson & Yao, 2005)

Davis’ model has been identified by Larson and Yao (2005; see Figure 2.4) as

having potential to further understanding of the process of clinical empathy, with

their adaptation focusing on the more salient aspects for this context. Retaining the

core aspects of antecedents, intrapersonal processes and outcomes of empathy, this

 39

Antecedents

• Physician characteristics

• Patient characteristics

• Situational
characteristics

Empathic Processes

• Non-cognitive processes
e.g. mimicry

• Cognitive Processes e.g.

association, role-taking

Intrapersonal Processes Interpersonal Processes

• Affective reactions • Helping behaviour

Parallel emotions
Reactive emotions • Conflict management

• Non-affective reactions • Social behaviour

Attributions

Physician Outcomes Patient Outcomes

• Job satisfaction • Satisfaction

• Burnout • Healthcare outcomes

Figure 2.4: Process model of clinical empathy (from Larson & Yao, 2005).
 40

direct adaptation by Larson and Yao (2005) illustrates the potential applicability of

Davis’ model to healthcare practitioners. The process model was developed to

provide a better understanding of the psychological and behavioural activities

involved in empathising and to help physicians incorporate empathy into daily

practice. It includes antecedents such as patient and practitioner characteristics as

well as situational characteristics. For example potential obstacles facing healthcare

professionals who aspire to develop empathy include workload, a lack of focus on

empathy and cynicism. Positive antecedents, if present, lead to the cognitive and

emotional processes specified by Davis. These go on to result in interpersonal

helping behaviours which include social behaviour and management of conflict. This

adaptation of Davis’ model provides guidance for practitioners wishing to develop

and sustain empathy, with the inclusion of both individual and situational

antecedents of empathy. Despite its strong theoretical support, the suggested

pathways remain untested to date.

2.4 Models of empathy: limitations and next steps

To summarise, chapter one concluded by arguing that to identify the best methods to

develop and sustain empathy in healthcare professionals there is a need to better

understand the antecedents and processes involved in empathic responding towards

patients. Chapter two has reviewed recent models of empathy in healthcare that

embrace a multidimensional approach, incorporating antecedents, cognitive,

emotional and behavioural components. Davis’ model (1983; 1996) is the most

comprehensive psychological model and the direct adaptation by Larson and Yao

(2005) highlights the potential application of Davis’ model to the healthcare context.
 41

However, to date there has been very little research that has tested the utility of the

model for understanding empathy between patients and healthcare professionals and,

as such, very little is known about the individual characteristics that might have an

impact on demonstrated empathic behaviour, or indeed what those specific

behaviours are. In order to answer these questions, this thesis now turns to a review

of the methods used to investigate empathy.

 42

Chapter 3. Measures of empathy

“In spite of the apparent difficulty involved in developing a valid and acceptable

measure of empathy, the theoretical import of the concept requires that continuing

efforts be made.”

Hogan (1969, p.308)

The predominant method of investigating empathy in adults is the self report

questionnaire (e.g., Davis, 1983; Hogan, 1969). As indicated by Hogan’s quote,

numerous measures have been developed, focusing on emotional, cognitive and

multidimensional aspects of empathy, including measures designed for use with the

general population as well as those targeted specifically at healthcare settings. Before

reviewing existing scales in more detail, however, it is first important to understand

two important psychometric properties, namely reliability and validity.

3.1. Reliability and validity

Reliability is concerned with the effectiveness of a test to measure something

consistently (Rust & Golombok, 1999). Types of reliability include internal and test-

retest reliability. Test-retest reliability is estimated by administering the scale to the

same group of respondents at two different time points and calculating the

correlation coefficient between the two resulting sets of scores (Kline, 2000).

Internal reliability can be estimated by either calculating Cronbach’s alpha, the

average correlation between all items on a test, or by calculating a split-half

reliability - the correlation between scores on one half of the scale items with the
 43

other half (Rust & Golombok, 1999). Generally, a minimum Pearson’s r of 0.7 is

required to demonstrate acceptable reliability (Kline, 2000).

Validity on the other hand is concerned with the effectiveness of the test in

measuring what it claims to measure (Kline, 2000). Cronbach and Meehl’s (1955)

seminal paper on test validity identifies four types: content; construct; predictive, and

concurrent. Content validity is evaluated subjectively, usually through consultation

with subject matter experts, to determine if the scale samples the domain of interest

satisfactorily (Rust & Golombok, 1999). Content validation is said to be particularly

challenging for constructs with debates or inconsistencies in definitions (Murphy &

Davidshofer, 2005). Furthermore, content validity is a state of an obtained scale

score (Messick, 1995) which can vary across populations. Content validity should

therefore be established for the population that will be sampled (Haynes, Richard &

Kubany, 1995). An evaluation of the content validity of existing measures of

empathy should therefore consider not only the complex nature of the concept but

also the context of healthcare practitioners.

The second type of validity identified by Cronbach and Meehl (1955) is construct

validity. Evidence of construct validity is generally accumulated from multiple

studies (Rust & Golombok, 1999). Construct validity is an umbrella term for types of

validity including convergent and discriminant validity (Domino & Domino, 2006).

Convergent validity is said to be established when consistent relationships are found

between the concept of interest and theoretically related concepts. Discriminant

validity is on the other hand established when no consistent relationships are found

with theoretically unrelated concepts (Kline, 2000). Factor analysis is a popular

 44

method of construct validation, having been used extensively in evaluations of

empathy measures (e.g. Carey, Fox & Spraggins, 1988; Pulos, Elison & Lennon,

2004). The third and fourth types of validity, predictive and concurrent, are known as

criterion-oriented validity (Cronbach & Meehl, 1955). In order to assess criterion-

oriented validity, a correlation is calculated between the predictor of interest and an

independent criterion measure. When the scale score and criterion score are

measured at the same time, this is known as concurrent validity, whereas predictive

validity studies involve the criterion score being taken at a later date. For a scale to

be deemed effective, it requires both reliability and validity for the intended purpose,

taking the healthcare context into account (Messick, 1995).

3.2. An overview of empathy scales

These concepts are of particular relevance when considering empathy measures; a

review of existing measures designed for use with the general population identified

only two questionnaires as having generated sufficient research to be able to

demonstrate satisfactory reliability and validity (Chlopan, McCain, Carbonell &

Hagen, 1985). These two measures were by Hogan (1969) and Mehrabian and

Epstein (1972). Other measures described as having inadequate supporting evidence

included, the George Washington Social Intelligence Test (Hunt, 1928; Moss, 1931;

Moss, Hunt, Omwake & Ronning, 1927; Moss, Hunt & Omwake, 1949); the Chapin

Social Insight Test (Chapin, 1942); the Dymond Rating Test of Insight and Empathy

(Dymond, 1949), and the Empathy Test (Kerr & Speroff, 1954). Since Chlopan et

al.’s review, however, the Interpersonal Reactivity Index developed by Davis (1983)

from his multidimensional model of empathy has also received support from
 45

researchers (e.g., Pulos et al., 2004) and this and the questionnaires by Hogan (1969)

and Mehrabian & Epstein (1972) remain in use today.

In addition to measures of empathy for use with general populations, significant

progress has been made in the development of empathy scales for use in healthcare

contexts. Within nursing, scales include the Empathy Construct Rating Scale (ECRS:

La Monica, 1981) and the empathic understanding subscale of the Barrett-Lennard

Relation Inventory (BLRI: 1964, 1978). More recently, the Jefferson Scale of

Physician Empathy (JSPE: Hojat et al., 2002) has been developed for use with

doctors although is now being used within other healthcare professions as well

(Chen, La Lopa & Dang, 2008). Publication of the Reynolds Empathy Scale

(Reynolds, 2000) and the Consultation and Relational Empathy measure (CARE:

Mercer, Maxwell, Heaney & Watt, 2004) have marked significant developments in

efforts to assess empathy based on patient perceptions of healthcare professionals.

The CARE measure is intended to be of use to a range of healthcare professionals,

while Reynolds’ scale is nursing-specific. In order to determine which measures are

likely to be of most use in empirical research, this chapter reviews each of these

eight measures and information is also summarised in tables 3.1 and 3.2.

3.2.1. Empathy Construct Rating Scale (ECRS)

One of the first specific measures of healthcare empathy was La Monica’s (1981)

Empathy Construct Rating Scale (ECRS) for nurses, which was originally developed

to evaluate an empathy training program (La Monica, Carew, Winder, Haase &

Blanchard, 1976). Initially, 259 items were generated by 25 female graduate nursing
 46

students and then reviewed by three experts skilled in psychometrics, psychology

and nursing respectively. These items were reduced to the final 100 via an item

facility analysis using three expert and 10 student ratings. To calculate reliability, La

Monica (1981) asked 103 nursing students to rate two colleagues, one high (Form A)

and one low (Form B) in empathy. Both forms were found to have high Cronbach’s

α coefficients and split-half reliabilities. However, no reliability statistics for the self

rating version were reported. Sample sizes are small in that experts recommend at

least 200 people for an item and reliability analysis (Rusk & Golombok, 1999).

Three hundred female nurses and nursing students then completed a battery of scales

in order to assess convergent and construct validity. The five subscales hypothesised

to be present within the ECRS included: non verbal behaviour; personality traits such

as openness and honesty; sensitivity to others; responding in ways such as

encouraging and supporting, and finally respect for self and others. However these

five subscales were not supported by a factor analysis. In analysis of the self ratings

of the 300 nurses and students, all five subscales loaded on to a single factor.

Analysis of self, patient and peer ratings concluded that the 100 items loaded on to

seven factors, although 84 of the items loaded significantly on to the first two

factors, one being positive and one negative indicators of empathy. Subsequent

studies using the ECRS have used these 84 items rather than the original 100 (e.g.,

Layton & Wykle, 1990).

With respect to construct validity, self and peer ratings showed a correlation of only

.20 (p<.001) while self and patient ratings were found to have an even lower

correlation of .10 (p<.05). In addition, only a small correlation (r = .16, p<.05)

 47

between self ratings and the Chapin Social Insight Test (Chapin, 1942) was found,

suggesting concerns over construct and convergent validity of the test. Criterion

related validity was also called into question when La Monica (1987) and Reynolds

(1986) failed to find significant changes in patient or self reported empathy

following empathy training.

Methodological concerns therefore include a lack of patient input into scale

development as well as failure to establish sound construct, convergent or criterion

related validity. Theoretically, the content of the scale does not appear to relate

clearly to any models of empathy discussed in chapter two, suggesting that there is

insufficient evidence to justify its continuing use (Reynolds, 2000).

3.2.2. The Barrett-Lennard Relation Inventory (BLRI)

Although an older measure than La Monica’s ECRS, some researchers have argued

that the strong theoretical underpinning of the Barrett-Lennard Relationship

Inventory (BLRI: Barrett- Lennard, 1964, 1978) makes it suitable for use in applied

settings beyond the counselling relationship for which it was originally developed

(Layton & Wykle, 1990). The most widely used form of the BLRI relies on the

patient describing the healthcare practitioner (the ‘other towards self’ or OS version).

There is also a ‘myself towards other’ (MO) version which is effectively a self rating

by the practitioner of their effectiveness in a specific relationship. The 64-item scale

has four subscales pertinent to effective interpersonal relationships, including:

empathic understanding; congruence; level of regard, and unconditionality of regard.

The first of these, empathic understanding, is of particular interest here. Several

 48

studies have employed the 16-item empathic understanding subscale as a measure of

empathy in its own right (e.g., Layton & Wykle, 1990) and the subscale has also

been adapted for use within medicine to assess the physician-patient relationship

(Simmons, Roberge, Kendrick & Richards, 1995). The empathic understanding

subscale of the BLRI is based on Barrett-Lennard’s cyclical model of empathy

described in chapter two. According to this model the empathy cycle starts with the

empathiser in an attentional set which allows them to pick up cues from the target

person. The cues are then perceived and empathy is expressed. The target person

receives this expression and responds in turn, thus the cycle continues. The OS

empathic understanding subscale taps into the empathy cycle at phase three, where

the patient receives empathy and judges the motivation and understanding of the

empathiser.

From development, the rationale of the BLRI has been "the logical presumption that

it is what the patient… himself [sic] experiences that affects him most directly"

(Barrett-Lennard, 1962, p. 2). In turn, this patient experience should be most closely

related to positive healthcare outcomes. Evidence of reliability and validity is

presented in Table 3.2. Numerous studies have reported results in which the OS

empathic understanding measure based on patient perceptions has successfully

predicted positive outcomes. Gurman and Razin (1977) for example reviewed 20

studies of actual help-seeking patients. In only one of these studies did the patient’s

perception of the relationship, including empathy, fail to predict therapeutic

outcomes.
 49

As noted by Reynolds (2000), however, the scale may have limited success when

applied to the less formal nurse-patient relationship. It has been found that when

used in conditions other than the counselling context for which it was designed, the

reliability of the BLRI is reduced (e.g. Polit & Hungler, 1983). In addition, content

validity was established by experts in patient-centred counselling (Gurman & Razin,

1977) rather than more general healthcare professions. According to Bennett (1995),

the range of issues discussed by nurses and patients is more diverse than the

therapist-patient interaction, and the content validity of the measure for the role of a

non-psychiatric nurse or doctor may therefore be limited. Whilst the BLRI appears to

capture an important aspect of empathy and is effective at including the patient’s

perspective, measures capable of capturing different aspects of empathy with items

more suited to general healthcare roles are needed (Bennett, 1995).

3.2.3. Jefferson Scale of Physician Empathy (JSPE)

The most commonly used measure for assessing practitioner empathy reported in the

medical literature is currently the Jefferson Scale of Physician Empathy (JSPE:

Hojat, Gonnella, Nasca, Mangione, Vergare & Magee, 2002). The authors began the

process of scale construction by reviewing the literature on general and clinical

empathy, with the authors concluding that empathy in medicine should be viewed as

a cognitive process. Qualitative research with subject matter experts (100 physicians)

was then conducted prior to scale construction, as recommended by psychometrics

experts (e.g., Rust & Golombok, 1999). Items for the scale were generated by these

experts using the definition of empathy as “An uncritical understanding of the

patient’s experiences, emotions, and feelings” (Hojat et al., 2002, p. 1563). This
 50

definition was explicitly contrasted with sympathy, which was defined as “feeling

with the patient or feeling similar emotions that the patient feels” (p. 1563). The

items were therefore immediately focused towards a cognitive view of empathy, and

away from sympathy, which was identified as a possible obstacle to objective

diagnosis and treatment (Hojat et al., 2002). The JSPE scale is widely used across

various healthcare roles apart from medicine, including nursing and pharmacy

(Chen, La Lopa & Dang, 2008; Fields, Hojat, Gonnella, Mangione, Kane & Magee,

2004). It has been found to have satisfactory psychometric properties, with high

reliability for samples of 56 nurses and 42 physicians (Fields et al., 2004). Tables 3.1

and 3.2 provide more information on the JSPE and its properties.

Despite the psychometric robustness of the JSPE it has theoretical limitations in that

it does not account for multidimensional components of empathy in recent models

described in chapter two. Notably, items lack assessment of the behavioural

dimension of empathy. In evaluating empathy training, Evans et al. (1993) found

that communicated empathy may be changeable whereas cognitive or affective

empathy is more stable over time (Feighny Arnold, Monaco, Munro & Earl, 1998),

suggesting a need to include a behavioural element in the assessment of empathy.

Of particular interest is the fact that although models of empathy emphasise the need

to focus on communication of empathy, patients were not typically involved in the

development of the empathy measures reviewed so far. In an attempt to address this

Kane, Gotto, Mangione, West and Hojat (2007) developed a five item patient version

of the JSPE, which is broadly in line with Davis’ multidimensional model of

empathy, and asks the patient to rate the physician’s empathic concern, perspective
 51

taking and behaviour. Analysis of data from 225 patients resulted in a single factor

with a Cronbach’s α of .58, which was deemed to be satisfactory given the small

number of items in the measure. However, the patient scale appears to suffer from a

ceiling effect as the mean rating was 23.8 while the maximum possible score is 25.

The patient scale may therefore not be useful in discriminating between physicians,

perhaps due to social desirability or that patients are unable to judge the internal

empathic processes of the physician.

In summary, although the JSPE is a popular measure of empathy, the definition used

to develop the scale is not multidimensional and neither version of the scale included

patient input to thoroughly understand the behaviours which communicate empathy

most effectively. There has been some progress in addressing these limitations in

other recent scales, which are now reviewed.

3.2.4. The Reynolds Empathy Scale

This scale was developed to ‘demystify’ the process of empathy between nurse and

patient, and allow nurses to better understand how to apply skills and appear

empathic in terms of behaviour (Reynolds, 2000). In developing a new nursing-

specific measure, Reynolds’ aim was to make progress in the measurement of

empathy in two respects, both of which were highlighted by a review of the

effectiveness of existing measures. First, he argued that a measure was needed that

the user could be confident of in terms of reliability and validity. Secondly, that an

effective measure should include patient perceptions in the development process,

something which other measures had typically failed to take into account and was
 52

considered a major weakness in light of efforts to include patients as active and

important collaborators in treatment (e.g., Barker, 1994; NHS,1996; The Scottish

Office, 1997).

In order to develop the scale, 30 patients were asked for their perceptions of effective

and ineffective interpersonal behaviours demonstrating empathy. These perceptions

were gathered from explanations of ratings of student nurses on La Monica’s ECRS

(Reynolds, 1986). Rather than use a method of qualitative inquiry purely focusing on

matters of importance to patients, the comments were therefore very much

influenced by the content of the ECRS. Comments were categorised into helpful or

unhelpful behaviours along four themes: creating an interpersonal climate where it is

possible to become aware of the patient’s emotions; listening to the concerns of the

patient; using a range of strategies to help the patient, and awareness of how the

process may be stalled by a failure to listen or the presence of unhelpful behaviour.

To be included in the scale, the item must have been mentioned by at least twenty of

the thirty patients and also be relevant to empathy as judged by reports in previous

literature. Twelve items resulted on the scale, six positive and six negative. Further

information can be found in Table 3.1.

Although the rationale for including patient perceptions in developing a measure of

empathy is sound (particularly for the behavioural dimension where patients are in

receipt of communicated behaviours), several criticisms can be levelled at the

approach taken by Reynolds. First, the qualitative data gathered from patients was

limited due to a focus on a previously constructed scale rather than the patients’ own

conceptualisations of empathy. Secondly, the actual words used by patients were not
 53

included in the items themselves, which potentially resulted in a loss of meaning. An

example is given to illustrate this point. Item 1 on the scale is ‘Attempts to explore

and clarify feelings’ whereas item 5 is ‘Explores personal meaning of feelings’. An

example of a patient explanation of item 5 is given as “When I don’t want to talk

about something she seems to recognise this mood and asks me about it. She won’t

persist if I am reluctant” (Reynolds, 2000, p. 56). However one could argue that this

statement reflects both items 1 and 5. Potentially, in moving from patient statements

to item generation, the researcher’s own interpretation may have altered the items

thus negating the argument to put the patient perceptions at the centre of the scale.

Indeed, content validation was checked in consultation with five experts from

nursing and clinical psychology rather than referring back to patients. Finally, in

focusing on behaviours rather than a multidimensional approach including cognitive

and affective processes involved with empathy, the scale again fails to incorporate a

multidimensional model of empathic processes. Reynolds’ scale is relatively new

and as yet little evidence is available of its usefulness, however, his central theme of

including patient perceptions is one which is growing in popularity and has also led

to the development of the CARE measure which is now introduced.

3.2.5. The CARE Measure

The CARE measure was developed on the basis of arguments that patients’ views

are central to the effectiveness of empathy in the clinical encounter (Mercer et al.,

2004). The authors also intended to build on the work of Reynolds by creating a

measure of empathy which could be relevant for clinical encounters other than

nursing. Their initial measure was piloted using a sample of general practitioners and
 54

patients, and qualitative and quantitative examinations of validity allowed for

appropriate revision of the CARE measure until the third version of the scale was

deemed satisfactory (details are provided in Tables 3.1 and 3.2).

Although sample sizes were small in testing the two pilot versions of the scale (20

patients in first pilot followed by 13 in the second), a key stage in ensuring that the

final scale represented patients’ perceptions was to validate the scale with patient

consultation. To further validate the CARE measure, 3044 patients then completed

the final version of the questionnaire for 26 GPs from different practices (Mercer,

McConnachie, Maxwell, Heaney & Watt, 2005). In total 76% of patients rated the

items within the measure as being very important to their consultation with the

doctor. In particular, patients with psychological or emotional distress and older

patients rated the items as more important. This suggests that the measure is a good

representation of the behavioural dimension of empathy within the clinical encounter

as it might reasonably be assumed that these patients would be even more in need of

empathy from their physician. Further analysis revealed that acceptable reliability of

ratings for a physician could be reached with a minimum of 50 patient ratings

(Mercer et al., 2005), and norms were established such that within the scoring range

of 10-50, a score of less than 39 represented a below average score while above 42

represented above average. This final stage of analysis has resulted in the measure

being adopted for use in appraisal of GPs practicing in Scotland. The measure

appears to demonstrate sound psychometric properties and development included

patient consultation, but it again focuses on the behavioural dimension (perhaps not

surprisingly as it was intended for use by patients appraising medical professionals)

rather than a multidimensional model of empathy.

 55

The final three measures of empathy reviewed here were developed for use with

general populations rather than the more specific healthcare context.

3.2.6. Hogan’s Empathy Scale

Hogan’s Empathy Scale was developed based on the dictionary definition of

empathy as: “the intellectual or imaginative apprehension of another’s condition

without actually experiencing that person’s feelings” (Hogan, 1969, p.308). It is

therefore rooted in an exclusively cognitive definition of empathy. Importantly,

Hogan argued that the definition of empathy involves apprehension of another’s

condition that does not imply (or therefore require) accuracy, consequently a self

report questionnaire is a suitable method of measurement. This definition was given

to nine psychologists and 14 other people who were asked independently to use

California Q-sort items (Block, 1961) to describe a highly empathic person. A high

degree of agreement in the items chosen by the psychologists and non-psychologists

suggests that this definition is understandable to people outside the realm of

psychology. Hogan therefore used it to develop a scale comprised of 100 items that

represented characteristics most and least representative of an empathic person.

Two samples of participants (100 military officers and 111 scientists/engineers)

completed this new scale alongside other personality scales. Reliability and validity

analyses resulted in 64 items being retained in the final scale. The choice of sample

for scale construction is an interesting one, as it could be argued that military officers

and those in scientific careers might not be the most representative on which to base
 56

a measure of empathy which may be used in the general population. In addition, no

information was presented in terms of gender of the samples. Perhaps not

surprisingly, Chlopan et al. (1985) note there is evidence to suggest that this scale is

more valid for males than females, possibly resulting from the participants being

drawn from traditionally masculine careers. Information regarding reliability and

validity is presented in Table 3.2. Of some concern, is that in an investigation of the

scale with a population of healthcare professionals (occupational therapy students)

internal consistency was found to be only .57 while test-retest reliability was only

.41 over 12 months (Froman & Peloquin, 2001). Using a restricted definition may

have limited the success of the Hogan empathy scale as a measure as reliability has

not been consistently established, particularly among healthcare professionals.

Although evidence regarding validity may be viewed positively, particularly for

males, concerns about using a purely cognitive definition remain and inclusion of an

affective component might be beneficial.

3.2.7. The Questionnaire Measure of Emotional Empathy

Mehrabian and Epstein (1972) addressed this need in the development of the

Questionnaire Measure of Emotional Empathy (QMEE). They argued that empathy

also contains an emotional response for which there was no adequate instrument, but

whilst Hogan’s scale can be criticised for being purely cognitive in nature, the

QMEE is purely emotional. The QMEE includes seven subscales: susceptibility to

emotional contagion; appreciation of the feelings of unfamiliar and distant others;

extreme emotional responsiveness; tendency to be moved by others’ positive

emotional experiences; tendency to be moved by others’ negative emotional

 57

experiences; sympathetic tendency, and willingness to be in contact with others who

have problems. Very little information is published about how items were

constructed, or the psychometric properties of the scale and the sample of individuals

used to develop it. However, the 33 final items were apparently selected as a result

of item analysis, validity analysis and a check of socially desirable responding. Each

item therefore had a significant item-total correlation; loaded on to a factor derived

from the data and showed no significant correlation with a measure of social

desirability. Information regarding reliability and validity is presented in Table 3.2.

Chlopan et al. (1985) concluded that of six measures reviewed, only the QMEE and

Hogan’s Empathy Scale had sufficient research to support their use. This is despite

the fact that theoretically the two purport to measure different components of

empathy, illustrated by low correlations between the two measures (Davis, 1983).

3.2.8. The Interpersonal Reactivity Index

Recognising a need for a multidimensional scale of empathy, Davis developed the

Interpersonal Reactivity Index (IRI; Davis, 1983) based on his multidimensional

model of the process of empathy. This has four distinct subscales and is based on the

rationale that “empathy can best be considered as a set of constructs, related in that

they all concern responsivity to others but are also clearly discriminable from each

other”(Davis, 1983, p.113). The four constructs measured by the scale are

perspective taking, empathic concern, fantasy and personal distress. Table 3.1

contains basic information regarding each of these subscales. Briefly, the Perspective

Taking subscale is representative of an individual’s tendency to adopt the views of

another. The Fantasy subscale is similar to this, although is based on imagining

 58

oneself in the role of characters in books, films or plays. The Empathic Concern

subscale asks about the individual’s own feelings of concern in response to another

person. The Personal Distress subscale is also emotional, but focuses on how much

one feels distress in response to another. According to Davis, while these four

subscales do not exhaust the possible range of reactions to others, previous theory

and research suggests that they reflect the variety of reactions to another that have at

some point been referred to as empathy.

To establish validity of the IRI and four subscales, Davis investigated relationships

between these and other psychological constructs. Findings from construct and

convergent validity studies provide support for the theoretical and psychometric

properties of the scale, with details of these investigations provided in Table 3.2.

Furthermore, given the theoretical accounts that the four subscales are related (Coke

et al., 1978; Hoffman, 1977), Davis argued that correlations between the subscales

would also provide evidence for the validity of the scale as a whole.

Although the IRI has proved a popular, reliable and valid instrument, based on a

multidimensional theory that includes both emotional and cognitive components, it

was not designed specifically for the healthcare context. However, Yarnold, Bryant,

Nightingale and Martin (1996), in a study of 114 physicians and 95 medical students,

found the IRI to have good structural integrity and convergent validity with a

measure of problem solving in the context of concern for the well being of others.

Evans, Stanley and Burrows (1993) also undertook a study using the IRI to assess

empathy in 55 medical students during clinical training. After completing the

measure students’ behaviours in a twenty minute consultation were scored, using

 59

five items from the 16-item History-taking Interview Rating Scale (Verby et al.,

1979) that assess behaviours relevant to empathy such as: eye contact; use of jargon;

tendency to interrupt patients; coverage of psychosocial and personal issues; warmth,

and ability to detect leads in what the patient is saying. Scores on the IRI and the five

items measuring empathic behaviours were positively correlated, suggesting that the

measure has the potential for use specifically with healthcare professionals.

3.3. Summary of empathy measures

Eight measures of empathy have been reviewed in this chapter; five designed

specifically for healthcare professionals and three for the general population. The

five healthcare measures were: the ECRS, the empathy subscale of the BLRI, the

JSPE, the Reynolds Empathy Scale, and the CARE measure. Questions arose over

the reliability of the ECRS and BLRI, particularly as the BLRI was designed in a

counselling context and use outside this setting led to reduced reliability (Polit &

Hungler, 1983). The JSPE is limited by its basis in a pure cognitive definition of

empathy and failure to accommodate emotional aspects or how empathy is

communicated to patients. Whilst the more recent scales by Reynolds (2000) and

Mercer et al. (2004) have attempted to include a patient perspective, both suffer from

the lack of a clear theoretical basis.

 60

Table 3.1: Descriptions of existing measures of empathy

Target Model and definition Number of items and Response format and scoring
Population of empathy sample items

Empathy Construct Rating Scale (ECRS: La Monica, 1981)

Nurses Cognitive/Behavioural: ‘Signifies a central focus and 100 item (e.g., ‘Listens carefully’, Self, peer and patient versions
feeling with and in the patient’s world. It involves ‘Checks to see if understanding of 6 point scale ‘extremely like’ to
accurate perception of the patient’s world by the helper, another’s experience is valid’) 20 ‘extremely unlike’
communication of this understanding to the patient, and item short version (La Monica, 1996) 46 negatively and 54 positively
the patient’s perception of the helper’s understanding’. worded items

Empathic understanding sub-test of the Barrett-Lennard Relation Inventory (BLRI: Barrett-Lennard, 1964)

Counselors Multidimensional: ‘To perceive the internal frame of 16 items (e.g., Usually senses and Self and patient versions
reference of another with accuracy, and with the realise what I am feeling - patient) 7 point scale ‘strong agreement’ to
emotional components...as if one were the other person ‘strong disagreement’
but without ever losing the ‘as if’ condition’ (Rogers, Half items negatively worded
1957). Extended to communicative aspects of empathy.

Jefferson Scale of Physician Empathy (JSPE: Hojat et al., 2002)

Medics Cognitive: ‘An uncritical understanding of the patient’s 20 items (e.g., I try to understand Self and patient versions
/general experiences, emotions, and feelings’. what is going on in my patients’ 7 point scale ‘strongly agree’ to
healthcare minds by paying attention to their ‘strongly disagree’
roles non-verbal cues and body language) Half items negatively worded
 61

Reynolds Empathy Scale (RES: Reynolds, 2000)

Nursing Cognitive /Behavioural: ‘Signifies a central focus and 12 items (e.g., ‘Attempts to explore Self, peer and patient versions
feeling with and in the patient’s world...’ from La Monica and clarify feelings’) 7 point scale. from ‘never like’ to
(1981). ‘always like’. Half items negatively
worded, reverse scored.
CARE Measure (CARE: Mercer et al., 2004)

General Cognitive / Behavioural: ‘Ability to: understand the 10 items (e.g., How was the doctor Patient ratings 5 point scale
healthcarepatient’s situation, perspective and feelings; to at… Making you feel at ease?) ‘poor’ to excellent’ plus ‘Does not
roles communicate that understanding and check its accuracy, apply’ No negatively worded items
and to act on that understanding with the patient in a
helpful way’.
Hogan Empathy Scale (HES: Hogan, 1969)

General Cognitive: The act of constructing for oneself another’s

64 items (e.g., I am usually rather True or false, negatively worded
Population mental state short-tempered with people who items reverse scored. Total summed
come around and bother me with score
foolish questions (-)
Questionnaire Measure of Emotional Empathy (QMEE: Mehrabian & Epstein, 1972)

General Emotional: An involuntary vicarious 33 items (e.g., The people around me 9 point scale +4 (very strong
Population experience of another’s emotional state have a great influence on my moods) agreement) to -4 (very strong
disagreement). Total summed score
Interpersonal Reactivity Index (IRI: Davis, 1983)
General Multidimensional. 4 subscales: Perspective Taking, 28 items , 7 per scale (e.g., I 5 point scale from 0 (does not
Population Fantasy, Empathic Concern, Personal Distress sometimes try to understand my describe me well) to 4 (describes me
friends better by imagining how very well)
things look from their perspective)
 62

Table 3.2: Reliability and validity information for existing empathy measures

Reliability Validity
ECRS (La Monica, 1981)
Cronbach’s α .97 for Form A, .98 for Form B. Construct: expected subscales not confirmed by factor analysis: single factor for self ratings (La
Split half reliability .89 for Form A, .96 for Monica, 1981). Convergent: large correlation with BLRI empathic understanding (r = .78, p<.001,
Form B. Peer ratings only, reliability of self Layton & Wykle, 1990); only small to moderate relationships between self, peer and patient ratings,
ratings not reported (La Monica, 1981). or self ratings and Chapin Social Insight Test (La Monica, 1981). Criterion: no significant changes
reported following empathy training (La Monica, 1981).

BLRI Empathic understanding sub-test (Barrett-Lennard, 1964, 1978)

Cronbach’s α from .64 - .92. Split half reliability Construct: factor analysis of 64 items yielded 3 factors, with items from the empathic understanding
.86. Test retest reliability from .66 - .91 (Gurman sub-test loading on two factors, acceptance of another person’s separateness (being open and non-
& Razin, 1977). Reliabilities reduced when used judgmental) and psychological insight (sensitivity and understanding), in line with pre-empathy
in contexts other than counselling (Polit & conditions and phase one empathy. Convergent: large correlation with ECRS (r = .78, p<.001, Layton
Hungler, 1983). & Wykle, 1990). Criterion: perceptions of empathic understanding predicted therapeutic change in
19/20 studies of patients in counselling interactions (Gurman & Razin, 1977).

Jefferson Scale of Physician Empathy (Hojat et al., 2001)

Cronbach’s α .87 and .89 for nurses and Predictive: Supervisor ratings significantly higher for top scorers on the measure than bottom scorers,
physicians respectively (Fields, Hojat, Gonnella, with a large effect size (.50) (Hojat, et al, 2005). Construct: Factor analysis found the major construct
Mangione, Kane & Magee, 2004). Cronbach’s α to be the physician’s view of the world from the patient’s perspective with a second significant factor
.58 for five item patient version (Kane, Gotto, defined as understanding the patients’ experiences, feelings and clues, both in line with the cognitive
Mangione, West & Hojat, 2007). definition adopted (Hojat et al., 2001). Concurrent: Students rated higher on clinical competence
were significantly higher scores on the JSPE, regardless of specialty. (n= 371, specialties included
family medicine, internal medicine, obstetrics/gynaecology, paediatrics, psychiatry and surgery
(Hojat et al., 2002). Divergent: Scale unrelated to a test of objective knowledge (Medical College
Admissions Test: Hojat et al., 2002)
 63

Reynolds Empathy Scale (Reynolds, 2000)

Cronbach’s α .90. Test-retest correlation .90 over Convergent: Reynolds Scale and La Monica’s ECRS positively correlated (n = 34, r = .84, p<.001).
4 weeks. No factor analysis reported.
Inter-rater: percentage agreement between 3
raters rose from 27.6% to 72.2% after training
(Reynolds, 2000).
CARE Measure (Mercer et al., 2004)
Cronbach’s α .93 (Mercer et al., 2004) Convergent: positive relationships with the BLRI (r = .85, p<.001) and with the Reynolds scale was
(.84, p<.001), however these correlations are Pearson’s r despite being based on sample sizes of just
10 patient ratings.
Hogan Empathy Scale (Hogan, 1969)
Test–retest reliability ranging from .60 to .84 Convergent: highest scores positively related to self awareness and social insight test scores (Mill &
(Cross & Sharpley 1982) Hogan, 1978), greater empathy is associated with lower anxiety (Hogan, 1969, Spielberger, Gorsuch
& Lushene, 1970). Predictive: highest scores associated with more positive prognoses from clinical
psychologists for their patients (Dubnicki, 1977). Highest scores of therapists associated with greater
improvement in hyperactive child patients (Kendall & Wilcox, 1980)
Questionnaire Measure of Emotional Empathy (Mehrabian & Epstein, 1972)
Split half reliability .94 (Mehrabian & Epstein, Construct: Females score higher than males: Males mean score = 23, SD = 22; females mean score =
1972) 44, SD = 21 (Mehrabian & Epstein, 1972). Divergent and convergent: scores correlate negatively
with aggressive behaviour and positively with helping behaviour (Mehrabian & Epstein, 1972). High
scores associated with high neuroticism measures (Eysenck & Eysenck, 1978).
Interpersonal Reactivity Index (Davis, 1983)
Internal consistency Cronbach’s α ranges from Divergent & Convergent: PT sub-scale correlates positively with the Hogan empathy scale (0.40) and
.71 to.77 for each subscale (Davis, 1983). Test – negatively with tests of social-dysfunction. EC sub-scale correlates positively with tests of sensitivity
retest ranges from.62 to .71 (Davis, 1983) to others and positively with the QMEE (0.60) (Davis 1983). Concurrent: IRI PT, EC and five-item
measure of empathy behaviours positively correlated for medical students (Evans et al., 1993).
Construct: Factor structure confirmed in medical students (Yarnold et al., 1996)
 64

Broadly therefore, none of the specific healthcare measures possess both satisfactory

theoretical underpinning and a robust method of scale construction.

Similarly, among the general population measures the Hogan Empathy Scale and

QMEE fail to accommodate the multidimensional conceptualisation of empathy

embraced by Davis’ IRI (1983). Research has found evidence of satisfactory

psychometric properties for the IRI and Evans et al. (1993) used the IRI measure

successfully with healthcare professionals, finding self-reported empathy to correlate

positively with observed behaviour. Therefore the measure that appears to escape

most criticism on both theoretical and psychometric grounds is the Interpersonal

Reactivity Index (IRI: Davis, 1983). More generally, the IRI questionnaire also has

benefits associated with a psychometric approach to data collection (Rust &

Golombok, 1999) in beings useful for theory and hypothesis testing, and convenient

to use due to its brevity and usability for the participant (Coolican, 2009). Based on

this review of available measures, the IRI is therefore chosen as the main focus for

investigation of empathy in this thesis.

3.4. Overview of Thesis and Research Questions

To summarise thus far, chapter one concluded by arguing that in order to identify the

best methods to develop and sustain empathy in healthcare professionals there is a

need for greater understanding of the antecedents and processes involved in

empathic responding towards patients. Chapter two reviewed models of empathy,

identifying Davis’ multidimensional model (1983; 1996) as the most comprehensive,

with Larson and Yao’s (2005) adaptation of this model emphasising its potential
 65

utility for the healthcare context. To date very little research has tested either model

in the context of empathy between patients and healthcare professionals. As such,

very little is known about the individual characteristics that might impact on how

empathic behaviour is demonstrated, or indeed, what behaviours impact on

judgments of empathy made by patients for healthcare professionals. In order to

address these questions, this chapter reviewed a range of existing measures and

identified the Interpersonal Reactivity Index as a suitable measure to investigate

empathy in the healthcare context, as it possesses both strong theoretical

underpinnings and psychometric properties. The specific research questions

addressed in this thesis are as follows:

1. What are the antecedents of empathy in healthcare practitioners, in terms of

individual differences and situational factors?

2. What behaviours are associated with empathy in the healthcare practitioner?

This thesis aims to further understanding of the antecedents and behaviours involved

in empathic interactions between patients and healthcare professionals. The

empirical studies utilise the IRI in an attempt to answer these questions. Studies one

and two begin by investigating the individual differences that are the antecedents of

empathy. Study one is a cross sectional, quantitative research study employing a

general population sample (n = 105). It utilises the IRI and the NEO Personality

Inventory (revised) [NEO PI-R] in order to investigate empathic disposition within

the framework of the five factor model of personality (Costa & McCrae, 1992).

Study two is a cross sectional, quantitative study using a sample composed of

medical professionals and students (n = 297). The study examines the link between

individual differences in emotional intelligence and empathy, using the IRI and Bar-
 66

On EQ-i. Study three investigates the relationships between emotional intelligence,

propensity to empathise and observed empathic behaviour. The study uses a cross

sectional, quantitative methodology to test the relationships between the IRI, EQ-i

and ratings of behaviour by assessors and simulated patients. Finally, study four

examines the antecedents of empathy in the healthcare context and identifies the

behaviours associated with empathy as judged by patients. In order to understand the

real experiences of patients currently in a hospital setting, study four involves a

qualitative research design, using semi structured interviews with twenty patients.

The studies are outlined in Figure 3.1.

 67

Study One
Individual differences in personality associated with a propensity to empathise
IRI and NEO PI-R, n=105, general population sample

Study Two
Individual differences in EI associated with a propensity to empathise
IRI and Bar-On EQ-in=297, general population and medical sample

Intrapersonal empathy
Empathic behaviour
Antecedents of empathy Reactive emotions
Helping behaviour
Practitioner characteristics (compassion and sympathy,
Social behaviour
Situational characteristics personal distress)
Conflict management
Perspective Taking

Study Three
Relationships between individual differences in propensity to empathise end empathic behaviour
IRI, EQ-i and ratings of behaviour (n=192, medical sample)

Study Four (a) Study Four (b)

Patient perspectives of situational factors Patient perspectives of empathic
affecting empathy in nurses behaviours
Semi-structured interviews (n = 20) Semi-structured interviews (n = 20)

Fig 3.1: Framework for investigating empathy addressed by the studies within this thesis
 68

Chapter 4 – Individual differences and empathy:

Personality and the IRI

Of all the empathy measures reviewed in Chapter two, Davis’ Interpersonal

Reactivity Index (IRI: 1983), with its strong theoretical background and good

psychometric properties, presents as the best option for investigating individual

differences associated with empathic processes. However, the Interpersonal

Reactivity Index is now almost thirty years old and evidence of construct validity

that could support its use in the healthcare setting is limited, particularly in light of

more recent developments in the understanding and measurement of personality

(Chamorro-Premuzic, 2007). In fact, there has been very little research

investigating the subcomponents of the IRI (perspective taking, empathic concern,

personal distress, and fantasy) in relation to more recent personality frameworks,

including the five factor model (FFM) of personality. The aim of study one was

therefore to investigate whether personality factors might be associated with

antecedents of empathy and, in doing so, to further investigate construct validity for

the IRI in terms of a broad spectrum five factor measure of personality. Before

reviewing previous evidence linking empathy to the FFM, the concept itself is

introduced.

4.1. The Five Factor Model (FFM) of Personality

Personality traits have been defined as “an individual’s characteristic pattern of

thought, emotion and behaviour” (Funder, 1997, pp1-2) that are relatively stable

across time and situations. Traits can be quantitatively assessed, using measures of
 69

personality that typically rely on self perceptions (McCrae & Costa, 1999). A

significant body of evidence now exists that supports claims that personality

descriptors can be accounted for almost completely by five robust factors (Digman,

1990). The five factors are: extraversion (or surgency); neuroticism (or emotional

stability); openness to experience (or intellect); agreeableness, and

conscientiousness. McCrae and Costa’s five factor theory, developed from

empirical findings, suggests that traits are organised hierarchically with these five

factors representing the highest level of the hierarchy (1999). Despite the recent

dominance of the five factor model (McCrae & Costa, 1999), however, most

existing research concerned with relationships between personality and empathy

has been based on Eysenck’s three factor model of personality comprised of

extraversion, neuroticism and psychoticism (e.g. Mehrabian, Young & Sato, 1988).

Evidence of a relationship between empathy and personality is derived from three

main sources: studies of emotional empathy, a small number of studies looking at

cognitive empathy, and Davis’ (1983) own exploration of his multidimensional

model of empathy. All three sources have used correlational analyses to investigate

relationships with three of the five factors of personality: neuroticism, extraversion

and agreeableness.

According to the five factor model, neuroticism is broadly concerned with an

individual’s propensity to experience negative emotions such as sadness, fear and

anger (Costa & McCrae, 1999). As such, studies examining links between empathy

and neuroticism have tended to focus on the emotional components of empathy.

This research has found that emotional empathy is positively related to trait anxiety.
 70

For example, in a study of 250 adult participants using the Questionnaire Measure

of Emotional Empathy (QMEE: Mehrabian & Epstein, 1972) emotional empathy

was found to have moderate positive correlations with neuroticism (r = .42, p<.05;

Rim, 1994). Similar findings were later reported by Eysenck and Eysenck (1978; r

= .35, p<.05) in a sample of adults and by Eysenck & McGurk (1980; r = .40,

p<.05) in a study of male adolescent delinquents.

In the IRI there are two subscales associated with emotional empathy: empathic

concern and personal distress. Empathic concern relates to feelings of sympathy

and concern for other people while personal distress refers to a tendency to

experience anxiety and unease in response to others’ distress. In Davis’ own initial

validation of the IRI (1983), several measures were used that relate to neuroticism,

including the public self consciousness and social anxiety subscales of the Self

Consciousness Scale (Fenigstein et al., 1975) and the Fearfulness scale of the

Emotionality, Activity, Sociability & Impulsivity (EASI) temperament measure

(Buss & Plomin, 1975). For both males and females, the two emotional subscales of

the IRI, personal distress and empathic concern, were found to be positively related

to these measures. This was explained by suggesting that emotional empathy and

trait anxiety both result from a higher level of emotional arousability (Eysenck,

1990). These results suggest that neuroticism will be positively correlated with the

two emotional subscales of the IRI, personal distress (Hypothesis 1a) and empathic

concern (Hypothesis 1b).

Agreeableness is the personality factor concerned with the nature of one’s

interpersonal interactions (Costa & McCrae, 1999). According to Graziano,

Habashi, Sheese & Tobin (2007) descriptors for the domain of agreeableness most
 71

closely resemble those associated with empathy, because they include words such

as sympathetic, forgiving and helpful. Graziano and Eisenberg (1997) have also

suggested that agreeableness is associated with prosocial behaviour, which is

typically viewed as an outcome of empathy (Eisenberg & Miller, 1987). Despite

this apparent link, evidence regarding relationships between the different

components of empathy and the factor of agreeableness is limited to Davis’ original

validation of the IRI (Davis, 1983). The four subscales of the IRI were examined in

relation to the F Scale of the Extended Personal Attributes Questionnaire (EPAQ:

Spence, Helmreich & Holohan, 1979) which indicates sensitivity to others’

feelings. Davis also used the Masculinity Scale of the EPAQ, which contains

descriptors such as arrogant and boastful that are negatively associated with

agreeableness. Both were found to be significantly related to the Perspective Taking

and Empathic Concern subscales of the IRI. Based on these initial findings it was

further hypothesised that perspective taking as measured by the IRI would be

positively associated with agreeableness, such that individuals who rate themselves

high on agreeableness will also rate themselves high on perspective taking

(Hypothesis 2a). Similarly, empathic concern was predicted to correlate positively

with agreeableness (Hypothesis 2b).

The third personality factor that has been found to relate to interpersonal

interactions is extraversion (e.g. Costa & McCrae, 1992). As extraversion is

concerned with the degree to which an individual seeks out social contact rather

than the nature of specific interactions (Costa & McCrae, 1999), individuals who

seek out social contact may also show a greater tendency to empathise with others.
 72

Findings from research investigating the relationship between emotional empathy

and extraversion have been mixed, however. For example, in a study using 250

participants, Rim (1994) found a small significant negative correlation between

emotional empathy measured by the QMEE and extraversion measured using

Eysenck’s EPQ, while Klis (1997) used the same measures with a sample of

teachers and found no significant correlation. However, the sample size of this

study was considerably smaller at just 79 participants. In addition, Mehrabian and

O’Reilly (1980) found a small significant positive correlation between emotional

empathy and extraversion (r = .25, p<.05) based on a sample of 211 psychology

undergraduates. Davis’ validation of the IRI also used several scales relating to the

domain of extraversion. These included shyness (Cheek and Buss, 1981), the

Masculinity Scale of the EPAQ (Spence et al., 1979) which contains descriptors

such as arrogant and boastful, and a scale of Extraversion composed of six items

from the Self Monitoring scale (Briggs, Cheek and Buss, 1980). The emotional

subscale of personal distress from the IRI was found to relate positively to shyness

and negatively to masculinity and extraversion. Apart from this, no consistent

findings relating to these scales and any of the other subscales of the IRI were

found for both males and females. Although less extensively studied, the picture is

similar for cognitive empathy and extraversion. Even with a small sample size of

79, Klis’s study (1997) found a significant positive correlation with extraversion (r

= .24, p<.05). However, Davis (1983) found no significant correlations between

perspective taking and scales related to extraversion.

The relationships between empathy and extraversion are therefore unclear. In a

recent study, De Young, Quilty and Peterson (2007) focused on ‘aspects’ of

personality, which they defined as a level of analysis between the factor and facet
 73

levels, arguing that extraversion could be meaningfully split into two such aspects:

‘enthusiasm’ and ‘assertiveness’. While it is possible that previous mixed findings

regarding emotional empathy and extraversion are due to sample size, existing

studies have explored relationships between empathy and the five higher factors,

rather than looking at aspect or even facet levels of personality. It is possible

therefore that these mixed findings can be accounted for by the fact that that

emotional empathy is related to certain facets of extraversion rather than the entire

domain. Theoretically, certain facets of extraversion appear to be more closely

related to the prosocial concepts of empathy, such as warmth and positive emotions.

For the purposes of this study it was hypothesised that extraversion would be

positively associated with perspective taking (hypothesis 3a) and empathic concern

(hypothesis 3b) as these are the two components of empathy associated with social

engagement. However this study extends previous research by exploring the

relationships between the empathic dimensions identified by Davis and the more

fine grained facets of the FFM of personality and addresses the construct validity of

the IRI.

4.2. Construct validity

According to Cronbach and Meehl (1955), “construct validity must be investigated

whenever no criterion or universe of content is accepted as entirely adequate to

define the quality to be measured” (p.281). Chapter two presented the ongoing

debates regarding definitions and models of empathy were presented, and as such

further evidence of construct validity for the IRI would be useful. Evidence of

construct validity is generally accumulated from multiple studies (Rust &

 74

Golombok, 1999). The method of construct validation used most extensively in

studies of the IRI is confirmatory factor analysis (e.g. Carey, 1988; Pulos, Ellison &

Lennon, 2004), which has provided evidence to support a four factor model in

general adult populations and healthcare workers (Carey, 1988; Pulos, Elison &

Lennon, 2004; Yarnold et al., 1996). An alternative approach to establishing

construct validity is to undertake a joint factor analysis of the measures to further

test the multidimensional model of empathy in relation to the five factor model of

personality (Kline, 2000).

4.3. Method

4.3.1 Participants and Procedure

A total of 105 volunteers took part in the study, 86 of whom were psychology

undergraduates. The remaining 19 were an opportunity sample of employed adults.

Mean age of the 105 participants was 22.83 years (S.D 8.41 years) while 75.7%

were female. Two pencil and paper questionnaires were administered in person to

participants for this part of the study. These were the 240 item NEO Personality

Inventory – Revised (NEO PI-R: Cost & McCrae, 1992) and the 28 item IRI

(Davis, 1983). No time limit was set.

4.3.2 Measures

Personality: NEO PI-R. The NEO PI-R (Cost & McCrae, 1992) is a measure of

personality, based on the FFM previously described. According to the measure,

 75

personality is made up of a pattern of preferences across the five factors. Within

each factor, a person’s traits are reflected in the fine grain ‘facets’. Each of the five

factors is made up of six facets, with eight items per facet. The study utilises a facet

level measure of the FFM to address the mixed findings from previous research

with respect to empathy and extraversion. The items within the questionnaire ask

the person to consider statements which are general rather than situation specific,

with the aim of measuring personality in a comprehensive manner so that it can be

used for multiple purposes. For each of the 240 items, participants respond to a

five-point Likert-type scale from “strongly disagree” through to “strongly agree”. A

brief description of the five factors is provided in Table 4.1, together with the facets

associated with each factor. According to the manual, each factor has acceptable

internal reliability, with αs ranging from .86 to .95. Facets have slightly lower

reliability statistics (αs from .56 to .81) which are deemed acceptable for scales

with only eight items (Costa & McCrae, 1992).

Empathy: The Interpersonal Reactivity Index. The Interpersonal Reactivity Index

(IRI; Davis, 1983) has been described extensively in Chapter Three. the Perspective

Taking subscale of the IRI is representative of an individual’s tendency to adopt the

psychological point of view of another with a sample item being “I sometimes try

to understand my friends better by imagining how things look from their

perspective”. The Fantasy subscale is similar to this, although is based on

imagining oneself in the role of characters in books, films or plays, with a sample

item being “I really get involved with the feelings of the characters in a novel”. The

Empathic Concern subscale asks about the individual’s own feelings of concern in

response to another person. A negatively worded sample item is “Other peoples’

 76

Table 4.1: Description of NEO PI-R five domains and facets

Factor Description Facets

Neuroticism Tendency to experience Anxiety
negative affect e.g. fear, Angry hostility
sadness, embarrassment, Depression
anger, guilt and disgust Self consciousness
Impulsiveness
Vulnerability

Extraversion The degree to which one Warmth

directs energy to the Gregariousness
external world Assertiveness
Activity
Excitement-seeking
Positive Emotions

Openness Having an active Fantasy

imagination, preference for Aesthetics
variety and intellectual Feelings
curiosity Actions
Ideas
Values

Agreeableness Concerned with Trust

interpersonal tendencies Straightforwardness
and how others find the Altruism
experience of being with Compliance
you Modesty
Tender-mindedness

Conscientiousness Tendency for self control Competence

and organisation; Order
associated with Dutifulness
achievement of potential Achievement Striving
Self-discipline
Deliberation

misfortunes do not usually disturb me a great deal”. The Personal Distress subscale

is also emotional, but is more self-oriented, focusing on how much one feels

distress in response to another. A sample item from this subscale is “I tend to lose

control during emergencies”. The four subscales are each composed of seven items,

to which participants are asked to respond using a five-point Likert-type scale

 77

(‘does not describe me well’ to ‘describes me very well’). All four sub-scales of the

IRI have been shown to have satisfactory internal reliability (α = 0.71 to 0.77) and

test – retest reliability (α = 0.62 to 0.71, Davis, 1983). A 0-4 scale is used for each

item, so the minimum possible score for each subscale is zero, with a maximum of

28 for each subscale. The scores are not intended to be combined, as each

represents a qualitatively different aspect of empathy, therefore no ‘total’ score is

calculated. The full questionnaire is included in Appendix 1.

4.3.3 Analysis

The hypotheses were tested by calculating subscale scores for each subscale of the

IRI and domains and facets of the NEO PI-R. Correlations were calculated using

the domain scores of the NEO PI-R and the IRI subscales. A joint factor analysis

was then conducted to examine similarities between Davis’ four dimensions of

empathy and the facets scores of personality measure. A joint factor analysis was

deemed appropriate for this because it avoids the risk of type I errors that are

associated with many correlational tests (Ferguson, 2001; Howell, 2002).

4.4. Results

Descriptive statistics for the NEO PI-R facets and the four IRI subscales are

provided in Table 4.2. For this sample of participants, reliabilities were comparable

to previously established reliabilities of the scales, with αs ranging from .50 to 82

for the NEO PI-R facet subscales and .69 to .80 for the IRI subscales. As

Cronbach’s α is sensitive to the number of items in the scale, these were deemed

acceptable reliabilities for the numbers of items within each scale. For shorter
 78

Table 4.2: Descriptive statistics for the NEO PI-R and Empathy (IRI) scales.

Scale Mean S.D α

Neuroticism Anxiety 18.44 5.40 .80
Angry-Hostility 14.67 5.34 .79
Depression 16.82 6.32 .84
Self-Consciousness 17.27 5.12 .72
Impulsivity 18.22 4.66 .66
Vulnerability 13.60 5.09 .76
Extraversion Warmth 23.01 4.01 .52
Gregariousness 19.63 5.35 .77
Assertiveness 15.19 5.46 .80
Activity 16.77 3.64 .45
Excitement-Seeking 18.96 4.87 .60
Positive Emotions 20.87 5.24 .79
Openness Openness to Fantasy 19.96 5.08 .75
Openness to Aesthetics 20.61 5.87 .81
Openness to Feelings 23.30 4.06 .71
Openness to Actions 16.84 4.67 .71
Openness to Ideas 21.00 5.08 .78
Openness to Values 21.68 3.88 .61
Agreeableness Trust 17.19 4.88 .77
Straightforwardness 19.44 5.19 .76
Altruism 23.46 3.48 .57
Compliance 16.35 4.90 .70
Modesty 19.45 4.73 .72
Tender-mindedness 21.17 3.21 .50
Conscientiousness Competence 18.62 4.04 .64
Order 15.92 5.21 .74
Dutifulness 20.16 4.17 .60
Achievement Striving 16.70 4.80 .75
Self-Discipline 16.11 5.72 .81
Deliberation 15.75 5.42 .82
Empathy [IRI] Fantasy 17.40 5.37 .80
Perspective Taking 21.30 3.76 .69
Empathic Concern 18.28 4.04 .70
Personal Distress 12.84 4.76 .79

Note: N=105 for all subscales. NEO PI-R facet scales all have 8 items each with a minimum score
of 5 and a maximum of 25, IRI scales have 7 items each with a minimum score of 0 and a maximum
of 28.
 79

scales, mean inter-item correlations (mics) should also be inspected as a measure of

reliability (Briggs & Cheek, 1986). For all scales, mics fell within the range .18 to

.45. These were within the acceptable range (Ferguson, 2001).

4.4.1 Correlations

To test hypotheses 1-3, first the correlations between the four empathy subscales

and five personality factors were calculated (see Table 4.3). A multi-stage adjusted

level of significance was adopted to control the family-wise error rate (Howell,

2002; Larzelere & Mulaik, 1977).

Hypothesis one: Neuroticism and empathy. Support was not found for hypothesis

1b, which predicted that a significant positive relationship would be found between

empathic concern, and neuroticism (r = .07, ns). However support was found for

hypothesis 1a as there was a significant large positive correlation between personal

distress and neuroticism (r = .47, p<.01). Furthermore, perspective taking was

found to be significantly negatively related to neuroticism (r = -.33, p<.01).

Hypothesis two: Agreeableness and empathy. Hypotheses 2a and 2b suggested that

perspective taking and empathic concern respectively would show significant

positive relationships with agreeableness. This was found to be the case (r = .38,

p<.01 and r = .29, p<.05 respectively).

 80

Table 4.3: Correlations between NEO PI-R factors and Empathy (IRI) scales.

Empathy
Perspective Empathic Personal
NEO PI-R Fantasy
Taking Concern Distress
Neuroticism (N) .13 -.33** .07 .47**

Extraversion (E) .26 .35** .30* -.25

Openness (O) .54** .24 .25 -.30*

Agreeableness (A) .10 .38** .29* .12

Conscientiousness (C) -.10 .14 -.04 -.23

Note: Larzalere and Mulaik (1977) adjusted *significant p<.05, ** p<.01.

Hypothesis three: Extraversion and empathy. Support was found for hypotheses 3a

and 3b, as significant moderate positive relationships were found between

extraversion and both perspective taking and empathic concern (r = .35, p<.01 and r

= .30, p<.05 respectively).

Openness and empathy. Finally, although no relationships were hypothesised

between the IRI and openness to experience, this domain score showed a

significant, large positive correlation with the IRI Fantasy subscale (r = .54, p<.01).

In addition, openness was significantly negatively correlated with personal distress

(r = -.30, p<.05).

4.4.2 Joint factor analysis

Correlations were only calculated between the four IRI subscales and the five

domain scores for the NEO PI-R. To produce a correlation matrix at a facet level
 81

would involve a very large number of tests and applying a correction for this would

result in very few interpretable results. However, Ferguson (2001) notes that joint

factor analysis (i.e. including subscale scores from the IRI and facet scores from the

NEO P-IR) is an appropriate method for investigating these types of relationships:

the simultaneous analysis allows for understanding of the IRI subscales within the

framework of the five factor model. This approach was followed to undertake a

facet level analysis of relationships between empathy and personality. Several pre-

analysis checks were conducted before embarking on the factor analysis.

Pre-analysis checks: To ensure that the IRI subscale scores and NEO PI-R facet

scores were suitable for an exploratory factor analysis, the data were checked

following Ferguson and Cox (1993). These checks include: examining that a stable

factor structure can be produced; that the variables are appropriately scaled and

distributed, and that there is systematic covariation within the data. In order to

determine the sample size required to produce a stable factor structure, statisticians

suggest a range of heuristics. Guadagnoli and Velicer (1988) reviewed these and

suggested that absolute sample size was the most important, as well as the

component saturation. The absolute sample size here of 105 participants is more

than the minimum of 100 suggested by Kline (1994). If the component saturation

(mean factor loading for a factor) is greater than 0.6, then according to Guadagnoli

and Velicer, increasing sample size will be of little value. This can only be

evaluated post-analysis and so this point will be returned to following the factor

analysis. Finally, skew and kurtosis of the variables were then explored. Muthen

and Kaplan (1985) identified three parameters as important when checking skew

and kurtosis. These were: the absolute magnitude of skew; the number of variables
 82

affected by skew, and the proportion of the initial correlations between variables

less than 0.2. None of the variables had skew of a magnitude of 2.0 (the modesty

subscale of the NEO PI-R demonstrated the greatest skew of -1.17 and was the only

variable to have a kurtosis value of more than 2.0: actual value 2.66).

Field (2005) also recommends calculating z scores for skew and kurtosis, with a

value of more than 1.96 marking significant skew or kurtosis. Using this method,

only four of the 34 variables were significantly skewed. These were from the IRI,

fantasy (z = -2.02) and empathic concern (z = -2.07) and from the NEO, excitement

seeking (z = 3.09) and modesty (z = 4.98). Only one variable showed significant

kurtosis (modesty, z = 5.70). Four of the 34 variables were therefore affected by

skew and/or kurtosis. According to Ferguson and Cox (1993), the cut off point for

acceptability is 25%. As only four out of 34 (12%) of the variables here were

affected, this should not adversely affect the solution. In addition, upon inspection

of the correlation matrix, 340 of a possible 561 (60.61%) correlations between all

34 variables were lower than 0.2. If the majority (more than 60%) of variables are

correlated less than 0.2, all variables can remain in the analysis regardless of skew

or kurtosis (Ferguson & Cox, 1993; Muthen & Kaplan, 1985). Therefore despite

there being some skew and kurtosis in four of the 34 variables, all were kept in this

analysis.

The final statistics recommended for inspection by Ferguson and Cox (1993) before

conducting exploratory factor analysis are those of the Kaiser-Meyer-Olkin [KMO]

test of sampling adequacy and Bartlett’s test of sphericity. The Bartlett’s test of

sphericity should be significant to indicate that relationships are present within the

data. A KMO statistic of at least 0.5 is required to indicate that the relationships
 83

between the variables can be accounted for by a reduced number of factors

(Dzubian & Shirkey, 1974). For this data, the KMO was found to be 0.73 and

Bartlett’s test of sphericity was significant (2176.38, p<.0001). The data were

therefore deemed suitable for exploratory factor analysis.

Factor extraction. The next stage is to choose a method for extracting factors

(Ferguson & Cox, 1993). The most widely used method is the Kaiser 1 (K1)

method, where all factors with eigenvalues greater than one are retained. A

principal components analysis using this method found seven factors with

eigenvalues greater than one. However, in a discussion of the various methods of

factor extraction, Zwick and Velicer (1986) concluded that this method leads to the

retention of too many factors. Instead the method of parallel analysis (Horn, 1965)

is recommended. Parallel analysis involves comparison of a randomly produced set

of eigenvalues, based on the same sample size, with those produced in the observed

data. A number of such randomly generated values are run and the average

eigenvalues are calculated. These average values and the observed values are then

plotted against the number of variables. The number of factors to retain is identified

immediately prior to the point where the two lines cross (Zwick & Velicer, 1986).

This method has been shown to be the most accurate method of factor extraction

and was therefore used here. Using syntax from O’Connor (2000), four analyses

were conducted at the 50th and 95th percentiles using 40 and 100 sets of randomly

generated data. The parallel analysis indicated that five factors should be retained.

Five factors were extracted using principal components analysis (Tabachnick &

Fidell, 2007) with varimax rotation.

 84

Table 4.4. Factor loadings for joint factor analysis of IRI subscales and NEO PI-R
facets

Component
Factor
1 2 3 4 5
1 Anxiety (N) 0.83 -0.01 -0.01 0.07 -0.01
Depression (N) 0.80 0.06 -0.21 -0.27 0.04
Vulnerability (N) 0.75 0.37 -0.10 -0.27 -0.11
Angry-Hostility (N) 0.72 -0.46 -0.17 0.00 -0.04
Self-Consciousness (N) 0.70 0.17 -0.27 -0.10 0.06
Personal Distress 0.54 0.31 0.05 -0.20 -0.29
Impulsivity (N) 0.52 -0.41 0.05 -0.32 0.17
Openness to Actions (O) -0.42 -0.19 0.18 -0.22 0.40
2 Compliance (A) -0.31 0.77 0.13 -0.06 0.06
Straightforwardness (A) 0.00 0.73 0.21 0.02 -0.05
Assertiveness (E) -0.29 -0.69 0.22 0.09 0.14
Modesty (A) 0.28 0.59 0.22 0.07 -0.02
Activity (E) -0.05 -0.49 0.22 0.07 0.14
Tender-mindedness (A) 0.13 0.47 0.34 -0.04 0.18
3 Warmth (E) -0.31 0.06 0.79 -0.02 0.19
Positive Emotions (E) -0.36 -0.01 0.73 0.01 0.17
Empathic Concern 0.24 0.06 0.67 0.04 0.17
Altruism (A) -0.01 0.37 0.66 0.17 0.13
Gregariousness (E) -0.23 -0.17 0.64 -0.33 -0.12
Perspective Taking -0.29 0.15 0.53 0.11 0.11
Trust (A) -0.46 0.43 0.47 0.05 0.07
4 Self-Discipline (C) -0.24 -0.23 0.20 0.78 -0.12
Dutifulness (C) -0.07 0.30 0.07 0.74 0.04
Achievement Striving (C) -0.13 -0.34 0.03 0.73 0.04
Competence (C) -0.36 -0.24 0.06 0.67 0.10
Order (C) 0.20 -0.08 -0.03 0.67 -0.17
Deliberation (C) -0.18 0.27 -0.05 0.65 -0.10
Excitement-Seeking (E) -0.08 -0.38 0.38 -0.52 0.11
5 Openness to Ideas (O) -0.15 -0.29 0.07 0.11 0.73
Openness to Aesthetics (O) -0.15 -0.10 0.20 0.00 0.67
Fantasy 0.19 0.02 0.22 -0.07 0.66
Openness to Fantasy (O) 0.06 0.05 0.11 -0.28 0.65
Openness to Feelings (O) 0.32 -0.04 0.45 0.14 0.60
Openness to Values (O) -0.18 0.20 -0.14 -0.06 0.57

Note: IRI subscales and factor loadings of 0.30 and greater are in boldface.
 85

Factor extraction using an oblique rotation was also performed, with five factors

being extracted. Factors two and three found to have a correlation coefficient of

0.24. Other than this, no factors showed substantial correlations and the pattern

matrix showed a highly similar solution to that of the varimax rotation. For

simplicity, the varimax rotated solution is reported.

The five factors extracted (eigenvalues = 6.72, 4.51, 4.28, 2.89, 1.94) accounted for

59.81% of the variance. Rotated factor loadings are presented in Table 4.4. The

acceptable magnitude of a factor loading for a variable to define a factor varies, but

the most commonly accepted level is 0.30 (e.g., Field, 2005). As mentioned

previously, factor saturation is defined as the mean factor loading for a factor: these

ranged from .62 to .73 for the five factors extracted in this analysis. As all were

found to be greater than 0.6, then according to Guadagnoli and Velicer (1988),

increasing sample size would be of little value. Therefore although the sample size

was relatively small (105), this was deemed to be sufficient for this analysis.

Interpretation of factors. Factor one contained all of the neuroticism facets of the

NEO PI-R. In addition, one of the openness facets (to actions) had a negative

loading on this factor but also cross-loaded positively on to Factor five with a

similar magnitude. This factor also included the Personal Distress subscale of the

IRI, with a factor loading of .54, suggesting that the tendency to become distressed

by another’s experiences is indeed associated with a greater tendency to experience

negative emotions in general.

 86

Factors two and three both contained a mix of facets from the extraversion and

agreeableness domains. Two subscales of the IRI, empathic concern and

perspective taking, both loaded positively on to Factor three with factor loadings of

.67 and .53 respectively. Factor Three also contains the Warmth, Positive Emotions,

Altruism, Gregariousness and Trust facets of the NEO PI-R. The Openness to

Feelings facet of openness also cross-loaded positively with this factor. This

suggests that Factor three is composed of several facets of personality which

indicate a compassionate disposition.

Factor Four contained all of the conscientiousness facets of the NEO PI-R along

with one extraversion facet: excitement seeking negatively loaded on to this factor.

None of the IRI subscales loaded significantly on to this factor. Finally, Factor five

was made up of the remaining five openness facets. The Fantasy subscale of the IRI

loaded positively on to this openness factor (.66).

4.5. Discussion

This study contributes in several ways to the theoretical understanding of individual

differences in empathy as measured by the IRI. Hypothesis one made predictions

regarding neuroticism and empathy. Support was not found for hypothesis 1b,

which predicted that a significant positive relationship would be found between

empathic concern, and neuroticism. However support was found for hypothesis 1a

as there was a significant large positive correlation between personal distress and

neuroticism. Hypothesis two predicted relationships between agreeableness and

empathy. Hypotheses 2a and 2b were both supported, with perspective taking and

empathic concern respectively showing significant positive relationships with

 87

agreeableness. Hypothesis three was also supported, with significant moderate

positive relationships were found between extraversion and both perspective taking

and empathic concern. Finally, although no relationships were hypothesised

between the IRI and openness to experience, this domain was found to be positively

correlated with fantasy and negatively correlated with personal distress. The joint

factor analysis provided a deeper, facet level understanding of the results for

perspective taking and empathic concern. Clear links were established between

these subscales and a single factor comprising a range of facets of Extraversion and

Agreeableness. These were warmth, positive emotions and gregariousness from

extraversion and altruism and trust from Agreeableness. Tender-mindedness also

cross loaded on to this factor. Personal distress was found to load positively on to

the Neuroticism factor as expected, while fantasy loaded on to openness to

experience.

This study aimed to investigate the antecedents of empathy in terms of individual

differences and more specifically the five factor model of personality. In order to do

this, the relationships between each subscale of the IRI and the facets of the NEO

PI-R have been investigated. Similar patterns of results were found for two of the

subscales, Perspective Taking and Empathic Concern in terms of relationships with

the five factor model. These two subscales are therefore discussed together,

followed by findings for Personal Distress and then the Fantasy subscale which

resulted in substantially different patterns of relationships.

 88

4.5.1 Perspective Taking and Empathic Concern

The Perspective Taking and Empathic Concern subscales were both found to

correlate with the factors of Agreeableness and Extraversion. Conducting a joint

factor analysis at the facet level has provided a greater understanding of these

relationships. It should be noted that it was not entirely unexpected to find that the

facets from extraversion and agreeableness appeared to form two new composite

factors in this analysis (Factors two and three). Other analyses have repeatedly

found these two factors to be related in their impact upon relationships and social

functioning (e.g., Noftle & Shaver, 2006; White, Hendrick & Hendrick, 2004).

Further evidence from Trapnell and Wiggins (1990) and McCrae and Costa (1989)

explains that these two factors comprise interpersonal aspects of personality and

that a circumplex approach combining the two factors complements the five factor

model. The two composite factors which emerged are in line with Mehrabian’s

Pleasure-Arousal-Dominance (PAD) Temperament Model of Personality (1996).

The facets of Factor two (Compliance, Straightforwardness, Assertiveness (-),

Modesty, Activity (-), and Tender-mindedness) all relate to Mehrabian’s

Submissive-Dominant dimension, with negatively loading facets representing

dominance. The two empathy subscales did not load significantly on to this factor.

Factor Three comprised the facets of Warmth, Positive Emotions, Altruism,

Gregariousness and Trust, and tender-mindedness also cross-loaded positively with

this factor. This factor is therefore representative of Mehrabian’s Pleasure

dimension. This was the factor onto which the subscales of Perspective Taking and

Empathic Concern were significantly positively loaded. In a recent study, De

 89

Young, Quilty and Peterson (2007) identified ‘aspects’ of personality which they

defined as a level of analysis between the factor and facet levels. They found that

both the extraversion and agreeableness factors could be meaningfully split into

two such aspects. The extraversion aspects were named ‘enthusiasm’ and

‘assertiveness’ while the agreeableness aspects were named ‘compassion’ and

‘politeness’. In line with their analysis, the two subscales of empathy were related

to the aspects of enthusiasm and compassion from De Young et al. (2007). These

findings suggest that an empathic person with a tendency for perspective taking and

empathic concern also reports being warm, caring, outgoing, compassionate and

trusting.

Perspective Taking and Empathic Concern showed the same patterns of

relationships with the other trait-based measure. Those relationships were in line

with hypotheses of expected descriptions for empathic individuals, providing

support for the construct validity of the two subscales as part of a measure of

empathy. Furthermore, the findings also provide evidence to include both cognitive

and affective processes in a definition and measure of empathy (Davis, 1996;

Vreeke & van der Mark, 2003). Researchers arguing for one or the other may well

be making artificial distinctions (Preston & de Waal, 2002). However, the subscale

of empathic concern is more related to congruent affect towards another person.

Shared affect, the process which is the usual topic of focus within the empathy

literature, is not captured by this subscale. The results of the analysis involving the

Personal Distress subscale are therefore relevant at this point. This subscale showed

very different relationships with the traits measured in this study, which are now

discussed.
 90

4.5.2 Personal Distress

The Personal Distress subscale of the IRI showed a significant and moderate

positive correlation with the Neuroticism scale of the NEO PI-R. In addition, a

moderate significant negative correlation was found with Openness. Two moderate

negative correlations with Extraversion and Conscientiousness did not reach

significance once a correction had been applied to reduce the risk of Type I error.

The joint factor analysis provided greater understanding of these relationships, as

Personal Distress was found to load positively on to the factor composed of all the

facets of Neuroticism. This factor also contained one facet of Openness, namely

Openness to Actions. The findings are in line with previous research that suggests

that emotional arousability, positively associated with the neuroticism factor,

underpins the emotional empathic response (Jabbi, Swart & Keysers, 2007). The

pattern of relationships found in this study suggest that people who become

distressed when others are in distress report greater experience of negative

emotions in general. In light of this, it will be interesting to investigate the criterion-

related validity of the IRI within the healthcare context. Before considering this

further, the results of the final section of the analysis require discussion.

4.5.3 Fantasy

The final subscale of the IRI, Fantasy, showed a different pattern of relationships to

any of the other subscales. In terms of its relationship to the five factor model,

Fantasy showed a large positive correlation with Openness as well as loading

 91

positively on to the fifth factor of the joint factor analysis, which all of the

Openness facets also loaded positively on to. This suggests that those people who

score highly on items in the Fantasy subscale concerning imagination and fantasy

of thought are more open to new experiences in general. There is little previous

research to explain this finding as the relationship between empathy and openness

had not previously been explored. The Fantasy subscale was originally included in

the IRI as it supposedly related to increased emotionality. Physiological evidence of

a link between fantasy and greater emotional responding to the emotions of others

was provided by Stotland et al., (1978), but in this study there was no significant

correlation between this subscale and Neuroticism. The factor loading of the

Fantasy subscale on to Factor one (Neuroticism) was positive, but only of a

magnitude of .19. This suggests that for this sample of participants, self reported

negative emotionality did not clearly relate to Fantasy. However, the Openness to

Experience factor did positively correlate with Fantasy (r = .54, p<.01), perspective

taking and empathic concern (although non significantly with Pearson’s r being .24

and .25 respectively). It may be that being open to new ideas, feelings and more

imaginative generally helps one to imagine the experience of another person, thus

facilitating greater depth of perspective taking and empathic concern. In particular,

in emotional labour roles such as healthcare practitioners, this openness may be

particularly relevant when encountering new people with whom one has not yet

developed relationships. The fantasy subscale, previously somewhat ignored, may

therefore be useful in this context. However, there are those who have used the IRI

without the Fantasy subscale, arguing for its lack of relevance to the topic of

research (e.g., Christopher, Owens & Stecker, 1993). It will therefore be important

to investigate the relationships between Fantasy and behaviour.

 92

4.5.4 Limitations and next steps

The study makes a useful contribution in that it constitutes the first known

comparison of the IRI with a broad five factor measure of personality, however,

whilst the factor saturations suggested that sample size would not have altered the

findings, the sample was relatively small size for this type of research (Field, 2005).

More importantly, given that the aim of this thesis is to investigate processes of

empathy in the healthcare setting, there is a need to extend findings to participants

from the healthcare professions and demonstrate the utility of the IRI in this

context. This is the focus of the next study.

 93

Chapter 5 – Emotional intelligence and empathy

Significant advances have been made in the assessment of individual differences

since the construction of the IRI. The last chapter focused on empathy within the

framework of the five factor model of personality (Digman, 1990). Also of

particular note is the emergence of the concept of Emotional Intelligence [EI],

which has received considerable attention from academics and practitioners in

organisational psychology (e.g., Goleman, 1995; Petrides & Furnham, 2000;

Salovey & Mayer, 1990). Empathy and EI share common theoretical and research

roots in social intelligence (Chlopan et al., 1985; Landy, 2005), but despite

conceptual similarities, no structured comparison of a multidimensional measure of

empathy and EI has been undertaken in an adult population. Indeed, researchers

have tended to assume that empathy is a component of EI (e.g., Goleman, 1995)

rather than investigate relationships between the two empirically, and the terms

empathy and EI (and their associated measures) have been referred to

interchangeably (e.g., Geher, Warner & Brown, 2001).

Consequently the first aim of this study is to undertake an empirical comparison of

self-assessed empathy and EI, in order to understand how Davis’ conceptualisation

of a multidimensional model of empathy fits with more recent research concerned

with the concept of EI. In addition, as the previous study used a sample of

participants from the general population and this thesis aims to further

understanding of empathy in the healthcare setting, a second aim of the study is to

investigate self-assessed empathy using the IRI among doctors currently employed

in the NHS. Before considering the relationships between empathy and EI, the

concept of EI itself is introduced.

 94

5.1 Emotional Intelligence

Emotional intelligence [EI] has received considerable attention in both popular and

academic literatures in recent years (e.g., Petrides & Furnham, 2000). Although

many different definitions have been proposed, a review by Ciarrochi, Chan and

Caputi (2000) identified four aspects that are included by most definitions of EI:

emotion perception, emotion regulation, emotion understanding and emotion

utilisation. These areas relate to emotions in both the self and others, for example

regulation of one’s own emotions in coping with stress and understanding of others

emotions in building social relationships (e.g. Bar-On,1997; Schutte et al., 2001).

The four aspects are similar across the definitions of many researchers such as

Davies, Stankov and Roberts (1998), Mayer, Caruso and Salovey (2000), and Law,

Wong and Song (2004). However, as in the empathy literature, debates continue

over definitions of EI. In fact, Pérez, Petrides and Furnham (2005) have argued that

whilst most studies of EI are carried out in a “theoretical vacuum” (p.182), this

does not appear to have hampered the development of measures of EI that are now

sold widely for commercial use (Schulze & Roberts, 2005).

A major dispute among EI researchers at present is whether EI should be

conceptualised as a trait, an ability, or a mix of the two (Pérez, Petrides & Furnham,

2005). Salovey and Mayer (1990) view EI as a cognitive ability, referring to it as

“the subset of intelligence that involves the ability to monitor one’s own and

others’ feelings and emotions, to discriminate among them and to use this

information to guide one’s thinking and actions” (p.189). In line with this, a ‘test’

of EI has been developed, the Mayer, Salovey & Caruso Emotional Intelligence
 95

Test (MSCEIT; Mayer Salovey & Caruso, 2002). However, in a review of evidence

pertaining to trait or ability EI, Pérez, Petrides and Furnham (2005) argue that such

tests of maximal performance are inherently problematical when investigating EI

due to the inability of researchers to objectively score responses. What constitutes a

correct response has either been determined through consensus or expert scoring,

neither of which can be deemed to be perfectly correct. Due to such difficulties,

greater progress has been made in developing measures which conceptualise EI as a

trait.

Petrides and Furnham (2001) emphasise the importance of the distinction between

tests of maximal performance to assess abilities and self-report measures of typical

performance for traits. Petrides, Perez-Gonzalez & Furnham, (2007) argue that EI is

better viewed as a construct which “encompasses self-perceptions and dispositions,

which accord with the subjective nature of emotions” (p, 274). One of the most

commonly used self-report measures, which has received support from empirical

studies, is the Bar-On Emotional Quotient Inventory (EQ-i: Bar-On, 1997). Bar-On

defines EI as “an array of non-cognitive capabilities, competencies, and skills that

influence one’s ability to succeed in coping with environmental demands and

pressures” (p.14). Although it has been claimed that the Bar-On EQ-i is a ‘mixed’

model of EI, assessing both abilities and preferences (e.g. Bar-On, 1997; Brackett

& Mayer, 2003), the self-report nature of the assessment aligns itself more closely

with a trait approach. Indeed, construct validation studies have been successful in

finding relationships between this measure and the five factor model of personality,

indicating that it can be viewed as a trait-based measure (e.g O’Connor & Little,

2003; Petrides & Furnham, 2001). The Bar-On EQ-i comprises fifteen different

subscales, scores from which are then combined into five composite factors:
 96

intrapersonal EQ; interpersonal EQ; adaptability; stress management, and general

mood. The exact mapping of the fifteen subscales to the five composite factors is

provided in Table 5.1.

Regarding construct validity of the EQ-i, the factor structure of the EQ-i has been

questioned. For example, the Bar-On EQ-i manual does not find support for the five

factors, instead reporting thirteen. However the method used for extracting factors

was the K1 method retaining all factors with eigenvalues greater than one. This

method has been found to lead to the retention of too many factors (Zwick &

Velicer, 1986). Using confirmatory factor analysis, Petrides and Furnham (2001)

found a single factor to be a better fit to data for a sample of 227 working adults.

Despite this, other evidence regarding the reliability and criterion-related validity of

the measure is positive (e.g. Dawda & Hart, 2000; O’Connor & Little, 2003). It is

therefore the measure used within this study. Having introduced the concept of EI,

evidence regarding the relationship between EI and empathy is now considered.

5.2 Empathy and Emotional Intelligence

Although empathy and EI share common roots in social intelligence (Chlopan,

1985; Landy, 2005), the two terms are often referred to interchangeably. Salovey

and Mayer argue that empathy is a component of EI (1990; Mayer & Salovey,

1997). Empathy is also named as a subscale of the EQ-i (Bar-On, 1997) and the

Trait Emotional Intelligence Questionnaire (TEIQue: Petrides, Perez-Gonzalez &

Furnham, 2007) Making assumptions about relationships between the empathy

 97

Table 5.1: Description of Bar-On EQ-i five factors and fifteen subscales

Composite factor Subscales Description

Intrapersonal EQ Self Regard High scorers are in touch
Emotional Self-Awareness with their feelings, feel
Assertiveness good about themselves,
Independence feel positive about their
Self Actualization lives, are independent
strong and confident

Interpersonal EQ Empathy High scorers are

Social Responsibility responsible and
Interpersonal Relationships dependable individuals
with good social skills,
understanding and
interacting well with
others

Adaptability EQ Reality Testing High scorers are flexible,

Flexibility realistic, effective in
Problem Solving understanding
problematic situations,
competent at arriving at
adequate solutions, find
good ways of dealing
with everyday difficulties

Stress Management EQ Stress Tolerance High scorers are able to

Impulse Control withstand stress without
falling apart or losing
control, calm, rarely
impulsive and work well
under pressure

General Mood Happiness High scorers are cheerful

Optimism positive and hopeful
individuals who know
how to enjoy life
 98

and EI can result in misleading interpretations of research findings. For example,

Geher, Warner and Brown (2001) investigated whether performance measures of

EI were more effective than trait based measures at predicting participant

performance in a video-based emotion detection task. The trait based ‘EI’

measures they used were in fact empathy measures – the QMEE and IRI, not EI

measures – a fact that was not acknowledged. The ability based measure of EI

and not the trait measures (of empathy) were found to predict performance on the

video-based task, and the authors interpret this finding as evidence that EI is

better conceptualised as an ability than a trait. As discussed in chapter three the

QMEE or the IRI do not assess emotion detection per se, but encompass a

broader assessment of the individual’s preferred response once emotions have

been detected. This example illustrates the need to be theoretically and

methodologically clear before interpreting findings. To further clarify these

relationships, a comparison of is needed of self-assessed empathy and emotional

intelligence. Comparing the IRI with a trait-based measure of EI could also

provide useful evidence of construct validity of the different IRI subscales. The

two concepts show overlap in definitions and measures and should therefore be

related. There are a small number of studies that have conducted correlational

analyses between measures of empathy as measured by the IRI and EI and shown

this to be the case.

Perspective Taking and EI. The empathy subscale of the TEIQue (Petrides,

Perez-Gonzalez & Furnham, 2007) contains items focusing on perspective

taking, implying that this is the key dimension of empathy relating to EI.

Charbonneau and Nicol (2002) investigated EI in a sample of 191 adolescents

 99

with a mean age of 14 years. Perspective taking was found to have a moderate

positive correlation with the Trait measure of emotional intelligence (TMEI:

Schutte et al., 1998). Schutte et al., (2001) also replicated this finding with two

samples of adults using the same measures. This was true for both a sample of 24

students (r = .35, p<.05) and 37 teaching interns (r = .59, p<.001). However, the

teaching interns were not given other subscales of the IRI and the sample size is

small, therefore generalisations from the findings of these studies are not

possible. Stratton, Elam, Murphy-Spencer and Quinlivan (2005) used the Trait

Meta Mood Scale of EI (TMMS: Salovey et al., 1995) with 165 medical students.

The TMMS scale is comprised of three subscales: attention to feelings (AF);

clarity of feelings (CF), and mood repair (MR). Perspective taking was found to

be positively related to all three subscales (AF r = .47; CF r = .28; MR r = .56,

all p<.05). Therefore hypothesis 1 states that perspective taking will be positively

related to a measure of EI.

Empathic concern and EI. Barchard (2003) attempted to assess the predictive

power of EI to explain academic success, asking students to complete 31

measures of EI. One of the measures chosen was the empathic concern subscale

of the IRI. All of the measures were used in a regression as separate predictors of

the criterion; no assessment was made of the relationships between empathic

concern and the other EI measures. Rather, it was assumed that empathic concern

represents EI. In their study of medical students, Stratton et al. (2005) found

positive relationships between the three TMMS subscales and empathic concern

(AF r = .71; CF r = .30; MR r = .49, all p<.05). Charbonneau and Nicol (2002)

also found moderate positive correlations between empathic concern and EI for a
 100

sample of adolescents (TMEI: Schutte et al., 1998). Therefore hypothesis 2 states

that empathic concern will be positively related to a measure of EI.

Fantasy and EI. The relationship between fantasy and EI is perhaps less clear on

first inspection. Indeed, in their investigation of medical students’ empathy,

Stratton et al. (2005) did not administer this subscale of the IRI to the students,

although no reason was given for this. Charbonneau and Nicol (2002) did find

moderate positive correlations between fantasy and EI with the sample of 191

adolescents. (TMEI: Schutte et al., 1998). Therefore hypothesis three states that

fantasy will be positively related to a measure of EI.

Personal distress and EI. One possible difference between EI and

multidimensional empathy is apparent. EI is defined as necessary for effective

social functioning, and trait-based measures such as the EQ-i have been

associated with lower neuroticism and greater well-being (Bar-On, 1997;

Petrides & Furnham, 2001). Evidence has been presented in the previous chapter

that emotional empathy, in particular personal distress, is associated with higher

neuroticism. Indeed, in their study of medical students Stratton et al. (2005)

reported significant negative correlations between personal distress and the three

subcsales of the TMMS measure of EI. (AF r = -.23; CF r = -.47; MR r = -.36,

all p<.05). Hypothesis four is therefore that personal distress will be negatively

related to a measure of EI.

By testing relationships between the IRI and EQ-i measure of emotional

intelligence, this study will aim to provide further construct validity evidence for
 101

the EQ-i. As in the previous chapter, one approach to establishing construct

validity is to undertake a joint factor analysis of the measures to understand the

multidimensional model of empathy in relation to the construct of emotional

intelligence. As the few other studies reported in this section have only used

correlational analyses, this study will aim expand understanding of the

relationships between the two constructs.

5.3 Method

5.3.1 Participants

Data were collected from two sources for the study. Two hundred and fifty six

applicants applying for GP specialty training in a UK deanery were invited to

attend assessment centres conducted over a one week period. On arrival at the

assessment centre, all applicants were invited to participate in the study on a

voluntary basis. Information sheets were disseminated to briefly explain the

purpose of the research study and applicants were assured that any information

from the psychometric questionnaires used for the research would not be made

available to those making selection decisions and as such the research did not

form part of the selection process. Consent forms were signed by all participants

to indicate their understanding and agreement to take part. (See Appendices 2

and 3 for the information sheet and consent form). Out of 256 applicants

attending the assessment centres, 192 doctors agreed to take part in the study and

completed the measures detailed below. Data were also collected from 105

students, to increase the sample size to an adequate number. The larger sample

size was required as a decision was taken to conduct an item level factor analysis
 102

of the emotional intelligence measure. The factor structure of the Bar-On EQ-i

has been a subject of debate (e.g. Petrides & Furnham, 2001) and currently there

is no clear agreement, therefore a factor structure for this sample was established.

Due to the high number of items (133), a large sample size was required. The

total sample for this study therefore comprised 297 participants. 59.1% of the

sample was female, with a mean age of 22.83 years (S.D 8.41 years).

5.3.2 Procedure and Measures

Participants were administered two pencil and paper questionnaires for this part

of the study. These were the 133 item Bar-On Emotional Quotient Inventory

(EQ-i: Bar-On, 1997) and the 28 item IRI (Davis, 1983). No time limit was set.

Empathy: The Interpersonal Reactivity Index. The Interpersonal Reactivity Index

(IRI; Davis, 1983) has been described extensively in Chapter Three. the

Perspective Taking subscale of the IRI is representative of an individual’s

tendency to adopt the psychological point of view of another with a sample item

being “I sometimes try to understand my friends better by imagining how things

look from their perspective”. The Fantasy subscale is similar to this, although is

based on imagining oneself in the role of characters in books, films or plays, with

a sample item being “I really get involved with the feelings of the characters in a

novel”. The Empathic Concern subscale asks about the individual’s own feelings

of concern in response to another person. A negatively worded sample item is

“Other peoples’ misfortunes do not usually disturb me a great deal”. The

Personal Distress subscale is also emotional, but is more self-oriented, focusing

 103

on how much one feels distress in response to another. A sample item from this

subscale is “I tend to lose control during emergencies”. The four subscales are

each composed of seven items, to which participants are asked to respond using a

five-point Likert-type scale (‘does not describe me well’ to ‘describes me very

well’). All four sub-scales of the IRI have been shown to have satisfactory

internal reliability (α = 0.71 to 0.77) and test – retest reliability (α = 0.62 to 0.71,

Davis, 1983). A 0-4 scale is used for each item, so the minimum possible score

for each subscale is zero, with a maximum of 28 for each subscale. The scores

are not intended to be combined, as each represents a qualitatively different

aspect of empathy, therefore no ‘total’ score is calculated. The full questionnaire

is included in Appendix 1.

Emotional Intelligence: the EQ-i (Bar-On, 1997). The Bar-On EQ-i (1997) is

currently one of the most widely used measures of emotional intelligence

available, in both academic research and practice. It measures an “array of non-

cognitive capabilities, competencies, and skills that influence one’s ability to

succeed in coping with environmental demands and pressures” (Bar-On, 1997,

p. 14). It is a self-report measure which assesses individuals on five composite

factors: (a) intrapersonal EQ, (b) interpersonal EQ, (c) adaptability (d) stress

management and (e) general mood. Each factor is comprised of several

subscales. Empathy is named as one of three subscales of the interpersonal EQ

factor. Respondents are asked to rate 133 items using a five-point likert scale (1

= ‘Very seldom or not true of me’ to 5 = ‘Very often true or true of me’). Only

117 of the 133 items relate to the five composite factors, the remaining 16 items

acting as ‘validity indicators’. Reliability coefficients for the five subscales range
 104

from α = 0.69 to 0.86 (Bar-On, 1997). Independent studies have also provided

support for the validity and reliability of the measure (e.g. Dawda & Hart, 2000).

EQ-i item level factor analysis. Although the manual for the EQ-i provides

information regarding the five composite factors, the decision was taken here to

undertake an exploratory factor analysis of the EQ-i. Two reasons are given for

this decision. First, 12 of the 117 items are used to calculate scores for more than

one of the subscales. This is not uncommon within psychology (e.g., the

Occupational Personality Questionnaire: Bartram, Brown, Fleck, Inceoglu &

Ward, 2006) but suggests that a more parsimonious model might be in order.

Secondly, the factor structure of the EQ-i is a subject of debate. For example, the

manual for the measure produces a 13 factor solution rather than five factors

(Bar-On, 1997), while Petrides and Furnham (2001) found a unifactorial solution

to be a better fit to the data. Therefore an item level factor analysis of the EQ-i

was conducted to determine the factors to use in the joint factor analysis with the

IRI. This process is described before reporting the main analysis to test the

hypotheses.

As new scales were to be calculated, a reliability analysis was also conducted.

Before conducting any analysis on the data from the EQ-i, 16 of the 133 items

were removed. These items are known as ‘validity indicators’, designed to detect

when respondents are giving overly positive or negative impressions of

themselves. While they may be a result of a lack of self awareness or issues with

self esteem, which may be related to one’s emotional intelligence, these scales do
 105

not form part of the emotional intelligence score (Bar-On, 1997) and so have

been removed from this analysis. A reliability analysis following guidance from

Rust and Golombok (1999) was then carried out.

The reliability analysis started with a full item analysis of the remaining 117

items. Item facilities were inspected to check that the items were able to

differentiate between respondents. Any items with a mean of less than 0.50 or

more than 4.50, or with an acceptable mean but a standard deviation of less than

0.75 were deemed to have insufficient facility indices and so were removed. Four

such items were removed. Next, item discrimination was checked to see if the

scale was discriminating between people on the construct of interest. Inter-item

correlations and item-total correlations were checked. These statistics should

indicate positive relationships between the items to be measuring aspects of the

construct in question. Any items with a majority of inter-item correlations or an

item-total correlation of less than 0.2 were removed as these would not appear to

be differentiating on the construct of interest. A further six items were removed

as a result of these checks. No items had a majority of inter-item correlations or

an item-total correlation of more than 0.8, suggesting that there was no issue of

multicollinearity.

Having removed the 16 ‘validity indicators’ and ten items which did not meet the

criteria of the item analysis, 107 items remained in the scale. Exploratory factor

analysis was then conducted to determine how these items would form the

subscales. A similar procedure to that of study one was followed. Again this

involved checking that a stable factor structure can be produced, appropriate

 106

scale and distribution and also that there was systematic covariation within the

data (Ferguson & Cox, 1993).

The absolute sample size of 297 participants is more than the minimum of 100

suggested by Kline (1994). In terms of skew and kurtosis, no items had a skew of

more than 2.0 and only 4 items had a kurtosis value greater than 2.0. This was

within the acceptable limit of 25% of items suggested by Ferguson and Cox

(1993), in line with Muthen and Kaplan’s parameters (1985). The solution should

therefore not be adversely affected by skew or kurtosis. Finally, the KMO

statistic of 0.88 and a significant Bartlett’s test of sphericity (17053.55, p<.0001)

suggested that there were discoverable relationships within the data and so

exploratory factor analysis was suitable.

Factors were then extracted, with parallel analysis determining the number of

factors to extract (Horn, 1965; Zwick & Velicer, 1986). Analyses were

conducted at the 50th and 95th percentiles using 40 and 100 sets of randomly

generated data. The parallel analysis indicated that eight factors should be

retained. Factors were extracted using principal components analysis

(Tabachnick & Fidell, 2007) using an oblimin rotation with delta set at 01. This

rotation was used as the factors were anticipated to be intercorrelated, which was

found to be the case (Field, 2005). The eight factors extracted (eigenvalues

ranged from 25.34 to 2.09) accounted for 45.44% of the variance. Eight items of

the 107 showed factor loadings of between .25 and .29.

1
Rotations with delta set at 1 and 2 were found to produce similar solutions.
 107

Table 5.2: Factor names, descriptions, means, S.D.s and reliabilities for eight subscales resulting from EQ-i factor analysis
Original EQ-i subscale Description and sample items N.
New Factor Name M S.D α
Items
Self esteem Self contentment, Having a positive self regard e.g. I have good self respect; I don't 21 3.47 .68 .94
enjoyment feel good about myself (-)

Self control Reality testing, impulse Having some control over one’s thoughts and behaviour e.g. I 11 3.37 .63 .78
control have strong impulses that are hard to control (-); I think its
important to be a law abiding citizen

Flexibility Flexibility Openness to change and adjustment e.g. its hard for me to change 10 3.30 .64 .80
my ways (-); I'm able to change old habits

Rationality Problem solving, stress Taking a logical and reasoned approach e.g. when trying to solve a 15 3.41 .56 .82
tolerance problem I look at each possibility and then decide on the best
way; I try to see things as they really are, without fantasizing or
daydreaming about them

Emotional Stress tolerance, anger Keeping control over one’s emotions e.g. I feel that its hard for 11 3.62 .75 .83
Regulation control me to control my anxiety (-); I can handle stress without getting
too nervous

Interpersonal Interpersonal Showing interpersonal respect and sensitivity e.g. I'm sensitive to 12 3.44 .59 .81
Sensitivity relationships, empathy, the feelings of others; I avoid hurting other people's feelings
social responsibility
Emotional Emotional self awareness Expression of one’s feelings e.g. it's fairly easy for me to express 10 3.47 .73 .84
Expression feelings; I'm unable to show affection (-)

Assertiveness Assertiveness/ Demonstrating resolve and decisiveness e.g. when I disagree with 17 4.08 .45 .86
Independence someone, I'm able to say so; Others think that I lack assertiveness
(-)
 108

These items have been kept in the analysis as their removal has the effect of

reducing reliability of the subscales. Rotated factor loadings are presented in

Appendix five. Mean factor loadings for this solution ranged from 0.32 to 0.54,

suggesting that component saturation is an issue and increasing the sample size

may be of benefit in future (Guadagnoli & Velicer, 1988). This was the primary

reason for using all of the data available in a single analysis. However, as the

absolute sample size was deemed to be satisfactory, results are interpreted.

For copyright reasons, the lists of items loading on to each of the eight

components cannot be given in full. They were given to three occupational

psychologists familiar with the topic of emotional intelligence for factor

interpretation and naming. Reliabilities and scores for the eight subscales were

then created in accordance with this factor analysis, details of which are provided

in Table 5.2. This table also shows the original EQ-i subscales that go into each

factor. The only subscale which was split over two of the new factors was stress

tolerance. Some of the items from this subscale loaded on to the rationality

factor, while others loaded on to emotional regulation. Apart from this, the new

factors were broadly represented by items from one or more of the original

subscales.

5.4 Results

To increase sample size for analysis, extra data were gathered from a sample of

the general population, composed largely of undergraduate psychology students.

To check that combination of the data was appropriate, the two groups were

compared on the subscales of both the EQ-i and the IRI. A MANOVA revealed
 109

no significant multivariate effect of group (F = 1.38, df = 284, p>.05). ANOVAs

(see Table 5.3) did reveal a significant main effect of group for two of the

subscales, with doctors reporting higher scores for emotional regulation and

empathic concern. However, these effect sizes were small (partial eta squared =

.01 in both cases). Therefore the two groups were deemed suitable to combine.

Table 5.3: Comparison of the medical and non-medical samples.

Subscale Group 1 – doctors Group 2 – other

(n = 192) (n = 105)
Mean SD Mean SD F Eta
squared
EQi
Self esteem 4.09 .59 3..96 .63
Self control 4.07 0.52 3.99 .64
Flexibility 3.51 .44 3.50 .45
Rationality 3.91 .43 3.89 .55
Emotional 3.95 .63 3.80 .60 3.89* .01
Regulation
Interpersonal 4.32 .46 4.23 .60
Sensitivity
Emotional 3.91 .63 3.92 .75
Expression
Assertiveness 3.75 .55 3.57 .68
IRI
Fantasy 14.22 5.96 14.84 6.34
Empathic concern 22.09 3.50 21.23 3.78 3.87* .01
Perspective 19.85 4.22 19.09 4.37
Taking
Personal Distress 9.98 4.51 10.45 3.73
Notes: EQi subscale scores range from 1-5; IRI subscale scores range from 0-28; *p=.05

5.4.1 Correlations

Using the newly created subscale scores for the EQ-i, and the four subscale

scores from the IRI, correlations and joint factor analysis were conducted to test

hypotheses one to four. Normal distributions were confirmed by inspecting

histograms for each variable. Correlations are presented in Table 5.4.

 110

Table 5.4: Correlations between EQ-i and Empathy (IRI) scales.

Empathy
EQ-i F EC PT PD
Self esteem -.11 .22** .37** -.38**

Self control -.34** .20* .33** -.31**

Flexibility -.17 .19 .40** -.37**

Rationality -.12 .25** .48** -.50**

Emotional regulation -.04 .13 .41** -.48**

Interpersonal sensitivity .02 .45** .46** -.21*

Emotional expression -.02 .30** .32** -.32**

Assertiveness -.13 .11 .29** -.54**

Note: Larzalere and Mulaik (1977) adjusted *significant p<.05, ** p<.01. Empathy (IRI) scales:
F = Fantasy, EC = Empathic Concern, PT = Perspective Taking, PD = Personal Distress.

Hypothesis one: Perspective Taking and EI. Support was found for hypothesis

one as significant moderate correlations were found between perspective taking

and all eight EQ-i subscales. Pearson’s r ranged from .29 to .48 (all p<.01).

Hypothesis two: Empathic Concern and EI. Partial support was found for

hypothesis two as small to moderate significant correlations were found between

empathic concern and six of the eight EQ-i subscales. Pearson’s r ranged from

.20 (p<.05) to .45 ( p<.01). The larger correlations were reported for

interpersonal sensitivity and emotional expression, both of which concern

interpersonal rather than intrapersonal EI.

Hypothesis three: Fantasy and EI. Support was not found for hypothesis three as

no significant positive relationships were reported between fantasy and the EQ-i

subscales. In fact, one significant moderate negative correlation was reported for

the subscale self control (r = -.34, p<.01). Self control was described in Table 5.2
 111

as ‘having some control over one’s thoughts and behaviour’ as opposed to the

fantasising associated with this IRI subscale.

Hypothesis four: Personal Distress and EI. Support was found for hypothesis four

as significant negative relationships were reported between personal distress and

all eight EQ-i subscales. P Pearson’s r ranged from -.21 (p<.05) with

interpersonal sensitivity to .54 ( p<.01) with assertiveness.

5.4.2 Joint factor analysis

To further understand the relationships between the IRI and EQ-i, a second joint

factor analysis of the two was conducted using the four IRI subscale scores and

eight EQ-i subscale scores. Using exactly the same procedure as for the joint

factor analysis in the previous study, pre-analysis checks were carried out to

check that a stable factor structure could be produced; that the variables are

appropriately scaled and distributed, and that there is systematic covariation

within the data (Ferguson & Cox, 1993). The absolute sample size here of 297

participants is again suitable (Kline, 1994). Skew and kurtosis of the variables

were found to be within the acceptable limits of the parameters suggested by

Muthen and Kaplan (1985). The Kaiser-Meyer-Olkin [KMO] test of sampling

adequacy was found to be 0.89 and Bartlett’s test of sphericity was significant

(1556.10, p<.0001). The data were therefore deemed suitable for exploratory

factor analysis.
 112

Using parallel analysis, four analyses were conducted at the 50th and 95th

percentiles using 40 and 100 sets of randomly generated data. The parallel

analysis indicated that two factors should be retained. However the two factor

solution produced cross loadings for six of the 12 variables in the analysis, with

only two of the variables loading primarily on to Factor 2. These two variables

were from the IRI (EC and F). Instead, a single factor solution was inspected and

accepted as making more theoretical sense. The single factor was extracted using

principal components analysis (Tabachnick & Fidell, 2007). With an eigenvalue

of 5.57, the factor accounted for 46.4% of the variance. Component saturation of

this single factor was 0.59. Factor loading are presented in Table 5.5.

Table 5.5: Factor loadings for IRI and EQ-i subscales

Factor Loadings
Rationality .84
Self Esteem .84
Flexibility .78
Emotion regulation .78
Interpersonal sensitivity .75
Assertiveness .74
Emotional Expression .74
Self control .69
Perspective Taking [IRI] .58
Personal Distress [IRI] -.56
Empathic concern [IRI] .35
Fantasy [IRI] -.17
 113

In support of hypotheses one, two and four, this analysis suggests that the eight

EQ-i subscales are best represented by a single factor which is positively

associated with Perspective Taking and Empathic Concern and negatively

associated with Personal Distress. Not in support of hypothesis three, the fantasy

subscale of the IRI did not load on to the EI factor.

5.5 Discussion

This study has investigated the relationship between each of subscale of the IRI

and emotional intelligence. Perspective taking and empathic concern were found

to be positively related to EI while personal distress was negatively related and

fantasy broadly unrelated to EI. Similar patterns of results were found for two of

the subscales, Perspective Taking and Empathic Concern in terms of

relationships with EI. These are discussed together, followed by findings for

Personal Distress and then the Fantasy subscale which, as in Chapter four,

resulted in substantially different patterns of relationships.

5.5.1 Perspective Taking and Empathic Concern

The joint factor analysis with the emotional intelligence measure found that

Perspective Taking and Empathic Concern significantly positively loaded on to a

single factor of emotional intelligence, suggesting that people who report high

Perspective Taking and Empathic Concern are also more socially oriented with

greater tendencies for emotion perception and regulation. Perspective Taking

most highly correlated with all of the EQ-i scales and had a stronger factor

loading of .58. Empathic Concern only correlated with significantly with 5 of the
 114

8 EI subscales and had a lower factor loading of .35. Perspective taking therefore

appears more strongly related to EI. This is in line with previous research which

relates perspective taking more closely to EI (e.g. Charbonnueau & Nicol, 2002;

Stratton et al., 2005).

As in Chapter four, these relationships were in line with hypotheses of expected

descriptions for empathic individuals, providing support for the construct validity

of the two subscales as part of a measure of multidimensional empathy including

both cognitive and congruent affective processes (Davis, 1996; Vreeke & van der

Mark, 2003). This study also found that the personal distress subscale showed

different relationships with EI compared to the perspective taking and empathic

concern subscales.

5.5.2 Personal Distress

In fact, the personal distress subscale was found to correlate negatively with all

of the subscales of the EQ-i and loaded negatively on to the EI factor in the joint

factor analysis. This is a new finding in terms of the literature on emotional

intelligence which has only described a positive relationship between empathy

and EI. The distinction between the different dimensions of empathy is therefore

useful in discussions of emotional intelligence. In this study, EI was found to be

related to those aspects of empathy which involve cognitive processes as well as

a congruent display of positive, compassionate emotions. However, empathy as

an automatic, emotional reaction of distress (Hodges & Wegner, 1997) was

found to be negatively related to EI, which requires more control over one’s

emotional reactions (e.g. Bar-On, 1997). This finding provides new evidence that
 115

the two concepts, while showing some areas of overlap, also show a key

difference. In light of this, it will be interesting to understand the implications of

this emotional response within the healthcare context. Before considering this

further, the results of the final section of the analysis require discussion.

5.5.3 Fantasy

The imaginative aspect of the IRI does not appear to be related to self reported

emotional intelligence as the fantasy subscale was neither correlated with, nor

loaded on to the factor of, emotional intelligence. This is unsurprising given

Davis’ original rationale for the subscale in the first place (1983). Both the

fantasy and personal distress subscales were defined as being ‘self oriented’ as

opposed to the ‘other oriented’ subscales of Perspective Taking and Empathic

Concern. Indeed, the ‘other oriented’ subscales are the only ones to show

positive relationships to emotional intelligence.

The fantasy subscale was originally included in the IRI as it supposedly related

to increased emotionality. Physiological evidence of a link between fantasy and

greater emotional responding to the emotions of others was provided by Stotland

et al., 1978. Furthermore, Cliffordson (2002), in confirming the four factors of

the IRI, also generated a one factor model of empathy which significantly

predicted scores on all four subscales including Fantasy. However, Cliffordson

also recognised the need to extend her findings as each subscale consists of

narrow dimensions and their relationships to other personality constructs were

not clear. Indeed, there are those who have used the IRI without the Fantasy

subscale, arguing for its lack of relevance to the topic of research (e.g.,
 116

Christopher, Owens & Stecker, 1993; Stratton et al., 2005). The mixed research

findings regarding this subscale will require further investigation.

5.5.4 Limitations

It should be noted that the original research design intended to gather data on the

IRI and the NEO PI-R for a sample of doctors, in order to confirm the results of

the factor analysis in the previous study. However this was not possible on this

occasion as participants were in an assessment centre context and the time and

effort required from participants gather this data was not realistic. This is one of

the limitations of attempting to conduct field research and so attempts were made

to overcome this by utilising a shorter measure of individual differences, namely

that of EI. In itself, this was a useful theoretical comparison to make. The further

point to note is that the choice of EI measure was limited somewhat within the

context of this research. Many existing measures of EI are commercially

marketed and for financial reasons were not practical to use within this thesis. Of

those that were available, as EI research is still very much in its infancy, limited

evidence was available to guide the choice of a suitable measure. While the Bar-

On EQ-i is still widely used in practice and research, it would be useful to

replicate findings with a further trait measure of EI such as the TEIQue (Petrides,

Perez-Gonzalez & Furnham, 2007).

5.5.5 Summary and next steps

This study makes several important contributions to our understanding of

empathy as measured by the IRI. Taken as a whole, it has been found that Davis’
 117

multidimensional model of empathy is shows overlap with, but also differences

between, emotional intelligence. First, the evidence regarding the construct

validity of the perspective taking and empathic concern subscales is positive,

with all hypotheses being supported for these subscales. Clear links were

established between these subscales and emotional intelligence. Secondly, a key

difference between empathy and emotional intelligence was established. Personal

distress was found to load negatively with EI while fantasy was found to be

unrelated to EI. Chapter three ended by identifying two key aims of this thesis as

follows:

1. What are the antecedents of empathy in healthcare practitioners, in terms

of individual differences and situational factors?

2. What behaviours are associated with empathy in the healthcare

practitioner?

The first two studies have therefore addressed the first aim only, by investigating

the individual differences that are associated with empathy. While empathy

involves automatic, emotional reactions, emotional intelligence involves control

over those reactions (Bar-On, 1997; Hodges & Wegner, 1997). It will be

important to investigate which concept, empathy or EI, has the greater utility in

terms of the healthcare context by examining the links between empathy, EI and

behaviour. This will begin to address the second aim of the thesis and is the focus

of the next study.

 118

Chapter 6 – Self-report empathy and other-rated

empathic behaviour

Chapters four and five investigated antecedents of empathy focusing on

personality and emotional intelligence. Findings suggest that (1) perspective

taking and empathic concern are closely associated with agreeableness and

extraversion, and also load positively onto the single factor of emotional

intelligence (2) fantasy is associated with openness to experience but not

emotional intelligence, and (3) personal distress was positively related to

neuroticism and negatively related to emotional intelligence. Having established

this, the thesis now turns to examine the relationship between individual

differences and empathic behaviour in the healthcare context, and explore the

potential utility of measures of empathy and emotional intelligence in the

selection and development of healthcare professionals. In order to test these

relationships, it is first necessary to revisit the model of empathy being tested in

this thesis, focusing on the potential pathways to empathic behaviour.

6.1 Empathy and behaviour

Although Davis (1996) identifies ‘interpersonal outcomes’ as the final stage in

his multidimensional model of empathy, in their adaptation of Davis’ model (see

Figure 6.1) Larson and Yao (2005) differentiate between ‘interpersonal

processes’ and the final stage of empathy as positive outcomes for patients and

practitioners. According to both models, a patient’s judgment of empathy in a

healthcare practitioner will depend not only on whether the practitioner

 119

Antecedents

• Physician characteristics

• Patient characteristics

• Situational characteristics

Empathic Processes

• Non-cognitive processes e.g.

mimicry

• Cognitive Processes e.g.

association, role-taking

Intrapersonal Processes Interpersonal Processes

• Affective reactions • Helping behaviour

Parallel emotions
Reactive emotions
• Conflict management

• Non-affective reactions
Attributions • Social behaviour

Physician Outcomes Patient Outcomes

• Job satisfaction • Satisfaction

• Burnout • Healthcare outcomes

Figure 6.1: Process model of clinical empathy (from Larson & Yao, 2005).
 120

understands the perspective of patients and their emotional response, but also the

extent to which the practitioner demonstrates this understanding in empathic

behaviour towards the patient. Therefore, whilst a doctor may rate themselves as

highly empathic, outcomes such as patient trust, compliance with medical

treatment, quality of care relationships and satisfaction with medical services will

depend on patients’ judgements that the doctor is empathic, which in turn will

depend on the doctor’s behaviour (Barnett, Howard, King & Dino, 1981; Becker

& Maiman, 1975).

There is a growing call for the assessment of such interpersonal behaviours as an

indicator of medical competence alongside clinical knowledge. For example,

Epstein and Hundert (2002) propose that professional competence in doctors

should be defined as “the habitual and judicious use of communication,

knowledge, technical skills, clinical reasoning, emotions, values and reflection in

daily practice for the benefit of the individual and the community being served”

(p.226). In addition, the CARE measure, reviewed in Chapter three, now forms

part of the method for appraising General Practitioners in Scotland. Further work

in the UK has sought to identify the behavioural competencies needed by

physicians in order to deliver good medical care, and among the competencies

for General Practitioners identified by Patterson, Ferguson, Lane, Farrell,

Martlew & Wells (2000) is ‘empathy and sensitivity’ which appears to relate

closely to the ‘interpersonal outcomes’ identified by Davis (1983) and Larson

and Yao (2005). Two further competencies, ‘communication skills’ and

‘professional integrity’ also reflect helping and social behaviours and may

therefore reflect components of empathy. This study tests

 121

whether interpersonal outcomes, specifically physician empathic behaviour, are

predicted by the empathic disposition as measured by the IRI. The study

therefore represents a test of the criterion-related validity of the IRI subscales.

6.2 Criterion-related validation

The third and fourth types of validity identified by Cronbach and Meehl (1955)

are predictive and concurrent (also known as ‘criterion-oriented’) validity. In

order to assess these, a correlation is calculated between the predictor of interest

and an independent criterion measure. When test scores and criterion scores are

measured at the same time, this is known as concurrent validity, whereas

predictive validity studies involve the criterion score being taken at a later date.

Landy and Conte (2007) note that the usual limitation of concurrent compared to

predictive validity studies is that if current employees in a particular role are

used, then range is restricted as only those with higher test scores are sampled.

However, an opportunity arose for concurrent validity data to be collected in a

sample of applicants for a role. This limitation was therefore not an issue and this

is the approach taken in this study.

Given the focus on empathy as an intra-psychic phenomenon it is not surprising

that comparatively few studies have investigated whether any of the self-report

empathy questionnaires are positively related to empathic behaviour as judged by

observers. Even less research has tested these relationships among practitioners

in the healthcare context. Of the few studies that have been conducted in the

healthcare setting, the predominant methodology has involved quantitative

 122

instruments with a relatively narrow or peripheral scope (Pedersen, 2009), and

often focusing on either cognitive or emotional components of empathy. For

example, Hojat et al. (2005) found that a measure of attitudes towards cognitive

empathy predicted later ratings of empathic behaviour during medical training.

Only one study has been located which aimed to compare self report with others’

judgements of empathy using the IRI. Carmel and Glick (1996) asked 324

physicians within an Israeli hospital to identify colleagues who were

‘Compassionate Empathetic Physicians’ [CEP], defined as those whose “pattern

of behavior reflects strong devotion to the welfare of patients on two crucial

dimensions of patient care: the scientific-technical and socio-emotional, or, as it

is often put, curing and caring” (p.1253). Those most frequently identified were

put into the high CEP category, while those least frequently identified were put

into the low CEP category. Results indicated that the high CEP group scored

significantly higher on perspective taking and significantly lower on the personal

distress subscale of the IRI. This supports research from other organisational

contexts which has also found a positive link between perspective taking and

interpersonal relating in the workplace (Parker, Atkins & Axtell, 2008). No

significant differences were found for the empathic concern or fantasy subscales

by Carmel and Glick. However this study did not use actual measures of

behaviour and so the reliability of the categorisation process is unclear.

Furthermore, the perspective taking and empathic concern subscales for the IRI

have been found to be closely related in the first two studies of this thesis, in line

with previous research. For example, Axtell, Parker, Holman and Totterdell

(2007) asked 347 agents from two UK call centres to complete self ratings of

perspective taking and empathic concern. Both of these significantly predicted

 123

managers’ ratings of helping behaviour. A similar result was found in a study of

manufacturing employees in helping internal customers, suggesting that the

results of the study may be generalisable (Parker & Axtell, 2001). In study three

it is therefore hypothesised that these two subscales (empathic concern and

perspective taking) will be positive related to ratings of empathic behaviour as

judged by others.

Hypothesis one: perspective taking will be positively correlated with

ratings of empathic behaviour.

Hypothesis two: empathic concern will be positively correlated with

ratings of empathic behaviour.

With respect to the fantasy scale of the IRI, Davis neither expected nor found any

relationship between this subscale and measures of interpersonal functioning,

commenting that ‘‘it is not apparent that a tendency to become deeply involved

in the fictitious world of books, movies, and plays will systematically affect one’s

social relationships’’ (Davis, 1983, p.123). However, Stinson and Ickes (1992)

found that those scoring higher on the fantasy scale also performed better on a

task of empathic accuracy in an interaction with a stranger. As healthcare

practitioners are often interacting with patients with whom they are not well

acquainted, it may be that a general tendency to imagine oneself in the shoes of a

stranger, as with a fictional character, could aid perspective taking and therefore

enhance the empathic process. The following hypothesis was therefore made:

Hypothesis three: fantasy will be positively correlated with ratings of

empathic behaviour.
 124

Turning to the fourth IRI subscale, Davis’ suggests that emotional arousability,

characterised by feelings of personal distress when observing the distress of

another, motivates the observer to help that person, and therefore acts as a

mechanism to relieve the observer’s distress. This subscale was found to be

positively related to neuroticism in study one, a personality trait that has

previously been associated negatively with work performance in general (e.g.

Salgado, 1997; Tett et al., 1991) and Carmel and Glick found that physicians

identified as more compassionate and empathetic by their colleagues were

actually found to score significantly lower on personal distress. In the healthcare

context, where objectivity is considered crucial to making accurate diagnosis and

treatment decisions, Hojat et al. (2001) argue that personal distress is the kind of

uncontrolled emotional response that could interfere with objectivity. Indeed, it

has been argued that patients will not perceive this type of emotional response as

helpful at a time when they are seeking reassurance (e.g. Morse et al., 2006).

This study therefore predicts that a self-rated tendency to experience personal

distress will negatively related to ratings of empathic behaviour in the healthcare

context.

Hypothesis 4: Personal Distress will be negatively correlated with

ratings of empathic behaviour.

As a key difference between empathy and emotional intelligence was found with

respect to personal distress in study two, it is relevant to consider whether

emotional intelligence may be a more appropriate concept to use in an applied

healthcare setting. Responsiveness to the emotions of others while controlling

one’s own emotions appears to be a potentially useful mechanism of combining a

 125

caring approach with the ability to maintain objectivity. Criterion-related validity

has been explored in some depth in relation to emotional intelligence in recent

years (e.g., Salovey & Mayer, 1990; Goleman, 1995), with EI found to play a

role in successful task performance (Lam and Kirby, 2002), academic

performance (Petrides, Frederickson & Furnham, 2004) and social competence

(Mavroveli, Petrides, Rieffe & Bakker, 2007). EI has also been suggested as an

important skill in medicine (Elam, Stratton & Andrykowski, 2001; Stratton et al.,

2005). However no study has been located to date that tests the relationship

between self-reported EI and empathic behaviour as judged by others in the

healthcare context. This study therefore investigated whether physicians who rate

themselves higher in emotional intelligence will also be evaluated more

positively in terms of their demonstrated empathic behaviours.

Hypothesis 5: Self-report emotional intelligence will be positively

associated with ratings of empathic behaviour.

6.3 Method

6.3.1 Context and participants

The data for study three were collected during an assessment centre to select

doctors applying to train as General Practitioners (GPs) in the NHS. Two

hundred and fifty six applicants applying for GP specialty training in one UK

deanery were invited to attend assessment centres conducted over a one week

period. On arrival at the assessment centre, all applicants were invited to

participate on a voluntary basis. Information sheets were disseminated to briefly

 126

explain the purpose of the research study and applicants were assured that any

information from the psychometric questionnaires used for the research would

not be made available to those making selection decisions and as such the

research did not form part of the selection process. Consent forms were signed by

all participants to indicate their understanding and agreement to take part. (See

Appendices 2 and 3 for the information sheet and consent form). Out of 256

applicants attending the assessment centres, 192 doctors agreed to take part in the

study and completed the measures detailed below. Half of the participants

completed the self-report measures before the assessment centre exercises, with

the other half completing the measures afterwards. This was primarily for

logistical reasons but also helpful in controlling for potential order effects. Of the

192 doctors taking part, 109 (56.77%) were male and 83 (43.23%) female. 32.8%

described themselves as White, 45.8% as Indian/Pakistani/Bangladeshi, 10.4% as

Asian/Chinese, 3.1% as Black and 7.8% as Multiracial/Other. Mean age was

30.55 years (SD 5.34 years). Sixty seven doctors (36.5%) had completed their

medical training within the UK and Ireland, while the other 125 were trained

overseas. Countries of qualification for those trained overseas were

predominantly India (35.4%) and Pakistan (10.3%) with the final 17.8%

qualifying in a wide range of countries including Sri Lanka, Nigeria, Iraq, Libya

and Myanmar.

The sample therefore provides an opportunity to gain greater understanding of

empathy and cross-cultural differences in a medical selection process. This is an

important area for consideration in a time when many medical professionals in

the UK have been recruited internationally. Participants in this study were

 127

doctors all currently resident and working in the UK. Many were however

trained in and native of different countries, particularly India, Pakistan and

Bangladesh. In 2008, the Home Office and Department of Health moved to

prevent overseas doctors from registering to work within the NHS, in an attempt

to preserve jobs for UK graduates. By this time however, some 277,000 overseas

doctors were already registered with the General Medical Council, with almost

half of these obtaining their original medical qualifications abroad (Hawkes,

2008). This is partly a result of overseas recruitment in response to a shortage of

qualified UK staff. In selecting overseas doctors, the only stipulation in addition

to a recognised medical qualification was that recruits were required to be

linguistically proficient, although those from within the European Economic

Area do not have to be assessed on language ability. (MacDonald, 2003). It will

therefore be interesting to see if there are cross cultural challenges in empathy for

overseas qualified doctors.

6.3.2 Measures

1. Empathy: The Interpersonal Reactivity Index (IRI; Davis, 1983): This is the

same measure used in Chapters four and five, chosen because of its positive

evaluations of reliability and validity. To recap, the Perspective Taking subscale

of the IRI is representative of an individual’s tendency to adopt the views of

another. The Fantasy subscale is similar to this, although is based on imagining

oneself in the role of characters in books, films or plays. The Empathic Concern

subscale asks about the individual’s own feelings of concern in response to

another person. The Personal Distress subscale is also emotional, but focuses on
 128

how much one feels distress in response to another. The four subscales

(Perspective Taking, Empathic Concern, Fantasy and Personal Distress) are each

composed of seven items, to which participants are asked to respond using a five-

point likert scale (‘does not describe me well’ to ‘describes me very well’). All

four sub-scales of the IRI have been shown to have satisfactory internal

reliability (α = 0.71 to 0.77) and test – retest reliability (α = 0.62 to 0.71, Davis,

1983). A 0-4 scale is used for each item, so the minimum possible score for each

subscale is zero, with a maximum of 28. The full questionnaire is included in

Appendix 1.

2. Emotional Intelligence: The Bar-On EQ-i (1997): Again, this measure was

used in Chapter five to investigate the construct validity of the IRI. It measures

an “array of non-cognitive capabilities, competencies, and skills that influence

one’s ability to succeed in coping with environmental demands and pressures”

(Bar-On, 1997, p. 14). Respondents are asked to rate items using a 5 point Likert-

type scale (where 1 = ‘Very seldom or not true of me’ and 5 = ‘Very often true or

true of me’). Chapter five produced an eight factor solution using 107 items of

the measure. Given that the participants in this study formed part of the sample

for Chapter five, it was deemed appropriate to use this factor structure for the test

of EI in this study. Eight subscale scores were therefore calculated: self esteem

(21 items); self control (11 items); assertiveness (17 items); rationality (15

items); sensitivity (12 items); emotional expression (10 items); emotional

regulation (11 items), and adaptability (10 items). Subscale scores were

calculated by reverse scoring negatively worded items and calculating the mean

score from the items for each scale. Mean scores were used to account for the
 129

differing numbers of items in each subscale. Reliabilities for all of the subscales

created in Chapter five were deemed to be satisfactory, with αs ranging from

0.78 to 0.94. An EQ-i total was calculated by summing these means. The

minimum possible score for each subscale is therefore one and the maximum is

five, while the minimum possible EQ-i total score is eight and the maximum is

40.

3. Cognitive ability: Ravens Advanced Progressive Matrices Set II (Raven, Court

& Raven, 1994): This was administered according to the standard instructions.

This questionnaire was considered suitable for exploring intellectual efficiency in

a cross-cultural context because it is relatively language free. Participants had 40

minutes to complete 36 items, each of which consisted of a pattern with a part

missing. The task is to identify the correct missing part from a range of eight

possible options. The internal consistency reliability of Ravens is estimated at

0.90 (Raven et al., 1994). Each correctly answered item receives one point, such

that the minimum possible total score is zero while the maximum is 36.

4. Empathic behaviour: Assessment centre ratings: Ratings of participant

behaviour were obtained from two sources, trained assessors and medical actors

playing the role of patients within a simulated consultation. These assessors

provided ratings of interpersonal behaviour during two of the assessment centre

exercises, a group discussion and a simulated patient consultation. The group

discussion exercise required participants to work as part of a team of four Senior

House Officers in a city hospital who must prioritise which patients should

receive surgery the following morning. In the simulated consultation exercise,

 130

participants were observed in a one-to-one dialogue with a ‘patient’, played by a

medical actor, who has just been diagnosed with cancer. ‘Patients’ followed a

pre-scripted dialogue in which they were instructed to express a range of

emotions including anger, fear and confusion, while responding naturally to the

interaction with the participant. In each of the exercises, the participant was

observed by a trained assessor who then rated them using behavioural indicators

from the competency framework for General Practitioners generated by previous

research (Patterson et al., 2000). Trained assessors were either medical

professionals or lay assessors, in this case a role fulfilled by Ofsted inspectors.

Three of the competencies included interpersonal behaviours: ‘empathy and

sensitivity’; ‘communication skills’, and ‘professional integrity’. Definitions and

sample indicators are provided in Table 6.1. After each exercise, assessors rated

participants on each of the three competencies using a 1-4 Likert-type scale

(where 1 = few positive behaviours, many negative behaviours, and 4 = no

negative behaviours and many positive behaviours). For each interpersonal

competency, scores from the two exercises were combined. The minimum

possible score for each competency is therefore two, while the maximum is eight.

After the simulated consultation, the ‘patients’ also rated the participants on two

of the interpersonal competencies, ‘empathy and sensitivity’ and ‘communication

skills’. For each competency, patients were given 5 statements and asked to rate

the participant on a 1-4 Likert scale (1 = strongly disagree, 2 = disagree, 3 =

agree, 4 = strongly agree). For Communication Skills, statements included ‘This

doctor communicated effectively with me’, ‘The nature of my problem was

explained/clarified clearly’ and ‘I had adequate opportunity to express my

 131

Table 6.1 Definitions and sample behavioural indicators of competencies related to empathy (from Patterson et al 2000).

Competency Definition Sample positive indicators Sample negative indicators

Empathy and Desire and ability to take in Demonstrated a caring manner towards Showed very little visible
sensitivity [ES] the perspective of others, and others interest/understanding.
sense associated feelings, Was clearly intent on establishing exactly Was quick to judge, make assumptions.
generating a safe, reassuring what others were thinking or feeling Appeared isolated or authoritarian.
atmosphere Was perceptive, responding to the Lacked warmth in voice/manner and
concerns of others with understanding failed to encourage patient
Clearly reassured others with appropriate Created uncomfortable atmosphere
words and actions

Communication Ability to engage others, Actively encouraged others through use of Failed to use supportive words or
Skills [CS] clearly and actively, in supportive words or comments comments to encourage others
constructive dialogue, Used open exploratory questions inviting Asked closed questions, restricting
adjusting language and non- others to become actively involved opportunities for others to become
verbal behaviour according to Adjusted language as appropriate to suit involved
the needs of differing particular needs of the situation Unable to adapt language to suit particular
situations needs of the situation

Professional Professional commitment (i) to Showed clear respect for others (whether Appeared to lack sufficient respect for
integrity [PI] provide equality of care for all, through words or actions) others (whether through words or actions)
(ii) to take responsibility for Was positive/enthusiastic during the Approached the exercise defensively,
own actions – while at the exercise, however challenging it seemed more as a problem than a challenge
same time recognising the When appropriate, was open and Appeared judgmental, not prepared to
parameters of one’s role and accepting of the particular situation of consider each situation on its merits
expertise, (iii) to act others
confidently but safely
 132

concerns’. For Empathy and Sensitivity, statements included ‘This doctor was

sensitive to my feelings’, ‘This doctor seemed to understand my

situation/concerns’ and ‘I felt at ease with this doctor’. A mean score was then

taken of these five statements, thus the patient ratings for each competency

ranged from 1-5. Unfortunately the assessment centre administrators did not

provide raw scores for every statement so a reliability analysis of these subscales

was not possible.

6.4 Results

Table 6.2 presents the means and standard deviations of the variables. Before

testing the relationships between the variables, distributions were considered.

Kolmogorov-Smirnov tests of normality for all variables were significant.

However, these tests are often significant with a large sample size and so

histograms should be inspected to assess suitability for parametric tests (Field,

2005). In addition, the skewness and kurtosis were converted to z-scores. As

large sample sizes give rise to small standard errors, z scores are large and so

significant z scores are found from small deviations from normality. In this case,

the criterion of 2.58 was used to represent a significant deviation from normality

(Field, 2005). All of the z scores for both skewness and kurtosis were less than

2.58, taken as support for the assumption of normality, for the subscales of the

IRI, EQ-i, RAPM scores and assessor ratings of behaviour from the three

interpersonal competencies. The histograms for these variables were deemed

acceptable for using parametric tests. However, the z scores were greater than

2.58 for age and the two patient competency ratings. The patient ratings for both
 133

competencies were significantly negatively skewed with a clear ceiling effect.

Age was also significantly positively skewed as most of the doctors were at the

beginning of their medical careers at the stage of choosing a specialty with the

minority coming to this point later in their careers. Transformations were not

successful in reducing this, so tests using these variables were non-parametric.

Table 6.2.Descriptives all variables (n=192).

Variable Mean SD

Age (Years) 30.55 5.39

Ravens Advanced Progressive Matrices 21.72 5.16

Empathy & Sensitivity-Assessor 5.91 1.53

Communication Skills- Assessor 5.92 1.43

Professional Integrity- Assessor 6.08 1.13

Empathy & Sensitivity-Patient 2.81 .82

Communication Skills-Patient 2.97 .79

IRI-Fantasy 14.22 5.96

IRI-Empathic Concern 22.09 3.50

IRI-Perspective Taking 19.85 4.22

IRI-Personal Distress 9.98 4.51

EQ-i Total 31.51 3.25

EQ-i Self Esteem 4.09 0.59

EQ-i Self Control 4.07 0.52

EQ-Assertiveness 3.75 0.55

EQ-i Rationality 3.91 0.43

EQ-i Interpersonal Sensitivity 4.32 0.46

EQ-i Emotional Expression 3.91 0.63

EQ-i Emotion Regulation 3.95 0.63

EQ-i Flexibility 3.51 0.44

 134

6.4.1 Correlations

Table 6.3 then presents correlations between the variables for all 192 doctors.

Correlations for age and the patient ratings are Spearman’s rho non-parametric

correlations, with all others reported being Pearson’s r. Age correlated

moderately negatively with intellectual efficiency as measured by the Ravens

Advanced Progressive Matrices [RAPM] (rho = -.37, p<.01) as well as assessor

ratings of all three competency scores (Empathy and Sensitivity [ES]: rho = -.36;

Communication Skills [CS]: rho = - .39; Professional Integrity [PI]: rho = -.40,

all p < .01). There were also small negative correlations between age and patient

ratings of ES (rho = -.14, p<.05) and CS (rho = -.21, p<.01). Age also showed a

small significant negative correlation with the Fantasy subscale of the IRI (r = -

.17, p<.05). The RAPM scores were significantly positively correlated with the

three competency scores as rated by assessors (ES: r = .36; CS: r = .37; PI r =

.43, all p < .01). These are all moderate positive correlations (Cohen, 1988). As

anticipated, the RAPM scores were not significantly correlated with any of the

subscales of either IRI. There was only one small negative correlations between

RAPM and one subscale of the EQ-i, self control (r = .19, p<.05). The positive

correlations between the Perspective Taking and Empathic Concern subscales of

the IRI with the various subscales of the EQ-i, again illustrate the overlap

between these elements of empathy and EI while the consistently negative

significant correlations between Personal Distress and EI demonstrate the key

difference between the two concepts.

 135

Table 6.3 Bivariate correlations for all variables (n = 192)

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

1. Age

2. RAPM -.37**

3. Empathy & Sensitivity-A -.36** .36**

4. Communication Skills-A -.39** .37** .85**

5. Professional Integrity I-A -.40** .43** .79** .76**

6. Empathy & Sensitivity-P -.14* .20** .47** .46** .36**

7. Communication Skills-P -.21** .25** .52** .46** .48** .74** .

8. IRI-Fantasy -.22** .13 .23** .19* .25** .14 .08

9. IRI-Empathic Concern .10 -.05 -.03 -.07 .01 -.05 .02 .08

10. IRI-Perspective Taking .14 -.03 -.01 -.07 .00 .09 .03 -.01 .39**

11. IRI-Personal Distress -.10 -.13 -.24** -.15 -.12 -.18* -.25** .10 .01 -.22**

12. EQ-i Self Esteem .09 -.05 -.04 -.05 -.02 -.04 -.03 -.03 .25** .35** -.31**

13. EQ-i Self Control .08 -.19* .01 .00 .02 -.02 -.07 -.22** .27** .30** -.29** .56**

14. EQ-Assertiveness -.11 -.03 .17* .16* .09 .16* .21** -.14 .05 .20** -.52** .60** .40**

15. EQ-i Rationality .32** -.12 -.02 .00 -.02 .01 -.05 -.06 .16* .41** -.36** .66** .36** .53**

EQ-i Interpersonal
16. -.09 .14 .14 .17* .13 .11 .16* .05 .47** .45** -.18* .69** .52** .45** .57**
Sensitivity

17. EQ-i Emotional Expression .08 -.11 .16* .13 .16* .05 .08 .06 .24** .27** -.31** .58** .46** .55** .48** .51**

18. EQ-i Emotion Regulation -.06 -.05 .14 .12 .10 -.11 -.13 -.02 .13 .42** -.46** .65** .57** .59** .54** .59** .52**

19. EQ-i Flexibility .07 .02 .03 .02 .01 .06 .05 -.15* .10 .29** -.33** .50** .42** .44** .47** .43** .31** .54**

Note: RAPM = Ravens Advance Progressive Matrices.For competency ratings, A indicates an assessor rating, P a ‘patient’ rating
 136

Individual differences and behaviour: assessor ratings

To assess the relationships between individual differences and behaviour,

correlations between the IRI, EQ-i and assessor ratings of the interpersonal

competencies were inspected. Few significant relationships were observed. From

the IRI, only the Fantasy subscale showed significant positive correlations with the

three competency scores (ES: r = .23, p < .01; CS: r = .19, p < .05; PI: r = .25, p <

.01). All of these correlations are small (Cohen, 1988). The Personal Distress

subscale showed a small negative correlation with one of the competency scores,

empathy and sensitivity (r = -.24, p<.01). Thus, hypotheses one and two

(perspective taking and empathic concern) were not supported while hypothesis

three regarding fantasy was supported and hypothesis 4 (personal distress) was

partially supported.

Some of the subscales of the EQ-i were correlated with assessor ratings of

behavior: Assertiveness showed small significant positive correlations with the two

of the three competency scores (ES: r = .17, CS: r = .16, both p < .05), and

Interpersonal Sensitivity was also positively correlated with the assessor rating of

Communication Skills (r = .17, p < .05). Emotional Expression was significantly

positively correlated with two of the three competency scores (ES: r = .16, PI: r =

.16, both p<.05), providing partial support for hypothesis five.

Individual differences and behaviour: patient ratings

To assess the relationships between individual differences and behaviour from the

patient perspective, the non-parametric correlations between the IRI, EQ-i and
 137

patient ratings of the interpersonal competencies were inspected. From the IRI, the

Personal Distress subscale was significantly negatively correlated with the patient

ratings of empathy and sensitivity (rho = -.18, p<.05) and communication skills

(rho = -.25, p<.01), providing support for hypothesis four. With regard to the EQ-i,

a similar pattern of results was found as with the assessor ratings. The

Assertiveness factor showed small significant positive correlations with the two

competency scores (ES: rho = .16, CS: r = .21, both p < .05). The Interpersonal

Sensitivity subscale was also positively correlated with the patient rating of

Communication Skills (rho = .16, p < .05). Partial support was therefore found for

hypothesis five from the patient ratings of behaviour.

6.4.2 Cultural differences: comparing groups

One of the potential reasons for the lack of a clear relationship between individual

differences and behaviour is the effect of cross-cultural interactions. As noted

previously, sixty seven doctors (36.5%) had completed their medical training within

the UK and Ireland, while the other 125 were trained overseas. The sample

therefore provided an opportunity to gain greater understanding of empathy and

cross-cultural differences in a medical context. This was not a hypothesised part of

the study, but as the opportunity arose, the decision was taken to investigate the

study hypotheses again, for both the sample of UK and Ireland trained doctors (n =

67) and those trained overseas (n = 125). The group of doctors trained overseas

were from a range of countries, predominantly India, Pakistan and Bangladesh.

Although not a homogenous group, they differ from the UK doctors in that they did
 138

not receive their initial medical training within the healthcare system in which they

are currently applying for GP training. This is the key difference being explored in

comparing the two groups.

Before doing this, tests for differences between the two groups on the key variables

were conducted. As age and RAPM scores were identified in Table 6.3 as

covariates, it was first necessary to check if applicants trained in the UK and

Ireland differed from overseas applicants on these variables. To test differences in

age, a non-parametric Mann Whitney U test was conducted as the assumption of

normality was not met. This found that applicants trained in the UK and Ireland

(Mdn = 26.00 years) were significantly younger than applicants trained overseas

(Mdn = 31.00 years), U = 1847.5, p<.001. To check differences in RAPM scores,

an independent t-test was conducted which revealed that applicants trained in the

UK and Ireland (M = 24.57, SE = .53) scored significantly higher than those trained

overseas (M = 20.20, SE = .44), t(190) = 6.10, p<.001.

In order to control for possible confounding effects of the differences in the

covariates, a randomised matched pairs design was used (Pedhazur & Schmelkin,

1991) to create two groups of doctors who had either trained in the UK and Ireland

(Group 1) or Overseas (Group 2). The groups were matched for age and RAPM

scores. In creating the two groups it was not possible to match all of the participants

therefore sample size for this set of analyses was 63 applicants in each group. To

confirm they were matched for age, a non-parametric Mann Whitney U test was

used because the assumption of normality was not met. This found that Group 1
 139

(Mdn = 27.00 years) and Group 2 (Mdn = 28.00 years) did not differ significantly in

age (U = 1708.5, ns). An independent t test also revealed no significant differences

in RAPM scores (Group 1: M = 24.13, SE = .50, Group 2: M = 23.13, SE =.43;

t(124) = 1.52, ns). To assess homogeneity of variance between the two groups, the

variance ratios were used rather than Levene’s test, which is sensitive to sample

size. All variance ratios were less than two indicating homogeneity of variance

(Field, 2005).

Group 1 was composed of 43 male and 20 female applicants. In terms of ethnicity,

90.5% described themselves as White, 4.8% as Indian/Pakistani/Bangladeshi, 3.2%

as Asian/Chinese and 1.6% as Black. Group 2 was composed of 44 male and 19

female applicants. 3.2% described themselves as White, 66.7% as

Indian/Pakistani/Bangladeshi, 14.3% as Asian/Chinese, 4.8% as Black and 11.1%

as Multiracial/Other.

To investigate differences between the two groups on the IRI, EQ-i and

competency scores, a MANOVA was conducted with one fixed factor (country of

qualification) and eight dependent variables (FS, EC, PT, PD, EQ-i total score,

assessor ratings of ES, CS and PI). All of the dependent variables were included in

a single analysis as there were hypothesised relationships between these variables,

which were partially supported in Table 6.3, therefore the variables were not

independent. A large multivariate effect of country of qualification was found (F =

26.40, df = 15, p<.001, η2 = .79). To investigate this effect further, a series of

 140

ANOVAs was conducted. Table 6.4 presents the means and standard deviations for

all variables for each group as well as F values from the ANOVAs.

Table 6.4 Descriptive statistics for UK and Ireland and Overseas trained doctors.

UK & Ireland Overseas

M SD M SD F
ES-A 7.14 1.01 5.54 1.27 57.62***
CS-A 7.13 .98 5.48 1.12 73.17***
PI-A 6.97 .90 5.83 .85 49.93***

F 16.48 6.09 12.63 5.51 11.96**

EC 21.78 3.41 22.27 3.48 .78
PT 19.59 3.75 20.30 4.59 1.02
PD 9.10 3.99 10.21 4.56 1.87

EQ-i total 23.00 1.28 23.29 1.62 1.21

Note: For each group, n = 63. ***sig p<.001, **sig p<.01

Large group effects were found for all three competency scores. Applicants who

trained in the UK and Ireland (Group 1) were rated significantly higher by assessors

on Empathy and Sensitivity (F = 57.62, df = 1, p<.01, η2 = .32), Communication

Skills (F = 73.17, df = 1, p<.01, η2 = .38) and Professional Integrity (F = 49.93, df

= 1, p<.01, η2 = .29) than doctors who trained overseas (Group 2). Despite the large

differences on the competency scores, the two groups did not differ significantly on

the three of the IRI subscales or the EQ-i. The only exception to this was the

Fantasy subscale where Group 1 scored significantly higher than Group 2 (F =

11.96, df = 1, p<.01, η2 = .09). This was a medium effect size. Additionally, non-

parametric tests were conducted to assess differences between the two groups on

the patient ratings of behaviour. For both competencies, applicants trained in the
 141

UK and Ireland (Mdn = 3) were rated significantly higher by patients than

applicants trained overseas (Mdn = 2). Non parametric tests confirmed that these

differences were significant (ES: U = 766; CS: U = 675.5, both p<.001).

6.4.3 Correlations by group

In order to explore the potential moderating role of the grouping variable on the

relationships between the self report questionnaires and the ratings of the

competency scores for the two groups, one method is to conduct hierarchical

regressions using interaction terms calculated by multiplying independent variables

with the grouping variable. However, due to the high number of independent

variables resulting and the relatively low sample size, there was insufficient power

to conduct this analysis (Field, 2005; Kline, 2000). Instead, partial correlations

were calculated controlling for age and RAPM scores. Age and RAPM scores were

controlled for because of the moderate correlations found between these variables

and some of the competency ratings (detailed in Table 6.3). The partial correlations

between IRI and EQ-i subscales and assessor ratings are reported in Table 6.5, and

patient ratings in Table 6.6. Fisher’s z transformation was then used to compare the

correlations (Howell, 2002). Pairs of correlations resulting in a z of at least 1.96 are

significantly different at the .05 level, with a z of 2.58 or more indicating a

significant difference at the .01 level. For this sample size, correlations needed to

differ by at least .35 to be deemed significantly different. Table 6.5 shows the z

values that reached significance.

 142

Individual differences and behaviour: assessor ratings

For the Perspective Taking subscale, the correlation for Group 1 was significantly

positive with Professional Integrity, (r = .29, p<.05) whereas for Group 2 the r

value was non-significant. This apparent difference was confirmed with a

significant Fisher’s z of 1.96 (p = .05). Partial correlations with Empathy and

Sensitivity and Communication Skills were non-significant for both groups,

although the correlation between perspective taking and empathy and sensitivity

approached significance (r = .20, p = .07). Only partial support was therefore found

for hypothesis one with group one and no support at all with group 2.

For Group 1, all correlations between the empathic concern subscale of the IRI and

the competencies were significantly positive (ES: r = .27, p<.05, CS: r = .28, p<.05,

PI r = .41, p<.01). Again for Group 2, none were significant. All pairs of

correlations were found to be significantly different (ES: z = 1.96, p = .05, CS: z =

2.35, p<.05, PI: z = 2.94, p<.01). Thus hypothesis two was fully supported for

group 1 only.

For Group 1, applicants trained in the UK & Ireland, the Fantasy subscale of the

IRI correlated significantly positively with Empathy and Sensitivity (r = .27,

p<.05), Communication Skills (r = .27, p<.05) and Professional Integrity (r = .37,

p<.01). For Group 2, this subscale did not correlate significantly with any of the

competency scores.
 143

Table 6.5 Partial Correlations controlling for age and intelligence between IRI, EQ-I and assessor-rated competency scores

Empathy & Sensitivity Communication Skills Professional Integrity

UK/Ir OS z UK/Ir OS z UK/Ir OS z
IRI
Fantasy .27* .02 .27* -.13 2.23* .37** -.05 2.40*
Empathic Concern .27* -.08 1.96* .28* -.14 2.35* .41** -.10 2.94**
Perspective Taking .20† -.03 .13 -.10 .29* -.06 1.96*
Personal Distress -.19 -.21 -.09 -.05 -.11 -.01

EQ-i Total
Self Esteem .21 -.03 .03 .01 .14 -.02
Self Control .13 -.07 .00 -.02 .07 -.07
Assertiveness .27* .14 .06 .15 .14 .05
†
Rationality .24 .01 .19 .10 .29* .10
Interpersonal Sensitivity .45** .06 2.33* .37** .16 .39** .16
Emotional Expression .29* .10 .18 .10 .32* .10
Emotional Regulation .23† .08 .04 .10 .13 .10
Flexibility .17 .11 -.01 .19 .01 .09
Note: For each group, n = 63; Larzalere and Mulaik (1977) adjusted *significant p<.05, ** p<.01, †p = .07
 144

For both Communication Skills and Professional Integrity, the correlations for Group

1 were significantly different from Group 2 (CS: z = 2.23, p<.05, PS: z = 2.40,

p<.05). Thus hypothesis three was fully supported for group 1 but not for group 2.

The final subscale of the IRI, Personal Distress, was not significantly positively

correlated with any of the competency scores for either group. Therefore hypothesis

four was not supported.

There were some significant findings when looking at the subscales of the EQ-i and

the assessor ratings of competency scores, but again only for Group 1. Assessor

ratings of Empathy and Sensitivity were significantly positively correlated with

Assertiveness (r = .27, p<.05), Interpersonal Sensitivity (r = .45, p<.01) and

Emotional Expression (r = .29, p<.05). The Interpersonal Sensitivity subscale of the

EQ-i was the only one for which this correlation differed significantly for Group 2 (z

= 2.33, p<.05). Again for Group 1, Communication Skills was significantly

positively correlated with Interpersonal Sensitivity (r = .37, p<.01). The last

competency, Professional Integrity was correlated significantly positively with

Rationality (r = .29, p<.05), Interpersonal Sensitivity (r = .39, p<.01) and Emotional

Expression (r = .32, p<.05). Apart from the significantly different correlation

between EQ-i Interpersonal Sensitivity and the Empathy and Sensitivity competency

score, none of the Fisher’s zs were significant looking at the pairs of correlations.

Thus partial support was found for hypothesis five with group 1 only.

.
 145

Table 6.6 Partial correlations controlling for age and intelligence between IRI, EQ-i and ‘patient’-rated competency scores

Empathy & Sensitivity-P Communication Skills-P

UK/Ir OS UK/Ir OS
IRI
Fantasy -.07 .04 -.04 -.12
Empathic Concern -.03 -.00 .10 .05
Perspective Taking .01 .20 -.02 .12
Personal Distress -.25* -.05 -.26* -.18

EQ-i
Self Esteem .15 -.08 .22 -.00
Self Control -.07 -.01 .03 -.07
Assertiveness .16 -.12 .04 -.01
Rationality .19 .11 .22 .08
Sensitivity .07 .11 .14 .19
Emotional Expression .08 .04 .31** .12
Emotional Regulation .14 .23 .07 .22
Adaptability .19 .12 .25* .15
Note: For each group, n = 63; Larzalere and Mulaik (1977) adjusted *significant p<.05, ** p<.01.
 146

Individual differences and behaviour: patient ratings

Partial correlations were calculated controlling for age and RAPM using rank scores

for the patient ratings (Pallant, 2007). The UK & Ireland group correlations did not

differ significantly from those in the overseas group, with all zs being less than 1.96.

It was found that personal distress was significantly negatively related to the two

competency ratings from patients for the UK and Ireland group only. For empathy

and sensitivity, r = -.25, p<.05 and for communication skills r = -.26, p<.05. Thus

partial support was found for hypothesis four but only for group 1. Also, in terms of

the EQ-i, emotional expression (r = .31, p<.01) and flexibility (r = .25, p<.05) were

positively related to communication skills for the UK and Ireland group only. Thus

hypothesis five was partially supported for group 1 but not for group 2.

6.5 Discussion

This study has built upon the first two studies to test the relationships between

individual differences in empathy, emotional intelligence and empathic behaviour as

rated by assessors and patients. Specifically, the study examined the relationship

between empathy as assessed using the Interpersonal Reactivity Index and the Bar-

on EQ-i measure of EI with assessor and patient ratings of interpersonal behaviours

demonstrated by doctors during an assessment centre. The main findings can be

summarised as follows:

1. Fantasy was positively associated with assessor ratings of empathic

behaviour and personal distress was negatively associated with patient ratings of

empathic behaviour. Few other relationships were apparent between self report
 147

measures of individual differences and others’ ratings of behaviours when

investigating the whole group of participants from the assessment centre.

2. Comparing doctors who had trained in the UK & Ireland with those trained

overseas revealed very little difference in terms of the self report measures of

individual differences. However, doctors who qualified in the UK and Ireland were

rated significantly higher on the interpersonal behaviours, by assessors and patients,

than doctors who trained overseas.

3. Self-reported empathy correlated significantly with observers’ ratings of

empathic behaviour for UK and Ireland trained doctors, but not for overseas trained

doctors. Specifically, fantasy, empathic concern and, to a lesser extent, perspective

taking were all positively related to assessor ratings of behaviours for the UK and

Ireland trained doctors.

4. Various aspects of self reported EI (Interpersonal sensitivity, emotional

expression) also correlated significantly with observers’ ratings of empathic

behaviour for UK and Ireland trained doctors, but not for overseas trained doctors.

5. There were more clear relationships between the self report measures of

empathy and EI and assessor ratings of behaviour than between the self report

measures and patient ratings of behaviour.

There are various potential explanations for this set of findings, related to both

method and theory. Theoretical considerations relate to situational factors such as the

challenge of empathising cross culturally, which has been investigated to some

degree in this study, as well as the content of the competency ratings.

Methodological reasons include the assessment context and the process of

 148

assessment as well as the measures chosen. Theoretical explanations are considered

first.

6.5.1 Empathy and behaviour

The first finding was that fantasy was associated with assessor ratings of empathic

behaviour. This may be surprising to some in the medical community who have

chosen to ignore this aspect of empathy (e.g. Elam, Stratton & Andrykovski, 2001).

However the relationship was predicted as there is some evidence to suggest that

fantasy is associated with empathy for a stranger (Stinson & Ickes, 1992), as is the

case in the assessment centre exercises where applicants consult with an unknown

simulated patient and interact in a group with other applicants generally unknown to

them. It may be that as relationships become well established, fantasy becomes less

important although this has yet to be explored. The weak findings for perspective

taking are perhaps surprising although it may be that this scale becomes more

relevant as relationships are established. Additionally, this subscale has been shown

to be susceptible to social desirability (Charbonneau & Nicol, 2002), which may be

of particular relevance when completed in a selection context.

The negative relationship between patient ratings of behaviour and personal distress

is consistent with the argument that automatic emotional reactions, while they may

be a motivator of helping behaviour (Davis, 1983), are not perceived as helpful by

patients in this context. Indeed, the more controlled responses of assertiveness and

interpersonal sensitivity from the EQ-i were positively related to patient ratings of
 149

behaviour. It would therefore seem possible that in this context, EI is a more

appropriate concept than empathy.

6.5.2 Empathy and culture

The study also provided an opportunity to investigate cross cultural displays of

empathy by comparing doctors trained in the UK and Ireland with those trained

overseas, predominantly in more collectivist cultures. A recent article identified a

need to investigate EI cross culturally as most of the research to date has been

conducted in westernised cultures (Walter, Cole & Humphrey, 2011). Although

medical education has begun to develop frameworks for assessing cultural

competence (Betancourt, 2003), in the US and Canada it has been claimed that

“medical schools do not employ effective methods of training and evaluation to

ensure culturally competent care” (Zabar et al., 2006, p.510). Whilst this is a

potential issue needing investigation, surprisingly little research has taken place. The

term ‘ethnocultural empathy’ was proposed by Wang et al. (2003) to describe the

process of empathising with people from racial, ethnic or cultural groups different

from one’s own. Wang et al. argue that in a multicultural environment, empathy is

not universally applicable and practitioners require training in order to raise

awareness and understanding of how care may need to adapt for diverse groups.

Thus far, studies have established that those with more open attitudes towards

diversity training are more likely to report intentions to empathise with people from

diverse backgrounds (e.g. Brouwer & Boros, 2010; Cundiff, Nadler & Swann, 2009).
 150

Appreciating the thoughts and concerns of a patient from another culture may require

the doctor to elicit extra information from that person, as comparisons from one’s

own experiences will become less appropriate. Also, from a patient perspective, the

behavioural cues that patients use to determine whether the doctor is empathising

with them may be different. If this is the case, there could be important implications

for doctors trained in one country operating successfully in another. Although there

has been little investigation of this in relation to doctor-patient relationships, there is

some evidence from the general population and other healthcare settings such as

counselling to suggest that it is more difficult to demonstrate empathy cross-

culturally (Chi-Ying Chung & Bemak, 2002). For example, in spontaneous

interactions outside the formal helping relationship, Webster Nelson and Baumgarte

(2004) found that American college students were less able to take the perspective of

targets who were dissimilar from U.S. cultural norms. In addition, less compassion

and sympathy were reported for targets from an unfamiliar cultural perspective.

Within a formal helping relationship, in an investigation of counselling across

cultures, Sue and Sundberg (1996) found that counsellors who demonstrated an

understanding of their patients’ family and societal backgrounds and acknowledged

them to be different from their own were evaluated more positively by those patients.

The findings of the present study, by comparing partial correlations from the

randomised matched groups, support the possibility that there is a moderating effect

of country of training on the relationship between empathic disposition and

demonstration of empathic behaviour.

The different patterns of associations between self report empathy and EI measures

and ratings of behaviours are consistent with previous research which suggests that,
 151

regardless of motivation to empathise, demonstration of empathy is more difficult

cross-culturally. It is possible that for those doctors trained overseas, whilst

motivation to empathise is equal, ability to demonstrate this effectively for a patient

from a different culture is reduced as a result of less familiarity with that person’s

experiences and background. It was noted during the assessment centre that all of the

medical actors were of White British origin. It would be interesting to see if the same

findings were apparent if there were more of a mix in ethnicity among the medical

actors for simulated consultation exercise.

6.5.3 Limitations

It is not possible to definitively explain the large group differences found between

the UK and Ireland doctors and the overseas trained doctors. An alternative

explanation for these findings could be that demonstrated behaviours are not in fact

different from those operating cross-culturally, but that they are being evaluated

differently when the assessor and candidate are from different cultural backgrounds.

However, in an examination of assessor characteristics on scores given in an

assessment centre, race of assessor was not found to have a significant impact on

scores (Lowry, 1993). Furthermore, the assessment centre in this study used only

trained assessors with very specific behavioural criteria, which has been found to

reduce any reliance on stereotypes or hence biases from occurring (London, 2001).

Future studies are clearly needed to provide a greater understanding of the impact of

culture on medical performance. For example, investigations could include analysis

of the behaviours associated with empathy as identified by practising GP doctors

cross-culturally.
 152

There are also some limitations to note with regard to the method of assessing

behaviours. First, within the assessment centre exercises, participants are required to

demonstrate maximal performance whereas the self report measures are assessments

of typical preferences for behaviour. It is therefore reasonable to expect that there

would not be large correlations between what participants can do and what they

normally do. Evidence of poor correlations between assessments of maximal and

typical performance is well established (e.g. Sackett, Zedeck & Fogli, 1988). The

second issue of note is that each assessment centre exercise lasted for no more than

thirty minutes and was a one off ‘snapshot’ of behaviour regarding interactions with

strangers. According to classical measurement theory, longer assessments tend to be

more reliable (Rust & Golombok, 1999). In addition, due to the context, sources of

error in all measures may have included test anxiety (Fletcher & Kerslake, 1993). It

would therefore be preferable to gain an assessment of behaviours within the normal

working environment over a greater number of interactions.

The final point to note is that few relationships were apparent between the self report

measures and the patient ratings of behaviour. This may have been because of the

skewed data on these variables, or because the behaviours being assessed aren’t

those that actually predict judgments of empathy in patients. There is some evidence

to suggest that patients use different cues from assessors in judging empathy

(Silvester et al., 2007). Ultimately, it is the patients who are the consumers of

practitioner empathy in this context and as such, this question requires further

research. Although patients did play a role in the creation of the competency model
 153

(Patterson et al., 2000), their role was relatively minor and not focused specifically

on empathy but on the overall understanding of performance for GPs.

6.5.4 Summary and next steps

Having developed a clear picture of the individual differences associated with a

propensity to empathise, this study has conducted an examination of the relationships

between empathic disposition and behaviour. In an assessment context, ratings of

behaviour were provided by trained assessors and patients and compared to self

report empathic disposition. For those doctors trained in the UK and Ireland, fantasy,

perspective taking and empathic concern were positively related to assessments of

interpersonal behaviour. However this relationship did not hold for doctors trained

overseas. Furthermore, personal distress was universally related to lower patient

ratings of interpersonal behaviour, again suggesting that this is not an effective

empathic process in this context. However this study also showed that patient ratings

of behaviour were highly skewed suggesting that there is no clear understanding of

the behaviours that patients associate with empathy. Therefore the final study of the

thesis will explore understanding of the specific behaviours associated with empathy

in the healthcare context, focusing on the patient perspective.

 154

Chapter 7 - Empathy from the patient perspective

Studies 1-3 utilised the IRI in understanding the individual differences associated

with the antecedents of empathy and aimed to explore the specific behaviours

connected with those individual differences in the healthcare context. While a clear

pattern of individual differences was apparent in terms of a broad measure of

personality and a specific measure of emotional intelligence, the behaviours

associated with these differences remain unclear. There are several reasons for this,

both methodological and theoretical, which were explored in detail in the previous

chapter. In terms of methodology, there is a need to explore typical empathic

behaviours from the perspective of the ultimate judge in this context, the patient.

Theoretically, the situational factors as well as the specific behaviours associated

with empathy in the healthcare context also need to be considered. Patient-

practitioner similarity in terms of culture was given as an example of this in the

previous study. The purpose of this final study is therefore twofold. First, it will aim

to expand current understanding of the typical behaviours associated with empathy

in this context, as judged by patients. Secondly, it will build upon the findings of

studies 1-3 to expand understanding of the antecedents of empathy by considering

situational factors as well as individual differences. This chapter will begin by

considering the theoretical basis of the present study, before moving on to justify an

alternative methodology.
 155

7.1 The patient perspective: judgements of empathy

Davis (1996) identifies ‘interpersonal outcomes’ as the final stage in his

multidimensional model of empathy, with the adapted model from Larson and Yao

(2005: see Figure 7.1) arguing that these ‘interpersonal processes’ are the indicators

by which empathy is judged by patients in this context. These judgments of empathy

will depend not only on whether the practitioner understands the perspective of

patients and their emotional response, but also the extent to which the practitioner

demonstrates this understanding in empathic behaviour towards the patient.

Outcomes such as patient trust, compliance with medical treatment, quality of care

relationships and satisfaction with medical services will depend on patients’

judgements that the doctor is empathic, which in turn will depend on the doctor’s

behaviour (Barnett, Howard, King & Dino, 1981; Becker & Maiman, 1975).

However, there is some evidence to suggest that the behaviours that patients use to

make these judgments are different to the ones identified as important by healthcare

professionals themselves. For example in a study of physicians, patient judgments of

empathy were predicted by reassurance, listening and being sensitive to needs,

whereas assessor ratings were influenced by the introduction of open communication

cues (Silvester et al., 2007). It is therefore important that research should focus on

the patient perspective, particularly in light of the fact that within the NHS there is an

ongoing commitment to acknowledging the patient perspective and developing

doctor-patient partnerships (Department of Health, 1996).

 156

Antecedents

• Physician characteristics
e.g. compassion (+),
cynicism (-)

• Patient characteristics

• Situational characteristics
e.g. patient-physician

similarity

Empathic Processes

• Non-cognitive processes e.g.

mimicry

• Cognitive Processes e.g.

association, role-taking

Intrapersonal Processes Interpersonal Processes

• Affective reactions • Helping behaviour

Parallel emotions
Reactive emotions • Conflict management

• Non-affective reactions • Social behaviour

Attributions Communication
Considerate social style

Physician Outcomes Patient Outcomes

• Job satisfaction • Satisfaction

• Burnout • Healthcare outcomes

Figure 7.1. Process model of clinical empathy (from Larson & Yao, 2005).
 157

7.2 Empathy and behaviour

While it is important to understand empathic behaviour from the patient perspective,

specification of the behaviours remains vague (Stepien & Baernstein, 2006). There is

a growing call for the assessment of interpersonal behaviours as an indicator of

medical competence alongside clinical knowledge (e.g Epstein & Hundert, 2002).

Methods of evaluating empathic behaviours have typically been developed without a

strong theoretical focus on empathy (e.g. Mercer & Reynolds, 2002). Although work

in the UK has sought to identify the behavioural competencies needed by physicians

in order to deliver good medical care (e.g. Patterson et al., 2000), these investigations

have also been broad in focus, looking at the full range of physician performance.

Although some more recent research has attempted to understand empathic

behaviour more specifically, examples are few. A study by Forchuk and Reynolds

(2001) asked 30 patients within a psychiatric unit to describe the behaviours of

nurses that they perceived as helpful in building an empathic relationship. Helpful

behaviours included exploration and clarification of feelings and their meaning to the

patient as well as helping clients to focus on future solutions rather than past

problems. Listening was an important indicator of this. While this research is a

useful starting point, it is limited to one context (psychiatric patients) and therefore

further research is needed to replicate these findings in alternative contexts. This

study therefore seeks to develop understanding of the specific interpersonal

behaviours which patients judge as empathic in the healthcare context.

 158

7.3 Antecedents: situational characteristics

In their adapted model, Larson and Yao (2005) also highlight the need to consider

not only the characteristics of the healthcare practitioner but also those of the patient

and the situation in discussing antecedents to empathic processes. For example, they

highlight patient-practitioner similarity as a potentially important variable. In the

previous study, similarity in terms of culture was raised as a possible important

factor. However, the majority of empathy research within the healthcare literature

focuses on the individual characteristics of the practitioner (e.g. Morse et al., 2006;

Stepien & Baernstein, 2006) and so very little is known about the situational factors

that may influence the relationship between individual differences and behaviour.

This final study will therefore pose two central research questions:

1. What are the specific behaviours identified by patients when making empathy

judgments in the healthcare context?

2. What are the situational antecedents of empathy that might impact on the

relationship between individual differences and behaviour?

In order to address these research questions, a different methodological approach is

used for this study, namely a qualitative one. It is therefore necessary to first justify

this decision.

7.4 Quantitative and qualitative methods

Quantitative approaches to research, such as those taken with the use of

psychometric questionnaires used in this thesis are useful for testing hypotheses via

measurement and control of a structured sample of variables (De Vaus, 2002). They
 159

are underpinned by a positivist assumption that ‘the truth is out there’, in that there

are consistent relationships which can be measured and observed. Indeed, the very

word ‘quantitative’ implies that measurement can be made on some numerical basis

(Rust & Golombok, 1999). Quantitative research aims to be as objective as possible,

in order to discover the generalised laws that apply to psychological constructs. As

Coolican (2009) notes, quantitative methods result in reliable, internally valid,

objective data. It is this kind of evidence which has been presented thus far.

However, this psychometric approach can have its limitations. For example, analysis

focuses on areas defined by the researcher, rather than exploring what is meaningful

to the participant (Robson, 2002). In addition, the method of investigation, namely

that of a self report pencil and paper questionnaire, could be seen as incongruent as a

method for examining a dynamic interpersonal concept such as empathy (Cassell &

Symon, 2004).

In order to address some of these limitations, this study uses a qualitative

methodology. For many years within psychology, a debate has existed with regard to

the relative benefits and limitations of quantitative and qualitative approaches.

Proponents of pure quantitative methods criticise qualitative research as lacking in

reliability and validity, being uncontrolled, subjective and biased (Coolican, 2009).

Qualitative researchers on the other hand would reply that a reductionist, quantitative

approach removes understanding of humans who are ‘laden with values and must be

understood in the context of their time and cultural setting’ (Bem & Looren de Jong,

1997, p.23).
 160

Rather than the positivist approach underpinning quantitative research, qualitative

researchers use a social constructionist approach (Smith, 2008). Instead of an

observable objective truth, the assumption is that reality is multiple and constructed

by individuals (Banister et al., 1994). Within this framework, participants’ accounts

are not simply representations of the world. Rather knowledge is actively created

between researcher and participant (Miles & Huberman, 1994). Importantly, rather

than investigating numbers and objective facts, qualitative research paradigms

emphasise meanings, descriptions and experiences, looking for the emergence of

themes or patterns. Raw data consists of what people have said in interviews or

recorded conversations, or a description of what has been observed (Smith, 2008).

Within the context of empathy research, this would involve allowing patients to talk

about their descriptions and experiences of empathy instead of guiding them with

reference to previous scales and definitions. It would seem then that empathy is a

suitable topic for qualitative research. It is complex concept with many dimensions

therefore qualitative investigation will allow the participants to focus on those areas

meaningful to them. As a dynamic interpersonal concept, a method of social

exchange rather than a pencil and paper questionnaire could be seen as more

authentic (Lincoln & Guba, 1985).

It should further be noted that qualitative and quantitative methods are not

necessarily incompatible. Multi-method approaches which aim to combine the two

perspectives are commonly found within theory and measure development

(Coolican, 2009). According to Bartunek & Seo (2002), qualitative research can

complement findings from research studies to develop understanding. Collection of

data using an authentic research method for the topic of interest can allow for a
 161

greater understanding of how phenomena are experienced in particular contexts, in

terms of depth and meaning from a smaller sample of participants (Lincoln & Guba,

1985). This is the therefore the approach taken in this study. Empathy from the

perspective of patients is explored, with a focus on real experiences.

7.5 Method

7.5.1 Participants

The study was located within three wards of the medical division of a large teaching

hospital in the north east of England. Approval was sought and finally gained from

the local ethics and research & development committees. Twenty patients were

interviewed over a two week period within private rooms on the wards. Of the 20

patients interviewed, 14 were female and six were male. Nineteen of the patients

were from the local area and of White (British) ethnic origin. One patient was from

Pakistan and of Pakistani ethnic origin. Age ranged from 28 to 78 years (Mean age

55 years, S.D. 15.62 years). Length of stay in the hospital ranged from one week to a

period of several months. In accordance with requests from the ethics committee, no

additional information was gathered as this was not deemed relevant to the study.

7.5.2 Procedure

Participation in the study was voluntary. Ward sisters on each of the three wards

involved were asked to identify current patients whom they felt were well enough to

take part in a 30 minute interview in a side room. Before starting the interview,
 162

information sheets were provided and the purpose of the study and format of the

interview were explained fully. Patients were asked to sign a consent form to

indicate that they understood this information and agreed to take part (see

Appendices 5 and 6 for the information sheet and consent form). Twenty two

patients were approached to take part in the study; only two patients declined to be

interviewed. All interviews were conducted face to face and audio-recordings were

made for accuracy.

A semi structured interview incorporating a critical incident method was adopted for

all patient interviews (Flanagan, 1954). Patients were asked to describe incidents of

when they felt a nurse had empathised with them and also when a nurse had not

empathised with them. Where patients spoke in general terms about a nurse, they

were then prompted to give a specific example of an incident to illustrate their point.

Within each critical incident, patients were asked to describe briefly what had

happened during a specific example and also the outcome for them (see Appendix 7

for the interview schedule). Confidentiality was assured by explaining that any

names mentioned would not be recorded in transcripts of interviews and that

information provided in the interviews would only be fed back to hospital staff in

general terms so that individuals could not be identified. Interviews lasted from 15 to

30 minutes and were transcribed verbatim for analysis.

7.5.3 Analysis

Template analysis was used to code the data from the patient interview transcripts,

using the approach specified by King (2004). In template analysis (King, 2004), a
 163

priori codes or categories are pre-selected according to the researcher’s particular

interests. Categories are usually organized in a hierarchical fashion, with several

lower order categories being grouped together to produce a more general, higher

order code. Production of the initial template was directly influenced by the model of

Larson and Yao (2005; Fig 7.1) in their adaptation of Davis’ (1996)

multidimensional model for the clinical encounter, but deliberately kept to only two

levels of code to allow for emergent themes from that data rather than taking a very

prescriptive coding approach.

According to the Larson and Yao model, six higher order (or level one) codes were

present in the initial template: antecedents to empathy; empathic processes;

intrapersonal processes; interpersonal processes; nurse outcomes, and patient

outcomes. Within these six higher order codes, one further lower level of code was

present. For example within ‘antecedents’, a level two code was ‘situation’. The

model could have been used to identify further levels of code at the initial stage, but

this was avoided so that the final template could be guided by the data rather than

being constrained by the initial template. The initial template is shown in Figure 7.2.
 164

Figure 7.2: Initial Template for analysing patient interviews

1. Antecedents of Empathy
1 Nurse
2 Patient
3 Situation
2. Intrapersonal processes
3. Empathic Processes
4. Interpersonal Behaviour
1 Helping Behaviour
2 Social Behaviour
3 Conflict management
5. Nurse Outcomes
6. Patient Outcomes

The initial template was used to code all 20 patient interview transcripts. It is

important to note that template analysis allows for parallel coding, in which the same

segment of text might be classified under two or more categories. A brief illustration

of the coding process is given for the following extract from the first patient who

was talking about the fact that a nurse had accompanied her for some tests because

she had been quite frightened:

“And when she came back up she thanked me [3 – interpersonal behaviour:

considerate social style] because she said she’d learnt a lot [5 - nurse

outcome: thanking the patient] and we’d had a nice chat on the way [3 –

interpersonal behaviour: communication, chatting], so we enjoyed it [5 -

nurse outcome, 6 – patient outcome: satisfaction].

 165

According to King (2004), it is necessary to consider modifications to the coding

template and make changes during this initial coding process, in order to develop a

final coding template: possible modifications inserting a theme; deleting a theme;

changing the scope of a theme to make it broader or narrower, and differentiating

between higher (broader) and lower (narrower) coding themes. It is also possible to

change the higher-order classification of a theme by moving it into a different

category. This illustrates an important advantage of template analysis in that, whilst

theories or previous research may suggest certain themes (and can include these to

formulate initial templates) the modification process allows for a degree of open

coding such that the final template fully represents emergent themes from the data.

Modifications made to the initial template are summarised below.

First, two higher order themes were combined. Specifically, there was little use of

the themes ‘intrapersonal processes’ and ‘empathic processes’ during coding, but

where they were, they appeared to fit together. This is unsurprising given that the

methodology focused on the patient perspective. Patients spoke relatively less about

the process of empathising within the nurse and more about the context and

behaviour.

Secondly, there was a need to differentiate several codes with the addition of a

number of lower level codes. It became apparent during initial coding that many of

second level codes needed to be redefined with the use of third and fourth level

codes in order to produce a useful final template. For example, the higher order

theme ‘interpersonal processes’ was originally split into three second level codes,

‘helping behaviour’, ‘conflict management’ and ‘social behaviour’. ‘Conflict

 166

management’ was not used at all in the coding process and this was deleted. The

other two codes ‘helping behaviour’ and social behaviour’ were used extensively.

Helping behaviour was split into four third level codes: responding to requests; being

quick to help; problem solving, and acting as a patient advocate. Social behaviour

was split into two third level codes, communication and considerate social style.

These were then split into a further five level four codes each, representing the fine

grained nature of the coding process for this area of the template.

Analysis was conducted fully by the researcher with modifications made throughout

the coding process. The final template was reached after four rounds of coding of the

20 transcripts. Appendix 8 includes a sample of a fully coded interview transcript,

which Coolican (1999) recommends providing as an indicator of transparency in a

rigorous coding process (King 2004). Once the final template had been developed,

transcripts were reviewed and recoded by the researcher to check that the coding

reflected the final template fully.

In terms of quality checks on the qualitative analysis process, there are several

options available. Respondent feedback is a useful method of checking the quality of

the analysis (King, 2004). In order to do this, those who participated in the research

area asked to comment on the analysis and interpretation. Unfortunately the hospital

participating in the research was not willing to allow this final stage of checking.

There are some researchers who recommend that it is necessary to consider inter-

rater reliability of identifying codes (Coolican, 2009). However this is not advocated

by King (2004) and, as noted by Coolican (2009), this may be more relevant where

teams of researchers are involved in coding the data from a single study, which was
 167

not the case here. Despite this, there remains some need to demonstrate the vigour

and transparency of the data analysis process. King (2004) suggests that a final

template may be satisfactory when: (a) no sections of the transcripts that are relevant

to the research question remain uncoded; (b) all data have been read through and the

coding checked at least twice, perhaps three or four times, and (c) collaborating

researchers (or in the absence of a collaborator, an outside expert) agree that the

template is sufficiently clear and complete. The final template was reached by the

researcher after four rounds of coding. At this point, independent scrutiny of the final

version of the template was conducted by two psychologists who were experienced

in the use of coding approaches to qualitative data. These psychologists used four of

the transcripts to consider whether there were: any themes that they found difficult to

employ; any aspects of text not covered by the template, and any other issues of note

when reading the text. No issues of concern were raised by either, therefore the

template was considered to be sufficient as a final version. The final template for all

transcripts from both positive and negative incidents of empathy is given in Figure

7.3.
 168

Figure 7.3 Final template from analysis of patient interviews

1. Antecedents
1. Nurse characteristics
1. Agreeable personality
2. Motivated and engaged
2. Patient characteristics
1. Negative feelings about medical condition or treatment
1. Sadness and loneliness
2. Frustration and anger
3. Anxiety and fear
4. Shock
2. Behaviours
1. Asking questions about condition or treatment
2. Withdrawal
3. Situational characteristics
1. Lack of time, nurses’ workload
2. Communication between staff
2. Intrapersonal processes
1. Perspective taking
2. Compassion and sympathy
3. Interpersonal Processes
1. Helping behaviour
1. Responding to requests
2. Acting as patient advocate
3. Quick to help
4. Problem solving for patient
2. Prosocial behaviour
1. Communication
1. More communicative
2. Initiating communication
3. Listening
4. Explaining and informing
5. Communicating with family
2. Considerate social style
1. Participative
2. Kind and considerate
3. Reassuring
4. Positive emotional display
4. Patient Outcomes
1. Satisfaction with care
1. Trust and confidence in nurses
2. Negative feelings alleviated
2. Healthcare outcomes
1. Confidence and optimism about medical condition
2. Compliance with treatment
5. Immediate Nurse Outcomes
1. Open communication from patients and family
2. Satisfaction with interactions
 169

7.6 Findings and Discussion

The interviews were successful in producing rich, detailed accounts of participants’

experiences of nurses’ empathy, in line with the complex multidimensional nature of

the concept. As a result, it is necessary to justify the nature of the presentation of the

findings. To go through every code within the final template in equal depth would

result in a rather superficial, descriptive account of the findings. Rather, the

discussion of the findings focuses first on the two central research questions for the

study before moving on to briefly consider other emergent findings. This selective,

interpretive presentation of findings is in line with previous articles employing

template analysis (e.g. King, Carroll, Newton & Dornan, 2002). Any participant

names are pseudonyms to emphasise individual participant’s experiences while also

maintaining confidentiality. The research questions for this study were:

1. What are the specific behaviours identified by patients when making empathy

judgments in the healthcare context?

2. What are the situational antecedents of empathy that might impact on the

relationship between individual differences and behaviour?

The findings from the template analysis with respect to ‘Interpersonal Processes’

will therefore first be discussed, in response to question one. This is followed by

exploration of the findings from ‘Antecedents – Situational characteristics’ in

response to question two.

 170

7.6.1 Interpersonal Processes

Perhaps understandably from interviews focusing on the patient perspective, the

most frequently identified codes fell within the higher order category of

interpersonal processes. These are of course much easier for the patient to observe

than intrapersonal processes. Patients identified two main categories of behaviour

with respect to interpersonal processes, namely helping and social behaviour. Each

of these second level codes was then split into several level three codes. These

findings are presented in considerable detail as they provide much specific

information regarding to behaviours associated with empathy in this context.

Interpersonal processes: helping behaviour

Helping behaviour was frequently characterised by patients as the nurse being

responsive to requests: “No matter what I’m asking about, she doesn’t give me the

brush off, she answers me properly. And she responds at the time that I’ve asked her

to. And she sits and listens to you. And I find that very helpful”. Many patients

identified other factors within helping behaviour as well as being responsive which

they saw as demonstration of the nurses’ understanding and caring. These included

acting as a patient advocate, helping quickly and resolving issues fully. All of these

themes are reflected in the quote from Ann, a 62 year old stroke victim talking about

Beth, a nurse she viewed as highly empathic:

“If you have a problem, she sorts it out. I was supposed to have an

appointment with the physio, the doctor said. I waited for days...but it didn’t

happen. I just spoke to Beth and she made it happen for me right away. She

gives me privacy when I want it, she’s the one that thinks about that sort of
 171

thing. She’s got more go. She sorts you out very quickly, even though she’s

very busy, she pays attention to everybody”.

Acting as an advocate is something which has been previously been identified by

nurses as an expression of empathic behaviour (Morse et al., 2006) and it appears it

is received in this way by patients.

Interpersonal processes: prosocial behaviour

Whereas helping behaviour referred largely to what the nurse did in response to the

patient, prosocial behaviour was seen as the style of responding. There were two

themes within this prosocial behaviour category which characterised incidents of

nurse empathy, namely communication and considerate social style. Both of these

themes were seen as resulting from the nurses’ personality and situational

characteristics. In particular, patients described nurses who were seen as

compassionate and understanding as more communicative in general. This was not

necessarily through spending more time with them, although this was sometimes the

case. The patients on these wards frequently spoke of empathic nurses as busy but

able to make the most of their time by chatting about social topics in addition to

clearly explaining and informing them of medical issues, as described by Ann, a

stroke victim:

“I was scared at first that I wasn’t going to be able to talk but she kept

talking to me and ...I was a lot better after a week. She would chat about

different things, how the stroke happened and explaining it to me so I

understood. But also just talking to me about the weather and my family, you

know, just to keep my spirits up”.

 172

In addition, empathic nurses were described as those who initiated and opened up

communication with patients, as a result of their responding to the patients’

emotions. Initiating conversation might involve going in to a side room to check on a

lonely patient or asking open questions to find out more about a situation:

“When I was in the side room, I didn’t see many people coming and going.

She would try to have a chat with you and buck you up a bit you know

because I was on my own, otherwise it was just very depressing”.

Patients also identified the importance of listening within incidents of nurse

empathy, as described by Susan, a young lady with an undiagnosed illness under

investigation:

“She listens to you, listens to the problems you’ve got. When she’s got time

then she’ll listen to you... To her, it’s more than my medical condition, I’m a

person”.

Finally within this theme, over half of the patients identified the nurses’

communication with the family as being demonstration of a caring and

understanding approach. In Susan’s interview she went on to say:

“My Mum and Dad are able to ask her things, that’s important. It’s difficult

for me because I’ve been due to go home a few times now and then at the last

minute there’s been a problem with my tests. They know how I desperate I am

to get home... she was disappointed for me as well. But at least I know Mum

and Dad can talk to her”.

 173

This theme was closely linked to the patient withdrawal as an antecedent to empathy;

where patients became less communicative, they judged communication with the

family as a demonstration of empathy towards them.

In addition to communication, the second theme within prosocial behaviour was the

presence of a considerate social style. Within this theme, a participative style

characterised several incidents of nurse empathy, demonstrated with patients being

asked and encouraged rather than told to do something:

“Things were explained fully, whether they could be treated or not. I wasn’t

made to do anything, I was encouraged. She didn’t tell me what to do”.

Helping and communication behaviours were generally characterised by kindness

and patience:

“she’s very patient and very kind. Like I had a stroke, and she was very

patient with me. Whatever you ask for, she helps you so nicely and you don’t

feel like you’re any trouble and you feel very good”.

In contrast, nurses who did not empathise were seen as being too quick and harsh in

their communication style: “I asked her something and she really turned round and

snapped at me like I was stupid.... And I was that upset”.

The provision of reassurance alongside other actions was also seen as important by

more than half of the patients, in particular those who reported feeling anxious,

frightened or worried. This is an example from a patient who was anxious about

requiring a hearing aid:

 174

“And she reassured me, she said “This is what we mean, we can’t see it on

you, no one knows you’ve got it on” I agreed and I was pleased, so I said yes,

I would have it put on”.

For those patients who reported feeling sad or down, rather than reassurance, they

frequently mentioned the use of humour as an effective social style in helping them

to feel supported and more optimistic:

“I felt completely suicidal last year when I was in for so long. She would

come and sit with me and let me talk, she was here when I was upset. She let

me talk and she tried to cheer me up, you know making jokes and that. It

makes a real difference to me”.

This use of humour and a general display of positive emotions is interesting because

it does not necessarily require the experience of a particular emotion from the nurse.

Larson and Yao (2005) identify clinical empathy as a form of emotional labour.

Whereas surface acting emotional empathy can refer to the appearance of emotions

that are not necessarily experienced, deep acting empathy involves the actual

experience and subsequent expression of these emotions (Grandey, 2003). These are

important issues to consider in developing and sustaining clinical empathy because

within the emotional labour literature, researchers have much to say about the

potential impact of care work on the practitioner and the patient. For example, it has

been argued that surface acting empathy might protect the practitioner from

becoming over-involved and emotionally exhausted (Maslach, 1978) but might not

be perceived by patients as genuine. However this did not seem to be the case for

these patients. Conversely, while deep acting emotional empathy provide the

practitioner with emotional harmony, potential negative consequences might include

 175

over-involvement, decreased objectivity and emotional exhaustion. Such issues are

important in understanding the difficulty of maintaining clinical empathy over time

(e.g. Hojat et al., 2004; Spencer, 2004). The discussion now turns to the antecedents

of empathy as identified by patients.

7.6.2 Antecedents of empathy

The patients interviewed in this study discussed several themes regarding the

antecedents of empathy. The situational factors that were the subject of the second

research question are discussed first.

Situational characteristics: workload

Most frequently, when patients were asked to describe an incident when a nurse had

been unable to empathise with them, they did not feel that the nurse was particularly

responsible for this situation occurring. Rather they saw it as a result of staffing

levels in the ward: “I wouldn’t say anyone doesn’t understand, they’re just busy”

and “They can’t spend too much time with each patient because there’s so many

patients and so few nurses”. Although workload might not prevent all nurses being

able to empathise, it certainly seemed to present an extra challenge for nurses

according to many of the patients. For example, Frank compared wards where nurses

were highly effective and less effective in empathy by saying:

“I don’t know, I suppose it’s just that they don’t have the time down there

[different ward]. Mind you, they don’t really have the time up here but they

always try to make time for you, just to pop in and see if you’re OK, if there’s

anything you need. You know you’re not going to get lonely up here”.
 176

Lindsey, a 33 year old female patient agreed with this, by describing incidents of

presence and absence of empathy within the same nurse whom she had previously

described as “a really caring person”, thus demonstrating that nurse personality is

not sufficient to ensure effective empathy in the nursing role: “If they’ve got time

then she’ll listen to you. Time is the main problem, because they have to look after so

many patients”.

Situational characteristics: communication between staff

Although the discussion of positive incidents of empathy did not identify this theme

within the data, three patients also mentioned an absence of empathy where there had

been poor communication between staff. “When you come to a shift change, the

nurse who comes on doesn’t usually understand what has happened to you during

the day - that is annoying”. Handover of information is not the only potential barrier

within this theme. Potential conflicts or poor team relationships create issues which

divert nurses’ attention away from patients. David, a 58 year old male patient in the

day unit described such a situation:

“It’s very frustrating not knowing what’s happening or why...I sometimes feel

like the nurses don’t know what the doctors are doing and the doctors don’t

trust the nurses. The doctor put me on antidepressants and the nurse said I

wasn’t depressed. I was just worried because I didn’t know what was going

on. The communication isn’t there between the doctors and the nurses...”.

These situational characteristics that patients described as barrier to empathy from

nurses on these wards are not entirely new findings. For example McCormack &

McCance (2006) identified similar themes in a review of the literature on the

 177

development of person–centred nursing, where key characteristics of the care

environment included culture of the workplace, effective staff relationships and

nursing leadership. However, it is an aspect of the process model of empathy that has

thus far been largely ignored, the focus instead placed centrally on the individual

practitioner in terms of skills and abilities.

So far, this discussion has focused on the areas of the template most relevant to the

two research questions. It is important to briefly describe the findings from the rest

of the template, in so much as they support previous literature regarding empathy in

healthcare settings.

Patient characteristics: negative feelings

Most of the patients began describing the incidents by discussing their own feelings

and behaviours which created the need for empathy from the nurse. This was to be

expected as nurse empathy is likely to be triggered by the patients’ expression of

negative feelings which ranged in both nature and intensity. These negative emotions

included anxiety about treatments or conditions, or about a perceived lack of control

over their situation. For other patients, there were feelings of sadness and upset. This

went as far as clinically diagnosed depression for two patients, associated to longer

and repeated stays in hospital and serious diagnoses (not recorded for confidentiality

reasons). Others reported feelings including frustration and shock. The large range

of negative feelings articulated by patients, which were either associated directly

with the medical conditions suffered by patients or indirectly with the experiences of

treatment and being in hospital all seemed to act as the trigger for the subsequent

empathic (or non-empathic) interaction.

 178

Patient characteristics: behaviours

Sometimes the patients also identified their own behaviours which resulted from

these feelings. Most frequently, patients reported that their feelings resulting in

questioning of nurses and other staff. The questions tended to be medical ones

stemming from uncertainty over the condition or treatment. Some patients on the

other hand reported the opposite of asking for help, in that they would withdraw

from interaction with others. This highlights the difficult task faced by nurses in

having to pick up on very different behavioural cues offered by different patients.

Nurse characteristics: agreeable personality

In discussing incidents of empathy from nurses, patients made judgements regarding

the nurses’ personalities, in particular with respect to the domain of agreeableness.

Words such as kind, gentle, caring and approachable were frequently used

throughout all twenty of the interviews. Some of the patients said more empathic

nurses were more tolerant and patient of their particular circumstances which they

appreciated. The patients who mentioned it definitely felt it was part of the nurse’s

personality which therefore facilitated an interaction in which the nurse

demonstrated empathy, in line with the findings of this thesis.

Nurse characteristics: motivated and engaged

In addition to personality, patients also identified nurses who were more motivated

and engaged as ones who were more approachable to share their feelings with:

“She’s always there for your needs. If you need anything, she’s prepared to go and
 179

do it for you”. This was highlighted particularly well by Andrew, a 55 year old male

patient, in his general observations of a particular nurse’s working style:

“She’s got more go. She sorts you out very quickly, even though she’s very

busy, she pays attention to everybody... If you need anything, she’s prepared

to go and do it for you. Even down to one someone getting the wrong tea,

she’ll go and get them the right meal. It’s not part of her job but she does

it...She does her job as a nurse, the obs and everything, but it’s more than

that”.

The importance of motivation and engagement in workplaces within the UK is

currently a topic for much investigation, with drivers of engagement being identified

as human resource management strategies and leadership (MacCleod & Clarke,

2009). This links back to the situational characteristics identified earlier and again

suggests that for nurses to be empathic, development interventions should take a

wider focus than the skills of individual practitioners.

7.6.3 Other areas of the template

Intrapersonal processes, nurse and patient outcomes

The nature of the intrapersonal processes of perspective taking and emotional

responsiveness are well documented and supported by the results of this analysis.

This is also the case for outcomes of empathy for both nurses and patients. As such,

these findings are not be explored in depth but are briefly described.

Most patients interviewed perceived that nurses who were more kind and caring

were more likely to engage in cognitive role taking such that they appeared to be
 180

thinking about what the individual patient might be feeling or needing. Indeed,

absence of this perspective taking was identified many times as being a factor in

negative incidents of empathy. Patients only talked about the affective reactions of

nurses in terms of appropriate reactive emotions observed in response to the patient’s

own negative feelings. Words such as sympathy and concern were used frequently

throughout examples of empathy. Correspondingly, a lack of compassionate

responding tended to be a major theme throughout interactions lacking in empathy:

“They did explain but they didn’t show much sympathy”. Although patients did not

identify parallel emotions or shared affect, this could have been because they are not

able to judge these kinds of processes within the nurse accurately. However, with

regard to the compassion and sympathy, there was a perception that this was

empathic because it was genuinely felt by the nurses. This was perceived to be a

genuine felt emotional reaction rather than merely the display of it, as described by

John, a 47 year old patient who had been in hospital for several months: “It bucks

you up, especially if you’re feeling low. It makes you think at least somebody cares,

you feel stronger you know and you don’t feel like you’re going to be just stuck in

here. She really does care... she was so good to me and my wife”.

The immediate outcomes identified by patients as a result of the interpersonal

processes described above included both satisfaction with the care they had received

and perceived improvement in health outcomes. Satisfaction with care fell into two

main themes: trust and confidence in the nurses, and the alleviation of negative

feelings that had acted as the trigger to the incidents described within the interviews.

Feelings of trust and confidence in the nurses were strongly reported by patients,

particularly resulting from the open communication and explanation of the current
 181

and future situation for them. The alleviation of negative feelings, crossing into the

generation of positive feelings for some patients, was the most frequently identified

outcome of nurse empathy: “It bucks you up, especially if you’re feeling low. It

makes you think at least somebody care”. As well as reporting satisfaction with

care, patients also reported perceived benefits in terms of health outcomes of nurse

empathy. This was characterised by an optimism and confidence in their ability to

recover or improve, particularly resulting from the encouraging style of nurses: “You

always feel nervous in hospital but when the nurses are good to you like that, you

feel like you’re being looked after and you’re going to get out”. The participative

style was also used to great effect in gaining compliance with treatment in those

occasions where the patient’s negative feelings were around a proposed treatment or

course of action.

Whilst literature suggests that the long term outcomes of empathising for nurses may

involve greater job satisfaction and burnout (e.g. Larson & Yao, 2005), as mentioned

previously patients were unable to comment on these broader issues. Although the

patients were able to identify positive experiences for the nurses, they were unable to

say whether or not this resulted in overall job satisfaction which can of course be

influenced by a large range of other factors such as pay or leadership (Larrabee et al.,

2003). Instead, the nurse outcomes identified revolved around the impact of patients’

own behaviour towards the nurses following a demonstration of empathy (or lack of

it). First, patients reported that once they had experienced nurse empathy, they were

more likely to communicate openly with that nurse. Secondly, many patients

perceived that the enhanced communication between nurse and patient did provide

nurses with positive experiences in their job.

 182

7.6.4 Reflections and limitations

There were several limitations within this study. First, it is important to note that the

patients interviewed as part of this study were identified by the Ward Sisters and as

such the sample was not random. In addition, interviewing current patients may in

itself be problematic as there were perhaps times when patients were reluctant to

speak openly about people upon whom they are currently relying for care. One way

around this for future research would be to go through GP surgeries to identify

patients who had recently experienced a hospital stay. However, the critical incident

technique and use of a private room were both effective in encouraging patients to

identify example of empathy and speak at some length. The interview skills required

for this kind of research are also important to note. It is unlikely that a patient would

open up to an interviewer on the topic of empathy if an empathic approach to

interviewing was not taken. In this case, having been a relative of a patient on one of

the wards involved in this study, it was possible for the researcher to understand and

relate to the patients experiences.

A further limitation was that the findings of the template analysis could not be shown

to the participants for feedback (King, 2004). However, the process of gaining

ethical approval was a long and drawn out one and in the end this was a part of the

research design which was not possible. It took eleven months, two committee

meetings and a letter of recommendation to finally obtain ethical approval. Hospital

ethics committees are possibly more used to medical research and so this kind of

proposal seemed to pose problems for them. Organisational access to collect data can
 183

be fraught with difficulties at the best of times (Robson, 2002). Perhaps this is one of

the reasons why the concrete experiences of patients with regard to this topic

remains so relatively unexamined (Forchuk & Reynolds, 2001).

7.6.5 Summary

This study adopted a different approach, to extend understanding of the situational

characteristics and specific behaviours associated with empathy in the healthcare

context. Importantly, the study aimed to gain the perspective of the receivers of

empathy, namely patients, whilst employing a congruent methodology with the

interpersonal nature of the topic. The themes resulting from the template analysis

supported the findings of the first two studies in terms of individual differences,

while adding to understanding of the situational characteristics that can act as

antecedents to empathy. These included issues of workload and communication

between staff. Specific behaviours were also identified in considerable detail,

including the provision of practical help as well as prosocial aspects of

communication and a participative, positive, considerate approach. This information

will be useful in guiding interventions for the development of empathy in the

healthcare setting, discussion of which is included in the final chapter.

 184

Chapter 8: Concluding discussion

A basic premise of this thesis has been that greater understanding is required of how

empathy can be developed in patient-healthcare professional interactions in order to

foster the development of more effective training and development. Despite evidence

that empathy in healthcare professionals can have an important impact on patient

care and professional satisfaction, surprisingly little progress has been made in

efforts to develop interventions capable of sustaining or significantly improving the

levels of empathy demonstrated by professionals. Interestingly, very few researchers

in the medical field have sought to develop an integrated theory of empathy based on

empirical research, but have relied instead on reviews of literature conducted in a

range of contexts. Chapter two identified a direct adaptation of a comprehensive

model of empathy for this context. This was the process model of clinical empathy.

The key aspects of this model which were identified for investigation were:

antecedents of empathy in terms of individual differences, patient and situational

characteristics, and the specific behaviours associated with these characteristics.

Chapters four to seven outlined four studies which examined these aspects. This

chapter will begin by summarising their findings before moving on to consider the

broader theoretical and practical implications of the findings.

8.1 Summary of the findings

The thesis began with a psychometric investigation of empathy and the five factor

model of personality in the general population, before moving on to investigate

emotional intelligence in a sample of healthcare professionals. The third study then

 185

examined the relationship between empathy and behaviour in the same sample of

healthcare professionals. The final study adopted a different methodology in an

attempt to triangulate the findings with respect to individual differences in addition

to expanding understanding of situational characteristics and empathic behaviours in

this context. In order to understand the implications of this stream of research, it is

first necessary to summarise the findings of each study.

Chapters four and five provided a psychometric investigation of the individual

differences that are the antecedents of empathising. The first study was conducted

using a general population sample and found a clear pattern of traits associated with

a propensity to empathise. At the domain level, correlations indicated that those

higher on agreeableness and extraversion were higher on self assessed perspective

taking and empathic concern, while personal distress was characterised by greater

neuroticism. A facet level analysis revealed a clearer picture of the empathic

disposition. Facets of extraversion positively related to empathic concern and

perspective taking included warmth, positive emotions and gregariousness. Facets of

agreeableness related to those same scales included altruism, trust and tender-

mindedness.

The second study then moved on to look at the individual differences associated with

empathy in a specific healthcare context. A sample of 192 doctors completed a

measure of emotional intelligence in addition to the measure of empathy, in addition

to data from a general population sample. A preference for perspective taking and

empathic concern was positively associated with a single factor of emotional

intelligence. Of interest, those who reported higher personal distress also reported
 186

lower emotional intelligence, suggesting that aspect of empathy may be problematic

in an emotional labour context where control of emotions to remain objective is

required.

Having developed a clear picture of the individual differences associated with a

propensity to empathise, an examination of the relationships between empathic

disposition and behaviour was then conducted. In an assessment context, ratings of

behaviour were provided by trained assessors and compared to self report empathic

disposition. For those doctors trained in the UK and Ireland, fantasy, perspective

taking and empathic concern were positively related to assessments of interpersonal

behaviour. However this relationship did not hold for doctors trained overseas.

Furthermore, personal distress was universally related to lower ratings of

interpersonal behaviour, again suggesting that this is not an effective empathic

response in this context. However study three also showed that patient ratings of

behaviour were highly skewed suggesting that there is no clear understanding of the

behaviours that patients associate with empathy.

Therefore the final study adopted a different approach to extend understanding of the

situational characteristics and specific behaviours associated with empathy in the

healthcare context. Importantly, the research aimed to gain the perspective of the

receivers of empathy, namely patients, whilst employing a methodology perhaps

more congruent with the interpersonal nature of the topic. Twenty semi-structured

interviews with patients within three medical wards were analysed. Themes resulting

from the template analysis supported the findings of the first two studies in terms of

individual differences, while adding to understanding of the situational

 187

characteristics that can act as antecedents to empathy. These included issues of

engagement, workload and communication between staff. Specific behaviours were

also identified in considerable detail, including the provision of practical help as well

as prosocial aspects of communication and a participative, positive, considerate

approach.

8.2 Implications for theory

The thesis used a multidimensional model of empathy (Davis, 1983; 1996) and an

adaptation of it by Larson and Yao (2005) as a framework for the examination of

empathy in the healthcare setting. Figure 8.1 summarises the implication of the

findings by incorporating them into a model extended from Larson & Yao (2005).

Developments of this model in comparison to the Larson & Yao model can be seen

particularly in the antecedents and interpersonal processes. Antecedents in terms of

individual differences of the practitioner as well as role engagement are new

findings, as are those of the situational characteristics around the work environment

and team communication. In terms of the intrapersonal empathic processes,

perspective taking and empathic concern were both found to fit with the model.

Personal distress, a kind of automatic emotional response to those in need, was not

seen as appropriate in this context where more control is required. Emotional

intelligence therefore appears to fit well with the emotional labour context. Finally,

the interpersonal processes that are likely to be judged as empathic have been
 188

Antecedents
• Practitioner characteristics
o Personality (emotional intelligence, personality)
o Role engagement

• Patient characteristics
o Negative emotions (sadness, frustration, anger, shock)
o Behaviours (asking questions, withdrawal)

• Situational characteristics
o Work environment
o Team communication
o Cultural similarity

Intrapersonal Processes
• Congruent reactive emotions (compassion and sympathy)

• Perspective Taking

Interpersonal Processes

• Helping behaviour
o Responding,
o Acting as an advocate
o Quick help
o Problem solving

• Prosocial behaviour
o Communication (initiating, listening, explaining, family)
o Considerate social style (participative, kind, reassuring, positive)

Practitioner Outcomes Patient Outcomes

• Satisfaction with care
• Open communication from
o Trust
patients and family
o Alleviation of negative
emotions
• Satisfying interactions
• Healthcare outcomes
o Confidence
o Compliance

Figure 8.1: A process model of empathy in the healthcare setting

 189

specified with much greater detail than previous models. The model as it now stands

has clear implications for practice, which are now discussed.

8.3 Implications for practice

As stated throughout this thesis, a basic premise of this research has been that greater

understanding of empathy in patient-healthcare professional interactions would be

useful to guide the development of more effective training and development. The

development of the model in Figure 8.2 has clear implications for the strategies

adopted by healthcare organisations in their attempts to foster empathy among staff.

These can be broadly aimed at three areas: training interventions; longer term

development interventions and work design interventions.

Findings from evaluations with regard to empathy training interventions remain

mixed at best (Stepien & Baernstein, 2006). Of particular note was the range of

theoretical definitions adopted within Chapter two, which have guided the design of

training content. Findings from this research clearly indicate that the concept of

empathy is a complex one, which perhaps it is unrealistic to change a great deal

through the provision of brief training interventions. It is possible that training

courses could form part of longer development interventions; the experiences shared

by patients in the final study of the thesis could be put to very good use in the design

of training materials.

As empathy is often seen as part of personality, support for which has been found

from this research, there are those who believe that training may only have a limited
 190

effect on practice as personality remains stable over time (e.g. Evans et al., 1993).

However, the focus on behaviour will be useful, as trait theory tells us that while our

natural preferences may not change over time, one can learn characteristics

adaptations to ensure a better fit to the requirements of the environment (McCrae &

Costa, 1999). As such, the greater knowledge of behaviours associated with

judgments of empathy should form the basis of training and development

interventions. It would also be useful to use psychometric questionnaires with

healthcare professionals to raise their self awareness of their natural preferences in

understanding how change might be achieved. This is consistent with Carper’s

(1978) description of personal knowledge in nursing, where self awareness enhances

empathy. The many commercially available emotional intelligence measures may

prove useful for this purpose, providing a method of assessing empathy but also

control of emotional responses (e.g. Bar-On, 1997).

Having considered direct interventions at the level of individual empathy, other

interventions that could create a work environment where empathy is more likely to

occur should also be considered. This is building on the findings from the research of

situational characteristics that are antecedents of empathy. The situational

characteristics included the work environment, employee engagement and cultural

similarity between patients and practitioners. Where employees are required to work

in multi-cultural environments, development interventions that focus on raising

awareness of cross-cultural issues in healthcare would be of value (Wang et al.,

2003). Given that more engaged nurses were identified by patients as the ones who

provide more empathic interactions, it will be important to assess the link between

engagement and empathy, currently a popular topic within organisational

 191

psychology. Issues of leadership and teamwork are likely to be key to this kind of

initiative (Robinson, Perryman & Hayday, 2004).

8.4 Future research

This thesis has raised many questions for future research. A limitation of all four

studies was the cross-sectional nature of data collection. Future research should aim

to take a longitudinal approach to assessing the development of empathy over time.

Furthermore, it would be useful to assess the effectiveness of any training and

development interventions via thorough evaluations. Within such investigations, it is

important to consider the perspectives of the healthcare professional, as emotional

labour roles are known to have impact on well-being over time (Maslach, 1978). It is

however fundamentally important to continue to include patient perspectives in

research, in order to understand the ultimate impact of empathy in the healthcare

setting. Although this thesis has focused entirely on the healthcare setting, it would

be very interesting to test the applicability of a multidimensional model of empathy

in other roles. Any roles in which interactions with people are part of the job are

likely to be relevant. Finally, the cross cultural challenge for empathy in healthcare

settings requires further investigation.

8.5 Closing points

Summarising the thesis, key messages arising from this research are as follows:
 192

1. There is a clear pattern of personality traits associated with empathy, which

can be used in personal development work. However, a wider focus than the

individual practitioner is needed with consideration of factors such as work

engagement, team communication and work design.

2. Emotional intelligence is potentially a more useful concept than empathy as it

is the management of one’s natural emotional responses that helps

practitioners to respond effectively in an emotional labour context.

3. Development interventions should include a patient perspective in their

design, as the central recipients of the care experience.

4. Future work should consider how to develop empathy in practitioners trained

in different cultures, including a more fine grained inspection of the impact

of culture on behaviour. As was noted by one of the nurses in the hospital

involved in study four in the North East of England:

She [nurse] sometimes doesn’t get what they [patients] are on about. She’s
not from here, she doesn’t understand some of our little expressions, so
maybe she doesn’t know when they’re upset

Where is she from?

London
 193

References

Archer, R. L., Foushee, H. C., Davis, M. H., & Aderman, D. (1979). Emotional
empathy in a courtroom simulation: A person-situation interaction. Journal of
Applied Social Psychology, 9(3), 275-291.

Aring, C. D. (1958). Sympathy and empathy. Journal of the American Medical

Association, 167(4), 448-452.

Avery, J. (1985). Lawyers Tell What Turns Some Patients Litigious. Medical
Malpractice Review, 2, 35-37.

Axtell, C. M., Parker, S. K., Holman, D., & Totterdell, P. (2007). Enhancing
customer service: Perspective taking in a call centre. European Journal of Work
and Organizational Psychology, 16(2), 141-168.

Bachrach, H. M. (1976). Empathy. We know what we mean, but what do we

measure? Archives of General Psychiatry, 33(1), 35-38.

Banister, P., Burman, E., Parker, I., Taylor, M., & Tindall, C. (1994). Qualitative
methods in psychology: A research guide. Buckingham: Open University Press.

Barchard, K. A. (2003). Does emotional intelligence assist in the prediction of

academic success? Educational and Psychological Measurement, 63, 840-858.

Bar-On, R. (1997). Bar-On Emotional Quotient Inventory: Technical Manual.

Toronto: Multi-Health Systems.

Barker, P. (1994). Locus of control in women with a diagnosis of manic-depressive

psychosis. Journal of Psychiatric and Mental Health Nursing, 1, 9-14.

Barnett, M., Howard, J., King, L., & Dino, G. (1981). Helping Behavior and the
Transfer of Empathy. The Journal of Social Psychology, 115(1), 125-132.

Baron-Cohen, S. (2003). The Essential Difference: Men, Women and the Extreme
Male Brain. London: Penguin Books.
 194

Barrett-Lennard, G. T. (1962). Dimensions of therapist response as causal factors in

therapeutic change. Psychological Monographs, 76, 1-33.

Barrett-Lennard, G. T. (1964). The Relationship Inventory: Forms OS-M-64/OS-F-

64 and MO-M-64/MO-F-64. Armidale, NSW: University of New England.

Barrett-Lennard, G. T. (1978). The relationship inventory: Later development and

adaptations. JSAS Catalog of Selected Documents in Psychology, 8, 68.

Barrett-Lennard, G. T. (1993). The phases and focus of empathy. The British Journal
of Medical Psychology, 66(1), 3-14.

Bartram, A., Brown, A., Fleck, S., & Inceoglu, I. (2006). OPQ32 Technical Manual.
Thames Ditton: SHL.

Bartunek, J. M., & Seo, M. G. (2002). Qualitative research can add new meanings to
quantitative research, 23(2), 237-242.

Batson, C. D., & Coke, J. S. (1981). Empathy: A source of altruistic motivation for
helping? In J. P. Rushton & R. M. Sorrentino (Eds.), Altruism and helping
behavior: Social, personality, and developmental perspectives (pp. 167-187).
Hillsdale, NJ: Erlbaum.

Batson, C. D., Fultz, J., & Schoenrade, P. A. (1987). Distress and empathy: two
qualitatively distinct vicarious emotions with different motivational
consequences. Journal of Personality, 55(1), 19-39.

Batson, C. D., O’Quin, K., Fultz, J., Vanderplas, M., & Isen, A. M. (1983). Influence
of self-reported distress and empathy on egoistic versus altruistic motivation to
help. Journal of Personality and Social Psychology, 45(3), 706-718.

Batson, C. D., & Shaw, L. L. (1991). Evidence for altruism: Toward a pluralism of
prosocial motives. Psychological Inquiry, 2(2), 107-122.
 195

Becker, M. H., & Maiman, L. A. (1975). Sociobehavioral determinants of

compliance with health and medical care recommendations. Medical Care,
13(1), 10-24.

Becker, M. H., & Rosenstock, I. M. (1984). Compliance with medical advice. In A.

Steptow & A. Mathews (Eds.), Health Care and Human Behavior (pp. 175-
208). London: Academic Press.

Beckman, H. B., Markakis, K. M., Suchman, A. L., & Frankel, R. M. (1994). The
Doctor-Patient Relationship and Malpractice: Lessons From Plaintiff
Depositions. Archives of Internal Medicine, 154(12), 1365-1370.

Bellini, L. M., Baime, M., & Shea, J. A. (2002). Variation of mood and empathy
during internship. Journal of The American Medical Association, 287(23),
3143-3146.

Bellini, L. M., & Shea, J. A. (2005). Mood change and empathy decline persist
during three years of internal medicine training. Academic Medicine, 80(2),
164-167.

Bem, S., & Looren de Jong, H. (1997). Theoretical Issues in Psychology: An

Introduction. London: Sage.

Bennett, J. A. (1995). "Methodological notes on empathy": Further considerations.

Advances in Nursing Science, 18(1), 36-50.

Betancourt, H. (1990). An attribution-empathy model of helping behavior:

Behavioral intentions and judgments of help-giving. Personality and Social
Psychology Bulletin, 16(3), 573-591.

Betancourt, J. R. (2003). Cross-cultural medical education: conceptual approaches

and frameworks for evaluation. Academic Medicine, 78(6), 560-9.

Blair, R. J. R. (2005). Responding to the emotions of others: dissociating forms of

empathy through the study of typical and psychiatric populations.
Consciousness and Cognition, 14(4), 698-718.
 196

Block, J. (1961). The Q- Sort Method in Personality Assessment and Psychiatric

Research. Springfield CA: Sage.

Brackett, M. A., & Mayer, M. D. (2003). Convergent, Discriminant, and Incremental

Validity of Competing Measures of Emotional Intelligence. Personality and
Social Psychology Bulletin, 29, 1-12

Briggs, S. R., & Cheek, J. M. (1986). The role of factor analysis in the development
and evaluation of personality scales. Journal of Personality, 54(1), 106-148.

Briggs, S. R., Cheek, J. M., & Buss, A. H. (1980). An analysis of the Self-
Monitoring Scale. Journal of Personality and Social Psychology, 38, 679-686.

Brouwer, M. A. R., & Boros, S. (2010). The influence of intergroup contact and
ethnocultural empathy on employees’ attitudes toward diversity. Cognition,
Brain and Behavior, 14(3), 243-260.

Buss, A. H., & Plomin, R. A. (1975). A temperament theory of personality

development. New York: Wiley.

Carey, J. C., Fox, E. A., & Spraggins, E. F. (1988). Replication of structure findings
regarding the interpersonal reactivity index. Measurement and Evaluation in
Counseling and Development, 21, 102-105.

Carmel, S., & Glick, S. M. (1996). Compassionate-empathic physicians: personality

traits and social-organizational factors that enhance or inhibit this behavior
pattern. Social science medicine, 43(8), 1253-1261.

Carper, B. A.(1978). Fundamental patterns of knowing in nursing. Advances in

Nursing Science, 1(1), 13-24.

Cassell, C., & Symon, G. (2004). Essential guide to qualitative methods in

organizational research. Oxford: Sage.

Chamorro-Premuzic, T. (2007). Personality and individual differences. Oxford:

Blackwell.
 197

Chapin, F. S. (1942). Preliminary standardization of a social insight scale. American

Sociological Review, 7, 214-225.

Charbonneau, D. & Nicol, A. A. M. (2002). Emotional intelligence and leadership in

adolescents. Personality and Individual Differences, 33, 1101–1113.

Cheek, J. M., & Buss, A. H. (1981). Shyness and sociability. Journal of personality
and social psychology, 41(2), 330-339.

Chen, J. T., La Lopa, J. M., & Dang, D. K. (2008). Innovations in teaching: Impact
of patient empathy modeling on pharmacy students caring for the underserved.
American Journal of Pharmacy Education, 72(2), 40.

Chi-Ying Chung, R., & Bemak, F. (2002). The relationship of culture and empathy
in cross-cultural counseling. Journal of Counseling and Development, 80(2),
154-159.

Chlopan, B. E., McCain, M. L., Carbonell, J. L., & Hagen, R. L. (1985). Empathy:
Review of available measures. Journal of Personality and Social Psychology,
48(3), 635-653.

Christopher, F. S., Owens, L. A., & Stecker, L, S. H. (1993). Exploring the darkside
of courtship: A test of a model of male premarital sexual aggressiveness.
Journal of Marriage and the Family, 55, 749-779.

Cialdini, R. B., Baumann, D. J., & Kenrick, D. T. (1981). Insights from sadness: A
three-step model of the development of altruism as hedonism. Developmental
Review, 1(3), 207-223.

Cialdini, R. B., Brown, S. L., Lewis, B. P., Luce, C., & Neuberg, S. L. (1997).
Reinterpreting the empathy-altruism relationship: when one into one equals
oneness. Journal of Personality and Social Psychology, 73(3), 481-494.

Ciarrochi, J., Chan, A., & Caputi, P. (2000). A critical evaluation of the emotional
intelligence construct. Personality and Individual Differences, 28(3), 539-561.
 198

Cliffordson, C. (2002). The hierarchical structure of empathy: Dimensional

organization and relations to social functioning. Scandinavian Journal of
Psychology, 43(1), 49-59.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale,
NJ: Lawrence Erlbaum Associates.

Coke, J. S., Batson, C. D., & McDavis, K. (1978). Empathic mediation of helping: A
two-stage model. Journal of Personality and Social Psychology, 36(7), 752-
766.

Coolican, H. (2009). Research Methods and Statistics in Psychology. London:

Hodder Arnold.

Costa, P. T., & McCrae, R. R. (1992). NEO PI-R professional manual. Odessa, FL:
Psychological Assessment Resources Inc.

Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests.

Psychological Bulletin, 52(4), 281-302.

Cross, D. G., & Sharpley, C. (1982). Measurement of empathy with the Hogan
Empathy Scale. Psychological Reports, 50, 62.

Cundiff, N. L., Nadler, J. T., & Swan, A. (2009). The Influence of Cultural Empathy
and Gender on Perceptions of Diversity Programs. Journal of Leadership &
Organizational Studies, 16(1), 97-110.

Damasio, A. R. (1994). Descartes’ error: Emotion; reason and the human brain.
New York: Grosset.

Davies, M., Stankov, L., & Roberts, R. D. (1998). Emotional intelligence: in search
of an elusive construct. Journal of personality and social psychology, 75(4),
989-1015.
 199

Davis, M. H. (1983). Measuring individual differences in empathy: Evidence for a

multidimensional approach. Journal of Personality and Social Psychology,
44(1), 113-126.

Davis, M. H. (1996). Empathy: A Social-Psychological Approach. Oxford:

Westview Press.

Dawda, D., & Hart, S. D. (2000). Assessing emotional intelligence: reliability and
validity of the Bar-On Emotional Quotient Inventory (EQ-i) in university
students. Personality and Individual Differences, 32, 37-48.

De Vaus, D. A. (2002). Surveys in social research (5th edition). London: Routledge.

Department of Health. (1996). Patient Partnership: Building a Collaborative

Strategy. London: NHS Executive.

Department of Health. (1999). Patient and public involvement in the new NHS.
London: NHS Executive.

DeYoung, C. G., Quilty, L. C., & Peterson, J. B. (2007). Between facets and
domains: 10 aspects of the Big Five. Journal of personality and social
psychology, 93(5), 880-96.

Di Blasi, Z., Harkness, E., Ernst, E., Georgiou, A., & Kleijnen, J. (2001). Influence
of context effects on health outcomes: a systematic review. Lancet, 357(9258),
757-762.

Digman, J. M. (1990). Personality Structure: Emergence of the Five-Factor Model

Annual Review of Psychology, 41, 417 -440.

Domino, G., & Domino, M. L. (2006). Psychological testing: An introduction.

London: Cambridge University Press.

Dubnicki, C. (1977). Relations among therapist empathy and authoritarianism and a

therapist’s prognosis. Journal of Consulting and Clinical Psychology, 45, 958-
959.
 200

Dymond, R. F. (1949). A scale for the measurement of empathic ability. Journal of

Consulting Psychology, 13, 127-133.

Dzubian, C. D., & Shirkey, E. C. (1974). When is a correlation matrix appropriate

for factor analysis? Psychological Bulletin, 81, 358-361.

Eisenberg, N. (1986). Altruistic emotion; cognition and behaviour. Hillsdale, NJ:

Erlbaum.

Eisenberg, N., Fabes, R. A., Schaller, M., Miller, P., Carlo, G., Poulin, R., et al.
(1991). Personality and socialization correlates of vicarious emotional
responding. Journal of Personality and Social Psychology, 61(3), 459-470.

Eisenberg, N., & Miller, P. A. (1987). The relation of empathy to prosocial and
related behaviors. Psychological Bulletin, 101(1), 91-119.

Elam, C., Stratton, T. D., & Andrykowski, M. A. (2001). Measuring the emotional
intelligence of medical school matriculants. Academic Medicine, 76(5), 507-8.

Epstein, R. M., & Hundert, E. M. (2002). Defining and Assessing Professional

Competence. The Journal of the American Medical Association, 287(2), 226-
235.

Eslinger, P. J. (1998). Neurological and neuropsychological bases of empathy.

European Neurology, 39(4), 193-199.

Evans, B. J., Stanley, R. O., & Burrows, G. D. (1993). Measuring medical students’
empathy skills. British Journal of Medical Psychology, 66(2), 121–33.

Eysenck, S. B. G., & Eysenck, H. J. (1978). Impulsiveness and venturesomeness:

Their position in a dimensional system of personality description. Psychological
Reports, 43, 1247-1255.

Eysenck, S. B. G., & McGurk, B. J. (1980). Impulsiveness and venturesomeness in

detention center population. Psychological Reports, (47), 1299-1306.
 201

Feighny, K. M., Arnold, L., Monaco, M., Munro, S., & Earl, B. (1998). In pursuit of
empathy and its relationship to physician communication skills:
multidimensional empathy training for medical students. Annual Behavioural
Science Medical Education, 5, 13-21.

Fenigstein, A., Scheier, M., & Buss, A. H. (1975). Public and private self
consciousness: Assessment and theory. Journal of Consulting and Clinical
Psychology, 43, 522-527.

Fenwick, A. (1995). On attribution theory: Challenging behaviour and staff beliefs.

Clinical Psychology Forum, 79, 29-32.

Ferguson, E. (2001). Personality and coping traits: A joint factor analysis. British
journal of health psychology, 6(4), 311-325.

Ferguson, E. & Cox, T. (1993). Exploratory Factor Analysis: A Users’Guide.

International Journal of Selection and Assessment, 1, 84–94.

Feshbach, N. D., & Roe, K. (1968). Empathy in six-and seven-year-olds. Child

Development, 39(1), 133-145.

Field, A. (2005). Discovering Statistics Using SPSS. London: Wiley.

Fields, S. K., Hojat, M., Gonnella, J. S., Mangione, S., Kane, G., & Magee, M.
(2004). Comparisons of nurses and physicians on an operational measure of
empathy. Evaluation & the health professions, 27(1), 80-94.

Flanagan, J. C. (1954). The Critical Incident Technique. Psychological Bulletin, 51,

327-358.

Fletcher, C., & Kerslake, C. (1993). Candidate Anxiety Level and Assessment
Centre Performance. Journal of Managerial Psychology, 8(5), 19 – 23.

Forchuk, C., & Reynolds, W. (2001) Clients’ reflections on relationships with

nurses: comparisons from Canada and Scotland. Journal of Psychiatric and
Mental Health Nursing, 8, 45–51.
 202

Frankel, R. M. (1995). Emotion and the physician-patient relationship. Motivation

and Emotion, 19(3), 163-173.

Frankel, R. M., & Beckman, H. B. (1989). Evaluating the patient’s primary

problem(s). In M. Stewart & D. Roter (Eds.), Communication with medical
patients. Newbury Park, CA: Sage.

Frasure-Smith, N., Lespérance, F., & Talajic, M. (1995). The impact of negative
emotions on prognosis following myocardial infarction: is it more than
depression? Health Psychology, 14(5), 388-398.

Froman, R. D., & Peloquin, S. M.(2001). Rethinking the use of the Hogan Empathy
Scale: a critical psychometric analysis. American Journal of Occupational
Therapy, 55(5), 566-72.

Funder, D. C. (1997). The personality puzzle. New York: W. W. Norton.

Garrity, T. F. (1981). Medical compliance and the clinician-patient relationship: a

review. Social science medicine Part E Medical psychology, 15(3), 215-222.

Geher, G., Warner, R. M., & Brown, A. S. (2001). Predictive validity of the
emotional accuracy research scale. Intelligence, 29(5), 373-388.

Gladstein, G. A. (1987). What It All Means. Empathy and counseling: explorations

in theory and research. New York: Springer-Verlag.

Goleman, D. (1995). Emotional Intelligence. London: Bloomsbury.

Grandey, A. A. (2003). When "The Show Must Go on": Surface Acting and Deep
Acting as Determinants of Emotional Exhaustion and Peer-Rated Service
Delivery. Academy of Management Journal, 46(1), 86-96.

Grattan, L. M., Bloomer, R. H., Archambault, F. X., & Eslinger, P. J. (1994).

Cognitive flexibility and empathy after frontal lobe lesion. Neuropsychiatry,
Neuropsychology and Behavioral Neurology, 7(4), 251-259.
 203

Graziano, W. G., & Eisenberg, N. (1997). Agreeableness: A dimension of

personality. In R. Hogan, J. Johnson, & S. R. Briggs (Eds.), Handbook of
Personality Psychology (pp. 795-824). San Diego, CA: Academic Press.

Graziano, W. G., Habashi, M. M., Sheese, B. E., & Tobin, R. M. (2007).

Agreeableness, empathy, and helping: a person x situation perspective. Journal
of Personality and Social Psychology, 93(4), 583-99.

Greenson, R. R. (1960). Empathy and its vicissitudes. International Journal of

Psychoanalysis, 41, 418-424.

Gruen, R. J., & Mendelsohn, G. (1986). Emotional responses to affective displays in

others: The distinction between empathy and sympathy. Journal of Personality
and Social Psychology, 51, 609-614.

Guadagnoli, E., & Velicer, F. W. (1988). Relation of sample size to the stability of
component patterns. Psychological bulletin, 103(2), 265-75.

Gurman, A., & Razin, A. (1977). The therapist’s handbook for effective
psychotherapy. New York: Pergamon Press.

Halpern, J. (2001). From Detached Concern to Empathy: Humanizing Medical

Practice. New York: Oxford University Press.

Hardee, J. T. (2003). An overview of empathy. The Permanente Journal, 7, 51-54.

Haynes, S. N., Richard, D. C. S., & Kubany, E. S. (1995). Content Validity in

Psychological Assessment: A Functional Approach to Concepts and Methods.
Psychological Assessment, 7(3), 247-238.

Hochschild, A. R. (1983). The Managed Heart. Los Angeles: University of

California Press.

Hodges, S., & Wegner, D. M. (1997). Automatic and controlled empathy. In W. J.

Ickes (Ed.), Empathic accuracy (pp. 311-339). New York: Guilford Press.
 204

Hoffman, M. L. (1984). Empathy; its limitations; and its role in a comprehensive

moral theory. In W. Kurtines & J. Gerwirtz (Eds.), Morality; Moral Behavior;
and Moral Development. New York: Wiley.

Hogan, R. (1969). Development of an empathy scale. Journal of Consulting and

Clinical Psychology, 33(3), 307-316.

Hojat, M., Gonnella, J. S., Nasca, T. J., Mangione, S., Veloksi, J., & Magee, M.
(2002). The Jefferson Scale of Physician Empathy: further psychometric data
and differences by gender and specialty at item level. Academic Medicine, 77,
58-60.

Hojat, M., Mangione, S., Nasca, T. J., Cohen, M. M., Gonnella, J. S., Erdmann, J. B.,
et al. (2001). The Jefferson Scale of Physician Empathy: Development and
Preliminary Psychometric Data. Educational and Psychological Measurement,
61(2), 349-365.

Hojat, M., Mangione, S., Nasca, T. J., Gonnella, J. S., & Magee, M. (2005).
Empathy scores in medical school and ratings of empathic behavior in
residency training 3 years later. The Journal of Social Psychology, 145(6), 663-
72.

Hojat, M., Mangione, S., Nasca, T. J., Rattner, S., Erdmann, J. B., Gonnella, J. S., et
al. (2004). An empirical study of decline in empathy in medical school. Medical
Education, 38(9), 934-941.

Horn, J. L. (1965). A rationale and test for the number of factors in factor analysis.
Psychometrica, 30, 179-185.

Hornsten, A., Lundman, B., Selstam, E. K., & Sandstrom, H. (2005). Patient
satisfaction with diabetes care. Journal of Advanced Nursing, 51(6), 609-617.

Howell, D. C. (2002). Statistical Methods for Psychology. London: Sage.

Hunt, T. (1928). The measurement of social intelligence. Journal of Applied

Psychology, 12, 317-334.
 205

Hynes, C. A., Baird, A. A., & Grafton, S. T. (2006). Differential role of the orbital
frontal lobe in emotional versus cognitive perspective-taking.
Neuropsychologia, 44(3), 374-383.

Ickes, W. (1993). Empathic Accuracy. Journal of Personality, 61, 587-609.

Jabbi, M., Swart, M., & Keysers, C. (2007). Empathy for positive and negative
emotions in the gustatory cortex. NeuroImage, 34(4), 1744-53.

Kane, G. C., Gotto, J. L., Mangione, S., West, S., & Hojat, M. (2007). Jefferson
Scale of Patient’s Perceptions of Physician Empathy: preliminary psychometric
data. Croatian Medical Journal, 48(1), 81-6.

Kaplan, S. H., Greenfield, S., & Ware, J. E. (1989). Assessing the effects of
physician-patient interactions on the outcomes of chronic disease. Medical
Care, 27(3), 110-127.

Kellett, J. B., Humphrey, R. H., & Sleeth, R. G. (2002). Empathy and complex task
performance: two routes to leadership. The Leadership Quarterly, 13(5), 523-
544.

Kendall, P. C., & Wilcox, L. E. (1980). Cognitive behavioural treatment for

impulsivity: Concrete versus conceptual training in non-self-controlled problem
children. Journal of Consulting and Clinical Psychology, 48, 80-91.

Kerr, W. A., & Speroff, B. J. (1954). Validation and evaluation of the empathy test.
Journal of General Psychology, 50, 269-276.

King, N. (2004). Using templates in the thematic analysis of texts. Essential guide to
qualitative methods in organizational research, 256–270.

King, N., Carroll, C., Newton, P., & Dornan, T. (2002). "You can’t cure it so you
have to endure it": The Experience of Adaptation to Diabetic Renal Disease.
Qualitative Health Research, 12(3), 329-346.

Kline, P. (1994). An easy guide to factor analysis. London: Routledge.

 206

Kline, P. (2000). The handbook of psychological testing. London: Routledge.

Klis, M. J. (1997). The relation of empathy to some dimensions of teachers

personality across their professional life: Polish Psychological Bulletin, 28(1),
71-82.

Kohler, W. (1929). Gestalt psychology. New York: Liverlight.

Kolata, G. (2005). When the doctor is in, but you wish he wasn’t. The New York
Times. Retrieved March 19, 2011, from
http://www.nytimes.com/2005/11/30/health/30patient.html.

La Monica, E. L. (1981). Construct validity of an empathy instrument. Research in

Nursing & Health, 4(4), 389-400.

La Monica, E. L. (1996). Empathy construct rating scale (pp. 1-8). Santa Clara, CA:
Xicom.

La Monica, E. L., Carew, D. K., Windsor, A. E., Haase, A. M., & Blanchard, K. H.
(1976). Empathy training as the major thrust of a staff development program.
Nursing Research, 25(6), 447-451.

La Monica, E. L., Wolf, R. M., Madea, A. R., & Oberst, M. T. (1987). Empathy and
Nursing Care Outcomes. Scholarly Inquiry in Nursing Practice, 1, 197-213.

Landy, F. J. (2005). Some historical and scientific issues related to research on

emotional intelligence. Journal of Organizational Behavior, 26(4), 411-424.

Landy, F. J., & Conte, J. M. (2007). Work in the 21st century: An introduction to
industrial and organizational psychology. Oxford: Blackwell.

Larson, E. B., & Yao, X. (2005). Clinical empathy as emotional labor in the patient-
physician relationship. Journal Of The American Medical Association, 293(9),
1100-1106.
 207

Larzelere, R. E., & Mulaik, S. A. (1977). Single-sample tests for many correlations.
Psychological Bulletin, 84(3), 557-569.

Law, K. S., Wong, C.-S., & Song, L. J. (2004). The construct and criterion validity
of emotional intelligence and its potential utility for management studies. The
Journal of applied psychology, 89(3), 483-96.

Layton, J. M., & Wykle, M. H. (1990). A validity study of four empathy instruments.
Research in Nursing & Health, 13(5), 319-325.

Lilja, J., Larsson, S., Hamilton, D., & Issakainen, J. (2000). Empathy as a
communication strategy in the pharmacy - a study based on cognitive and
behavioural analysis. International Journal of Pharmacy Practice, 8(3), 176-
187.

Lincoln, Y., & Guba, E. (1985). Naturalistic inquiry. Beverly Hills: Sage.

MacDonald, R. (2003). The lowdown on international recruitment. British Medical

Journal Career Focus, 362(2), 197-199.

Mackay, R. C., Hughes, J. R., & Carver, J. (1990). Empathy in the Helping
Relationship. New York: Springer.

Macleod, D. & Clarke, N. (2009). Engaging for Success: enhancing performance

through employee engagement. Report for the Department for Business,

Innovation and Skills.

Makoul, G. (2001). Essential elements of communication in medical encounters: the

Kalamazoo consensus statement. Academic Medicine, 76(4), 390-393.

Mansfield, E. (1973). Empathy: Concept and Identified Psychiatric Nursing

Behavior. Nursing Research, 22(6), 525-530.

Maslach, C. (1978). Job burnout: How people cope. Public Welfare, 36, 56-58.
 208

Mayer, J. D., Salovey, P., & Caruso, D. (2000). Models of Emotional Intelligence.
London: Cambridge University Press.

McCormack, B., & McCance, T. V. (2006). Development of a framework for

person-centred nursing. Journal of Advanced Nursing, 56, 472–479.

McCrae, R. R., & Costa, P. T. (1989). Reinterpreting the Myers-Briggs Type

Indicator from the perspective of the five-factor model of personality. Journal
of personality, 57(1), 17-40.

McCrae, R. R., & Costa, P. T. (1999). A Five-Factor Theory of personality. In L. A.

Pervisn & O. P. John (Eds.), Handbook of personality: Theory and research
(pp. 139-153). New York: Guilford.

McGilton, K., Irwin-Robinson, H., Boscart, V., & Spanjevic, L. (2006).

Communication enhancement: nurse and patient satisfaction outcomes in a
complex continuing care facility. Journal of Advanced Nursing, 54(1), 35-44.

Mehrabian, A. (1996). Pleasure-arousal-dominance: A general framework for

describing and measuring individual differences in Temperament. Current
Psychology, 14(4), 261-292.

Mehrabian, A., & Epstein, N. (1972). A measure of emotional empathy. Journal of

Personality, 40(4), 525-543.

Mehrabian, A., & O’Reilly, E. (1980). Analysis of personality measures in terms of

basic dimensions of temperament. Journal of Personality and Social
Psychology, 38(3), 492-503.

Mehrabian, A., Young, A. L., & Sato, S. (1988). Emotional empathy and associated
individual differences. Current Psychology, 7(3), 221-240.

Mercer, S. W., & Reynolds, W. J. (2002). Empathy and quality of care. The British
Journal of General Practice, 52, 9-12.
 209

Mercer, S. W., Maxwell, M., Heaney, D, & Watt, G. (2004). The consultation and
relational empathy (CARE) measure: development and preliminary validation
and reliability of an empathy-based consultation process measure. Family
practice, 21(6), 699-705.

Mercer, S. W., McConnachie, A., Maxwell, M., Heaney, D, & Watt, G. (2005).
Relevance and practical use of the Consultation and Relational Empathy
(CARE) Measure in general practice. Family practice, 22(3), 328-334.

Messick, S. (1995). Validity of psychological assessment: validation of inferences

from persons’ responses and performances as scientific inquiry into score
meaning. American Psychologist, 50(9), 741-749.

Meyer, J. P., & Mulherin, A. (1980). From attribution to helping: An analysis of the
mediating effects of affect and expectancy. Journal of Personality and Social
Psychology, 39(2), 201-210.

Miles, M. B., & Huberman, A. M. (1994).Qualitative data analysis: an expanded

sourcebook. Beverly Hills: Sage.

Moore, P. J., Adler, N. E., & Robertson, P. A. (2000). Medical malpractice: the
effect of doctor-patient relations on medical patient perceptions and malpractice
intentions. The Western Journal of Medicine, 173(4), 244-50.

Morse, J. M., Anderson, G., Bottorff, J. L., Yonge, O., O’Brien, B., Solberg, S. M.,
et al. (1992). Exploring empathy: a conceptual fit for nursing practice? Journal
of Nursing Scholarship, 24(4), 273-280.

Morse, J. M., Bottorff, J., Anderson, G., O'Brien, B., & Solberg, S. (2006). Beyond
empathy: expanding expressions of caring. Journal of Advanced Nursing, 53,
75–87.

Moscrop, A. (2001). Empathy: a lost meaning? The Western Journal of Medicine,

175(1), 59-60.
 210

Moss, F. A. (1931). Preliminary report of a study of social intelligence and executive

ability. Public Personnel Studies, 9, 2-9.

Moss, F. A., Hunt, T., & Omwake, K. T. (1949). Manual for the social intelligence
test, revised form. Washington, DC: Center for Psychological Service.

Moss, F. A., Hunt, T., Omwake, K. T., & Ronning, M. M. (1927). Social Intelligence
Test. Washington, DC: Center for Psychological Service.

Murphy, K. R., & Davidshofer, C. O. (2005). Psychological testing: principles and

applications. London: Prentice Hall.

Muthen, B., & Kaplan, D. (1985). A comparison of some methodologies for the
factor analysis of non-normal Likert variables. Journal of Mathematical and
Statistical Psychology, 38(2), 171-189.

O’Connor, R. M., & Little, I. S. (2003). Revisiting the predictive validity of

emotional intelligence: Self-report versus ability-based measures. Personality
and Individual Differences, 35, 1893–1902.

Odom-Forren, J. (2007). Nurses: angels of mercy or competent professionals.

Journal of Perianesthesia, 22(2), 81-83.

Omdahl, B. L. (1995). Cognitive appraisal, emotion, and empathy. Mahwah, NJ:

Lawrence Erlbaum Associates.

O’Connor, B. P. (2000). SPSS and SAS programs for determining the number of
components using parallel analysis and Velicer’s MAP test. Behavior Research
Methods, Instruments, & Computers, 32(3), 396-402.

Pallant, J. (2007). SPSS Survival Manual (3rd ed). Maidenhead: Open University
Press.

Parker, S. K., & Axtell, C. M. (2001). Seeing Another Viewpoint: Antecedents and
Outcomes of Employee Perspective Taking. Academy of Management Journal,
44(6), 1085-1100.
 211

Patterson, F, Ferguson, E., Lane, P., Farrell, K., Martlew, J., & Wells, A. (2000). A
competency model for General Practice: Implications for selection and
development. British Journal of General Practice, 50, 188-193.

Patterson, F, Ferguson, E., Norfolk, T., & Lane, P. (2005). A new selection system to
recruit general practice registrars: preliminary findings from a validation study.
British Medical Journal, 330(4), 711-714.

Pedhazur, E. J., & Schmelkin, L. (1991). Measurement, Design, and Analysis: An

Integrated Approach. New York: Lawrence Erlbaum Associates.

Peplau, H. E. (1952). Interpersonal Relations in Nursing. New York: G.P. Putnam &
Sons.

Perez, J., Petrides, K. V., & Furnham, A. (2005). Measuring Trait Emotional
Intelligence. In R. Schultz & R. Roberts (Eds.), Emotional Intelligence: An
International Handbook (pp. 181-201). Gottingen: Hogrefe.

Peterson, C., Semmel, A., Baeyer, C., Abramson, L. Y., Metalsky, G. I., &
Seligman, M. E. P. (1982). The attributional Style questionnaire. Cognitive
Therapy and Research, 6(3), 287-299.

Peterson, C., & Villanova, P. (1988). An expanded attributional style questionnaire.

Journal of Abnormal Psychology, 97, 87-89.

Petrides, K, & Furnham, A. (2000). On the dimensional structure of emotional

intelligence. Personality and Individual Differences, 29(2), 313-320.

Petrides, K. V. & Furnham, A. (2001). Trait emotional intelligence: Psychometric

investigation with reference to established trait taxonomies. European Journal
of Personality, 15, 425-448.

Petrides, K. V., Pérez-González, J. C., & Furnham, A. (2007). On the criterion and
incremental validity of trait emotional intelligence. Cognition and Emotion, 21,
26-55.
 212

Plank, R. E., Minton, A. P., & Reid, D. A. (1996). A short measure of perceived
empathy. Psychological Reports, 79(3), 1219-1226.

Polit, D., & Hunglar, B. (1983). Nursing research. Principles and methods.
Philadelphia: Sage.

Pound, P., Gompertz, P., & Ebrahim, S. (1998). A patient-centred study of the
consequences of stroke. Clinical Rehabilitation, 12(4), 338-347.

Preston, S. D., & de Waal, F. B. M. (2002). Empathy: Its ultimate and proximate
bases. Behavioral and Brain Sciences, 25(01), 1-20.

Pulos, S., Elison, J., & Lennon, R. (2004). The hierarchical structure of the
interpersonal reactivity index. Social Behavior and Personality: an
international journal, 32(4), 355-359.

Ramjan, L. M. (2004). Nurses and the “therapeutic relationship”: Caring for

adolescents with anorexia nervosa. Journal of Advanced Nursing, 45(5), 495-
503.

Raven, J. C., Court, J. H., & Raven, J. (1994). Raven’s Advanced Progressive
Matrices. Pearson Assessment.

Regan, D. T., & Totten, J. (1975). Empathy and attribution: turning observers into
actors. Journal of personality and social psychology, 32(5), 850-6.

Reynolds W. (2000). The Measurement and Development of Empathy in Nursing.

Aldershot: Ashgate.

Reynolds, W. J., Scott, B., & Jessiman, W. C. (1999). Empathy has not been
measured in clients’ terms or effectively taught: a review of the literature.
Journal of Advanced Nursing, 30(5), 1177-1185.

Reynolds, W. J., & Scott, B. (2000). Do nurses and other professional helpers
normally display much empathy? Journal of Advanced Nursing, 31(1), 226-234.
 213

Rietveld, S., & Prins, P. J. (1998). The relationship between negative emotions and
acute subjective and objective symptoms of childhood asthma. Psychological
Medicine, 28(2), 407-415.

Rim, Y. (1994). Impulsivity, venturesomeness, empathy and schizotypy. Personality

and Individual Differences, 17, 853–854.

Robinson, D., Perryman, S., & Hayday, S. (2004). The Drivers of Employee
Engagement. Brighton: Institute for Employment Studies.

Robson, C. (2002). Real world research: A resource for social scientists and
practitioner –researchers. Oxford: Blackwell Publishing.

Rogers, C. R. (1957). The necessary and sufficient conditions of therapeutic

personality change. Journal of Consulting Psychology, 21(2), 95-103.

Russell, D., & McAuley, E. (1986). Causal attributions, causal dimensions, and
affective reactions to success and failure. Journal of Personality and Social
Psychology, 50(6), 1174-1185. doi: 10.1037/0022-3514.50.6.1174.

Rust, J., & Golombok, S. (1999). Modern Psychometrics: The science of

psychological assessment. London: Routledge.

Salovey, P. & Mayer, J.D. (1990). Emotional intelligence. Imagination, Cognition,

and Personality, 9, 185-211

Salovey, P., Mayer, J. D., Goldman, S. L., Turvey, C., & Palfai, T. P. (1995)
Emotional attention, clarity, and repair: Exploring emotional intelligence using
the Trait Meta-Mood Scale. In J. W. Pennebaker (Ed .), Emotion, Disclosure, &
Health. Washington: American Psychological Association.

Schmidt, G., & Weiner, B. (1988). An Attribution-Affect-Action Theory of

Behavior: Replications of Judgments of Help-Giving. Personality and Social
Psychology Bulletin, 14(3), 610-621.
 214

Schulze, R., & Roberts, R. D. (Eds). (2005). Emotional Intelligence: An

International Handbook. Gottingen: Hogrefe

Schutte, N. S., Malouff, J. M., Bobik, C., Coston, T. D., Greeson, C., Jedlicka, C.,
Rhodes, E., & Wendorf, G. (2001). Emotional intelligence and interpersonal
relations. The Journal of Social Psychology, 141(4), 523-536.

Schutte, N.S., Malouff, J.M., Hall, L.E., Haggerty, D.J., Cooper, J.T., Golden, C.J.,
et al. (1998). Development and validation of a measure of emotional
intelligence. Personality and Individual Differences, 25, 167-177.

Scotland, E., Mathews, K. E., & Sherman, S. (1978). Empathy, fantasy, and helping.
Beverly Hills, CA: Sage.

Silvester, J., Patterson, Fiona, Koczwara, A., & Ferguson, E. (2007). "Trust me...":
psychological and behavioral predictors of perceived physician empathy.
Journal of Applied Psychology, 92(2), 519-527.

Simmons, J., Roberge, L., Kendrick, S. B., & Richards, B. (1995). The Interpersonal
Relationship in Clinical Practice: The Barrett-Lennard Relationship Inventory
as an Assessment Instrument. Evaluation & the Health Professions, 18(1), 103-
112.

Smith, J.A. (2008). Qualitative Psychology: A Practical Guide to Research Methods.

London: Sage.

Sparr, L. F., Gordon, G. H., Hickam, D. H., & Girard, D. E. (1988). The doctor-
patient relationship during medical internship: the evolution of dissatisfaction.
Social Science & Medicine, 26(11), 1095-1101.

Spence, J. T., Helmreich, R. L., & Holohan, C. K. (1979). Negative and positive
components of psychological masculinity and femininity and their relationships
to self reports of neurotic and acting out behaviors. Journal of Personality and
Social Psychology, 37, 1673-1682.
 215

Spencer, J. (2004). Decline in empathy in medical education: how can we stop the
rot? Medical Education, 38(9), 916-918.

Spielberger, R. L., Gorusch, R. E., & Luschene, P. (1970). State-trait anxiety

inventory (self-evaluation questionnaire). Consulting Psychologists Press.

Squier, R. W. (1990). A model of empathic understanding and adherence to

treatment regimens in practitioner-patient relationships. Social Science &
Medicine, 30(3), 325-339.

Stepien, K. A., & Baernstein, A. (2006). Educating for Empathy. Journal of General
Internal Medicine, 21(5), 524-530.

Stewart, M. A. (1995). Effective physician-patient communication and health

outcomes: a review. Canadian Medical Association, 152, 1423-33.

Stinson, L., & Ickes, W. (1992). Empathic accuracy in the interactions of male
friends versus male strangers. Journal of Personality and Social Psychology,
62, 787-797.

Stotland, J. (1978). Empathy, fantasy and helping. California: Sage.

Stratton, T., Elam, C. L., Murphy-Spencer, A. E., & Quinlivan, S. L. (2005).

Emotional intelligence and clinical skills: Preliminary results from a
comprehensive clinical performance examination. Academic Medicine, 80, S34-
S37.

Strayer, J., & Eisenberg, N. (1987). Empathy viewed in context. In N. Eisenberg & J.
Strayer (Eds.), Empathy and its development (pp. 389-398). New York:
Cambridge University Press.

Suchman, A. L., Markakis, K. M., Beckman, H. B., & Frankel, R. M. (1997). A

model of empathic communication in the medical interview. Journal Of The
American Medical Association, 277(8), 678-682.
 216

Sue, D., & Sundbery, N. D. (1996). Research and research hypotheses about
effectiveness in intercultural counseling. In P. B. Pedersen, J. Draguns, W.
Lonner, & J. Trimble (Eds.), Counseling across cultures (pp. 323-352).
California: Sage.

Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics. Boston:

Pearson.

Tarrant, C., Windridge, K., Boulton, M., Baker, R., & Freeman, G. K. (2003).
Qualitative study of the meaning of personal care in general practice. British
Medical Journal, 326, 1310-1316.

The Scottish Office. (1997). Designed to care: Renewing the National Health
Service in Scotland. London: The Stationary Office.

Titchener, E. (1909). Experimental psychology of the thought process. New York:

Macmillan.

Trapnell, P. D., & Wiggins, J. S. (1990). Extension of the Interpersonal Adjective

Scales to Include the Big Five dimensions of personality. Journal of Personality
and Social Psychology, 59, 781-790.

Verby, J. E., Holden, P., & Davis, R. H. (1979). Peer review of consultations in
primary care: The use of audio-visual recordings. British Medical Journal, 1,
1686-1688.

Vreeke, G., & van der Mark, I. L. (2003). Empathy, an integrative model. New Ideas
in Psychology, 21(3), 177-207.

Walter, F. H., Cole, M. S., & Humphrey, R. H. (2011). Emotional Intelligence: Sine
Qua Non of Leadership or Folderol? Academy of Management Perspectives,
25(1), 45-59.

Wang, Y.-W., Bleier, J., Davidson, M., Savoy, H., Tan, J., & Yakushko, O. (2003).
The scale of ethnocultural empathy. Development, validation, and reliability.
Journal of Counseling Psychology, 2, 221-234.
 217

Wasserman, R. C., Inui, T. S., Barriatua, R. D., Carter, W. B., & Lippincott, P.
(1984). Pediatric clinicians’ support for parents makes a difference: an
outcome-based analysis of clinician-parent interaction. Pediatrics, 74(6), 1047-
1053.

Webster Nelson, D., & Baumgarte, R. (2004). Cross-Cultural Misunderstandings

Reduce Empathic Responding1. Journal of Applied Social Psychology, 34(2),
391-401.

Weiner, B. (1980). A cognitive (attribution)-emotion-action model of helping

behaviour: An analysis of judgements of help giving. Journal of Personality
and Social Psychology, 39, 1142-1162.

Weiner, B. (1985). An attributional theory of motivation and emotion. Psychological

Review, 2, 543-571.

Weiner, B. (1986). An Attributional Theory of Motivation and Emotion. Berlin:

Springer-Verlag.

Winefield, H. R., Murrell, T. G., & Clifford, J. (1995). Process and outcomes in
general practice consultations: problems in defining high quality care. Social
science medicine, 41(7), 969-975.

Woloschuk, W., Harasym, P. H., & Temple, W. (2004). Attitude change during
medical school: a cohort study. Medical Education, 38(5), 522-534.

Yardley, L. (2008). Demonstrating validity in qualitative psychology. In J. Smith

(Ed.), Qualitative psychology: a practical guide to research methods. London:
Sage.

Yarnold, P. R., Bryant, F. B., Nightingale, S. D., & Martin, G. J. (1996). Assessing
physician empathy using the Interpersonal Reactivity Index: A measurement
model and cross-sectional analysis. Psychology, Health and Medicine, 1, 207-
221.
 218

Zabar, S., Hanley, K., Kachur, E., Stevens, D., Schwartz, M. D., Pearlman, E.,
Adams, J., Felix, K., Lipkin, M., & Kalet, A. (2006), “Oh! She Doesn't Speak
English!” Assessing Resident Competence in Managing Linguistic and Cultural
Barriers. Journal of General Internal Medicine, 21, 510–513.

Zahn-Waxler, C., & Robinson, J. (1995). Empathy and guilt: early origins of feelings
of responsibility. In K. Fischer & J. Tangney (Eds.), Self conscious emotions:
Shame, guilt, embarrassment and pride (pp. 143-173). New York: Guilford
Press.

Zderad, L. (1970). Empathy – from cliché to construct. Third Nursing Theory

Conference (pp. 46-75). University of Kansas Medical Center Department of
Nursing, Kansas City.

Zwick, W. R., & Velicer, F. W. (1986). Comparison of five rules for determining the
number of components to retain. Psychological Bulletin, 99, 432-442.
 219

Appendix 1: Development of Research

This thesis began with the aim of designing an intervention to increase empathy in
doctors. For several reasons, this aim altered. The first was that, upon reading
evaluations of empathy training interventions, very little was found by way of detail
or consensus on what empathy training or development interventions should look
like in this context. The heart of this problem was a lack of consensus on a definition
of empathy in healthcare settings. At times, literature took a very broad approach,
but at times it was conflicting, with disagreement over the role of an emotional
response to patients. Some (e.g. Hardee, 2003) would view detached concern for
patients as insufficient, with a genuine emotional response required to demonstrate
empathy. Others, including researchers at the Jefferson Medical School in the US
(Hojat et al., 2004), view an emotional response as an interference in the process of
accurate diagnosis.

After reviewing the literature, the decision was taken that this thesis should therefore
take a step backwards. The research therefore seeks to build a model of empathy in
medicine that came from empirical research as opposed to opinion or literature
review. This model would then provide a solid foundation for designing
interventions in the future. In deciding on a model and measure to test as the central
theme of the thesis, Davis’ multidimensional model and corresponding measure of
empathy were chosen as an appropriate starting point.

The first data obtained were in fact the self report ratings of empathy and emotional
intelligence (study two) alongside the behavioural ratings within an assessment
centre context (study three). The aim of this study was to compare empathy with
emotional intelligence, then comparing their relative criterion-related validities by
examining the relationships with behaviour.

On completion of this study, it was noted that the measure being used would benefit
from greater evidence of construct validity in terms of recent personality
frameworks, namely the five factor model. At this stage then, the self report empathy
and personality questionnaires were administered (study one). It was a shame not to
 220

get access to detailed personality data for doctors; instead study one used a general
population study. However, this was an issue of access at the time, in taking the
decision to conduct a facet level analysis. It is certainly something for future
research.

Study four represented a break in the research in several ways. Primarily, the
methodology switched from quantitative to qualitative. This stemmed from a desire
to employ a more authentic method for the topic in question: the research conducted
using psychometric questionnaires was felt insufficient in fully capturing such an
interpersonal dynamic concept. In addition, the study focused on patient perceptions
of nurses’ empathy. Again this was not ideal as the other research was more focused
on empathy in doctors. However the hospital that agreed to take part in the project
and granted me ethical approval were more interested in the question of empathy in
nurses as they felt this was a more important issue for their practice. This is the
reality of conducting applied research – designs must be agreed in partnership with
organisations and as such need to be adapted to take their needs into consideration.
In the end, the study was valuable and there is much research to suggest that it is not
necessarily the role of the healthcare practitioner but the needs of the patient that
dictate the characteristics of an empathic interaction.

Overall, while the studies did not progress as originally intended, they do show clear
progression in the exploration of a model of empathy in healthcare roles. They have
hopefully built a solid foundation for the design and evaluation of empathy
development interventions, as well as identifying many further opportunities for
worthwhile and interesting research.
 221

Appendix 2: The Interpersonal Reactivity Index

The following statements inquire about your thoughts and feelings in a variety of situations.
For each item, indicate how well it describes you by choosing the appropriate letter on the
scale: A, B, C, D, or E. When you have decided on your answer, circle the appropriate letter.
Read each item carefully before responding, answering as honestly as you can.

Does not Describes

describe me very
me well well

1. I daydream and fantasize, with some regularity,

A B C D E
about things that might happen to me.

2. I often have tender, concerned feelings for

A B C D E
people less fortunate than me.

3. I sometimes find it difficult to see things from

A B C D E
the "other guy's" point of view.

4. Sometimes I don't feel very sorry for other

A B C D E
people when they are having problems.

5. I really get involved with the feelings of the

A B C D E
characters in a novel.

6. In emergency situations, I feel apprehensive

A B C D E
and ill-at-ease.

7. I am usually objective when I watch a movie or

play, and I don't often get completely caught up A B C D E
in it.

8. I try to look at everybody's side of a

A B C D E
disagreement before I make a decision.

9. When I see someone being taken advantage of,

A B C D E
I feel kind of protective towards them.

10. I sometimes feel helpless when I am in the

A B C D E
middle of a very emotional situation.

11. I sometimes try to understand my friends

better by imagining how things look from their A B C D E
perspective.

12. Becoming extremely involved in a good book

A B C D E
or movie is somewhat rare for me.

13. When I see someone get hurt, I tend to remain

calm. A B C D E
 222

Does not Describes

describe me very
me well well
14. Other people's misfortunes do not usually
A B C D E
disturb me a great deal.

15. If I'm sure I'm right about something, I

don't waste much time listening to other A B C D E
people's arguments.

16. After seeing a play or movie, I have felt

A B C D E
as though I were one of the characters.

17. Being in a tense emotional situation

A B C D E
scares me.

18. When I see someone being treated

unfairly, I sometimes don't feel very much A B C D E
pity for them.

19. I am usually pretty effective in dealing

A B C D E
with emergencies.

20. I am often quite touched by things that I

A B C D E
see happen.

21. I believe that there are two sides to every

A B C D E
question and try to look at them both.

22. I would describe myself as a pretty soft-

A B C D E
hearted person.

23. When I watch a good movie, I can very

easily put myself in the place of a leading A B C D E
character.

24. I tend to lose control during emergencies. A B C D E

25. When I'm upset at someone, I usually try

A B C D E
to "put myself in his shoes" for a while.

26. When I am reading an interesting story or

novel, I imagine how I would feel if the A B C D E
events in the story were happening to me.

27. When I see someone who badly needs

A B C D E
help in an emergency, I go to pieces.

28. Before criticizing somebody, I try to

imagine how I would feel if I were in their A B C D E
place.
Thank you for your participation.
 223

Appendix 3: Information Sheet (doctors)

My name is Helen Wilkin. I am a PhD student in the Psychology department of Goldsmiths

College, working with Professor Jo Silvester. We are currently investigating empathy in the
role of doctors. Empathy is an important concept within medical roles as it has been shown
to be an important factor in developing relationships to gain trust and confidence. This can
enhance delivery of care and increase patient satisfaction.

As part of my PhD, I am conducting a study to investigate empathy to better understand the

meaning of the concept and how it is measured. Information obtained from this study will be
used to build a greater understanding of how empathy can be measured in a medical setting.

Within the Wales Deanery psychometric tests are currently being piloted as part of the
selection centre process. This pilot work is entirely separate from the actual selection and
recruitment process. None of the results from the psychometric tests will be used in the
decision-making process.

Your Participation
I would be very grateful if you could spare around two hours to be involved in this piloting
of psychometric materials as part of the selection centre. Your participation is entirely
voluntary. Your decision about whether or not to participate will have no bearing whatsoever
on the selection and appointments process. If you choose not to participate, you will not be
adversely affecting your chances of being selected.

The Use of Information Collected

None of the information from the psychometric tests will be used to make decisions about
appointments to posts.

For the project, I will be asking you to complete 2 questionnaires, the Interpersonal
Reactivity Index which is a measure of empathy and the Bar-On EQ-I measure of emotional
intelligence. I would like to look at how your performance on the psychometric tests relates
to your performance in the selection centre exercises.

It should be stressed that all questionnaires will be anonymous. Findings will be discussed in
general terms only. Participation is voluntary and all information will be kept strictly
confidential.

If at any stage you wish to withdraw from the project, you are entirely free to do so. This
will in no way have any bearing on the outcome of this assessment centre. If you wish to
receive further information about the project following your participation, please provide
details on the following page so that I can send this to you.

I would very much appreciate your participation in this project.

Yours faithfully,

Helen Wilkin
 224

Appendix 4: Consent Form (doctors)

Piloting of Psychometric Tests

Wales Deanery GP selection centre

Record of Consent

Print Name: Date:

Signed
I confirm that I have volunteered to
participate in the piloting of psychometric
tests.

I understand that my performance on the

psychometric tests will not be used to
make decisions in today’s selection
process.

I consent that information about my

performance on the psychometric tests
may be used in Helen Wilkin’s PhD
regarding the measurement of empathy in
a medical setting. I understand that this
may involve looking at how my
performance on psychometric tests today
relates to my performance in other
selection centre exercises.

You may withdraw your consent at any time during or after completion of the
psychometric tests.

If you would like to receive a summary of the findings of this project, please leave an
address where this can be sent (this will not be used for research purposes and will be kept
strictly confidential).

Address (email or postal):

 225

Appendix 5: Factor loadings for the Bar On EQ-i items

Factor Loadings
Items 1 2 3 4 5 6 7 8
Factor 1
Item 56 0.82 0.08 0.03 0.07 -0.09 -0.10 -0.01 -0.01
Item 100 0.70 0.00 -0.01 -0.08 0.06 0.14 0.12 -0.25
Item 114 0.68 0.05 -0.05 -0.09 0.05 0.05 0.02 -0.21
Item 70 0.68 0.19 0.08 0.19 -0.04 -0.07 -0.05 -0.03
Item 85 0.67 -0.01 0.09 0.08 0.14 -0.07 -0.11 -0.08
Item 129 0.65 0.13 0.01 -0.15 0.09 -0.08 0.11 -0.05
Item 47 0.65 0.13 0.02 0.09 0.05 -0.20 -0.07 0.08
Item 40 0.63 -0.02 0.09 -0.15 -0.01 -0.16 0.04 0.08
Item 106 0.59 -0.03 -0.08 -0.05 0.26 0.05 -0.10 0.12
Item 91 0.57 -0.01 0.11 0.18 0.06 -0.22 -0.14 -0.03
Item 26 0.51 -0.10 -0.10 -0.17 0.02 -0.11 -0.17 -0.21
Item 77 0.49 0.22 0.09 0.01 -0.12 -0.02 -0.17 0.11
Item 02 0.49 -0.02 0.02 0.21 0.01 -0.28 -0.27 0.09
Item 24 0.46 0.00 0.16 -0.15 -0.26 -0.21 0.01 -0.19
Item 54 0.45 0.02 -0.09 -0.07 0.36 0.09 -0.04 0.02
Item 11 0.44 -0.10 0.07 -0.28 -0.09 0.05 -0.11 -0.09
Item 51 0.39 0.21 0.16 0.08 0.17 -0.15 0.04 -0.14
Item 80 0.38 -0.13 0.09 -0.18 0.17 -0.04 0.08 -0.13
Item 81 0.36 -0.13 0.16 -0.20 0.33 -0.02 0.00 -0.09
Item 21 0.33 0.08 0.11 0.02 -0.08 -0.16 -0.04 -0.11
Item 127 0.30 0.17 0.08 -0.10 0.10 0.05 -0.08 -0.14
Factor 2
Item 86 0.06 0.64 0.04 -0.09 -0.04 -0.11 -0.15 0.08
Item 102 0.06 0.61 -0.14 -0.09 0.03 0.02 -0.03 -0.11
Item 83 0.13 0.55 0.19 -0.20 -0.22 -0.10 0.02 0.06
Item 58 -0.06 0.51 0.02 0.03 0.16 0.13 -0.13 -0.12
Item 42 0.10 0.50 -0.08 -0.04 0.11 -0.17 -0.17 0.07
Item 76 0.15 0.50 -0.16 -0.03 0.09 -0.07 0.17 -0.14
Item 97 0.04 0.45 0.15 -0.04 0.07 0.26 -0.11 -0.23
Item 53 0.02 0.39 0.14 0.23 0.07 -0.37 0.00 -0.07
Item 104 0.16 0.36 -0.11 -0.24 0.10 0.00 0.14 -0.18
Item 38 -0.09 0.35 0.29 0.09 0.03 -0.20 -0.12 0.13
Item 39 0.26 -0.33 0.01 -0.11 0.08 -0.24 -0.30 -0.04
Factor 3
Item 107 0.10 0.07 0.58 0.09 -0.06 0.12 -0.05 -0.15
Item 48 0.14 0.12 0.56 -0.17 -0.05 0.09 0.07 -0.04
Item 92 -0.04 -0.31 0.55 -0.04 -0.14 -0.12 -0.10 -0.09
Item 19 0.04 0.00 0.55 -0.07 0.14 -0.01 0.16 0.07
Item 111 -0.03 0.00 0.55 0.00 0.03 -0.17 0.06 -0.16
Item 118 0.06 0.10 0.54 -0.14 0.05 0.09 -0.16 -0.09
Item 126 0.13 -0.07 0.54 -0.07 0.11 -0.24 0.02 0.00
Item 03 0.00 -0.12 0.52 0.07 0.05 0.07 -0.05 0.02
Item 32 0.00 -0.01 0.51 -0.20 -0.15 -0.07 0.06 -0.20
Item 67 0.15 -0.24 0.42 -0.06 -0.01 -0.14 -0.05 -0.08
Item 121 0.31 0.12 0.38 -0.06 0.03 0.11 0.03 -0.03
Item 82 -0.07 0.12 0.36 0.05 -0.27 -0.29 -0.13 -0.14
Item 93 -0.09 -0.01 0.36 -0.08 -0.14 -0.10 -0.21 -0.31
Item 75 -0.02 0.11 0.35 -0.16 -0.11 -0.07 -0.19 -0.22
Item 46 -0.01 0.12 0.34 -0.03 0.19 -0.17 0.03 0.07
Item 68 0.04 0.22 0.32 0.08 0.15 -0.23 -0.07 0.02
Item 66 0.09 0.00 0.25 0.19 0.23 -0.12 -0.19 -0.11
 226

Factor Loadings
1 2 3 4 5 6 7 8
Factor 4
Item 45 0.04 0.16 -0.03 -0.64 -0.04 -0.07 -0.01 -0.14
Item 15 -0.20 0.21 0.10 -0.55 0.13 -0.08 -0.04 -0.06
Item 29 0.00 0.20 0.04 -0.51 0.01 -0.14 0.05 -0.16
Item 60 0.06 -0.03 0.13 -0.47 0.27 -0.15 0.19 -0.26
Item 89 -0.07 0.03 0.13 -0.45 0.24 -0.05 0.15 -0.25
Item 20 0.20 -0.09 0.27 -0.41 0.04 0.14 -0.18 0.05
Item 04 0.00 -0.05 0.18 -0.41 0.01 -0.03 -0.20 -0.04
Item 108 0.25 -0.09 0.18 -0.40 0.13 -0.05 -0.10 -0.11
Item 78 0.15 0.07 0.21 -0.39 0.16 0.18 -0.23 -0.01
Item 08 0.25 0.26 0.02 -0.39 -0.21 -0.07 0.12 0.10
Item 30 -0.04 0.16 0.10 0.36 0.33 0.03 -0.06 -0.15
Item 88 0.13 -0.07 -0.09 -0.31 0.27 -0.04 -0.22 0.00
Item 06 0.25 -0.11 -0.05 -0.29 0.18 -0.06 0.12 -0.28
Item 63 0.02 -0.12 0.07 -0.29 0.29 -0.15 -0.28 0.15
Item 36 0.23 0.10 0.19 0.26 0.08 -0.21 -0.13 -0.22
Factor 5
Item 98 0.05 0.06 -0.02 0.01 0.65 -0.05 0.08 -0.10
Item 72 0.06 -0.04 -0.12 0.05 0.58 -0.08 -0.05 -0.16
Item 124 0.05 0.19 -0.16 0.01 0.55 0.09 -0.01 -0.21
Item 84 -0.06 0.01 0.11 0.10 0.49 -0.17 -0.19 0.08
Item 55 0.05 0.03 0.01 -0.14 0.49 -0.19 0.16 0.15
Item 90 0.06 0.09 0.11 -0.03 0.45 0.05 -0.07 -0.07
Item 110 0.31 -0.07 0.18 -0.10 0.43 0.03 -0.09 -0.07
Item 95 0.34 -0.08 0.18 -0.06 0.40 -0.01 -0.09 0.11
Item 44 -0.03 0.01 -0.05 -0.27 0.35 -0.05 -0.20 0.01
Item 128 0.14 0.00 0.24 0.08 0.28 -0.24 0.10 0.01
Item 112 0.08 0.06 0.18 -0.23 0.27 0.03 -0.10 0.02
Item 69 0.18 0.02 0.17 0.11 0.26 -0.20 -0.23 -0.06
Factor 6
Item 23 0.02 0.13 -0.05 0.02 0.05 -0.68 0.05 -0.18
Item 07 -0.05 -0.16 -0.10 -0.25 0.02 -0.68 -0.08 0.01
Item 52 0.02 0.07 0.11 0.11 0.01 -0.61 -0.03 -0.14
Item 116 0.12 0.19 -0.13 -0.07 -0.07 -0.61 -0.04 -0.18
Item 10 0.04 0.05 0.01 0.07 0.30 -0.56 0.11 0.08
Item 37 0.09 -0.11 0.11 -0.22 -0.13 -0.48 0.10 0.09
Item 96 0.17 -0.14 0.14 -0.18 0.08 -0.45 -0.06 -0.14
Item 35 0.23 0.26 0.01 -0.04 -0.01 -0.40 -0.14 -0.08
Item 31 0.30 -0.08 -0.20 -0.08 0.05 -0.39 -0.35 -0.06
Item 113 0.18 -0.21 0.12 -0.10 0.30 -0.31 -0.15 0.02
Factor 7
Item 13 0.03 0.21 -0.26 0.06 0.10 0.00 -0.60 -0.14
Item 64 0.28 0.06 0.14 0.04 -0.13 -0.01 -0.53 -0.13
Item 130 0.03 0.27 -0.18 -0.01 0.25 0.06 -0.51 -0.14
Item 117 0.08 0.35 -0.21 -0.02 0.14 0.05 -0.50 -0.22
Item 33 0.21 0.00 0.12 -0.44 -0.15 0.09 -0.45 0.02
Item 14 -0.02 -0.18 0.18 0.05 -0.08 -0.15 -0.44 -0.27
Item 122 0.37 0.01 0.10 0.04 -0.14 0.06 -0.43 -0.18
Item 49 0.15 0.12 0.26 -0.25 -0.12 0.10 -0.40 -0.08
Item 17 0.05 0.05 -0.03 0.12 0.17 -0.25 -0.39 0.00
Item 18 -0.08 0.18 0.14 -0.06 0.18 -0.03 -0.34 -0.01
Item 62 0.07 -0.29 0.27 -0.09 0.22 -0.20 -0.30 0.12
 227

Factor Loadings
1 2 3 4 5 6 7 8
Factor 8
Item 28 0.03 -0.03 -0.07 -0.02 -0.05 -0.19 -0.15 -0.61
Item 103 0.04 0.07 0.06 -0.03 0.16 -0.12 -0.08 -0.55
Item 87 0.07 0.14 0.08 -0.04 0.10 -0.20 -0.02 -0.52
Item 131 -0.01 -0.11 0.11 -0.03 -0.07 0.07 -0.06 -0.48
Item 74 0.16 -0.01 0.09 -0.06 0.12 -0.05 0.12 -0.46
Item 73 0.05 0.15 -0.13 -0.04 -0.03 0.04 -0.17 -0.45
Item 43 -0.02 0.20 0.12 0.27 0.13 0.07 -0.18 -0.42
Item 59 0.28 -0.27 0.18 -0.05 0.03 0.00 -0.01 -0.38
Item 01 -0.04 0.13 0.03 -0.29 0.11 -0.16 0.09 -0.32
Item 125 0.25 0.09 0.07 -0.04 0.09 -0.19 0.10 -0.28

Note: Factor loadings of 0.30 and greater are in boldface

 228

Appendix 6: Information sheet (patients)

Title of Project: Investigating empathy in the nurse-patient relationship

My name is Helen Wilkin. I am a PhD student in the Psychology department of Goldsmiths

College, working with Professor Jo Silvester. We are currently investigating empathy in the
role of nurses. I am inviting you to take part in a research study as part of my PhD. Before
you decide it is important for you to understand why the research is being done and what it
will involve. Please take time to read the following information carefully and discuss it with
others if you wish. Ask me if there is anything that is not clear or if you would like more
information. Take time to decide whether or not you wish to take part. Thank you for
reading this.

Purpose of the study

Empathy is important for nurses as it is needed for patients to have trust and confidence in
them. However very little research has looked at differences in people’s ability to empathise.
One possible reason for this is that research sometimes takes a narrow approach and doesn’t
focus on the things which are meaningful to nurses or patients.

As part of my PhD, I am conducting a study to better understand empathy, by talking

patients. Information obtained from this study will be used to build a greater understanding
of how empathy is shown to patients by nurses. I aim to contribute to the training and
development of nurses in a way which focuses on the things that mean most to patients.

What is involved
I am talking to a range of patients, to find out how and when nurses are able to show
empathy to them. Participation is voluntary. It is up to you to decide whether or not to take
part. If you do decide to take part you will be given this information sheet to keep and be
asked to sign a consent form. If you decide to take part you are still free to withdraw at any
time and without giving a reason. A decision to withdraw at any time, or a decision not to
take part, will not affect the standard of care you receive.

For the project, I would like to interview you for 30 minutes to talk about nurses who are
good at empathising with patients as well as those who find it more difficult. I will ask for
examples of how these people differ from each other. You will not need to mention names
and all interviews will be confidential. Following the interviews, if you have any questions
or concerns, you will able to contact me to discuss them. I will be tape recording the 30
minute interview to make sure that I record all of the information accurately. These tapes
will be kept securely until I have written up the information which I need, all of which will
be anonymous. I will then destroy the tapes so that they are unusable.

If you have had a bad experience or find it upsetting to talk about this, please do not feel you
need to take part. If you wish to complain, or have any concerns about any aspect of the way
you have been approached or treated during the course of the study, the normal National
Health Service complaints mechanism should be available to you.

If you consent to take part in this study I will not have access to your medical records. Your
name will not be disclosed, you will not be recognised from the written information and all
information will be kept strictly confidential. If you take part, a copy of the consent form
will be kept on your hospital notes.
 229

The results of the research will be written up as part of my PhD. They may also be published
in a journal for other psychologists. You will not be identified in any written or published
reports.

The research is being funded by Goldsmiths College, University of London. The James
Cook University Hospital will not be paid for the study, nor will I.

The Research Ethics Committee here at the James Cook University hospital has reviewed
this study to make sure it is ethical.

You can keep this copy of the information sheet and also a copy of your signed consent form
if you agree to take part.

I would very much appreciate your participation in this project. If you are interested in
taking part, please contact me by either telephone or email (details above) and we will
arrange a time to meet. I look forward to hearing from you,

Yours faithfully,

Helen Wilkin
 230

Appendix 7: Consent form (patients)

CONSENT FORM - Patients

Title of Project: Investigating Empathy in the nurse-patient

relationship
Name of Researcher: Helen Wilkin, Goldsmiths College, University
of London

Please initial box

1. I confirm that I have read and understand the information sheet for
the above study and have had the opportunity to ask questions.
2. I understand that my participation is voluntary and that I am free to
withdraw at any time, without giving any reason, without my medical
care or legal rights being affected.
3. I understand that the researcher will not need to have access to my
medical notes as these are not relevant to the research.

4. I agree to take part in the above study.

_________________________ ________________ _____________

Name of Patient Date Signature

_________________________ ________________ _____________

Name of Person taking consent Date Signature
(if different from researcher)

_________________________ ________________ _____________

Researcher Date Signature

1 for patient; 1 for researcher; 1 to be kept with hospital notes.

 231

Appendix 8: Study four interview schedule

1. Welcome, Introduction [My role, confidentiality, right to withdraw, Interviews

will be recorded and will last approximately 30 minutes].

2. Can you describe what you understand by the term empathy?

[If absolutely unable to do this, interview will not continue]

3. Please can you tell me about an incident you have experienced here when one of
the nurses was able to empathise with you?
Prompts: Can you give a specific example?
What was the situation?
What did the nurse do?
Why did that happen?
What was the outcome?

3. Please can you tell me about an incident you have experienced here when one of
the nurses was NOT able to empathise with you?
Prompts: Can you give a specific example?
What was the situation?
What did the nurse do?
Why did that happen?
What was the outcome?

4. Finally, I will be sending summary of the research to anyone who wishes to see it.
Would you like to receive this? If so, please leave contact details which will be kept
strictly confidential.

5. Thanks for participating.

 232

Appendix 9: Coded Interview

233