See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/313429953

The cognitive emotion regulation questionnaire: Factorial, convergent and

criterion validity analyses of the full and short versions

Article in Personality and Individual Differences · February 2017

DOI: 10.1016/j.paid.2017.01.035

CITATIONS READS

36 3,335

3 authors:

Michael James Ireland Bonnie A Clough

University of Southern Queensland Griffith University
78 PUBLICATIONS 1,059 CITATIONS 47 PUBLICATIONS 1,273 CITATIONS

SEE PROFILE SEE PROFILE

Jamin J Day
The University of Newcastle, Australia
37 PUBLICATIONS 1,165 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Working Well: Investigation and Intervention into Drs Wellbeing View project

Psychometric Research Program View project

All content following this page was uploaded by Michael James Ireland on 25 March 2020.

The user has requested enhancement of the downloaded file.

 1

The Cognitive Emotion Regulation Questionnaire: Factorial, convergent, and criterion

validity analyses of the full and short versions

 2

Abstract

Aim: Emotion regulation (ER) theories and measurement scales have received considerable

attention in clinical and research settings. However, there is a need for independent validation

of these scales to ensure rigor within this field. The aim of the current study was to examine

the factorial, convergent, and criterion validity of one of the most popular measures of ER,

the Cognitive Emotion Regulations Questionnaire (CERQ), both short and long form.

Methods: The CERQ (and CERQ-short), positive and negative affect schedule, and

difficulties in emotion regulation scale were administered to 795 participants (70% female, M

age = 36.36).

Results: Confirmatory factor analysis supported the 9-factor structure of the CERQ-short, but

not the full CERQ. Adequate fit for the full CERQ was achieved after the removal of three

poorly performing items. Correlations supported the convergent and criterion validity of both

scales, although the CERQ-short demonstrated weaker associations than the full scale.

Conclusions: The factorial, convergent, and criterion validity of the CERQ and CERQ-short

were generally supported. However, future research may wish to examine several high inter-

factor correlations that were observed among the full CERQ, as well as the weaker validity

demonstrated by the CERQ-short.

 3

1. Introduction

Theories of emotion regulation have become increasingly popular and useful in

modern accounts of psychological functioning and psychopathology. Emotion regulation has

been defined as the (conscious or unconscious) processes of monitoring, evaluating, and

modulating emotional experiences in goal-directed ways (Thompson, 1994). Broadly

speaking, the goals of this control process relate to which emotions are experienced, when

and how they are experienced, and how they are expressed (Gross, 1998).

A broad range of behavioral and cognitive strategies can be employed to achieve

these regulatory goals. However, different strategies have different implications for adaptive

functioning (John & Gross, 2004). Importantly, individuals who fail to regulate their

emotions, or who use ineffective or maladaptive strategies, are at greater risk of suffering

emotional disorders (Aldao & Nolen-Hoeksema, 2010). In fact, emotion dysregulation has

received attention as a transdiagnostic risk factor implicated in a range of psychological

disorders (Kring & Sloan, 2009).

Of particular clinical interest are the ways in which people regulate and adapt to

stressful or negative emotional states. To facilitate research into these issues, Garnefski,

Kraaij, and Spinhoven (2001) constructed a self-report multidimensional instrument (the

Cognitive Emotion Regulation Questionnaire or CERQ) designed to assess a range of

cognitive emotion regulation strategies. The full instrument has 36-items evenly divided

among nine subscales: four are adaptive strategies (Putting into perspective, Positive

refocusing, Positive reappraisal and Planning) and five are maladaptive strategies (Self-

blame, Other-blame, Rumination, Catastrophizing, and Acceptance). The CERQ has

demonstrate promising psychometric properties and has become widely adopted with the

original derivation article amassing 835 citations according to Google Scholar.

 4

Garnefski and Kraaij (2006) also developed a fast screening instrument for use among

psychiatric patients and to enable easier inclusion in large questionnaire batteries. The

strategy employed was to halve the number of items on each subscale through the use of a

stepwise omission process based on the highest ‘alpha if item deleted’ criterion. The

hypothesized factor structure of the instrument remained unchanged with principle

components analysis (PCA) producing the expected nine factors accounting for 82.7% of

variance. While this approach provides a simple way of culling items, it risks creating scales

that are too narrow and with attenuated reliability and validity (Loevinger, 1954).

Specifically, removal of components based on ‘alpha if item deleted’ has been shown to

inflate coefficient alpha while leading to a deterioration of actual reliability and validity

(Raykov, 2007, 2008). Furthermore, although the derivation study of the CERQ-short

reported preliminary evidence for its validity (via PCA and correlations with depression and

anxiety), novel replication in an independent sample remains crucial.

Confirmatory factor analysis (CFA) provides a powerful tool to test the hypothesized

latent structure and construct validity of test instruments. Only one CFA of the full 36 item

English instrument could be located (Garnefski & Kraaij, 2007). While model fit was

adequate for a sample at two time points (time 1 AASR = .045, χ2 [546] = 591.58, p =.008,

CFI = .92 and time 2 AASR = .049, χ2 [560] = 558.79, p =.510, CFI = .97), there was a lack

of detail in reporting the results particularly regarding the number and magnitude of error

covariances that were freely estimated (the AASR or Average Absolute Standardized

Residual measures degree of misfit and values <.06 indicate good fit). The 9-factor structure

of the full CERQ has also been examined and supported by investigations using other

language groups such as French (Jermann, Linden, d'Acremont, & Zermatten, 2006), Spanish

(Domínguez-Sánchez, Lasa-Aristu, Amor, & Holgado-Tello, 2011), Chinese (Zhu et al.,

2008), Turkish (Tuna & Bozo, 2012). However, confirmatory factor analyses of the English
 5

version of full CERQ as well as the CERQ-short are required to validate the hypothesized

structure of these instruments.

The aim of this study was to test the theoretical structures of the CERQ and CERQ-

short in a mixed non-clinical sample. A secondary aim was to compare the extent that CERQ

and CERQ-short scales co-vary with subscales from the Difficulties in Emotion Regulation

Scale (DERS) and predict positive and negative affect, thus determining their convergent and

criterion validity. These validity scales were selected for several reasons. The choice of the

DERS for converging validity was based on the need for a well-validated emotion regulation

instrument that captured a broad range of emotion regulation facets. The DERS and the

Emotion Regulation Questionnaire (ERQ: Gross & John, 2003) are the best validated

emotion regulation instruments, however the ERQ is restricted to measuring only two facets

of emotion regulation. Regarding criterion validity, it is well established that emotion

regulation has a direct influence on emotional experiences, however given the current sample

was not a clinical sample, we were concerned that commonly used emotional disturbance

scales such as measures of anxiety and depression may suffer a restriction in range.

Therefore, the PANAS was selected as the criterion measure as it is one of the most well-

established instruments to assess positive and negative affective states in non-clinical

samples.

It was hypothesized that the CERQ subscales representing maladaptive strategies

would correlate positively with greater Difficulties in Emotion Regulation Strategies and

Negative affect, and would correlate inversely with Positive affect. Conversely, the CERQ

subscales representing adaptive strategies would correlate inversely with greater Difficulties

in Emotion Regulation Strategies and Negative affect, and would correlate positively with

Positive affect
 6

2. Method

2.1. Participants and Procedure

The sample consisted of 795 community residents (68.3% in full- or part-time

employment) and students (17.4%). Most participants were female (70%) and the mean age

was 36.36 (SD = 13.86).

Following ethics approval by the host institution, data was collected anonymously

online, using the research platform Qualtrics. Community participants were recruited via

social media advertisements, personal networks, discussion forums, email discussion lists,

and snowball sampling. Student participants were recruited through the online research

participation portal at the host institution, and could apply for course credit in exchange for

their participation. No other incentives were offered.

2.2. Measures

2.2.1. Cognitive Emotion Regulation Questionnaire. The CERQ is a 36 item self-

report measure of cognitive emotion regulation strategies used when responding to a stressful

life event. There are nine four-item subscales with five of them assessing maladaptive and

adaptive regulatory responses. Participants rate how often statements apply to them on a five

point scale from 1 “almost never”, to 5 “almost always”. Each subscale demonstrates

acceptable internal consistency (Cronbach’s α >.70) and the CERQ has been shown to have

acceptable test-retest reliability (ranging from .41 to .59; Garnefski et al., 2001).

The CER-short is a brief version of the CERQ that halves the number of items for

each of the nine subscales. Alpha coefficients for the short scales range from .68 (Self-blame)

to .81 (Positive appraisal and Catastrophizing; Garnefski & Kraaij, 2006).

2.2.2. Positive and Negative Affect Schedule. The Positive and Negative Affect

Schedule (PANAS; Watson, Clark, & Tellegen, 1988) is a 20-item scale (10 for each
 7

construct) designed to measure positive and negative affectivity. Participants rate how often

statements apply to them on a five-point scale from 1 “very slightly or not at all”, to 5

“extremely”. Since its introduction the PANAS has accumulated a compelling body of

evidence attesting to its reliability and validity.

2.2.1. Difficulties in Emotion Regulation Scale. The Difficulties in Emotion

Regulation Scale (DERS; Gratz & Roemer, 2004). The DERS is a 36-item measure that

evaluates the individual’s response to negative emotional states. Participants rate how often

statements apply to them on a five-point scale from 1 “almost never”, to 5 “almost always”.

Items combine to form six subscales: Lack of acceptance of emotions (example item, “When

I’m upset, I feel guilty for feeling that way”), Inability to engage in goal-directed behavior,

Impulse control difficulties when upset; (4) Limited access to strategies for emotion

regulation; (5) Lack of awareness of emotions; and (6) Lack of clarity of emotions. The total

DERS demonstrated high internal consistency (α = .93), with all six subscales also

demonstrating adequate internal consistency (Cronbach’s α >.80).

2.3. Analysis

Mplus version 7.4 (Muthén & Muthén, 2012) was used for CFAs and IBM SPSS

version 22 was used for the other analyses. Due to non-normality (Mardia’s normalized

coefficient of multivariate kurtosis exceeded recommended cut-off critical ratio < 3.00), the

measurement models were fit using the Satorra-Bentler scaling correction (MLM estimator;

Satorra & Bentler, 1994), which estimates a mean-adjusted chi-square to account for non-

normal data.

CFA model fit was assessed against multiple criteria. The chi-square test is reported

though interpreted with caution given it is heavily influenced by sample size. The relative χ2

(χ2/df) will also be used since it is adjusted for sample size, and alternative fit indices
 8

including the Root Mean Squared Error of Approximation (RMSEA: Steiger, 1990),

Standardized Root Mean Square Residual (Hu & Bentler, 1999), and Comparative Fit Index

(CFI; Bentler, 1990) are reported to assess model fit. Following convention (Hu & Bentler,

1999), good model fit is indicated by non-significant chi-square, a relative χ2 < 3 (Kline,

2011), RMSEA and SRMR less than approximately .07 (Steiger, 2007), and CFI > .95

(Hooper, Coughlan, & Mullen, 2008).

3. Results

3.1. Confirmatory Factor Analyses

Indices of fit for the specified models are displayed in Table 1. For the full 36 item

CERQ, χ2 was significant (p <.001) and large, relative to the degrees of freedom (> 3).

Despite sub-optimal CFI and SRMR values, the RMSEA was within the acceptable range (<

.07). Model fit was better for the CERQ-short with all fit indices except for a significant χ2 (p

<.001) suggesting good fit to the data. Post-hoc exploratory analysis of a modified full CERQ

model (with the same estimation method) was attempted after removing items with low

estimated coefficients. These modifications produced acceptable fit (reported in Table 1 as

Modified Full CERQ).

Table 1

Fit indices for all measurement models

Relative RMSEA
Model χ2 df χ2 (90% CI) SRMR CFI
Full CERQ 1958.13 558 3.51 .06 (.055, .060) 0.101 .88
Modified Full CERQ 1051.52 459 2.29 .04 (.038, .044) 0.047 .95
8-factor CERQ 1 2033.04 566 3.59 .06 (.056, .061) 0.100 .88
CERQ-short 260.70 99 2.63 .05 (.039, .053) 0.060 .96
1
NOTE: full CERQ model estimated after collapsing items from Rumination and Catastrophizing
 9

Item loadings for both models are displayed in Table 2. While all estimates are

statistically significant, there were a few weak coefficients (bolded). For the full CERQ,

relatively low weights (< .50) were observed for one item on the Acceptance scale and two

items on the Rumination scale. For the CERQ-short, only one of the items on the Rumination

scale had a low weighting.

Correlations among the estimated latent factors are displayed in Table 3. In general

inter-correlations were weaker among the CERQ-short factors. The association between

Acceptance and Positive reappraisal was significant for the CERQ-short, but not for the full

CERQ. Conversely, correlations between Rumination and Planning and Positive reappraisal

became non-significant for the CERQ-short, in contrast to the full CERQ. The same was true

of the correlation between Positive reappraisal and Self-blame.

The high correlation between Rumination and Catastrophizing suggests that

collapsing these might improve fit and lead to a more parsimonious measurement model. As

seen in Table 1, this 8-factor model demonstrated slightly worse fit to that of the full 9-factor

model.
 10

Table 2

Factor loadings for the CFA of full CERQ and the CERQ-short

Full CERQ CERQ-short

95% CI 95% CI
Factor Item Estimate Low Hi Estimate Low Hi
Acceptance I think that I have to accept that this has happened 0.67 0.60 0.73 0.70 0.62 0.78
Acceptance I think I have to accept the situation 0.76 0.70 0.82 0.75 0.67 0.84
Acceptance I think that I cannot change anything about it 0.26 0.19 0.34
Acceptance I think that I must learn to live with it 0.54 0.47 0.60
Catastrophizing I keep thinking about how terrible it is what I have experienced 0.67 0.62 0.71 0.64 0.59 0.69
Catastrophizing I continually think how horrible the situation has been 0.82 0.78 0.85 0.85 0.80 0.89
Catastrophizing I often think that what I have experienced is much worse than what other 0.48 0.42 0.54
Catastrophizing I often think that what I have experienced is the worst that can happen 0.53 0.47 0.59
Other blame I feel that basically the cause lies with others 0.84 0.80 0.87 0.64 0.63 0.78
Other blame I feel that others are responsible for what has happened 0.77 0.73 0.80 0.82 0.82 0.96
Other blame I feel that others are to blame for it 0.87 0.84 0.91
Other blame I think about the mistakes others have made in this matter 0.53 0.48 0.59
Planning I think about a plan of what I can do best 0.67 0.63 0.71 0.69 0.64 0.74
Planning I think about how to change the situation 0.76 0.73 0.80 0.78 0.73 0.83
Planning I think of what I can do best 0.76 0.73 0.79
Planning I think about how I can best cope with the situation 0.78 0.75 0.81
Positive reappraisal I think that I can become a stronger person as a result of what has happened 0.56 0.51 0.60 0.68 0.62 0.73
Positive reappraisal I think I can learn something from the situation 0.68 0.64 0.72 0.81 0.75 0.86
Positive reappraisal I think that the situation also has its positive sides 0.79 0.76 0.82
Positive reappraisal I look for the positive sides to the matter 0.85 0.83 0.88
Positive refocusing I think of pleasant things that have nothing to do with it 0.79 0.76 0.82 0.81 0.76 0.85
Positive refocusing I think of something nice instead of what has happened 0.86 0.83 0.88 0.86 0.83 0.90
Positive refocusing I think of nicer things than what I have experienced 0.80 0.77 0.83
Positive refocusing I think about pleasant experiences 0.78 0.74 0.83
Putting into perspective I think that it hasn’t been too bad compared to other things 0.80 0.77 0.84 0.82 0.77 0.85
Putting into perspective I tell myself that there are worse things in life 0.82 0.79 0.85 0.78 0.74 0.83
Putting into perspective I think that it all could have been much worse 0.63 0.59 0.67
Putting into perspective I think that other people go through much worse experiences 0.79 0.76 0.82
Rumination I often think about how I feel about what I have experienced 0.29 0.23 0.35 0.33 0.26 0.41
Rumination I am preoccupied with what I think and feel about what I have experienced 0.76 0.72 0.80 0.90 0.80 1.01
 11

Rumination I want to understand why I feel the way I do about what I have experienced 0.14 0.08 0.20
Rumination I dwell upon the feelings the situation has evoked in me 0.78 0.74 0.82
Self-blame I think that basically the cause must lay within me 0.81 0.77 0.84 0.83 0.77 0.89
Self-blame I feel I am the one who is responsible for what has happened 0.74 0.69 0.78 0.73 0.67 0.79
Self-blame I feel that I am the one to blame for it 0.82 0.79 0.86
Self-blame I think about the mistakes I have made in this matter 0.53 0.48 0.58
Note: All estimates significant at the p < .001 level. Indicators with poor factor loadings are shown in bold.

Table 3

Inter-correlations among latent factors

Full CERQ CERQ-short

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9
1. Acceptance .26 .21 .01 ns .01 ns
-.01 ns .17 .33 .38 .14 .15 .02 ns .14 .05 ns .24 .23 .33
2. Catastrophizing .55 -.50 -.60 -.46 -.46 .92 .61 .52 -.45 -.43 -.48 -.41 .76 .55
3. Other blame -.13 -.23 -.17 -.16 .51 .32 -.10 -.16 -.19 -.14 .44 .24
4. Planning .87 .73 .72 -.28 -.25 .73 .67 .70 -.14 ns -.19
5. Positive reappraisal .82 .78 -.37 -.25 .55 .66 -.12 ns -.08 ns
6. Positive refocusing .72 -.36 -.24 .75 -.33 -.22
7. Putting into perspective -.26 -.17 -.21 -.18
8. Rumination .66 .53
9. Self-blame
Note. ns = not significant at α = .05
 12

3.2. Validity and Reliability Analysis

To investigate the convergent and criterion validity of the CERQ and CERQ-short,

associations with the DERS (convergent validity) and PANAS (criterion validity) were

calculated (see Table 4). With the exception of the Unaware subscale of the DERS, the

pattern of correlations were largely significant and in the directions expected. At the level of

the individual associations, only one convergent and one criterion correlation differed in

significance between the full CERQ and CERQ-short. The correlation between the CERQ

Acceptance subscale and the Impulse control difficulty subscale of the DERS was moderate

and significant for the full CERQ, but non-significant for the CERQ-short. The correlation

between the CERQ Acceptance subscale and Positive affect was small and significant for the

full CERQ, but non-significant for the CERQ-short. For the whole measure, R2 values

indicated a statistically significant small to medium (following Cohen's 1988 benchmarks)

drop in variance explained by the CERQ-short when compared with the full CERQ.

Regarding reliability, alpha coefficients for most CERQ and CERQ-short scales were

‘respectable’ (> .7; DeVellis, 1991). The full CERQ Acceptance and Rumination scales had

‘undesirable’ coefficients (< .65), however the Acceptance scale from the CERQ-short was

‘minimally acceptable’. Coefficient alpha for CERQ-short Planning was minimally

acceptable, however the Rumination scale was ‘unacceptable’ (< .50).

Finally, the degree of overlap between the long (unmodified) and short CERQ factors

was calculated. These correlations ranged from .87 (for Positive reappraisal) to .97 (for

Acceptance, Catastrophizing, Positive refocusing, and Putting into perspective). This

represents between 75% and 94% of variance shared between long and short version factors.
13
 14

Table 4

Alpha coefficients, and convergent and criterion validity correlations with DERS and PANAS Subscales

DERS Subscales PANAS Scales

CERQ α Non-accept Goals Impulse Unaware Strategies Clarity Positive Negative
Full .62 .41 .27 .27 .05 ns .38 .25 -.14 .31
Acceptance ns ns ns
Short .68 .26 .11 .08 -.01 .14 .11 -.01 .17
Full .73 .51 .52 .56 .07 ns .75 .35 -.25 .42
Catastrophizing
Short .70 .46 .57 .50 .02 ns .74 .29 -.27 .40
Full .83 .32 .32 .40 -.01 ns .44 .19 -.16 .27
Other blame
Short .76 .30 .31 .42 .02 ns .43 .18 -.16 .22
Full .83 -.31 -.34 -.36 -.43 -.52 -.38 .42 -.23
Planning
Short .69 -.27 -.27 -.27 -.40 -.43 -.32 .36 -.16
Full .82 -.28 -.37 -.35 -.37 -.53 -.33 .44 -.23
Positive reappraisal
Short .71 -.18 -.26 -.22 -.35 -.38 -.27 .36 -.16
Full .88 -.25 -.39 -.34 -.25 -.48 -.28 .39 -.25
Positive refocusing
Short .81 -.23 -.38 -.32 -.21 -.45 -.26 .37 -.23
Full .85 -.19 -.26 -.32 -.29 -.40 -.27 .31 -.19
Putting into perspective
Short .78 -.21 -.27 -.32 -.31 -.42 -.31 .31 -.21
Full .63 .35 .50 .37 -.46 .45 .08 -.06 .37
Rumination
Short .47 .34 .49 .35 -.40 .41 .08 -.08 .34
Full .81 .64 .48 .42 .06 ns .58 .38 -.23 .48
Self-blame
Short .75 .59 .42 .38 .06 ns .52 .37 -.20 .45
Full .51 .50 .46 .45 .76 .30 .51 .58
R2
Short .46 .47 .39 .37 .70 .25 .48 .55
ΛR2 , p - .06, < .001 .03, <.001 .06, < .001 .08, < .001 .07, < .001 .05, < .001 .03, < .001 .04, < .001
Note. = not significant at α = .05.
ns
Non-accept = Non-acceptance of emotional responses; Goals = Difficulties engaging in goal directed behaviour; Impulse

= Impulse control difficulties; Unaware = Lack of emotional awareness; Strategies = Limited access to emotion regulation strategies; Clarity = Lack of

emotional clarity.
 15

4. Discussion

The CERQ is one of the most widely used instruments to assess emotion regulation.

The 9-factor measurement model for the CERQ-short fit the data well. However, the model

fit for the full CERQ was inadequate. Adequate fit was achieved for the full CERQ after the

removal of three poorly fitting items. Following these modification, model fit observed in the

current sample was similar to that mentioned in Garnefski and Kraaij (2007). While their

analysis and results are not fully described, they reported CFI values of .92 and .97 (assessed

at two time-points). Positive support for the measurement model are also consistent with

those reported from investigations using non-English-speaking samples (Domínguez-Sánchez

et al., 2011; Jermann et al., 2006; Tuna & Bozo, 2012; Zhu et al., 2008).

Inspection of the poorly performing items suggests possible reasons for the misfit of

the full CERQ. For example, “I think that I cannot change anything about it”, was also found

to be the lowest loading item on the Acceptance factor previously (Garnefski & Kraaij, 2006;

Garnefski et al., 2001). This item may be tapping something closer to a sense of futility rather

than acceptance. Likewise, the worst fitting item on the rumination factor (i.e., “I often think

about how I feel about what I have experienced”) also had the lowest loading in prior

analyses (Garnefski & Kraaij, 2006; Garnefski et al., 2001), and may be tapping something

closer to tendencies for curiosity or speculation rather than rumination. However, this is

purely speculation and any conclusions may be spurious and sample-specific. Given the

removal of these items was the determining factor in producing adequate model fit, they

should be targeted for future revision.

For six of the nine factors (all except Acceptance, Catastrophizing, and Putting into

perspective, different items for the CERQ-short would have been selected if they were based

on the highest factor loadings from the current results. While these loadings may be sample

specific, it is noteworthy that the ‘alpha if item deleted’ criterion resulted in the retention of
 16

one of the worst fitting items “I often think about how I feel about what I have experienced”.

This highlights the need to use multiple criteria in scale refinement and for cross-validating of

the resulting instrument.

The current results generally support the factorial validity of the CERQ and the

CERQ-short, and also provide a novel replication of the converging and criterion validity

evidence. One finding of note is the number of high correlations among the latent CERQ

factors. For the full-scale, seven of these associations were >.70 (indicating >50% overlap in

the measured variance). The high degree of overlap between Planning, Positive reappraisal,

and Positive refocusing suggests some potential redundancy. Similarly, Catastrophizing and

Rumination share 85% of variance and thus, as measured by the CERQ are likely

conceptually redundant at the latent-factor level. In fact, a modified model that collapses

Ruination and Catastrophizing produced only slightly poorer fit and, for some researchers, it

may be worth combining them to achieved increased parsimony. Regardless, future scale

revision would be beneficial to further demarcate these factors. In general, the pattern of high

inter-factor correlations were fewer and less pronounced for the CERQ-short.

The converging and criterion correlations were largely as hypothesized except for the

Unaware scale of the DERS, which did not significantly correlate with four of the

maladaptive strategies of the CERQ (Acceptance, Catastrophizing, Other blame, and Self-

blame). The CERQ and CERQ-short differed with regard to the association of Acceptance

with Impulse control difficulties and Positive affect. The items removed to form the CERQ-

short (“I think that I cannot change anything about it” and “I think that I must learn to live

with it”) may represent content that is critically relevant to the overlap of the full Acceptance

scale with Impulse control difficulties and its inverse relationship to Positive affect. This

impacts the construct validity of the resulting scale and researchers choosing to use the

CERQ-short need to be aware of these potential trade-offs. Overall, the pattern of correlations
 17

showed that the CERQ-short exhibited similar, though generally weaker, associations with

DERS and PANAS subscales. The multiple coefficient of determination estimates indicate

the average discrepancies in variance explained is small (M ΔR2 = .05). Whether this reflects

a deterioration in general construct validity as a result of the item trimming is an important

issue for future investigation. Attenuated correlations may have also resulted from generally

weaker internal consistency of the CERQ-short scales.

4.1. Limitations

The inability to calculate recruitment response rate and the high proportion of females

within the sample are limitations of the present research. The convergent and criterion

validity analyses may be degraded by validity issues with the PANAS and DERS. The

evidence largely supports the validity of these instruments, however, there is some evidence

suggesting revisions to the measurement model improve their fit (Bardeen, Fergus, & Orcutt,

2012; Merz et al., 2013). Current space limitations prevent us from testing and refining these

measurement models of these instruments and this also lies outside the scope of the current

study.

4.2. Conclusions

Validation of scales using independent samples is crucial for understanding the body

of evidence they produce, and allowing clinicians and researchers to make informed

decisions regarding appropriate use of scales among treatment and research populations. This

process was deemed to be of particular importance for the CERQ and CERQ-short, given

their popularity. The present study has generally found support for the construct, convergent,

and criterion validity of these scales. The current results provide confirmation that the 9-

factor model is appropriate for the CERQ-short, and after removing poorly performing items,

for the full CERQ as well. However, several very high inter-factor correlations among the full

CERQ and a trend for the CERQ-short to produce weaker associations with converging and
 18

criterion constructs will be important avenues to be addressed in future research. The high

overlap among factors also suggests that alternative hierarchical models warrant testing with

an independent sample, and this is will be an important task for future research. Validation of

these instruments within clinical samples is also required, especially as the CERQ-short was

purportedly developed for use in psychiatric screening.

 19

References

Aldao, A., & Nolen-Hoeksema, S. (2010). Specificity of cognitive emotion regulation

strategies: A transdiagnostic examination. Behaviour research and therapy, 48(10),
974-983.
Bardeen, J. R., Fergus, T. A., & Orcutt, H. K. (2012). An Examination of the Latent Structure
of the Difficulties in Emotion Regulation Scale. Journal of psychopathology and
behavioral assessment, 34(3), 382-392. doi:10.1007/s10862-012-9280-y
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin,
107(2), 238.
Cohen, J. (1988). Statistical power analysis for the behavioural sciences. Hillside. NJ:
Lawrence Earlbaum Associates.
DeVellis, R. F. (1991). Scale development: Theory and applications. Thousand Oaks, CA,
US: Sage Publications, Inc.
Domínguez-Sánchez, F. J., Lasa-Aristu, A., Amor, P. J., & Holgado-Tello, F. P. (2011).
Psychometric Properties of the Spanish Version of the Cognitive Emotion Regulation
Questionnaire. Assessment, 20(2), 253-261. doi:10.1177/1073191110397274
Garnefski, N., & Kraaij, V. (2006). Cognitive emotion regulation questionnaire–development
of a short 18-item version (CERQ-short). Personality and Individual differences,
41(6), 1045-1053.
Garnefski, N., & Kraaij, V. (2007). The cognitive emotion regulation questionnaire.
European Journal of Psychological Assessment, 23(3), 141-149.
Garnefski, N., Kraaij, V., & Spinhoven, P. (2001). Negative life events, cognitive emotion
regulation and emotional problems. Personality and Individual differences, 30(8),
1311-1327.
Gratz, K. L., & Roemer, L. (2004). Multidimensional assessment of emotion regulation and
dysregulation: Development, factor structure, and initial validation of the difficulties
in emotion regulation scale. Journal of psychopathology and behavioral assessment,
26(1), 41-54.
Gross, J. J. (1998). The Emerging Field of Emotion Regulation: An Integrative Review.
Review of General Psychology, 2(3), 271-299. doi:10.1037/1089-2680.2.3.271
Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes:
Implications for affect, relationships, and well-being. Journal of personality and
social psychology, 85(2), 348-362. doi:10.1037/0022-3514.85.2.348
Hooper, D., Coughlan, J., & Mullen, M. (2008). Structural equation modelling: Guidelines
for determining model fit. Articles, 2.
Jermann, F., Linden, M. V. d., d'Acremont, M., & Zermatten, A. (2006). Cognitive Emotion
Regulation Questionnaire (CERQ). European Journal of Psychological Assessment,
22(2), 126-131. doi:10.1027/1015-5759.22.2.126
John, O. P., & Gross, J. J. (2004). Healthy and unhealthy emotion regulation: Personality
processes, individual differences, and life span development. Journal of personality,
72(6), 1301-1334.
Kline, R. B. (2011). Principles and Practice of Structural Equation Modeling. New York,
NY: Guilford Press.
Kring, A. M., & Sloan, D. M. (Eds.). (2009). Emotion regulation and psychopathology: A
transdiagnostic approach to etiology and treatment. New York, NY: Guilford Press.
Loevinger, J. (1954). The attenuation paradox in test theory. Psychological Bulletin, 51(5),
493-504. doi:10.1037/h0058543
Merz, E. L., Malcarne, V. L., Roesch, S. C., Ko, C. M., Emerson, M., Roma, V. G., & Sadler,
G. R. (2013). Psychometric properties of Positive and Negative Affect Schedule
 20

(PANAS) original and short forms in an African American community sample.

Journal of affective disorders, 151(3), 942-949. doi:10.1016/j.jad.2013.08.011
Muthén, L. K., & Muthén, B. O. (2012). Mplus User’s Guide. Los Angeles, CA.
Raykov, T. (2007). Reliability if deleted, not ‘alpha if deleted’: Evaluation of scale reliability
following component deletion. British Journal of Mathematical and Statistical
Psychology, 60(2), 201-216.
Raykov, T. (2008). Alpha if item deleted: A note on loss of criterion validity in scale
development if maximizing coefficient alpha. British Journal of Mathematical and
Statistical Psychology, 61(2), 275-285.
Satorra, A., & Bentler, P. M. (1994). Corrections to test statistics and standard errors in
covariance structure analysis. In A. von Eye & C. C. Clogg (Eds.), Latent variables
analysis: Applications for developmental research (pp. 399–419). Thousand Oaks,
CA: Sage Publications.
Steiger, J. H. (1990). Structural model evaluation and modification: An interval estimation
approach. Multivariate behavioral research, 25(2), 173-180.
Steiger, J. H. (2007). Understanding the limitations of global fit assessment in structural
equation modeling. Personality and Individual differences, 42(5), 893-898.
Thompson, R. A. (1994). Emotion regulation: A theme in search of definition. Monographs
of the society for research in child development, 59(2‐3), 25-52.
Tuna, E., & Bozo, Ö. (2012). The Cognitive Emotion Regulation Questionnaire: Factor
Structure and Psychometric Properties of the Turkish Version. Journal of
psychopathology and behavioral assessment, 34(4), 564-570. doi:10.1007/s10862-
012-9303-8
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief
measures of positive and negative affect: the PANAS scales. Journal of personality
and social psychology, 54(6), 1063.
Zhu, X., Auerbach, R. P., Yao, S., Abela, J. R. Z., Xiao, J., & Tong, X. (2008). Psychometric
properties of the Cognitive Emotion Regulation Questionnaire: Chinese version.
Cognition and Emotion, 22(2), 288-307. doi:10.1080/02699930701369035

View publication stats