Dissertations and Theses

7-2016

The Effects of Safety Culture and Ethical Leadership on Safety

Performance
Kevin O’Leary

Follow this and additional works at: https://commons.erau.edu/edt

Part of the Aviation Safety and Security Commons, and the Leadership Studies Commons

Scholarly Commons Citation

O’Leary, Kevin, "The Effects of Safety Culture and Ethical Leadership on Safety Performance" (2016).
Dissertations and Theses. 201.
https://commons.erau.edu/edt/201

This Dissertation - Open Access is brought to you for free and open access by Scholarly Commons. It has been
accepted for inclusion in Dissertations and Theses by an authorized administrator of Scholarly Commons. For
more information, please contact commons@erau.edu.
THE EFFECTS OF SAFETY CULTURE AND ETHICAL LEADERSHIP ON
SAFETY PERFORMANCE

by

Kevin O’Leary

A Dissertation Submitted to the College of Aviation

in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy in Aviation

Embry-Riddle Aeronautical University

Daytona Beach, Florida
July 2016
© 2016 Kevin O’Leary
All Rights Reserved
 ABSTRACT

Researcher: Kevin O’Leary

Title: THE EFFECTS OF SAFETY CULTURE AND ETHICAL

LEADERSHIP ON SAFETY PERFORMANCE

Institution: Embry-Riddle Aeronautical University

Degree: Doctor of Philosophy in Aviation

Year: 2016

This dissertation investigated the effects of safety culture and ethical leadership on safety

performance in Fractional jet pilots in the United States. The primary objective was to

develop a well-fitted model linking these constructs. A composite survey instrument was

developed from instruments previously validated in the literature.

There were 305 complete and valid responses from Fractional pilots. The

hypothesized factor structure consisted of seven factors. The exogenous factor of safety

culture was made up of four sub-factors. The endogenous factors included ethical

leadership, pilot commitment, and safety performance. Safety performance was a second

order factor consisting of errors and attitudes to violations. The hypothesized model was

not well fit for the data; therefore, an exploratory factor analysis was conducted. The

new model consisted of three factors: safety culture new, ethical leadership new, and not

following procedures.

A structural equation model was developed to test the relationships between

constructs. Safety culture new demonstrated a strong and significant positive effect on

ethical leadership new. Safety culture new, unexpectedly, did not have a significant

negative relationship with not following procedures. Additionally, ethical leadership new

did not have a significant negative effect on not following procedures. These findings

iii
conflicted with previous studies in the literature that confirmed a significant relationship

between both safety culture and ethical leadership with safety behavior. The main

finding illuminates the influence of safety culture new on ethical leadership new.

Additional findings showed the factor structure for most of the previously validated

survey instruments was not maintained in this study with the Fractional pilot data.

iv
 DEDICATION

I dedicate this work to my parents, wife, and children.

v
 ACKNOWLEDGEMENTS

I would like to thank my parents for their countless sacrifices. I would like to

thank Ms. Cavallo and my other teachers who combined learning with both

encouragement and discipline. I would like to thank Major Ed Donnelly and Rich

Heckman for their mentorship and friendship.

I would like to thank my ERAU student friends for their encouragement. I would

like thank Dr. Goodheart, Dr. Freiwald, Dr. M.J. Smith, and many others for their help. I

would like to thank my former and present staff for their encouragement, assistance, and

dedication.

I would like to thank my committee chair, Dr. Alan Stolzer, for his assistance,

insights, quick revisions, and patience. I would like thank both Dr. Truong and Dr.

Goodheart for their many hours of encouragement and coaching. I would like to thank

Dr. O’Toole for his safety insights as well as the remainder of my committee. I would

like to thank the ERAU faculty, staff, and Susie for all the help over the years.

I would like to thank those who assisted with the collection of data. The ERAU

alumni relations department and several key ERAU alumni contributed immensely to

data collection. I would like to thank Chad, Amy, Bill, Curt, Dave, and many others who

assisted in soliciting respondents.

I would finally like to thank my family and especially my wife, Julie, who has

acted as a single parent for most weekends for the last six years. I cannot thank you

enough for allowing me the opportunity to complete this degree. You are a great mom

and wife and for you I am now Dr. “Owe”. I would also like to thank my daughter,

Shannon, who allowed me to be away for her so much time during a very important time

vi
in her life. I would like to thank my boys Connor and Trevor for their sacrifices during

this journey.

vii
 TABLE OF CONTENTS

Page

Committee Signature Page ................................................................................................. ii

Abstract .............................................................................................................................. iii

Dedication ............................................................................................................................v

Acknowledgements ............................................................................................................ vi

List of Tables .................................................................................................................... xii

List of Figures .................................................................................................................. xiii

Chapter I Introduction ...............................................................................................1

Significance of the Study .................................................................6

Purpose Statement ............................................................................6

Hypothesis [Research Question or equivalent] ................................7

Delimitations ....................................................................................7

Limitations and Assumptions ..........................................................8

Definition of Terms..........................................................................9

List of Acronyms .............................................................................9

Chapter II Review of the Relevant Literature ...........................................................11

Introduction ....................................................................................11

Accident Investigation ...................................................................13

Pilot Error Causing Accidents .......................................................14

Human Error ..................................................................................15

Human Factors Analysis and Classification System (HFACS) .....16

Culture............................................................................................17

viii
 Safety Culture ................................................................................19

Safety Climate ................................................................................20

Safety Climate and Culture as Predictors of Safety Performance .21

Employee Commitment to the Organization .................................28

Ethics..............................................................................................29

Ethical Leadership .........................................................................30

Ethical Leadership Scale (ELS) .....................................................31

Ethics as a Predictor of Behavior ...................................................33

Criterion or Self-Reported Outcomes ............................................33

Criterion Based Outcomes in Aviation ..........................................35

Criterion Measurement Variability and Reliability .......................36

Self-Reporting Outcomes...............................................................37

Consistent Methodology ................................................................38

Hypotheses .....................................................................................38

Hypothesized SEM Model .............................................................40

Summary ........................................................................................40

Chapter III Methodology ..........................................................................................42

Research Approach ........................................................................42

Design and Procedures .......................................................44

Apparatus and Materials ....................................................46

Population/Sample .........................................................................46

Sources of the Data ........................................................................49

Data Collection Device / Survey Design .......................................50

ix
 Safety culture (SC) .............................................................50

Pilot commitment to AMC (PC) ........................................51

Ethical leadership (EL) ......................................................51

Pilots’ own attitude to violations (AT) ..............................51

Pilot error behavior (ER) ...................................................52

Construct validity ...............................................................52

Convergent validity............................................................52

Reliability...........................................................................53

Discriminant validity .........................................................54

Nomological & face validity..............................................54

Treatment of the Data ....................................................................54

Descriptive Statistics ..........................................................54

Missing data .......................................................................55

Outliers...............................................................................55

Normality ...........................................................................55

Confirmatory Factor Analysis (CFA) ............................................56

Exploratory Factor Analysis (EFA) ...............................................58

Model 2 (M2) Confirmatory Factor Analysis (CFA) ....................59

Post hoc analysis ............................................................................59

Structural Equation Model & Hypotheses Testing ........................60

Chapter IV Results ...................................................................................................61

Demographic Data .........................................................................61

Normality & Outlier Checks ..........................................................65

x
 Confirmatory Factor Analysis........................................................66

Exploratory Factor Analysis ..........................................................69

Confirmatory Factor Analysis Model 2 (M2) ................................72

Construct Reliability ......................................................................75

Convergent Validity .......................................................................75

Discriminant Validity.....................................................................75

Structural Equation Model .............................................................77

Hypothesis Testing.........................................................................79

Chapter V Discussion, Conclusions, and Recommendations ……………………..82

Discussion ......................................................................................82

Conclusions ....................................................................................85

Contributions to the Literature .......................................................87

Study Limitations ...........................................................................90

Practical Implications.....................................................................91

Future Research .............................................................................93

References ..........................................................................................................................97

Appendices

A Permission to Conduct Research .............................................................106

B Data Collection Device ............................................................................109

C Descriptive Statistics ................................................................................134

D Tables ......................................................................................................150

E Revised Survey Instrument .....................................................................153

xi
 LIST OF TABLES

Page

Table

1 Study Variables ......................................................................................................43

2 Fractional Pilots in U.S. .........................................................................................47

3 SEM Sample Size Requirements ...........................................................................49

4 Completed Responses ............................................................................................62

5 Pilot Age ................................................................................................................62

6 Position at AMC (Aircraft Management Company). .............................................63

7 Aircraft Type Flown ..............................................................................................64

8 Pilot Experience .....................................................................................................65

9 Tenure at AMC ......................................................................................................65

10 Fit Criteria. .............................................................................................................68

11 Model Factors for Hypothesis Testing ...................................................................71

12 Model Fit History...................................................................................................73

13 Discriminant Validity Test .....................................................................................76

14 Final SEM Model Fit .............................................................................................78

15 SEM Regression Weights ......................................................................................79

xii
 LIST OF FIGURES

Page

Figure

1 U.S. Jet Accident Ratio ............................................................................................4

2 Commercial Aviation Safety Survey Factor Structure ..........................................26

3 Safety Management Continuum ............................................................................35

4 Hypothesized SEM Model .....................................................................................40

5 Average Variance Extracted Formula ....................................................................53

6 Construct Reliability Formula................................................................................54

7 Hypothesized CFA Model .....................................................................................67

8 EFA Final Scree Plot .............................................................................................70

9 Final CFA Model ...................................................................................................74

10 Final SEM Model ...................................................................................................77

11 Proposed Future SEM Model.................................................................................94

xiii
 1

CHAPTER I

INTRODUCTION

Flying on U.S. registered private jets for hire (U.S. jets) is considered a very safe

endeavor, especially compared to flying on private jets in many other countries (Robert

Breiling Annual Aircraft Accident Review, 2014). However, some research states the

accident rate in general aviation remains too high and the Federal Aviation

Administration (FAA) has lagged in its responsibility to regulate general aviation to

improve safety outcomes (Kuhn, 2009). As evidence of the FAA’s failure to effectively

ensure safety in General Aviation, Kuhn (2009) points to the fact that the FAA has yet to

mandate the use of Safety Management Systems (SMSs), with their associated reporting

requirements, for either type of for-hire U.S. jet operation: fractional aircraft ownership

programs (Fractionals) or 14 CFR air-taxi operations (Charter).

Over the 25-year period from 1990 through 2014, U.S. jets experienced 410

accidents, with only 96 (23%) of those having fatalities (Breiling, 2014). Over the period

from 2007 through 2014, inclusive, there were 126 accidents involving U.S. jets with 27

(21%) of those resulting in fatalities. According to the research firm JetNet’s

website(www.jetnet.com) , the number of U.S. jets at the beginning of 1990 was 7,336,

while by the end of 2007 that number had risen 63% to 11,961. Despite the increase in

the number of U.S. jets, the average annual rate of both non-fatal accidents and fatal

accidents has been on a downward trend. During the period from 1990 through 2006, the

annualized mean number of accidents was 16.6 per year with 4.3 of those being fatal

accidents. From 2007 through 2014, those rates had declined to 15.8 and 3.4 per year,

respectively.
 2

During the period 2007 through 2014, for domestic flights U.S. jets had an

average of 1.8 million departures and 2.8 million flight hours. This total does not include

the flights taken by U.S. jets abroad. Therefore, since the accident data includes all

flights of U.S. jets, the accident per flight hour rate is presumably lower than reported.

The average accident rate per 100,000 flight hours for U.S. jets was 0.55 during this

period. The fatal accident rate during the same period was 0.12. This equates to one fatal

accident involving a U.S. jet about every 800,000 flight hours.

A traveler can arrange for flights on U.S. jets in three predominant service

models: chartering a jet for hire (Charter), fractional ownership (Fractional), and

ownership. Charter, which is similar to using a taxi or car service, is where an aircraft

manager supplies the pilot and aircraft. In U.S. aviation, the operator responsible for

these Charter flights is called the aircraft manger. The aircraft manager must maintain a

Federal Aviation Regulation 14 CFR part 135 (FAR 135) certificate with the FAA in

order to offer charter flights to the public for hire.

A second option, Fractional, is a model in which a consumer buys a share of a

specific aircraft and the designated aircraft management company flies the owner

whenever a trip is requested. Though regulated under its own section of 14 CFR, namely

part 91(k) (FAR 91(k)), these Fractional manager’s flights are often flown under the

arguably more stringent FAR 135 rules and regulations, where the management

company, rather than the owner, maintains operational control of the majority of the

flights.

The final option to fly a jet privately is full ownership, where a person or entity

purchases a private aircraft. The private jet owner is responsible for the operation of the
 3

aircraft. Many of the owner’s responsibilities can be delegated to an aircraft management

company; however, the owner maintains operational control under 14 CFR part 91 (FAR

91).

Both Fractional and Charter managers hold the same type of FAR 135 certificate,

operate under similar rules and regulations, maintain operational control of the majority

of their flights, and are subject to similar scrutiny by the FAA. However, the annual

accident totals and accidents per hour flown rates are substantially different between

these two groups as shown in Figure 1. Over the 25-year period from 1990 through 2014,

the U.S. jet Charter operators have been involved in 188 accidents with 46 (24%) of those

being fatal. The U.S. jet Fractional operators were involved in just 26 accidents over the

same period with zero fatal accidents. In the period from 2007 through 2014, the U.S. jet

Charter operators have averaged a rate of 6.0 accidents with 1.4 (23%) fatal accidents per

year, while the Fractional operators have averaged 1.4 accidents per year and zero fatal

accidents (Robert Breiling Annual Aircraft Accident Review, 2014).

In the period from 2007 through 2014, the U.S. jet Charter accident rate per

100,000 flight hours averaged .71 (TRAQPak Report, 2014). The fatal accident rate

during the same period was .16. The U.S. jet Fractional accident rate during the same

period was .27 per 100,000 flight hours with zero fatal accidents. The Charter rate of

accidents per 100,000 flight hours of 0.71 is 0.16 (29%) higher than the U.S. jet fleet

average of 0.55; conversely, the Fractional rate is 0.28 (51%) lower at 0.27. The fatal

accident rate per 100,000 flight hours for Charter (0.16) is .04 (33%) higher than the U.S.

jet average of 0.12, while Fractional did not have a fatal accident during this period.
 4

Fractional did not had a single fatal accident during the period of 1990 through 2014

(Robert Breiling Annual Aircraft Accident Review, 2014).

0.8
0.71

0.55

Accidents
0.4
Fatal Accidents
0.27
0.16
0.12

0.00
0.0
U.S. jet Average Frax Charter

Figure 1. U.S. Jet Accident Rate. The U.S. jet fleet average accident rate per 100,000
flight hours for the period of 2007 through 2014. (Breiling 2014; TRAQpak 2014).

The focus on causation of aircraft accidents has shifted since the early 1990s. The

previous research on accident causation concentrated on a very granular search for the

final causal or contributing factors that lead to the accident. This causation research often

pointed to the last line of defense in the entire safety system: the pilot. Accident

investigators diligently searched for the smoking gun or the last item in a chain of events

that, had it been corrected likely would have changed the course of events and prevented

the accident. Because pilots are the last line of defense in the safety system, they were
 5

indicated as the main causal factor in the vast majority of aviation accidents (Vincoli,

1990).

In the last 25 years, safety has evolved into its own discipline where processes are

designed and implemented to make flying safer (Stolzer & Goglia, 2015). Historically,

pilots were blamed as the cause of most aviation accidents; however, in the 1990’s, this

trend started to evolve. This paradigm shift was the result of the growing understanding

of safety as a system and consideration of the multiple causal interactions of accidents.

These multiple causal interactions include those that reside within the flight organization,

such as group behaviors and culture. Many human factors researchers, such as von

Thaden, Wiegmann, & Shappell (2006); Jennings (2008); and Li, Harris, and Yu (2008)

have revisited aviation accidents dating back many years and have persuasively

demonstrated that the organization and its characteristics strongly influenced the causal

factors of the majority of accidents. The aforementioned research results were important

because they illuminated the key interrelationships within an organization. This

increased understanding of these key interrelationships provided the opportunity to make

organizational changes that were likely to further enhance safety.

As a result of this shift in understanding of the importance of organizational

characteristics in maintaining and improving safety, the effort to measure the safety

culture, organizational commitment, and even ethics of the organization has gained

momentum in the literature. Researchers have attempted to develop and validate survey

instruments to take these measurements in order to better understand how they influence

safety outcomes, such as occupational accidents and safety performance (Alsowayigh,

2014; Freiwald, 2013; Zohar, 1980). If the safety culture or ethics of an organization can
 6

be accurately measured and shown to have a predictable influence on future safety

outcomes or performance, this could create an opportunity for comparatively low cost

interventions that would significantly improve safety in the system (Freiwald, 2013).

Statement of the Problem

To date, the relationship of safety culture, ethical leadership, pilot commitment to

the organization, and safety performance has not been measured or investigated in U.S.

jet Fractionals. Though these constructs have been studied in many airlines, the

Fractionals differ operationally from airlines in many ways. The Fractionals, for

example, fly exclusively point to point and do not fly in the hub-and-spoke flight patterns

common to most airlines. The historical differences in the total number and rate of both

fatal and non-fatal accidents are strong quantitative evidence that suggests there are

operational and likely cultural differences between the U.S. jet Fractional and U.S. jet

Charter operators.

Purpose Statement

The purpose of this study was to examine: (1) the Fractional pilots’ perceptions of

their organizations’ level of safety culture and ethical leadership, and (2) the potential

influence of these perceptions on the pilot’s commitment to the organization and their

safety performance. Since the Fractional operators have fewer accidents than the Charter

operators in the U.S. during the period under review, the practical application of this

research could be the identification of a baseline model for safety culture. Future studies
 7

would be required to research the safety culture of the Charter companies and compare

results.

Research Questions

This research addressed four questions that were derived from the research

conducted by Alsowayigh (2014) on Saudi Airline pilots and Freiwald (2013) on aviation

and healthcare personnel.

1. How does safety culture influence safety performance at U.S. jet Fractionals?

2. How does safety culture influence ethical leadership at U.S. jet Fractionals?

3. How does safety culture influence pilot commitment to the organization at U.S.

jet Fractionals?

4. How do ethical leadership and pilot commitment to the organization influence

safety performance at U.S. jet Fractionals?

Delimitations

The survey data collected in this study were comprised of responses from the

pilots of major U.S. jet FAR 135 Fractional operators with more than 25 jets under

management. The 25 jet minimum was selected because only three companies exceed 25

jets (NetJets, FlexJet, and Executive AirShare) and are estimated to operate 97.6% of the

Fractional jet aircraft in the U.S. (www.JetNet.com; November 7, 2015).

It was not within the scope of this research to investigate safety outcomes from

the NTSB accident investigation reports, Flight Operations Quality Assurance (FOQA),

or other criterion-based data to search for relationships or causation. This is due to the
 8

concern in the literature that accident rates are too low to make valid predictions

(O’Connor et al., 2011), and criterion based data such as FOQA are not consistently

recorded across general aviation aircraft (Cistone et al., 2011); data recording systems are

expensive to install and therefore inconsistently deployed in the fleet (Mitchell, Sholy, &

Stolzer, 2007); and data that were recorded are not publically available.

This research was not intended to develop the appropriate path to improvement of

U.S. jet FAR 135 operations, but rather to determine the relationships between safety

culture, ethical leadership, pilot commitment, and safety performance of U.S. jet

Fractionals.

Limitations

This study was intended to measure and investigate the relationships between the

constructs of safety culture, ethical leadership, and safety performance for U.S. jet

Fractional operators. It was assumed that due to the fact the pilots were notified through

their unions, nearly all pilots had the opportunity to complete the survey, and therefore

the results will likely be generalizable throughout the organization. Additionally, since

these pilots represent over 97% of the Fractional pilots in the U.S., the results are likely

to be generalizable to all U.S. Fractional pilots. Non-response bias was tested through a

comparison of the results between different survey collection dates. The comparison

included an analysis of the responses by similar demographic groups across various

survey collection dates.

 9

The construct for safety performance was self-reported items describing pilot

errors and their attitudes to violations. There are concerns in the literature about the

potential inaccuracy due to the nature of self-reported items (O’Connor et al., 2011).

Definitions of Terms

AMC Aircraft Management Companies are those companies managing

jet aircraft and offering flights to the public for hire. Both

fractional jet managers (Fractional) and U.S. jet FAR 135 aircraft

management companies (Charter) are considered AMCs.

Charter Charter refers to the companies where flights are

offered to the public for hire by a certificated FAR 135 aircraft

management company.

charter When not capitalized, this term refers to flights flown by Charter

companies for hire.

Fractional(s) Fractional aircraft management company(ies)

Micro-
accidents These are small workplace accidents such as cuts and bruises.

U.S. jets Refers to U.S. registered private jets that are used in a Fractional

aircraft program or flown by a duly certificated FAR 135 aircraft

management company for hire.

List of Acronyms

AMC Aircraft Management Company (both Fractional & Charter)

ELS Ethical Leadership Scale

 10

FAA Federal Aviation Administration

FOQA Flight Operations Quality Assurance

NTSB National Transportation Safety Board

PCAMC Pilot Commitment to AMC

SEM Structural Equation Model

SCFSS Safety Culture Formal Safety System

SCISS Safety Culture Informal Safety System

SCOC Safety Culture Organizational Commitment

SCOP Safety Culture Operations Personnel

SPATV Safety Performance Attitude To Violations

SPERR Safety Performance Pilot Error Behavior

ZCSQ Zohar Client Safety Questionnaire

 11

CHAPTER II

REVIEW OF THE RELEVANT LITERATURE

Introduction

The review provided in this chapter begins with a brief history of aircraft accident

investigations and how the conduct of these investigations has evolved over the last 40

years. Accident investigation is considered one of the initial steps in aviation history

directed toward improving safety through better understanding the causal factors in

accidents and applying that knowledge to preventing similar accidents in the future

(Stolzer & Goglia, 2015). Accident investigators have, in both past and current

investigations, conducted a very granular analysis of each accident to determine the

proximate causal factors. Once the causal factors are determined, the results are

categorized and analyzed across many accidents to identify themes. The knowledge

gained from these accidents and subsequent analyses or themes has inspired the

development of new technologies, equipment, and procedures that have contributed to the

continued improvements in aviation safety (Stolzer & Goglia, 2015).

With improvements in technology, equipment, and procedures, accident

investigators began to find fewer and fewer causal factors attributable to equipment

failures (Vincoli, 1990). These improvements in the reliability of both the equipment and

procedures led investigators to label the main causal factor in the majority of accidents as

pilot or human error (Vincoli, 1990). Since the majority of accidents were and continue

to be determined to be pilot error, and the goal in aviation was to continue to improve

safety, aviation practitioners needed to better understand the causes of human error, and
 12

more specifically the active and latent conditions that contributed to the malfunction of

the pilot (Reason, 1990).

As the construct of human error became more fully understood, aviation accident

investigators and practitioners still needed to further adapt these concepts to an aviation

setting to continue the improvements in safety outcomes. The study of human error

provided a framework for scholars to adapt those, along with other concepts, to develop

the human factors classification system (HFACS). HFACS provided a common

taxonomy that enabled accident investigators, aviation practitioners, and researchers to

both identify and categorize human errors (Shappell & Wiegmann, 1997). The errors

were labeled active (human mistakes), latent, or organizational factors (training, over

scheduling, or procedures errors, etc.) that contributed to accidents that had been labeled

as just pilot error in the past (Shappell & Wiegmann, 1997).

Along with the study of HFACS, organizational culture began to emerge as an

important construct in the literature as a possible antecedent to safety outcomes (Cox &

Flin, 1998; Helmreich & Merritt, 1998; Zohar, 1980). Studies focusing on safety culture,

communication, cockpit resource management, employee commitment to the

organization, and company leadership began to emerge in the literature as possible

constructs that could be measured and had the potential to influence safety outcomes.

This study builds upon previous research focused on the constructs of safety

culture (Alsowayigh, 2014), ethical leadership (Freiwald, 2013), and their potential

influence on self-reported safety performance, such as a pilot admitting to making

occasional errors. If these relationships exist and are significant, this research has the
 13

potential to provide insight into a possible safety culture model for Charter operators to

follow that could improve safety in U.S. jet FAR 135 operations overall.

Accident Investigation

When aviation accident investigation began, there was a “fly-crash-fix-fly”

approach (Stolzer & Goglia, 2015, p.15). The investigator’s mission was to determine

the cause of the accident, publicize the results, and adopt new regulations to prevent

future re-occurrences with the same cause. The causes sometimes were related to

unforeseen weather conditions, design flaws, structural/mechanical failures, or human

error (most often by pilots and sometimes by mechanics) (Stolzer et al., 2011).

An article by Walter Tye published in 1980 demonstrates the major concerns of

the day with commercial aviation. Tye wrote, to improve aviation’s upward trend in

safety, the industry had to focus on new aircraft designs, improvements in avionics to

avoid mid-air collisions and controlled flight into terrain (CFIT), and, ultimately, better

procedures (Tye, 1980). Tye’s research estimated that approximately one-half of the fatal

commercial aviation accidents from 1972 until 1980 were the result of CFIT and,

additionally, almost 25% were from mid-air collisions (Tye, 1980).

Some examples of accidents include: In 1987 a Learjet 35A sustained substantial

damage after a hard landing in rain and heavy winds. The National Transportation Safety

Board (NTSB) named wind shear as one of the main contributing factors (NTSB Brief

MIA88LA026, 1989). In 1990, a Lear 24 experienced a fire in the cockpit when the

wires from the map light chafed together, causing the wires to arc, and resulting in a

cockpit fire that precipitated a forced landing (NTSB Brief ATL90LA080, 1992). A
 14

different accident which resulted from a gear failure on a Challenger in 1997 (NTSB

Brief: ATL96LA073, 1997) could have been avoided through better organizational

procedures. The NTSB report suggested that improved procedures at the aircraft

management company requiring use of the emergency gear extension checklist may have

prevented the accident. The NTSB recommendation centered on the pilot neglecting to

verify the gear was down and locked after an initial indication that the gear was not

locked in place, which is the proper procedure as published in the aircraft’s operating

handbook.

Tye’s suggestions from 1980 have all been adopted; first by the commercial

aircraft manufacturers and later by the private jet manufacturers. Avionics improvements

included ground proximity warning systems, traffic collision avoidance, and ground

based and cockpit based wind shear detection systems. Additionally, procedural

improvements were made such as the adoption of crew resource management (CRM)

programs. Aircraft designs improved structural soundness and systems reliability. As

Tye’s published suggestions have been implemented in aviation, the accident rates have

continued to decline.

Pilot Error Causing Accidents

The reliability of aircraft as well as of the air transportation system itself

improved in the 1980s (Vincoli, 1990). As suggested in later research, the main causal

factor in most aviation accidents was pilot error (Vincoli, 1990). In the previous 20

years, the NTSB had identified pilot error as the primary cause for 66% of aviation

accidents (Vincoli, 1990). The U.S. Army conducted a study and concluded that over
 15

80% of Army aviation accidents during the years 1958-1976 were the result of pilot error

(Vincoli, 1990). This led to the NTSB seemingly declaring pilot error as its default

finding, as evidenced by two cases where independent investigators reviewed the

evidence and found conclusive proof of mechanical failures previously missed by the

NTSB that were major causal factors (Vincoli, 1990).

Vincoli went on to warn the industry and the investigators that safety of flight is

the responsibility of the aircraft manager or airline, and this responsibility cannot be

delegated to the pilot (Vincoli, 1990). Vincoli also warned that if the trend of

disproportionately identifying pilot error as the primary cause in the vast majority of

accidents continued, the industry would not be able to move forward to improve safety,

nor to prevent future accidents effectively.

Human Error

In 1982, Rasmussen wrote his seminal paper describing human error, attempting

to bring structure to the construct and foster proper collection of data. In his work, he

described the characteristics and definitions associated with human failure. It was

asserted that most inadequate results or outright systems failures could be traced back to

human failure in design, operation, or maintenance (Rasmussen, 1982). The author also

pointed out that quite often the system failure was the result of a latent condition that

existed prior to the actual system failure (Rasmussen, 1982).

Reason’s 1990 book Human Error furthered the body of knowledge on the topic

of human malfunctions and continued to provide understanding of where humans are

likely to fail in a complex safety system such as those comprising aviation. Reason
 16

postulated that there were two main types of human errors: active and latent. Active

errors occur when the operator of a system, such as the pilot of an aircraft has the wrong

reaction to a stimulus or situation and proximately causes the system failure. Conversely,

a latent error may occur far away from and long before the system failure, such as an

aircraft manager over-scheduling a crew which contributes to the pilot’s fatigue and

reduced effectiveness (Reason, 1990). Since most pilots overestimate their personal

capabilities, they are unlikely to acknowledge or admit their reduced abilities when

stressed or fatigued (Helmreich & Merritt, 1998).

Human Factors Analysis and Classification System (HFACS)

Building upon the research from Rasmussen and Reason, Shappell and Wiegmann

published Human Error Approach to Accident Investigation: The Taxonomy of Unsafe

Operations in 1997. This research contributed to what is now known as HFACS. The

authors’ objective was to develop a common taxonomy for accident investigators to use

when classifying types of human errors. A common taxonomy allows researchers and

practitioners to communicate more effectively. The goal of HFACS was to determine

both the active (human) errors and the latent (organizational) errors. Shappell &

Wiegmann attempted to determine the true root cause of aviation accidents in order to

take the next step toward improving aviation safety (Shappell & Wiegmann, 1997).

In the 1990s, there was a paradigm shift in the literature in which aviation

accidents were considered to be the result of a chain of events rather than being due to a

single, proximate cause. The root causes, which had often been blamed on just the pilots,

were expanded to include the latent failures of the aviation organization (McFadden &
 17

Towell, 1999). Aviation accidents that were classified as pilot error have been re-

examined using the HFACS perspective, and many latent or organizational errors have

been identified (Wiegmann & Shappell, 2001). These findings have motivated a

fundamental shift toward proactive system improvement to enable aviation organizations

to reduce the incidence of latent errors and thereby forestall accidents (McFadden &

Towell, 1999).

Culture

Culture is commonly associated with national culture and has its roots in

anthropology. It is concerned with the core values of a group (Cox & Flin, 1998). Pilots

experience three distinct cultures in their work: national, professional, and organizational

(Helmreich & Merritt, 1998). In January of 1990, Avianca Flight 52 crashed in New

York as a result of fuel starvation. The flight engineer was aware of the criticality of the

situation but failed to make those concerns known to the captain. In this situation, all

three forms of culture, national (deference to authority), professional (not questioning the

higher ranking captain), and organizational (lack of CRM) contributed to the chain of

events that resulted in an otherwise avoidable aviation accident (Helmreich & Merritt,

1998).

Aviation professionals have a distinct culture. In that professional culture, pilots

have a specialized skill that provides prestige and high pay, which encourages some

pilots to feel overconfident (Helmreich & Merritt, 1998). This feeling of overconfidence

can lead to poor decision making, such as skipping routine checklists and taking

unnecessary risks (Helmreich & Merritt, 1998). In the Avianca case, the crew had many
 18

options to divert the aircraft; however, poor crew communication led to the continuation

of the flight to the point of fuel exhaustion.

The development of CRM was motivated by a desire to address both

organizational and pilots’ professional culture factors that had been shown to contribute

to accidents. As it has been implemented in aviation organizations, CRM has

demonstrated success at increasing communications in the cockpit and breaking down

several barriers to optimally safe and efficient aircraft operation (Helmreich & Merritt,

1998). CRM is implemented in part by creating a subculture in the overall organizational

culture comprised of a set of values and norms required to support the effective use of

CRM operational practices.

National culture is a broader term related to those values, norms, and beliefs held

by particular nationalities (Helmreich, 1998). The Avianca flight is an example of the

consequences of poor or absent CRM practices. The flight engineer knew the aircraft

was critically low on fuel; however, the flight engineer neglected to communicate that

situation clearly to the captain. A combination of the flight engineer’s national culture,

Avianca’s organizational culture, and the flight engineer’s professional culture did not

provide the flight engineer with the confidence to communicate a critical safety issue to

the captain (Helmreich & Merritt, 1998). Though this flight’s mishap can correctly be

assigned a proximate cause of pilot error, HFACS would identify the latent

organizational, professional, and cultural issues as major contributing factors.

Subsequent research into culture asserted that culture surrounds the organization

and is intertwined with leadership and its behavior (Schein, 2004). Therefore, a leader

can engineer culture by attempting to insert values into the organization that will
 19

influence and govern employee behavior and interactions (Schein, 2004). Because of the

stable nature of the values set forth in organizational culture, it has been called the

personality of the organization (Cox & Flin, 1998, Schein, 2004).

Safety Culture

“A safety culture is more than a group of individuals promulgating a set of safety

guidelines, it is a group of individuals guided in their behavior by their joint belief in the

importance of safety (Helmreich & Merritt, 1998, p. 133).”

Safety culture is a subset of the overall culture in an organization. The term

safety culture first came to prominence from the report on the Chernobyl nuclear disaster

from the International Atomic Energy Agency (Cox & Flin, 1998). The report discussed

the poor safety culture that was present in the Russian nuclear plant. Safety culture is

comprised of beliefs and values held in an organization regarding employee safety,

hazard reduction, and a safe work environment (Cox & Flin, 1998). These values are

stable, meaning they do not fluctuate in the short term (Cox & Flin, 1998). Initially,

some researchers expressed concerns that the importance of safety culture was overstated

and that it was not a proven theoretical concept (Cox & Flin, 1998). In contrast, other

research in CRM fully supported the concept of culture as relevant to understanding and

motivating positive change in the larger organizational culture, and showing that changes

in culture had the ability to improve or reduce safety (Helmreich et al., 1997).
 20

Safety Climate

The concept of organizational climate dates back to the 1930s; however, the

measurement of the character of an organization did not start until the 1960s (Cox & Flin,

1998). Safety climate is the subset of the organizational climate that focuses on safety

(Neal et al., 2000). The literature often treats the constructs of culture and climate

interchangeably (Mearns & Flin, 1999). The difference between culture and climate has

been compared to the differences between personality and mood of a person. A person’s

personality is based on the person’s own core values and principles, and though it can be

changed, it cannot be changed quickly; like culture, it is stable and enduring.

Organizational climate, conversely, is more closely associated with a person’s mood,

which can change quickly based on the environment and the day’s activity; therefore, it is

short term and more variable, and measurements of climate are similar to a snapshot at

one point in time (Cox & Flin, 1998).

The construct of safety climate was enhanced by the research of Zohar in the early

1980s. The research included a 40-item survey that was randomly distributed to 20

workers in 20 different industrial organizations (Zohar, 1980). The researcher then

compared the results of the survey with the results of an independent safety inspector’s

evaluation of the safety effectiveness of each industrial organization. There was a high

correlation between the inspector’s evaluations of the effectiveness of safety programs at

the different companies and the survey results from the workers (Zohar, 1980). The

highest level of correlation was between the worker’s perceptions of management’s

attitudes about safety and the rated effectiveness by the inspectors (Zohar, 1980).
 21

Safety Climate and Culture as Predictors of Safety Performance (Outcomes)

The Zohar safety climate research was instrumental in developing the concept of

safety climate though the use of an independent measurement to validate the results.

Helmreich et al. used a similar validation technique in 1986; the research measured pilot

attitudes and compared those responses to their performance evaluations from

experienced check airmen. The study showed an attitude-performance linkage

(Helmreich et al., 1986).

The Zohar and Helmreich et al. studies were important because they not only

validated the construct of safety climate, they also established there was a link to

performance. The accident rate in aviation is very low; therefore, it lacks the sensitivity

to establish the predictor variables for safety performance or accidents (O’Connor et al.,

2011). The importance of measuring both safety climate and safety culture lies in the

potential to harness their predictive capability to improve safety performance and reduce

accidents.

Before 2000, there were few research studies on the connection between safety

climate and safety behavior, though many studies have shown a correlation between

safety climate and safety outcomes (Neal et al., 2000; O’Toole, 2002). Researchers

hypothesized that organizational climate would exert influence on safety climate, and

safety climate would exert influence on safety performance (Neal et al., 2000). Neal et

al. (2000) defined safety performance as compliance with procedures and promotion of

safety. It should be noted this research relied on self-reporting of safety performance,

which has been criticized in the literature as potentially biased (Barling et al., 2002).

Zohar asserted “safety climate research has been hampered by a lack of criterion data”
 22

(Zohar, 2000, p. 589). O’Connor et al. (2011) suggested using objective data such as

FOQA to evaluate safety performance.

The findings of the Neal et al. (2000) research support the hypothesis that

organizational climate had a significant impact on safety climate. Safety climate had a

significant impact on self-reported safety compliance, and safety climate is a predictor of

safety performance (Neal et al., 2000).

Additional criterion-based safety climate research was conducted to predict the

effect of group climate on micro-accidents in the manufacturing industry (Zohar, 2000).

This research used a newly developed scale to estimate the perception of safety climate of

factory workers. The data on micro-accidents was recorded during the five-month period

following the safety climate survey. The results established an empirical link between

safety climate and micro-accidents where the group safety climate predicted the safety

outcomes (Zohar, 2000). Zohar’s research suggested that an increase in micro-accidents

was a predictor of larger or catastrophic accidents (Zohar, 2000). In 2004, there was a

study conducted in Japan on the track maintenance train operators’ attitudes versus

objective accident data. The findings suggested that operator attitudes were significantly

correlated with accidents, and the recommendation called for proactive improvements in

attitudes in order to improve safety (Itoh et al., 2004).

Cooper & Phillips (2004) conducted a safety climate study before and after a

behavioral safety initiative. Their findings concluded the relationship of safety climate to

safety behavior though the relationship between safety behavior and accidents was not as

strong as other similar findings in the literature. Though the researchers concluded that

the statistical relationship between safety climate and accidents was neither direct nor
 23

significant, the research suggested that safety climate measurements are useful in

assessing the effectiveness of how safety is operationalized in an organization (Cooper &

Phillips, 2004).

A case study was undertaken to evaluate the safety culture of a large construction

company and its influence on safety performance. The construction company had

implemented safety initiatives that had varied in success across different regions. The

case study employed a mixed method analysis consisting of in-depth interviews, safety

surveys, and qualitative observations. The results indicated that safety culture had a

mediating role over safety performance (Cai, 2005). One main concern that was

identified was the construction company was found to be taking the human error position

when determining the cause of accidents rather than an organizational error approach,

which is harmful to safety culture and safety reporting (Cai, 2005).

Clarke published a meta-analysis of criterion-based research of the relationship

between safety climate, safety performance, and accidents in 2006. The research showed

that, in all studies, the relationship between safety climate was found to be positive,

though weak, and with a large standard deviation; therefore, the safety climate link to

accidents was not strongly supported (Clarke, 2006). In the case of prospective research

designs where the safety climate measurement takes place before the safety data were

collected, the link between safety climate and accidents was found to be valid and

generalizable (Clarke, 2006). The link between safety climate and safety performance

was positive, and overall the research supported the concept that improving safety

climate would improve safety performance (compliance and participation) and help to

reduce accidents (Clarke, 2006).

 24

Few multi-year studies have been conducted, but one exception is the research by

Neal and Griffin in 2006. This study was conducted over a five-year period with safety

climate measures from two separate sampling frames compared with criterion accident

data. The researchers were attempting to determine a link between safety climate and

safety motivation as well as the link between safety motivation and behavior, under the

hypothesis that safety motivation plays a mediating role between safety climate and

safety performance. The researchers found that there is a reciprocal relationship between

safety motivation and safety participation (safety participation is a component of safety

performance), which indicates that participating in safety tasks that benefit the

organization leads to higher motivation (Neal & Griffin, 2006). Additional findings

showed that, at the group level, self-reported safety behavior has predictive validity for

accidents (Neal & Griffin, 2006).

Despite all of the positive results cited above, Johnson opined that the predictive

validity of safety climate had not yet been firmly established in the literature (Johnson,

2007). Johnson conducted a study that used the 16 item Zohar Safety Client

Questionnaire (ZCSQ) on 292 workers at three manufacturing facilities and subsequently

monitored the accident experience data for the following five-month period. The results

showed that the ZCSQ could be reduced to 11 items with little loss of explanatory power,

and the predictive validity of safety climate to predict accidents was confirmed (Johnson,

2007).

The research result of safety climate as a valid predictor of safety performance

was further supported by Chang and Lu (2009) and then by Kao et al. (2009). However,

the predictive validity of safety climate and patient outcomes were not supported in
 25

Wilson’s (2007) and Lyon’s (2007) dissertations. Lyon’s dissertation on the relationship

between safety culture and infections found contrary evidence that safety climate was low

when infections were low (Lyon, 2007). Goodheart & Smith (2014) suggested that safety

climate predicting safety performance might not be generalizable to aviation from other

industries.

O’Connor et al. (2011) conducted an in-depth meta-analysis of safety climate

studies in the aviation industry. The research analyzed 23 studies conducted in aviation.

Pilots and mechanics made up nearly 65% of the respondents, while 17% had a mixed

target, and the remainders were either cabin crew or ground handlers. Half of the

respondents were military personnel. O’Connor argues that safety climate research needs

to continue to focus less on developing and validating new survey instruments and more

on the ability of the existing instruments to discriminate among groups (O’Connor et al.,

2011). The construct validity of safety climate as a social measure is reasonable, though

there is a lack of agreement in themes across aviation safety climate questionnaires

(O’Connor et al., 2011). There would be a benefit to consolidating the themes in the

literature and to have more consistency. The greater problem with the extant research is

the lack of testing of discriminant validity (O’Connor et al., 2011). If the existing

instruments are not able to discriminate among groups with differing safety performance

scores, the instruments will be of little usefulness as a leading indicator of safety issues

(O’Connor et al., 2011).

Gibbons, von Thaden, and Wiegmann designed a survey instrument in 2006 with

the intention of being more comprehensive than the existing safety climate and safety

culture instruments available. The authors named this improved survey the Commercial
 26

Aviation Safety Survey (CASS). The questionnaire started as an 84 item tool but after

confirmatory factor analysis (CFA), was later revised to 55 items with four general

factors (Organizational Commitment, Operations Personnel, Informal Safety System, and

Formal Safety System) and 12 sub-factors as shown in Figure 2 (Gibbons et al., 2006;

O’Connor et al. 2011; Alsowayigh, 2014). The CASS has been chosen for this research

because it has been deployed in several airlines worldwide, including Saudi Airlines in

2014 and has maintained consistent results.

Figure 2. Commercial Aviation Safety Survey Factor Structure (Alsowayigh, 2014

p.30).
 27

The Saudi Airlines study used the CASS and compared it with self-reported safety

performance, which was measured by pilot attitude to violations and pilot error behavior

(Alsowayigh, 2014). The study included 247 voluntary responses which represented a

29% response rate from active Saudi Airlines pilots. The results were validated with

CFA, and the relationships among variables were analyzed using structural equation

modeling (SEM) (Alsowayigh, 2014).

The Saudi Airline results showed that safety culture had a direct and significant

influence over pilot’s own attitudes to violations and had a mediating role on pilot error

behaviors (Alsowayigh, 2014). Safety culture was found to have neither a direct nor a

significant influence over pilot error behavior, though this relationship was mediated by

pilot’s attitude to violations (Alsowayigh, 2014). Pilot’s commitment to the airline did

not have a significant relationship with either pilot error behavior or attitude to violations,

which suggests that a pilot’s safety performance, as measured by these self-reported

variables, is not strongly related to the characteristics of the organization where the pilot

is employed (Alsowayigh, 2014).

The CASS was designed to be a comprehensive instrument to measure the safety

culture for aviation organizations (Gibbons et al., 2006). Other multi-use instruments,

such as Zohar’s safety climate scale are significantly shorter than the CASS and were

designed to take a quick view or snap shot of safety climate of many types of

organizations, whereas the CASS was developed specifically for the aviation industry.

Additionally, the CASS has also been deployed in many airlines worldwide, and the

constructs have remained stable. The comprehensive nature of the CASS does make it
 28

longer than other instruments, which requires respondents to spend more time completing

the survey.

Employee Commitment to the Organization

In the past, the commitment to the organizations was measured to determine the

likelihood of employee retention. In a longitudinal study over a six-year period, Sheridan

(1992) studied the organizational commitment by young accountants entering the

profession. The researcher controlled for changes in the economy and for labor market

fluctuations to evaluate the role of organizational culture and its relationship to employee

retention.

At about the same time the Sheridan (1992) six-year longitudinal study was

concluding, Meyer & Allen (1991) were researching the causal implications of employee

commitment to an organization. Their research showed that employee commitment to an

organization was related to how the employee was involved in decision making (Meyer

& Allen, 1991; Walton, 1985) in the organization and how their company decisions

aligned with their own values (Meyer & Allen, 1991). The researchers during this period

began to analyze the construct of employee commitment to the organization as a causal

factor; the researchers agreed that the existing structural equation models only showed

evidence of directional relationship without any conclusive findings (Meyer & Allen,

1991).

Researchers interested in the construct of employee commitment to the

organization continued to search for directional relationships. Alsowayigh (2014)

researched the pilot’s commitment to the Saudi Airlines, not as a casual factor, but as a
 29

mediator between safety culture and safety performance (Alsowayigh, 2014). The pilot

commitment to Saudi Airlines was measured with the Porter et al. (1974) nine-item

Organizational Commitment Questionnaire (OCQ). The OCQ has a 14-item version and

nine-item version; the nine-item version was suggested in the literature (Commerias &

Fournier, 2001) and was used in the Alsowayigh (2014) study. The OCQ measured the

employees’ willingness to go above and beyond for their organization and to what extent

employees associated themselves with the company’s success (Commerias & Fournier,

2001).

Alsowayigh’s results (2014) showed that the pilot’s commitment to Saudi Airlines

did not play a mediating role between safety culture and safety performance as measured

by self-reports of pilot error behavior and pilot attitude to violations (Alsowayigh, 2014).

However, it did reveal that safety culture was a statistically significant predictor of the

pilot commitment to the airline (Alsowayigh, 2014).

Ethics

“Ethics is the area of philosophy that deals with values and customs of a person or

society—essentially how one determines what is right or wrong. As far back as Aristotle,

ethics has been considered a fundamental driving force of human behavior” (Kapp &

Parboteeah, 2008, p. 28). Despite being labeled a fundamental driving force of human

behavior, there are relatively few studies about ethics as a construct and the role it plays

in the behavior of employees (Freiwald, 2013; Kapp & Parboteeah, 2008).

The question of what is and what is not ethical is often judged by others. There

are numerous popular media references to stories of politicians, professionals, athletes,

 30

and average citizens who commit acts that are judged by the writers to be wrong or

unethical (Brown, Treviño, & Harrison, 2005). The concept used in this research to

determine what is right/ethical or wrong/unethical is closest to the rule-based

utilitarianism concept (Rachels, 2002). Those acts that are considered wrong or unethical

are the ones that primarily benefit the person committing the acts while at the same time

actually or potentially harming others (Rachels, 2002). Those acts that are considered

altruistic and benefit others or society as much as or more than the person committing the

acts are considered right or ethical (Rachels, 2002).

There are rare acts that may benefit others far more than, or even risk injury to,

the person committing the acts; these acts are considered supererogatory, such as entering

a burning building to search for those in need of help (Craig & Gustafson, 1998;

Freiwald, 2013). Supererogatory acts are considered above and beyond what society

considers socially responsible, just, or ethical behavior; therefore, acts do not have to be

supererogatory to be considered ethical or right for the purposes of this research.

Ethical Leadership

Ethical leadership is “the demonstration of normatively appropriate conduct

through personal actions and interpersonal relationships, and the promotion of such

conduct to followers through two-way communication, reinforcement, and decision-

making” (Brown, Treviño, & Harrison, 2005, p. 120). Ethical leadership is a dimension

of both ethics and leadership. In the literature, there has been little empirical research

into either the construct of ethical leadership or the outcomes influenced by ethical

leadership (Brown, Treviño, & Harrison, 2005; Craig & Gustafson, 1998; Freiwald,
 31

2013). The construct of ethical leadership was researched by Howell and Avolio in 1992,

though their research focused primarily on charismatic leadership. Their results

supported the theory that ethical leaders were those willing to listen to subordinates, and

unethical leaders refused to listen to them (Howell & Avolio, 1992). Other research

studies have showed that employees who perceive their leaders to have high ethical

standards are more willing to report problems without fear of reprisal (Brown et al.,

2005).

Ethical leaders are considered to be altruistic as judged by their employees; these

ethical leaders are the ones acting for the betterment of others, such as other employees

(Brown et al., 2005). The literature has shown that leaders should be concerned with

their employees’ view of their ethics (Craig & Gustafson, 1998). If their employees view

these leaders as “attractive, credible, and legitimate” (Brown et al., 2005, p. 120), their

actions and behaviors will be emulated by their subordinates. A separate article stated

these leaders need to have and maintain a high level of integrity (Craig & Gustafson,

1998). If leaders maintain these qualities, they will hold their employees’ attention and

influence their behavior (Brown et al., 2005).

Ethical Leadership Scale (ELS)

The ethical leadership scale (ELS) is a survey instrument that was developed by

Brown, Treviño, and Harrison in 2005. Their hypothesis stated ethical leadership was an

important component of both transformational and charismatic leadership (Brown et al.,

2005). The ethical leadership component is the one that relates to the ability of the leader

to inspire, and influences to what degree employees want to emulate the leader’s behavior
 32

(Brown et al., 2005). Brown et al. demonstrated that the construct of ethical leadership

influenced behavioral outcomes such as job satisfaction, dedication, or commitment to

the organization and the employee’s willingness to communicate issues (Brown et al.,

2005).

Brown, Treviño, and Harrison developed the ELS by initially researching the

existing literature for extant measurement instruments of charismatic, transformational,

and ethical leadership. The researchers independently developed two versions of a

measurement instrument before subsequently comparing them and eliminating their

overlap (Brown et al., 2005). The researchers then conducted in-depth interviews with 20

MBA students with professional work experience (Brown et al., 2005) to further refine

the ELS. The initial result was a 48-item survey instrument on a five point Likert scale

that measured ethical leadership.

Brown, Treviño, and Harrison conducted seven studies with the ELS. Study one

was conducted on 154 MBA students that were, on average, 29.3 years of age, 68.9%

male, and had 6.3 years of professional work experience (Brown et al., 2005). After

Brown et al. conducted an Exploratory Factor Analysis (EFA), principal factor analysis,

with an oblique rotation (direct oblimin), and scree plot, the eigenvalues showed one

primary factor accounted for 60.1% of the variation (Brown et al., 2005). Further

analysis and consultation with construct experts revealed the ELS could be reduced to a

10-item scale with little loss of explanatory power (Brown et al., 2005). Studies three

through six were conducted with the revised 10-item ELS. The tests included CFA and

discriminant analysis that contributed to the confirmation that the ELS had both construct

and discriminant validity. Study seven was conducted with the ELS and included
 33

structural equation modeling (SEM) for the analysis of in-group agreement. The results

indicated the ELS predicted several items, including the employees’ willingness to report

problems to leadership (Brown et al., 2005).

Ethics as a Predictor of Behavior

The literature on the relationship between ethics and safety performance has not

been clearly defined or well researched (Freiwald, 2013; Kapp & Parboteeah, 2008).

There is a belief that management has an ethical obligation to maintain safety (Erikson,

1997). Research has suggested that if employees believe that management values safety,

then safety performance is enhanced (Erikson, 1997). Other studies have asserted that

ethical climate has a strong influence on safety behavior (Kapp & Parboteeah, 2008).

Freiwald’s (2013) research showed a strong positive relationship between ethical

leadership and workplace injuries. The results of the survey and subsequent SEM

showed a statistically significant relationship between employees’ perceptions of ethics in

their company leadership and fewer injuries (Freiwald, 2013). Additionally, Brown,

Treviño, and Harrison’s ELS (2005) demonstrated the ability to predict the employee’s

willingness to discuss problems with organizational leadership (Brown et al., 2005).

Criterion or Self-Reported Outcomes

There are many studies in the literature that support the theory that safety climate

influences safety behavior, though some concerns exist about possible confounding

variables. Theoretically, the relationship between safety climate and safety behavior may

be caused by other factors such as the social exchange theory (Vroom, 1964) where the
 34

company’s concern for the employees is reciprocated through the employees trying to

provide value in return by adhering to safety policies or alternately, by the expectancy-

valance theory where the employees want to participate in the safety program due to a

belief that it will lead to an outcome valuable to themselves (Neal & Griffin, 2006).

Additionally, there were other concerns in the literature about reverse causality in the

relationship between safety climate and safety behavior / safety performance, though the

reverse causality concerns were rejected by both Clarke (2006) and Neal and Griffin

(2006).

Despite the aforementioned concerns, there have been a series of safety climate or

safety culture studies that indicate a strong and statistically significant relationship

between safety climate or culture and safety behavior (Neal & Griffin, 2006; O’Toole,

2002). These results have led to an ongoing debate on the superiority of criterion-based

safety outcomes versus self-reported safety outcomes.

In 2000, Zohar wrote that safety climate research was being hampered by a lack

of criterion data (Zohar, 2000). Johnson’s study in 2007 supported the predictive validity

of safety climate as characterized by criterion data. More recently, both Freiwald (2013)

and Alsowayigh (2014) supported the concept that safety culture influenced directly or

indirectly self-reported injuries and safety performance, respectively. Both

methodologies have their merits and their issues. The concern with criterion-based

reports is that there is bias in the reporting, where many minor occurrences such as

smaller injuries or minor violations can go unreported, therefore tainting the results

(Thompson et al., 1998). These minor occurrences have the potential to be leading

indicators for a decline in safety performance, but only if reported (Thompson et al.,
 35

1998). Self-reported survey results on safety climate also may contain bias from the

respondents based on having been in an accident or witnessing one (Neal & Griffin,

2006).

Criterion Based Outcomes in Aviation

Figure 3. Safety Management Continuum. (Stolzer et al., 2011, p. 235).

There are several scholars such as Zohar and O’Connor et al. that support

quantitative criterion data superiority versus forms of data such as survey results from

self-reports of errors or violations. Zohar’s (2000) research on micro-accidents was

evidence of the predictive value of safety climate, though the researcher relied upon

smaller accidents that were properly documented. Thompson et al. (1998) suggests that
 36

many smaller accidents go unreported, which has the potential to bias future studies

without the controls employed by Zohar.

In Figure 3, Stolzer and Goglia’s Safety Management Continuum illustration

(Stolzer & Goglia, 2015, p. 2015) shows that in an SMS, many of the sources of data are

criterion based. Examples include data from flight data analysis / FOQA, most of the

predictive sources of data from data mining, probabilistic risk assessment, and modeling

are inherently criterion-based data that are quantitative and not self-reported. O’Connor

et al. also suggested FOQA would be a possible criterion data source for the prediction of

aviation accidents (O’Connor et al., 2011). Despite the potential benefits, FOQA data in

general aviation aircraft can be very expensive (Mitchell, Sholy, & Stolzer, 2007), and

the use of data from those devices would raise many privacy and autonomy concerns.

O’Connor, et al. stated the accident event rate in aviation is already too low to

generate valid predictive models based solely on accidents themselves (O’Connor et al.,

2011); therefore, aviation needs reliable and affordable measures of the deterioration of

safety performance before the chain of events that leads to accidents begins.

Criterion Measurement Variability and Reliability

Criterion, or hard quantitative based data, is unlikely to be comprised of

comparable measurements across diverse aviation organizations. The measurements of

parameters will be calibrated differently and therefore have different meaning from

organization to organization. For example, the accelerometer is designed to measure the

amount of gravity or g-forces applied to the aircraft upon landing. During one study

conducted by Cistone et al. (2011), many inconsistencies were discovered in the

 37

measurement of the g-forces experienced by one airline’s fleet. The variability of the

measurements, even within a single aviation organization, was such that it made it

difficult to derive valid results. Sources of variability included that accelerometers were

not all placed on the aircraft in the same location, the levels of calibration varied from

accelerometer to accelerometer, and the manufacturer of the accelerometers varied.

Additionally, the variation among aircraft types and the different levels of g-force

tolerance for those different types made cross comparisons of the importance of specific

g-force measurements significantly more difficult. This example illustrates the challenge

of deriving useful comparable data even when measurements were all conducted within

the same aviation organization. The same type of research, if attempted across many

diverse aviation organizations with over 100 different aircraft types, would suffer even

more from this problem. Therefore, a useful cross comparison of hard data on some

measures may be nearly impossible.

Self-Reporting Outcomes

Many studies have shown that safety climate either directly or indirectly

influences both self-reported and criterion-measured safety behavior. Alsowayigh (2014)

and Freiwald’s (2013) research results supported safety culture / climate and ethical

leadership as a viable mechanism to predict self-reported safety outcomes. Clarke (2006)

concluded that safety culture predicted safety performance, and safety performance was a

valid and generalizable predictor of accidents when accident involvement was measured

after the safety climate measurement.

 38

Consistent Methodology

O’Connor et al. (2011) have suggested as a best practice that researchers use

consistent measurements in order to compare results with similar themes. Yet, there are

few replicated studies in the literature conducted regarding safety culture and self-

reported safety performance of different organizations such as Fractionals. This research

has the potential to re-confirm the relationship of safety culture, pilot commitment to

their organization, and safety performance of similar organizations. This cross

comparison would be an inexpensive measure to implement and monitor, yet the findings

could have a meaningful impact on improving safety in other U.S. jet FAR 135

companies.

Hypotheses

A structural equation model was used to evaluate the relationship among the

variables used in this study. Previous studies found in the extant literature were analyzed

to develop the conceptual framework for the model. This study augments previous work

by evaluating the relationship of safety culture with pilot commitment to the

organization, ethical leadership and self-reported safety performance. The assumptions

were based on the findings from the more recent studies by Alsowayigh (2014) and

Freiwald (2013), though the foundations of the assumptions date back to long established

constructs. The hypotheses shown in Figure 4 were tested in this research.

𝐻1 : A positive safety culture has a positive influence on pilot commitment to the

organization.
 39

Safety culture was found to have a direct and significant influence over pilot

commitment to the airline in the Alsowayigh (2014) study. This relationship is likely to

remain consistent with the pilots of the U.S. jet Fractionals.

𝐻2 : A positive safety culture has a positive influence on ethical leadership.

𝐻3 : A positive safety culture has a negative influence on safety performance.

The findings from Alsowayigh (2014) showed there was no significant direct

effect between safety culture and pilot error behavior. Previous research (Alsowayigh,

2014) has shown a significant and direct negative relationship between safety culture and

own attitude to violations. The same research also demonstrated the relationship between

safety performance and safety culture was not mediated by pilot commitment to the

airline (Alsowayigh, 2014). The relationship in this study is unlikely to be mediated by

the Fractional pilot commitment to the organization.

𝐻4 : A positive pilot commitment to the organization has a positive influence on

safety performance.

Previous research (Alsowayigh, 2014) has shown that pilot commitment to the

airline did not have a significant relationship with the pilot’s performance in the cockpit.

Alsowayigh (2014) suggested that safety performance in the cockpit was driven by their

professionalism as a pilot.

𝐻5 : A positive ethical leadership has a negative influence on safety performance.

Ethical leadership has been shown to be related to the safety outcomes

subcomponent of safety climate construct (Freiwald, 2013). This study has the potential

to find a relationship between ethical leadership and safety performance.

 40

𝐻6 : A positive ethical leadership has a positive influence on pilot commitment to

the organization.

Ethical leadership has been correlated to employee commitment to the

organization (Trevino et al., 1998). This study has the potential to find a relationship

between ethical leadership and pilot commitment to the organization.

Hypothesized SEM Model

H1 H4

H3
H6

H5
H2

Figure 4. Hypothesized SEM Model.

Summary

There exists a material gap in the literature of research focused on Fractional and

Charter jet operations. Fractional and Charter operations are dissimilar to airline
 41

operations in several key areas. One such area is the amount of airports served by

Fractional and Charter far exceeds those served by the airlines. This means that

Fractional and Charter operators often use second and third tier airports that have shorter

runways with less safety equipment and possibly no operating control tower. Another

area that is dissimilar to most airline operations many Fractional and Charter flights

encounter is autonomy. This means the pilots for many Fractionals and Charters perform

the majority of their duties autonomously without the benefit of direct supervision.

There is an opportunity to advance aviation research using consistent

methodologies (O’Connor et al., 2011) through the study of corporate jet operations.

There are distinct differences in the historical safety performance between Fractionals

and Charters despite operating under similar FAA regulations. This study determined a

baseline of safety culture and ethical leadership for the Fractionals. These baselines can

be used in future research to search for differences between Fractionals and Charters to

begin to draw inferences of causation. If causal inferences can be drawn and operational

changes enacted, the historical safety gap between these two groups can potentially be

narrowed. In addition to safety in corporate jets being enhanced, the lessons learned may

be applied to other sectors of aviation.

 42

CHAPTER III

METHODOLOGY

A review of the available literature on safety culture, ethical leadership, and safety

performance supports that structural equation modeling (SEM) is an appropriate method

to determine the relationships among variables and is an effective means of investigating

the hypotheses of this study. Freiwald (2013) used this approach in the determination of

the relationship among ethical workplace climate, safety climate, and occupational

injuries. SEM was also employed by Alsowayigh (2014) when establishing the

relationship among safety culture, pilot commitment to the airline, and safety

performance.

Research Approach

SEM is a methodology that tests hypotheses in a confirmatory manner. The

underlying regression equations in SEM determine a structure to the relationships under

study and display these relationships graphically for better understanding. SEM tests

these hypothesized relationships simultaneously. If the model is adequate, the underlying

relationships may be determined to be both directional and possibly causal. SEM is used

for confirmatory analysis and not for exploratory analysis (Byrne, 2010).

The naming of the factors was based on the previous construct names used in the

literature, abbreviated due to the space constraints, and adapted for improved recognition.

As shown in Table 1, the exogenous variable is Safety Culture, and the endogenous

variables are Pilot Commitment to Aircraft Management Company (AMC), Ethical

Leadership, and Safety Performance.

 43

Table 1

Study Variables

Variable Dimension Abbreviation Description

Exogenous Variable
How the AMC values
safety and if the AMC
Organizational
OC goes above and beyond
Commitment
the minimum
requirements.
This evaluates AMC
Operations
OP personnel (chief pilot,
Personnel
dispatch, trainers).
Safety
This evaluates the
Culture
support and
Informal
IS encouragement among
Safety System
AMC pilots toward
safety.
This rates the safety
Formal Safety reporting and feedback
FS
System loop and AMC's safety
personnel.
Endogenous Variables
This evaluates the
Ethical Ethical perception of AMC
EL
Leadership Leadership leadership's moral and
ethical behavior.

This evaluates the

Pilot Pilot
pilot's willingness to go
Commitment Commitment PC
above and beyond for
to AMC to AMC
the AMC.

This is a self-report of
Pilot Error
ER mistakes made by AMC
Behavior
pilots during operations.
Safety
This is a self-report of
Performance Pilot Own
AMC pilot's attitude
Attitude
AT toward the regulations
Toward
and their willingness to
Violations
bend the rules.
 44

Design and Procedures

The survey instrument was modeled after the instrument in the Alsowayigh

(2014) study with minor adaptations to adapt from commercial aviation to general

aviation vernacular. The ELS was added to the end of the survey to preserve the question

order from the Alsowayigh (2014) study. The survey was constructed and facilitated in

Survey Monkey® online service. The Survey Monkey® online service was selected based

on previous studies found in the literature.

All pilots who were invited to take the research survey and allowed access to the

research survey were verified with FAA records to hold an Airline Transport Pilot

certificate (ATP), a current First Class Medical certificate, and a type rating consistent

with those aircraft types flown by U.S. Fractional companies. The prequalification

process (Pre-Qual) included verifying the credentials of each respondent before the

respondents were allowed access to the survey.

Prior to employment at Flight Options, Flexjet, and Net Jets, each pilot was

required to meet the aforementioned minimum pilot standards. Therefore, all Fractional

pilots on the union message boards met the Pre-Qual standard and were allowed

immediate access to the research survey.

A separate pre-qualification survey was set up in Survey Monkey® requiring

pilots who did not undergo the Pre-Qual process to provide their name, home town, level

of medical certificate, level of pilot certificate, and type ratings held. A research assistant

verified the credentials for each pre-qualification survey respondent with the FAA

database. If the respondent’s answers were not verified, the respondent was not sent the

research survey.
 45

The Fractional pilots who were invited to take the research survey by direct mail

and ERAU alumni emails were pre-qualified by a research assistant prior to receiving the

invitation to participate. These pre-qualified pilots who opted to participate were allowed

immediate access to the research survey.

There were three other sources of pilots who volunteered to participate in the

research study. Aviation International News (AIN) has a bi-weekly newsletter that ran

three solicitations in its newsletter asking Fractional pilots to participate in a research

study. Of the estimated 1,000 plus Fractional pilots who may have seen the solicitation,

50 pilots were verified through the Pre-Qual process and invited to take the research

survey. Of the pilots who passed the Pre-Qual process, 37 completed the research

survey. This process was repeated in the Flight Safety Information (FSI) newsletter,

where 20 additional fractional pilots volunteered to participate, 8 pilots passed the Pre-

Qual process, and 6 pilots completed the survey. In addition to the newsletter

solicitations, a former Flight Safety Instructor for Net Jets invited several current Net

Jets’ pilots to take the survey. The pilots who responded were required to go through the

Pre-Qual process before taking the research survey.

All pilots who volunteered to participate were directed to an informed consent

form (see Appendix B) prior to taking the survey. The pilots who consented were

prompted to also confirm their position as a current Fractional pilot for a U.S. based

Fractional program. The survey was constructed to terminate if the pilot did not confirm

his or her current status as a pilot at a U.S. Fractional AMC. The pilots then continued to

the demographics portion of the survey and were then asked to provide their perceptions
 46

of their company’s safety culture, their own commitment to the organization, ethical

leadership qualities of their organization, and their safety performance.

The survey software was constructed to limit the pilots to one answer for each

item within the instrument. All incomplete surveys were excluded from the study. The

data received through the Survey Monkey® software were exported directly to IBM SPSS

23 software for further analysis. A confirmatory factor analysis and full structural

equation model were conducted with IBM AMOS 23.

Apparatus and Materials

The survey was facilitated electronically and could be taken on most smart

phones, tablets, or computers. The survey was developed, delivered, and data were

collected through the Survey Monkey® online platform. The survey consisted of 93 total

questions. The response to the first question determined if the respondent was qualified

to participate in the study. The subsequent five questions were demographic questions

referring to the primary aircraft flown, year of birth, company position, flight experience,

and tenure with the AMC. The remaining 87 questions were adapted from previously

validated surveys with necessary modifications to adapt from commercial aviation

vernacular to that of general aviation. The last question was added based on a question

inserted in the Alsowayigh (2014) research, and because it applied similarly to this study.

Population/Sample

The population of Fractional jet pilots in the United States, as shown in Table 2, is

estimated to be 3,660, with 3,425 of those pilots being unionized. This estimate is based
 47

on a ratio of 6.1 pilots per aircraft managed by the Fractional companies. These figures

are derived from the ratio of union members to aircraft managed by their respective

Fractional companies. NetJets is the largest Fractional company with 429 aircraft in the

United States (JetNet Fractional Program Summary, 2015) with an estimated 2,700 pilots.

Net Jets’ pilots are unionized, and an estimated 2,690 (99.8%) are represented by the Net

Jets Association of Shared Aircraft Pilots (NJASAP). Flight Options has 60 aircraft in

the United States with an estimated 385 pilots. Flight Options’ pilots are unionized with

an estimated 380 (99%) that are represented by the International Brotherhood of

Teamsters #1108. FlexJet was recently acquired by Flight Options and has 66 aircraft in

the United States with 350 pilots. FlexJet and Flight Options’ pilots voted to unionize in

December of 2015, and the FlexJet pilots became members of the Flight Options’ union

(IBT 1108). The remaining Fractional pilots are employed at Executive AirShare and

several small regional Fractional programs, which have an estimated total of 150

additional non-union Fractional jet pilots.

Table 2

Fractional Pilots in U.S.

Jets in Union
*Pilots
Fleet Members
NetJets 429 2,700 2,690
FlexJet 63 380 375
Flight Options 60 370 360
Executive Airshare 27 167 0
Others in U.S. 7 43 0
Total 586 3,660 3,425
* Estimated based on 6.1 average pilots per jet ratio
 48

The sampling frame consisted of an estimated minimum of 3,460 Fractional

pilots. There were the 3,425 union pilots who have access to their union message boards

plus an additional 35 Fractional pilots who were contacted directly through U.S. mail

(See Appendix F) or email. Each of the 3,460 pilots had a non-zero chance of

participating in the research survey. The sampling frame, therefore, consisted of 95.2%

or more of U.S. Fractional pilots. The remaining 4.8% (175 pilots) of Fractional pilots

may have seen the multiple invitations in both Aviation International News (AIN) alerts

and / or the Flight Safety Information Newsletter. Due to these newsletter invitations,

many of the remaining non-union Fractional pilots had a non-zero chance to participate in

the survey, therefore minimizing coverage error (Dillman et al., 2009).

The SEM methodology requires the sample size to vary with the complexity of

the model under study (Westland, 2010). Determination of the appropriate sample size

for the SEM model is non-trivial (Westland, 2010) and must meet the requirements

considered acceptable in the available literature. Presented in Table 3 are several

researchers and their suggested sample sizes based on the hypothesized SEM in this

study. This study has 87 observed variables and 10 latent variables with a targeted

significance level of .05 (p = .05), effect size of .1, and statistical power of .8. The

sample size based on the majority of the literature is 200 respondents or greater. The

current study has over 300 completed and valid responses (n = 305). The current study’s

sample size of 305 responses satisfies the requirements of Ding et al.’s (1995) (n = 150),

Kline’s (2005) (n > 200), and Boomsma & Hoogland (2001) (n > 200) as shown in Table

3.
 49

Table 3

SEM Sample Size Requirements

Researcher(s) Year
N: 100-150 Ding, Velicer, and Harlow, 1995
N: > 200 Kline 2005
N: > 200 Boomsma & Hoogland 2001
N: 579 to 3,231 Westland 2010

Sources of the Data

The data used in this study were obtained through the online survey responses

received by pilots who volunteered to complete the survey. The survey is a compilation

of five different instruments. The survey questions seen by the respondents are displayed

in Appendix B.

The respondents from electronic solicitations were presented a link in a

newsletter, email, or on their union message board. The respondents from the post card

in Appendix F were directed to a web domain (www.safetyculturesurvey.org) that

connected them to the research survey. All respondents provided their informed consent,

shown in Appendix B, before advancing to the research survey. No direct emails of any

of the recipients were provided by any of the organizations targeted for this study. Union

members posted a direct link to the survey on their union message boards. Additional

controlled invitations were sent via direct email, email, a posted link on controlled

websites, or electronic newsletters.

Prior to conducting this research, initial training from the Collaborative

Institutional Training Initiative (CITI) was completed and an application was submitted
 50

to the Institutional Review Board (IRB) at Embry-Riddle Aeronautical University. The

application received approval prior to start of data collection. The IRB approval letter is

presented in Appendix A.

Data Collection Device / Survey Design

The study included six demographic variables plus 87 observed variables (see

Appendix B) that represented ten constructs that were derived from five instruments that

had been used extensively in the literature. The instruments were:

Safety culture (SC). The Commercial Aviation Safety Survey (CASS) was

developed and validated by Gibbons et al. (2006). Initially, the instrument was an 84

item scale that consists of five constructs; however, during validation, the instrument was

reduced to a 55 item scale with four constructs. Each question is measured using a 7-

point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). The main

factors include organizational commitment (OC), operational personnel (OP), formal

safety systems (FS), and informal safety system (IS) (Gibbons et al. 2006).

Organizational commitment (OC) items include, “management expects pilots to push for

on-time performance, even if it means compromising safety.” Operational personnel

(OP) items include, management “inappropriately uses the MEL (e.g., use when it would

be better to fix equipment).” Formal safety systems (FS) items include, “the safety

reporting system is convenient and easy to use.” Informal safety system (IS) items

include, “management shows favoritism for certain pilots.”

 51

Pilot commitment to AMC (PC). The Organizational Commitment

Questionnaire (OCQ) was initially developed by Porter et al. (1974) and has two

versions: a long and short version. The long version has 15 questions and is multi-

dimensional, whereas the short version, which is recommended by Commerias and

Fournier (2001), has 9 questions and is considered uni-dimensional. The questions are

measured on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly

agree). Items include, “I talk up this organization to my friends as a great organization to

work for,” and this aircraft management company “inspires the best in me in the way of

job performance.”

Ethical leadership (EL). The ethical leadership scale (ELS) was developed by

Brown et al. (2005) and originally consisted of 48 items. After Brown et al. (2005)

conducted Exploratory Factor Analysis (EFA), the ELS was reduced to a 10 item

instrument. This instrument used a 5-point Likert scale from 1 (Strongly Disagree) to 5

(Strongly Agree). Items include, management “makes fair and balanced decisions” and

management “can be trusted.”

Pilots’ own attitude to violations (AT). The own attitude to violation scale was

developed by Fogarty (2004) as a self-reported scale and included nine items. These

items were measured on a 5-point Likert scale. Items include, “bending a procedure is

not the same as breaking it” (Fogarty, 2004).

 52

Pilot error behavior (ER). The error scale questionnaire was developed by

Fogarty (2004) and included three items. This survey was initially developed as a self-

reported scale for airline maintenance personnel. Alsowayigh stated, “The questions are

general and can be applied to airline pilots” (Alsowayigh, 2014, p .38). The questions are

measured on a 5-point Likert scale. Items include, “I make errors in my job from time to

time” (Fogarty, 2004).

Construct Validity

The items in the study were measured to confirm they represented the latent

constructs they were expected to measure based on the available literature (Hair et al.,

2010). The four components of construct validity are Convergent, Discriminant, Face,

and Nomological (Hair et al., 2010). The model diagnostics of each component was

tested in this study.

The five instruments selected to create the composite instrument in this research

have all have been used repeatedly in the literature. Each instrument has had its construct

validity demonstrated in the literature, and many of these instruments have been used in

multiple studies.

Convergent validity. There are several measures used to estimate the convergent

validity of the items in a research study (Hair et al., 2010). The factor loadings and

average variance extracted (AVE) were each checked in the model (Hair et al., 2010).
 53

The AVE is a summary measure of convergent validity, and the formula is shown

in Figure 5. The standardized factor loadings for each item on each construct were

squared and then a construct average variance was established (Hair et al., 2010).

𝑛
∑ 𝜆2𝑖
𝑖=1
AVE =
𝑛

Figure 5. Average Variance Extracted (Hair et al., 2010).

Reliability. Reliability was tested using Cronbach’s alpha (1951). Before

inclusion in the study, each of the five instruments employed to create the composite

survey was previously tested for internal consistency. In each case, the instruments used

in this study satisfied the minimum suggested value of .7 (Hair et al., 2006) as measured

by Cronbach’s alpha (1951), with the exception of the pilot error scale, which had been

measured at .6 in one study (Fogarty, 2004).

In recent SEM studies, construct reliability (CR) has been tested by comparing the

square of the summed standardized factor loadings with the error variances (Hair et al.,

2010) for each factor as shown in Figure 6. CR values over .7 suggest good reliability

(Hair et al., 2010).

 54

Figure 6. Construct Reliability Formula (Hair et al., 2006, p. 777).

Discriminant validity. The discriminant validity is a measure by which each

construct is truly distinct (Hair et al., 2010). This is tested through a comparison of the

variance-extracted percentages of two constructs with the squared correlation between the

two constructs. (Hair et al., 2006). Kline (2005) suggested that a model has discriminant

validity if no two factors have correlations higher than .85.

Nomological & face validity. Nomological validity was analyzed by reviewing

the correlations between the constructs to determine if they made sense (Hair et al.,

2010). The face validity was analyzed by a review of the content of the items in each

construct to ensure they measured what was intended. Face validity of the items of each

construct was also analyzed by two experienced general aviation pilots. These two pilots

had a combined experience of more than 40 years and had both been employed in a

Fractional program.

Treatment of the data

Demographic Data. Descriptive statistics were computed from the survey data

based on pilot tenure at the AMC, weight of equipment flown, position, and age. The
 55

pilot demographic data were also collected for potential inclusion in future research to

compare group differences.

Missing data. The survey was constructed to require one answer for each

question prior to continuing the survey. A not applicable choice was not presented in the

instrument. All 52 incomplete responses were excluded from the analysis; therefore,

there were no surveys with missing data used in the study.

Outliers. The Mahalanobis distance (D2) was calculated for each of the variables

searching for significant outliers. The literature suggests that outliers should be retained

unless their retention is particularly detrimental to the model (Hair et al., 2006). The

model was tested with and without the outliers, and the model fit deteriorated with the

outliers removed. The determination was made to retain all significant outliers in the

model.

Normality. Multivariate normality was analyzed with particular consideration for

kurtosis because SEM is sensitive to kurtosis (Byrne, 2010). In the assessment of

multivariate normality, items that were determined to be more than slightly skewed

(>1.0) or kurtotic (> 7.0) (Byrne, 2010) were evaluated. The content of these non-normal

items was reviewed and a determination of their importance to the model was made.

Items that were non-normal, contributed little to the model, and their temporary removal

benefitted the model fit were permanently removed from the study.
 56

Confirmatory Factor Analysis (CFA)

CFA was used to confirm the latent variables for each of the 10 factors in the

model (Byrne, 2010). The CFA was conducted with IBM SPSS AMOS 23 software in

order to validate the measurement model and confirm the factors measured as intended

(Byrne, 2010). The model was checked for covariance, outliers, and cross-loading.

Model re-specification was conducted by changing one item per iteration.

The model was evaluated using Normed Fit Index (NFI), Goodness of Fit Index

(GFI), Adjusted Goodness of Fit Index (AGFI), Comparative Fit Index (CFI), Root Mean

Square Error of Approximation (RMSEA), and normed Chi-square (CMIN/df) (Byrne,

2010). According to Vandenberg and Scarpello (1990), the fitness of a model should be

analyzed with more than one fitness index, so the NFI, GFI, AGFI, CFI, RMSEA, and

CMIN/df were used in the present study.

The first analysis of model fit was conducted with the Normed Fit Index (NFI).

The NFI is a non-centrality based index (Byrne, 2010) that tests the hypothesized model

against the null hypothesis (Byrne, 2010). If the NFI analysis returns a value close to .95

(Byrne, 2010; Hu & Bentler, 1999), it is considered a good fit, with values from .90 to

.949 still considered acceptable. The NFI has been known to underestimate fit in smaller

sample sizes (Byrne, 2010); therefore, the Comparative Fit Index (CFI) was also used to

evaluate the model fit.

The subsequent analysis of model fit was conducted with both the Goodness of

Fit Index (GFI) and the Adjusted Goodness of Fit Index (AGFI). The GFI measures the

relative amount of variance and covariance in the sample data that the hypothesized

model can explain (Byrne, 2010). The GFI was developed to be less sensitive to large
 57

sample sizes (Hair et al., 2006). The AGFI is very similar, except that the AGFI accounts

for the degrees of freedom in the model (Byrne, 2010). If the GFI and AGFI indices are

greater than .9 (> .9), then model fit is considered acceptable (Hair et al., 2006). The

closer the value is to 1.0, the better the fit (Byrne, 2010; Jöreskog & Sörbom, 1993)

Additional analysis of model fit was conducted with the comparative fit index or

CFI which, like the NFI, is a non-centrality based index (Byrne, 2010) that tests the

proposed model against the null hypothesis (Byrne, 2010). The CFI is chosen frequently

in studies because it demonstrates insensitivity to model complexity (Hair et al., 2010).

As with the NFI, if the CFI analysis returns a value close to .95 or greater, it is considered

a good fit (Byrne, 2010; Hu & Bentler, 1999). If the CFI returns values from .90 to .949,

the fit is still considered acceptable.

A further metric employed was the Root Mean Square of Error Approximation

(RMSEA). The RMSEA is considered a badness of fit index, which means that lower

values indicate a better fitting model (Byrne, 2010). RMSEA is recommended for studies

with a large number of observed variables because other 𝜒 2 Goodness of Fit (GOF) test

statistics tend to reject acceptable models with a large number of observed variables, such

as the current study (Hair et al., 2010). A value of the RMSEA of .6 or below is

considered a good fit for the data (Byrne, 2010; Hu & Bentler, 1999).

The final fit metric was the 𝜒 2 statistic (CMIN/df), which computes the model’s

distance from a theoretically perfectly fitted model divided by the degrees of freedom

(Hu & Bentler, 1999). The lower the CMIN/df value is, the better the model fitness. The

chi-square is sensitive to sample size (Hair et al., 2006). The CMIN/df is a comparative
 58

ratio and is considered to be acceptable if value is below three (Byrne, 2010; Hair et al.,

2006).

Exploratory Factor Analysis (EFA)

The model did not achieve the fit criteria in Table 10; therefore, an EFA was

conducted on the data (Byrne, 2010). A principal component analysis (PCA) was

conducted with Varimax rotation. The PCA was chosen because the results were

considered easier to interpret. The PCA is designed to reduce the number of variables

down to the items that explain the largest amount of variance in a given model (Grimm et

al., 2000). An oblique rotation was considered due to its advantage with cross-loading

items (Hair et al., 2006); however, the Varimax rotation was selected because it was more

frequently chosen in the safety culture and safety climate literature, such as Freiwald’s

(2013) study.

The EFA was run, and the Kaiser-Meyer-Oklin (KMO) measure of sampling

adequacy was analyzed (Hair et al., 2006). This is the measure of the ratio of squared

correlations between variables and the partial squared correlations between variables.

KMO measures above .9 (> .9) are considered very good (Field, 2009).

The Measure of Sampling Adequacy (MSA) was analyzed for the appropriateness

of conducting an EFA. All variables (> .5) were considered appropriate (Hair et al.,

2005). The variables below .5 were removed from the model, and the model was re-run.

The EFA was conducted with IBM SPSS 23 software. All factors that returned

eigenvalues greater than 1.0 (> 1.0) and had a contribution percentage of greater than 1%

(> 1%) of the variance in the model (Grimm et al., 2000) were analyzed. The EFA
 59

results displayed many more than the eight first order factors in the proposed model;

therefore, after evaluation, the model was re-run with a constraint for seven factors. The

seven-factor constraint was chosen based on grounded theory to reduce the complexity in

the model. All items with similar factor loadings on multiple factors were evaluated for

removal. Factors with no basis in grounded theory were analyzed for removal from the

study.

Model 2 (M2) Confirmatory Factor Analysis (CFA)

The CFA was conducted on M2 model. Based on a review of the available

literature, the M2 constructs were evaluated against the validated instruments chosen for

the study. Based on grounded theory of the latent factor structure, items that were

loading near or below .7 (Hair et al., 2006), non-normal, or loading on a latent factor not

supported by previous studies were evaluated for removal.

Post hoc analysis. Post hoc analysis was conducted based on the Modification

Indices (MIs). Model re-specification is by nature exploratory because the researcher is

re-specifying the hypothesized model for methods to improve the model (Byrne, 2010).

A model with good fit indices and also with high MIs can be an indication of multi-

collinearity in the model (Kline, 2005) rather than causal significance. MIs were

reviewed, and those that exceeded 5.00 were co-varied when on the same factors.

The CFA for the M2 required additional regressions constraints on each of the

items in the ERN and ATN constructs. Hair et al. (2006) recommend the use of at least

three items for each factor when the sample size is below 300 (n < 300). There is a
 60

concern that factors with less than three items will not have the appropriate level of

degrees of freedom to determine a solution that fits the data (Hair et al., 2006). The

current research study had over 300 (n = 305) completed and valid responses; therefore,

additional regression constraints were added before conducting the SEM.

Structural Equation Model & Hypotheses Testing

The previously mentioned model fit indices were re-evaluated by comparing them

to the model fit in the final CFA and additionally to the fit criteria in Table 10. The

model fit in SEM was similar to the final CFA and met all the criteria in Table 10. The

AGFI was the only fit criteria below the target level (> .9). As previously stated, it was

determined to be acceptable.

The six hypotheses were evaluated by reviewing the SEM regression weights,

standardized estimates, and p values. The analysis was conducted using IBM SPSS

AMOS 23 software. The maximum likelihood estimation was employed for the analysis

(Byrne, 2010). The elimination of the PC factor in the EFA precluded the testing of three

of the six hypotheses. The model fit was determined to be adequate, and the remaining

three hypotheses were tested.

 61

CHAPTER IV

RESULTS

This study explored the relationship between Safety Culture, Ethical Leadership,

Pilot Commitment to the AMC, and Safety Performance. Based on the available

literature, a model was developed to determine the effect of Safety Culture on Ethical

Leadership, Pilot Commitment, and Safety Performance. Additionally, the effect of

Ethical Leadership on Safety Performance was also tested.

This chapter shows the results of the CFA on the proposed model, subsequent

EFA, final CFA, and SEM. The model fit history of the CFA is shown with nine

revisions in Table 12 and the SEM model fit shown in Table 14. The results of the

hypothesis testing are included in this chapter. The descriptive statistics for each of the

items is displayed in Appendix C. The SC & PC constructs were measured on a seven-

point Likert scale. The remaining constructs of ER, AT, and EL were each measured on

a five-point Likert scale.

Demographic Data

Three hundred fifty-seven respondents participated in the research survey; all

respondents completed the survey electronically. Table 4 shows there were 305 (n = 305)

complete and valid responses used in the study, representing 8.3% of the estimated 3,660

Fractional jet pilots in the United States.

 62

Table 4

Completed Responses

Estimated Completed
Source Views Pre-Qual Surveys Percentage
Direct Mail to Prequalified Pilots 1,759 All 111 36.4%
NJASAP Message Board 2,660 All 80 26.2%
FlexJet/FO Message Board 780 All 46 15.1%
Aviation International News 1,000 50 37 12.1%
Embry-Riddle Alumni Email 249 All 16 5.2%
Flight Safety Instructor 180 9 9 3.0%
Curt Lewis Newsletter 160 8 6 2.0%
Total 6,788 305

Table 5 shows the pilots’ ages ranged from 28 years old to 74 years old,

representing a range of 46 years between the youngest and oldest pilot. The median age

was 49, and the mean age was 49.14 years old. The proximity of the mean age to the

median age of the data showed the age data was not skewed. The mode was 43 years of

age.

Table 5

Pilot Age (Years)

Cumulative
Frequency Percentage Percentage
20-29 years 1 0.3% 0.3%
30-39 years 37 12.1% 12.5%
40-49 years 131 42.9% 55.4%
50-59 years 97 31.8% 87.8%
60-69 years 35 11.5% 98.7%
70-79 years 4 1.3% 100.0%
Total 305
 63

The most frequent position held by 54.8% of the respondents was Pilot In

Command (PIC), often called Captain, followed by First Officer or Second in Command

(SIC), which represented 27.8% of the respondents. Table 6 shows there were 15.4% of

pilots who were Captains with additional duties such as Check Airman, and 2% of the

respondents were part of the management team at the AMC.

Table 6

Position at AMC

Cumulative
Frequency Percentage Percentage
0-4 years 11 3.6% 3.6%
5-9 years 39 12.8% 16.4%
10-14 years 121 39.7% 56.1%
15 or more years 134 43.9% 100.0%
Total 305

The type of equipment flown by the pilots in Table 7 was split evenly among

Light Jet (29.5%), Mid-Sized Jet (25.6%), Super Mid-Sized Jet (24.3%), and Large Jets

& Long Range Jets (20.7%). The data contained a well-balanced mix of pilots flying a

wide range of equipment.

 64

Table 7

Aircraft Type Flown (Max Takeoff Weight)

Cumulative
Frequency Percentage Percentage

Light Jet (up to 19,999 lbs) 90 29.5% 29.5%

Mid-sized Jet (20,000 -
29,999 lbs) 78 25.6% 55.1%
Super Mid-sized Jet
(30,000 - 39,999 lbs) 74 24.3% 79.3%
Large Jet (40,000 - 49,999
lbs) 32 10.5% 89.3%
Long Range (50,000 lbs or
greater) 31 10.2% 100.0%
Total 305

Table 8 shows the majority of respondents (51.5%) had over 10,000 hours of

flight experience with 27.5% having between 7,500 and 9,999 hours of flight experience,

18.7% had between 5,000 and 7,499 hours, and just 2.3% had below 5,000 hours. In

contrast to commercial pilots, general aviation pilots do not accumulate flight hours at the

same pace; therefore, having the majority of pilots with over 10,000 hours of flight

experience is uncommonly high for a general aviation organization.

Table 9 shows that 3.6% of respondents had been with their AMC less than 5

years, 12.8% had been with their AMC between 5-9 years, 39.7% between 10-14 years,

and 43.9% had been with their respective AMC for 15 years or more. The tenure with

the AMC indicates that the Fractional pilots that completed the survey stay with their

respective companies for many years.

 65

Table 8

Pilot Experience (Hours)

Cumulative
Frequency Percentage Percentage
2,500 - 4,999 hours 7 2.3% 2.3%
5,000 - 7,499 hours 57 18.7% 21.0%
7,500 - 9,999 hours 84 27.5% 48.5%
10,000 hours or more 157 51.5% 100.0%
Total 305

Table 9

Tenure at AMC

Cumulative
Frequency Percentage Percentage
0-4 years 11 3.6% 3.6%
5-9 years 39 12.8% 16.4%
10-14 years 121 39.7% 56.1%
15 or more
years 134 43.9% 100.0%
Total 305

Normality & Outlier Checks

The outliers were checked by analyzing the Mahalanobis D2. There were 57 cases

that were considered outliers that were significant to the .05 level (p < .05). The model

fit was checked with the outliers, and the model fit indices were CMIN/df = 1.777, NFI =

.715, GFI = .669, AGFI = .649, CFI = .85, and RMSEA = .051. After the outliers were

removed, the model fit indices deteriorated with CMIN/df = 1.704, NFI = .701, GFI =
 66

.637, AGFI = .615, CFI = .849, and RMSEA = .053. The outliers were retained in all

future models.

The multivariate normality was analyzed, and it was determined there were

several variables that had a skewness over 1.0 and/or a kurtosis greater than 7.0 (See

Appendix C). The content of the items was reviewed, and items critical to the model

were retained. ER62 (3.844) (I make errors in my job from time to time.) and ER64

(4.553) (I have made errors that have been detected by other pilots.) had acceptable,

though noticeably high kurtosis values. The content of both questions led to one

common answer; therefore, kurtosis was to be expected, and the items were retained.

The remaining non-normal items were retained until the CFA was conducted and the

model fit analyzed. If an item was determined to have a combination of loading below .5

(< .5) (Hair et al., 2006) and high skewness or kurtosis, it was temporarily removed from

the model. If the model fit improved after the item was removed, and it was determined

that the content of the item was not critical to the model, it was removed permanently

from the model.

Confirmatory Factor Analysis

In Figure 7, the proposed CFA factor structure is shown with OC, OP, FS, IS, PC,

EL, AT, and ER. The proposed model consists of the original 55 items of the CASS

(Gibbons et al., 2006). The CASS was hypothesized to have a four-factor structure (OC,

OP, FS, IS) with a second order factor for SC. The 9 items of Porter et al.’s PC scale

(1974), 10 items from the Brown et al. (2005) ELS, and Fogarty’s (2004) Maintenance

Environment Survey comprised the items in both AT and ER.

 67

Figure 7. Proposed CFA Model.

The proposed model had model fit indices of CMIN/df = 2.019, NFI = .675, GFI

= .626, AGFI = .605, CFI = .803, and RMSEA = .058 as displayed in Revision 1 of Table

12. The CMIN/df and RMSEA were considered acceptable as shown in the fit criteria in

Table 10; however, the GFI of .626 was less than the .90 targeted fit criteria, AGFI of

.605 was less than .90 targeted fit criteria, and CFI of .803 was less than .95 targeted fit

criteria (Hair et al. 2006). The Modification Indices (MIs) were checked for values over

20. For each of the MI values over 20 that loaded on the same factor, a covariance was

established. There were 20 iterations conducted, and the model fit improved, though the

model fit remained unacceptable. The model fit indices were CMIN/df = 1.777, NFI =
 68

.715, GFI = .669, AGFI = .649, CFI = .85, and RMSEA = .051. The model was then

tested with the outliers removed from the data. After outliers were removed, the model

fit further deteriorated with CMIN/df = 1.704, NFI = .701, GFI = .637, AGFI = .615, CFI

= .849, and RMSEA = .053. The outliers were returned to the data and remained in the

model.

Table 10

Fit Criteria

Model Fit Fit Criteria Reference Acceptable

CMIN/df 1.399 below 3.00 (Byrne, 2010; Hair et al., 2006) Yes
NFI 0.939 close to 0.95 (Byrne, 2010; Hu & Bentler, 1999) Yes
GFI 0.905 close to 1.00 (Byrne, 2010; Jöreskog & Sörbom, 1993) Yes
AGFI 0.879 close to 1.00 (Byrne, 2010; Jöreskog & Sörbom, 1993) Yes
CFI 0.982 close to 0.95 (Byrne, 2010; Hu & Bentler, 1999) Yes
RMSEA 0.036 less than 0.60 (Byrne, 2010; Hu & Bentler, 1999) Yes

The items with low factor loadings (< .4) were removed from the model (Byrne,

2010). There were 14 additional model revisions conducted to improve the model fit.

The model fit improved, though the model fit remained unacceptable with values of

CMIN/df = 1.778, NFI = .77, GFI = .705, AGFI = .683, CFI = .884, and RMSEA = .051.

The model fit for the proposed model was determined to be unacceptable based on the

target model fit indices in Table 10. It was determined that an exploratory factor analysis

(EFA) should be conducted based on the poor model fit.

 69

Exploratory Factor Analysis

The measurement model was analyzed with the survey data collected, and the

model fit remained unacceptable due to a poor model fit indices. An EFA was initiated

on the full dataset. Before the EFA was conducted, the data was confirmed to meet the

assumptions for an EFA. A review of the Kaiser-Meyer-Olkin Measure of Sampling

Adequacy (KMO) showed that it was strong at .953. The Bartlett’s Test of Sphericity

was significant (p < .000). The Measure of Sampling Adequacy (MSA) was analyzed.

After the removal of one item (A93); the MSA was determined to be satisfactory because

a review of the Anti-Image Matrix showed all items were above .5 (>.5). The KMO also

improved to .965 after the removal of item A93.

Based on Hair et al. (2010), a Principal Components Analysis (PCA) with

Varimax rotation was conducted on all items. The initial result showed the items loading

on 16 different factors with eigenvalues greater than 1.0 which explained 68.3% of the

variance in the model.

Based on the proposed model developed from the research conducted by

Alsowayigh (2014) and Brown et al. (2005), the PCA was run again with a factor

constraint of seven. The scree plot in Figure 8 shows the results of the CFA with the

constraint of seven factors. The eigenvalues, located in Appendix D1, shows the seven

factor model explained 67.959% of the variance in the model. The first component was

named Safety Culture New (SCN), and it consisted of 24 items from the original Safety

Culture (SC) second order factor. The second component was named Ethical Leadership

Pilot Commitment (ELPC) due to 13 of the 18 items coming from the previous factors of

Ethical Leadership and Pilot Commitment to the AMC (PC). The remaining five items
 70

were from SC. The third factor was labeled Pilot Commitment New (PCN) with four low

loading items exclusively from the previous PC factor. The fourth component consisted

of three low loading items from SC and PC. The fifth component was labeled Reporting

(REP) and consisted of two items from the original SC factor. The sixth component was

labeled Safety Performance 1 (SP1), which consisted of five items from the original

Attitude To Violations (AT). The seventh component was labeled Safety Performance 2

(SP2) and consisted of five items from ER and AT.

Figure 8. EFA Final Scree Plot.

After reviewing the loadings below .7 (< .7) alongside item content, further model

revisions were made. The third factor (PCN) was removed because the average loading
 71

was (below .7) .573, with 25% of the items cross-loading to ELPC. Factor 4 was

removed due to low average loading of .566. Factor 5 (REP) was also removed due to

poor average factor loading of .573. Additionally, several items were removed with low

loading (below .6) or cross-loading concerns. Cross-loading concerns arise when one

item has similar loading values on multiple components; this may cause model fit and

discriminant validity issues. Items with cross-loading issues were reviewed and removed

from the model.

The original PC factor was eliminated from the model due to poor factor loading

and cross-loading concerns. The elimination of PC reduced the hypotheses in the study

from six to three. The remaining factors shown in Table 11 were SCN (20 items), ELPC

(11 items), ATN (2 items), and ERN (2 items). The model could still test hypotheses H2,

H3, and H5.

Table 11

Model Factors for Hypothesis Testing

Proposed Model Model 2 (M2)

Second
Order Post EFA Final
First Order Factors Factors Factors
OC
OP
SC SCN SCN
FS
IS
EL
ELPC ELN
PC
AT
SP SP1 NFP
ER
 72

Confirmatory Factor Analysis Model 2 (M2)

Model 2 (M2) was analyzed with the survey data collected and the model fit

improved from the model fit in the CFA conducted prior to the EFA; though the model fit

shown in Table 12 was still not acceptable with CMIN/df = 2.237, NFI = .865, GFI =

.793, AGFI = .766, CFI = .92, and RMSEA = .064. The M2 was checked for normality,

and five items were slightly skewed with skewness values above 1.0. There was one item

(OP31) with a skewness of 1.3 that was removed from the model after review of the

content. Two items (ER62, ER64) had elevated kurtosis values (> 7.0). After a review of

the content, it was determined the format of both items led to a justifiable common

answer; therefore, the items remained unchanged in the model. A review of the

Mahalanobis D² values indicated there were 57 cases where the respondents’ answers

were outliers and were significant (p < .05). The model was checked with the outliers

removed and the model fit eroded; therefore, the outliers remained in the model

permanently.

The M2 went through four additional iterations to improve the model fit with

CMIN/df = 1.93, NFI = .885, GFI = .828, AGFI = .804, CFI = .941, and RMSEA = .055.

The model fit remained unacceptable. The proposed factor structure in the literature was

reviewed, and based on grounded theory, the ELPC factor was reduced to more closely

match the original EL factor. The items loading from the former factors of SC and PC

(PC75, IS48, IS49) were deleted from the ELPC construct. ELPC was renamed ELN and

maintained 80% of the items from the EL construct. After the deletion of these three

items in ELPC, the model fit continued to improve with CMIN/df = 2.026, NFI = .891,

GFI = .837, AGFI = .812, CFI = .941, and RMSEA = .055.

 73

Three additional items with standardized estimates below .65 were removed from

the model, and the overall fit improved with CMIN/df = 2.059, NFI = .903, GFI = .848,

AGFI = .766, CFI = .947, and RMSEA = .059. The CMIN/df increased slightly from

2.026 to 2.059, and the RMSEA increased from .055 to .059, though both values were

still considered good after the items were removed.

Table 12

CFA Model Fit History

Revision CMIN/df NFI GFI AGFI CFI RMSEA

1 2.019 0.675 0.626 0.605 0.803 0.058
2 1.777 0.715 0.669 0.649 0.850 0.051
3 1.778 0.770 0.705 0.683 0.884 0.051
4 (M2) 2.237 0.865 0.793 0.766 0.920 0.064
5 1.930 0.885 0.828 0.804 0.941 0.055
6 2.026 0.891 0.837 0.812 0.941 0.058
7 2.059 0.903 0.848 0.822 0.947 0.059
8 1.390 0.940 0.906 0.880 0.982 0.036
9 1.399 0.939 0.905 0.879 0.982 0.036

The MIs were analyzed further and adjustments were made to co-vary appropriate

error terms that exceeded 4.0. The standardized regressions were analyzed for each of

the subsequent 29 model revisions to improve the model fit. The final model fit values

were CMIN/df = 1.39, NFI = .94, GFI = .906, AGFI = .88, CFI = .982, and RMSEA =

.035. According to Byrne (2010), each of the model fit values were acceptable. The

AGFI = .88 remained marginal, though concerns with the AGFI under-reporting in

complex models similar to the model in this current study allowed for the AGFI to be

deemed acceptable.
 74

In Figure 9, the final factor structure is shown with SCN, ELN, and NFP, which is

a second order factor comprised of ERN and ATN. The final M2 model consists of one

first order factor for SCN, which is made up of 17 of the original 55 items of the Gibbons

et al. (2006) CASS. The CASS was hypothesized to have a four-factor structure with a

second order factor for SC. ELN is made up of 80% of the items from the Brown et al.

(2005) ELS. The PC factor was completely removed. The NFP second order factor

consists of the remaining four items from the original 12 items in Fogarty’s (2004)

Maintenance Environment Survey. A Heywood case (Hair et al., 2006) was discovered

in the CFA model. The regression weights for the ERN and ATN were equalized (Hair et

al., 2006) to allow for the model to run properly.

Figure 9. Final CFA Model.

 75

Construct Reliability

Each factor was analyzed for construct reliability (CR) using the formula in

Figure 6. The CR values for the factors in the model were SCN = .905, ELN = .945,

ATN = .919, and ERN = .795. Due to reverse worded items, SCN values were converted

to absolute numbers prior to calculating the CR value. The factors in this model all have

achieved acceptable construct reliability with values greater than .7 (> .7) (Hair et al.,

2010). The Cronbach’s alpha (1951) for the factors were SCN = .911, ELN = .950, ATN

= .903, and ERN = .788.

Convergent Validity

Convergent Validity was calculated using the Average Variance Extract (AVE)

by taking the standardized factor loading squared for each item in each factor and then

calculating the average. The AVE values for the factors in the model were SCN = .599,

EL = .710, ATN = .823, and ERN = .650. According to Hair et al. (2010), any factors

with an AVE greater than .5 are considered to have convergent validity; therefore, all the

factors in the final model had convergent validity.

Discriminant Validity

Discriminant Validity was assessed using two methodologies. The first, shown in

Table 13, was assessed by comparing the squared factor correlations with the AVE for

each factor. The AVE for SCN = .599, and the squared correlations between SCN and

EL = .677, SCN and ERN = .024, and SCN and ATN = .063. The AVE for ELN = .710,

and the squared correlations between SCN and ELN = .677, ELN and ERN = .012, and
 76

ELN and ATN = .079. The AVE for ERN = .650 and the squared correlations between

ERN and SCN = .024, ERN and ELN = .012, and ERN and ATN = .011. The AVE for

ATN = .823 and the squared correlations between ATN and SCN = .063, ATN and ELN

= .079, and ATN and ERN = .011. According to Hair et al. (2010), discriminant validity

within the model was confirmed between all factors except between SCN and ELN. A

subsequent methodology was employed to confirm discriminant validity between SCN

and ELN. According to Kline (2005), correlations below < .85 are considered to have

discriminant validity. The correlation between SCN and ELN was below .85 at .824;

therefore, the model has discriminant validity (Kline, 2005).

Table 13

Discriminant Validity Test

Factor AVE. Squared Correlations Confirmed

0.677 (SCN:ELN) N*
SCN 0.599 0.024 (SCN:ERN) Y
0.063 (SCN:ATN) Y
0.677 (ELN:SCN) Y
ELN 0.710 0.120 (ELN:ERN) Y
0.079 (ELN:ATN) Y
0.024 (ERN:SCN) Y
ERN 0.650 0.120 (ERN:ELN) Y
0.011 (ERN:ATN) Y
0.063 (ATN:SCN) Y
ATN 0.823 0.079 (ATN:ELN) Y
0.011 (ATN:ERN) Y
*Discriminant validity confirmed with alternate
methodology
 77

Structural Equation Model

The SEM displayed in Figure 10 shows the proposed relationships of SCN on

ELN, SCN on NFP, and ELN on NFP. Due to the removal of the PC factor, three other

hypotheses were no longer testable in the study and were removed from the SEM.

Figure 10. Final SEM Model.

Table 14 shows the model fit values for the SEM were acceptable with CMIN/df

= 1.387, NFI = .94, GFI = .906, AGFI = .881, CFI = .982, and RMSEA = .036. (Hair et

al., 2010). These model fit values are similar to the final CFA and as mentioned

previously, determined to be acceptable.

 78

Table 14

Final SEM Model Fit

Revision CMIN/df NFI GFI AGFI CFI RMSEA

SEM 1.399 0.939 0.905 0.879 0.982 0.036

The results of the EFA reduced the number of factors in the proposed model from

eight first order factors to four. The proposed model consisted of four first order factors

(FS, IS, OP, OC) loading onto one second order factor SC. After the EFA, SC was

reduced to one first order factor renamed SCN. SCN is one first order factor made up of

17 of the original 55 items from SC. Of the seventeen items, eight items were from OC,

five items were from OP, four items were from FS, and zero items remained from IS.

Two of the items from IS loaded onto the ELPC factor; however, after review of the

extant research, the two IS items were removed from the factor ELPC. ELPC was re-

named ELN after the removal of two IS (IS48, IS49) items and removal of one PC item

(PC75).

PC was eliminated from the model due to low to moderate loading and cross-

loading on many different factors. The factor was determined to no longer be testable;

therefore, it was eliminated. This elimination of PC from the model precluded the testing

of Hypotheses H1, H4, and H6 in the SEM model.

The 33% in ER and 78% in AT factors led to the renaming of the SP second order

factor to NFP (Not Follow Procedures) based on the content of the items remaining. EL

was reduced by 20% and was renamed ELN in the final model.

In Figure 10, the final factor structure is shown with SCN, ELN, and NFP, which

is a second order factor comprised of ERN and ATN. The final SEM model tests the
 79

direct relationship between SCN on ELN (𝐻2 ), SCN on NFP (𝐻3 ), and ELN on NFP

(𝐻5 ).

Hypothesis Testing

Hypothesis 1

𝐻1 A positive Safety culture has a positive influence on pilot commitment to the

organization.

This hypothesis can no longer be tested due to the elimination of the PC factor

during the EFA.

Hypothesis 2

𝐻2 : A positive safety culture (SCN) has a positive influence on ethical leadership (ELN).

As shown in Table 15, this hypothesis is supported.

Table 15

SEM Hypothesis Testing

VAR DIR VAR Std Est S.E. C.R P Supported

H1 n/a
H2 ELN <--- SCN 0.824 0.036 11.565 *** Yes
H3 NFP <--- SCN -0.330 14.910 -1.442 0.149 No
H4 n/a
H5 NFP <--- ELN -0.317 30.471 -1.327 0.184 No
H6 n/a
 80

The results of the SEM analysis confirmed the relationship between SCN and

ELN was both strong (Estimate = .824) and significant (p < .001). This study supports

that there is a significant relationship and positive relationship between SCN and ELN.

Hypothesis 3

𝐻3 : A positive safety culture (SCN) has a negative influence on safety performance

(NFP).

As shown in Table 15, this hypothesis is not supported. The results of the SEM

analysis determined SCN does not have a negative influence on NFP, and that

relationship is not significant. The relationship between SCN and NFP did not

materialize as hypothesized; the relationship between SCN and NFP had a significance

level of .149.

Hypothesis 4

𝐻4 : A positive pilot commitment to the organization (PC) has a positive influence on

safety performance (NFP).

This hypothesis could no longer be tested due to the elimination of the PC factor

during the EFA.

Hypothesis 5

𝐻5 : A positive ethical leadership (ELN) has a negative influence on safety performance

(NFP).
 81

As shown in Table 15, this hypothesis is not supported. The results of the SEM

analysis confirmed ELN had a non-significant (p = -.184) and negative relationship to

NFP. This result was unexpected based on a review of the literature.

Hypothesis 6

𝐻6 : A positive ethical leadership (ELN) has a positive influence on pilot commitment to

the organization (PC).

This hypothesis can no longer be tested due to the elimination of the PC factor

during the EFA.

 82

CHAPTER V

DISCUSSION, CONCLUSIONS, AND RECOMMENDATIONS

This study analyzed the relationship between safety culture (SC), ethical

leadership (EL), pilot commitment to the AMC (PC), and safety performance (SP) for

U.S. based Fractional jet pilots. The proposed factor model structure derived from the

literature could not attain an adequate model fit during the initial CFA; therefore, an EFA

was conducted. After the EFA, a second CFA was conducted on M2 followed by the

development and testing of a SEM. The SEM developed allowed for hypothesis testing

based on the new factor structure.

The objective of this chapter is to discuss the results of the study and how these

results compare with the findings in the available literature. Additionally, this chapter

will interpret these results, discuss how these results may impact general aviation in the

future, and discuss recommendations for future research.

Discussion

Hypotheses. There were six hypotheses planned for this research study. After

the EFA, three (H1, H4, H6) of the six hypotheses could no longer be tested due to the

removal of the PC factor.

(𝐻1 ) A positive safety culture has a positive influence on pilot commitment to the

organization. This hypothesis (𝐻1 ) could not be tested because of the low and cross

loading of the PC items as a stand-alone factor.

(𝐻2 ) A positive safety culture (SCN) has a positive influence on ethical

leadership (ELN). This hypothesis was tested and supported. The results showed H2 had
 83

both a significant (p =.001) and strong (estimate = .824) relationship. These results

confirm Schein’s (2004) assertion that corporate culture is intertwined with

organizational leadership. The high correlation and the inability to confirm one of the

two discriminant validity tests performed between the SCN and ELN constructs suggest a

deep relationship between ELN and SCN. One of the important revelations in this study

is that in Fractional pilots there exists a strong correlation between ELN and SCN. There

is a need for discrimination between these two constructs to better understand how to

measure, monitor, and improve them respectively, if needed. Many studies have

concluded that both EL (Freiwald, 2013) and SC (Alsowayigh, 2014) influence the safety

of an organization, though the current study did not confirm those conclusions.

As noted above, the current study results do not match Freiwald’s (2013) findings

that ethical leadership (EL) did not have a significant relationship with proactive safety

climate. Freiwald’s (2013) reasoning suggested that EL is merely a subset of the larger

construct of leadership, and Freiwald stated that the narrowness of the EL construct might

explain the lack of a relationship in the 2013 study (Freiwald, 2013). Additionally, the

Freiwald study included EL as the exogenous variable and safety climate as the

endogenous variable, whereas the present study reverses the direction of that relationship.

(𝐻3 ) A positive safety culture (SCN) has a negative influence on safety

performance (NFP). The SEM analysis showed that H3 is not supported, and SCN does

not have a significant influence on NFP. This result was unexpected due to the support in

previous studies (Alsowayigh, 2014; Fogarty, 2004) showing a significant relationship

between safety culture or safety climate and self-reported safety performance. Due to the

infrequency of aviation accidents or incidents potentially leading to invalid conclusions

 84

(O’Connor et al., 2011), the current study relied on self-reported safety behavior as did

Alsowayigh (2014) and Fogarty (2004). In contrast, research by Zohar (2000) relied on

quantitative outcome variables, such as employee micro-accidents. This micro-accident

research also concluded there was a significant relationship between safety climate and

safety performance. Zohar hypothesized that micro-accidents were a leading indicator to

a decline in safety climate that could lead to larger accidents. General aviation needs to

develop a methodology that includes identifying and monitoring quantifiable data that is

considered a leading indicator of a decline in safety to augment self-reported data.

Future research should continue to test the relationship between SCN and NFP

because the results are likely to be more consistent with past research from Alsowayigh

(2014), Fogarty (2004), and Zohar (2000). Freiwald (2013) suggested that the

narrowness of the EL construct in the 2013 study was a potential cause for the

unexpected lack of support for the relationship between EL and employee injuries. In the

current study, the major reduction in the SP items from 13 original items to 4 items could

have also narrowed the NFP construct in a similar manner, thereby altering the

significance of the relationship.

(𝐻4 ) A positive pilot commitment to the organization (PC) has a positive

influence on safety performance (NFP). Alsowayigh (2014) found that PC did not

mediate the relationship between ER and AT. Alsowayigh (2014) also determined that

PC did not influence a professional pilot’s behavior in the cockpit. The inability of the

PC items to maintain integrity as a factor combined with the results of previous research

suggests that PC is not essential for future research attempting to predict pilot safety

behavior.
 85

(𝐻5 ) A positive ethical leadership (ELN) has a negative influence on safety

performance (NFP). The SEM results did not support that positive ELN reduces the

likelihood of pilots not following procedures (NFP). In 1998, Craig and Gustafson

(1998) warned managers that ethical leadership should be a priority. The study by Kapp

and Parboteeah (2008) concluded that ethical climate had a strong influence over safety

behavior. Freiwald (2013) concluded that ethical leadership led to fewer occupational

accidents. The present study did not match these other studies and did not support the

construct that ethical leadership plays a significant role in safety behavior and outcomes.

There is ample evidence in the literature suggesting that future studies continue to test the

relationship between ELN and safety behaviors. The positioning of ELN as the

exogenous variable in future studies is likely to influence the level of significance

between ELN and safety behaviors.

(𝐻6 ) A positive Ethical leadership (ELN) has a positive influence on pilot

commitment to the organization (PC). This hypothesis (𝐻6 ) could not be tested because

of the low, cross, and sporadic loading of the PC items during the EFA.

Conclusions

This study analyzed the relationship between safety culture (SCN), ethical

leadership (ELN), and safety performance (NFP). Schein (2004) stated that corporate

culture was the personality of the organization and that corporate culture was strongly

connected with leadership and employee behavior (Schein, 2004). James Reason (1997)

wrote that when employees of an organization hold similar beliefs, those beliefs will
 86

govern behavior. In 1979, Butler warned that leaders who distanced themselves from

tasks may contribute to accidents.

The present study tested the nature of this relationship between safety culture and

ethical leadership. It was concluded that SCN and ELN had a strong and significant

relationship. In addition to this strong and significant relationship, these two factors were

also highly correlated. The constructs of SCN and ELN also had discriminant validity

concerns based on one conservative test of discriminant validity (Hair et al., 2010). The

cross-loading of many of the items between the SC and EL factors also suggested a

strong relationship between the constructs.

In the perceptions of the Fractional pilots, the constructs of SC and EL are closely

related. Stolzer et al. (2015) confirmed this by suggesting the need for safety mandates to

have the complete support of the company leadership. Though these findings re-confirm

the conclusions by other studies and subject matter experts, there exists a new concern

about the ability to discriminate between the two constructs in future research. If SC and

EL are so closely perceived by Fractional pilots, the construct of SC may be too wide and

the CASS too broad in scope. The CASS did not retain the expected factor structure and

lost 69% of the original items during the study of Fractional pilots. In contrast to the

CASS, the ELS (Brown et al., 2005) was concise, and 80% the items remained together

throughout the EFA and multiple CFA processes.

The unexpected result from this study was the non-significant relationship

between SCN and NFP. Research from Alsowayigh (2014), Fogarty (2004), and Zohar

(2000) supported that safety culture or safety climate has a significant effect on safety

performance. The number of items in the second order factor SP in the proposed model
 87

was reduced from 13 items to 4 (NFP) in the final model. It is plausible that this

narrowing of the items may have affected this relationship. Future research is

recommended, as it is likely to re-confirm the research from Alsowayigh (2014), Fogarty

(2004), and Zohar (2000) that safety culture or safety climate influences safety

performance or safety behavior.

The positioning of the ELN factor as the exogenous variable in the recommended

future model shown in Figure 11 is likely to influence the significance of these

relationships. The shifting of the ELN scale to the exogenous position is also consistent

with the SEM model presented in the Freiwald (2013) research.

Contributions to the Literature

This study contributed to the literature by re-confirming several previous studies

and opening the discussion to re-examine the validity and reliability of four survey

instruments in the literature.

This research supports the O’Connor et al. study (2011) which concluded that, in

aviation, there are too many different instruments attempting to measure similar

constructs, and called for future studies to begin confirming the reliability and

discriminant validity of the existing instruments rather than testing new instruments. The

O’Connor et al. (2011) study stated that studies are needed that re-confirm both the

predictive ability of the instruments and their discriminant validity from other constructs.

In the current study with Fractional pilot data, the factor structure of most of the

instruments used did not maintain their proposed factor structure during the EFA. This

lack of factor structure integrity causes a concern that these instruments will not maintain
 88

their integrity when tested on various aviation groups in future research. As suggested by

O’Connor et al. (2011), confirming predictive capability from unreliable instruments will

not be possible. Additionally, if the constructs cannot maintain their discriminant validity

from other constructs when measured together, the results will be difficult to interpret,

easily challenged, and have little practical benefit.

The CASS (Gibbons et al., 2006) was a very broad instrument and the proposed

factor structure did not hold up to the Fractional pilot survey data. The CASS had four

first order factors with one second order factor for SC. The post EFA structure was

reduced to one first order factor (SCN). It may be argued the CASS was originally

designed for commercial airline pilots; therefore, the questions were developed for a

different pilot group. During this research, there were only minor adaptations needed for

the CASS to be applicable to Fractional pilots. The survey was tested with multiple

experienced pilots before deployment. Fractional companies and airlines in the U.S. both

operate very large fleets and face many of the same challenges. Both pilot groups are

mostly unionized; therefore, the CASS should be adaptable to the Fractional pilot group.

The CASS, in the form used for this study, was arguably overly complex and too

large in scope for this research. The items in the CASS overlapped with other

instruments in the study; however, the main concern was the factor structure was not

maintained with the data from the Fractional pilots. The result of the first EFA showed

16 components with eigenvalues over 1.0 that explained 68% of the variance in the

model. The subsequent EFA was constrained to seven factors that explained 67.959% of

the variation in that model. The final three components from the EFA model constrained
 89

to seven components, made up just 7% of the remaining variance; therefore, those items

would have added minimal value to the study had they been retained.

Of the original 55 items in the CASS, only 17 items were retained in the final

model due to low, cross, and sporadic loading. This major reduction in the CASS items

due to cross-loading combined with the high correlation with the ELN construct suggests

the CASS is a comprehensive survey instrument and is likely broader in scope than the

construct of safety culture. In Appendix E, the 17 remaining CASS items are presented

for consideration for the measurement of SCN for future research on pilot groups similar

to Fractional pilots. The aviation industry needs to agree on a standard set of instruments

that measures the intended construct and maintains both reliability and discriminant

validity. This set of instruments must also possess the ability to predict declines in safety

behavior or the instruments will be of minimal value.

The prediction of safety performance should be forecasted from a combination of

qualitative and quantitative data. Survey data may reveal the perception of a decline in

safety culture which could be the antecedent to a decline in safety performance. The

weakness in qualitative data is that self-reported survey data have the potential to be

biased by the respondent. Conversely, accurately compiled quantitative data can provide

unbiased data that can forecast a decline in safety performance. The weakness in

quantitative data can be the inability to accurately measure or interpret the data. The

weaknesses in both qualitative and quantitative measurements should compel safety

practitioners to rely on a combination of both qualitative and quantitative data to forecast

declines in safety performance.

 90

Study Limitations

The data collected in the study was collected through the voluntary participation

of Fractional jet pilots in the U.S. The responses by the participants were based on their

perception of ELN, SCN, and NFP. The perceptions of the Fractional pilots may have

been affected by the challenges between the unions and management during the data

collection process. NJASAP completed their negotiation of a new collective bargaining

agreement (CBA) after years of negotiations in December 2015. Flight Options pilots

had been unionized for many years while Flexjet pilots were non-union. After the merger

of Flight Options and Flexjet, there was a vote to continue a company-wide union or

disband the union. The union passed by a narrow margin. The total affirmative votes

were less than the number of existing Flight Options union members; therefore, many

union members did not vote for the union. The results were so close they were

challenged by Flight Options / Flexjet management.

Each of the aforementioned issues had the potential to influence the responses

provided by the Fractional pilots. Additionally, these situations could have influenced

which pilots were motivated to participate in the survey. Nearly all of the Fractional

pilots in this study were protected by their respective unions; therefore, they would have

been able to answer the questions in this study without fear of repercussions.

One limitation included the inability to confirm the discriminant validity between

ELN and SCN in one of two tests of discriminant validity conducted. According to the

more conservative method from Hair et al. (2010), the AVE for each factor should be

higher than the squared correlation between factors. The AVE of SCN was .599;

however, the squared correlation between SCN and ELN was .677. In an alternative
 91

method for confirming discriminant validity, the correlation coefficient between SCN and

ELN did pass the standard set by Kline (2005) of <.85 with a correlation of .824. Based

on the extensive existing literature demonstrating the factors as distinct and achieving

Kline’s (2010) <.85, both SCN and ELN were retained. The relatively high correlation

and inability to confirm discriminant validity by one methodology may have been due to

the broad scope of questions in the CASS and the question content being similar between

these factors. Several of the original CASS items loaded better on the ELPC variable

than the SCN during the CFA.

In the final revisions of the CFA and the SEM, there was a negative variance

discovered in the model. This issue was determined to be a Heywood case and may have

been caused by the M2 not meeting the suggested minimum of three items loading on

ATN and three items loading on ERN (Hair et al., 2006). The solution suggested by Hair

et al. (2006) was to equalize the regression weights in the model for the ATN and ERN

items. The ATN items were both set to 1.0 and the ERN items were both set to .005, and

the issue was resolved. The model fit worsened from revision 8 to revision 9 by a

minimal amount as shown in Table 12.

Practical Implications

The practical implication of this research may be far reaching for general aviation

and for AMCs. New and inexpensive survey programs can be implemented and

monitored that could improve the understanding of the relationship between the AMC

and their pilots. Additionally, these monitoring programs may prove to have the ability

to predict a decline in safety behavior.

 92

The conclusion that SCN predicts ELN should encourage AMCs to monitor these

factors within their organizations. The implementation of a survey-based measurement

program is inexpensive and easy to both implement and interpret. A survey-based

measurement program may also be considered part of the requirement for their AMC’s

SMS to continually improve safety (Stolzer & Goglia., 2015). The AMC would be able

to identify and react to any declines in the SCN and or ELN. This identification and

reaction has the potential to improve the organization’s culture and relationship with their

pilots. A positive safety culture and a positive perception of leadership have been

demonstrated in other studies to reduce accidents and improve safety behavior.

The other important implication of this research is that AMC owners and

organizational leaders may realize their leadership is an important aspect for both the

financial success and the safety of their organization. Brown et al. (2005) stated that if

leaders are attractive, credible, and legitimate, they will govern employee’s behavior.

Schein (2004) stated that a strong positive culture leads to better financial performance.

This research study concluded that SCN and ELN are highly correlated and, therefore,

both are of critical importance to the success of the organization. The leaders of AMCs

must be ethical and strong leaders who create a just and blame free organization that

encourages open communication. AMC leaders must be committed to safety initiatives

to realize any long-lasting effects of their efforts (Helmreich et al., 1997). Strong and

ethical AMC leaders may enjoy a financially sound and safe operation.
 93

Future Research

O’Connor et al. (2011) called for the repeated use of common survey instruments

that could withstand rigorous discriminant validity and predict reliable results. This

study re-confirmed the need for survey instruments that can be applied across different

groups and maintain both construct integrity and discriminant validity. In aviation, there

needs to be a reliable instrument or small set of instruments that are open for use across

diverse groups. This common group of survey instruments needs to have the ability to

detect a decline in safety behavior or their antecedents early enough to implement

solutions before these declines become safety issues.

The IS, PC, and AT items used from the literature did not load strongly on their

hypothesized factors and, therefore, may not be reliable instruments for future research

with Fractional pilots or similar groups, or the questions would need to be revised.

Future instruments need to be concise and measure the intended construct efficiently.

The IS, PC, and AT factors may not provide enough benefit for future studies on similar

pilot groups.

Future research may include the following alternative SEM model based on the

existing literature from Brown et al. (2005) and Freiwald (2013). The Brown, Treviño, &

Harrison (2005) and Freiwald (2013) studies suggested ethical leadership has an

influence on safety behavior and outcomes. These studies suggest that future research

may be conducted with ethical leadership or the wider construct of leadership as the

exogenous or predictor variable in a causal model with safety culture and safety

performance as the endogenous variables. The following model for future SEM research
 94

has the potential for strong and significant relationships of both hypotheses (see Figure

11).

Figure 11. Proposed Future SEM Model.

Conducting the revised study on similar pilot groups with varying historical safety

records may yield actionable group differences. The Fractional companies have achieved

a superior safety record when compared with Charter operators; therefore, conducting the

same study for random Charter pilots in the U.S. has the potential to both test the revised

model and identify group differences. If significant, these group differences may lead to

strategies to improve general aviation safety.

 95

Future studies should include a reliable and quantifiable data source to augment

the self-reporting data. Zohar’s study (2000) used quantifiable data as the endogenous

variable from which to draw conclusions. Zohar has advocated the use of quantifiable

data such as micro-accidents as the endogenous variable in a safety climate research. In

general aviation, the accident and incident rates are so low that drawing valid conclusions

about antecedents to accidents and incidents may not be valid (O’Connor, 2011). In an

unpublished study using quantifiable data in commercial aviation, Cistone et al. (2011)

encountered issues with the reliability and validity of the accelerometer measurements for

hard landings at one Middle Eastern airline. The accelerometers had both measurement

errors and instrument calibration issues across the fleet that made drawing conclusions

from the data difficult.

Self-reported data will remain an important part of aviation safety due to

infrequency of accidents and or incidents; however, augmenting survey data with reliable

and quantifiable data would be recommended to create a more comprehensive

methodology to predict declines in aviation safety. In 2000, Zohar used micro-accidents

to illuminate declines in safety before more serious accidents could occur. The Quick

Access Recorder (QAR) installed in many aircraft, records operational data, such as pilot

inputs. This QAR data can be analyzed and used as an indication that safety is declining.

For example, in May 2014, a G-IV crashed while departing Bedford, MA (KBED). In its

report, the NTSB reviewed the QAR data and determined the crew had not performed a

proper check of the flight controls on 89.8% of the previous 176 flights (NTSB AAR-

15/03, 2015). If the QAR data had been monitored, it would have demonstrated this
 96

crew’s disregard for standard pre-flight checks, and corrective actions could have been

implemented that would have likely prevented this accident.

Finally, the instruments used in aviation need to be more reliable, freely available

for use in other studies, and must maintain discriminant validity when used with other

instruments. These instruments need to be concise and measure the intended construct.

Without the open and repeated use of a distinct and reliable instrument or a small set of

instruments, aviation is unlikely to realize the potential benefits of forecasting a decline

in safety behavior. Reliable forecasting of declines in safety behavior has the potential to

prevent catastrophic aviation accidents.

 97

REFERENCES

Alsowayigh, Mohammad (2014). Assessing safety culture among pilots in Saudi Airlines:

A quantitative study approach. (Doctoral dissertation). Retrieved from

Electronic Theses and Dissertations. Paper 4594.

http://stars.library.ucf.edu/etd/4594

Boomsma, A., & Hoogland J.J. (2001). The robustness of LISREL modeling revisited. In

R. Cudek, S. DeToit, & D. Sörbom (Eds.), Structural equation modeling: Present

and future. (pp. 111-129). Lincolnwood, IL: Scientific Software International.

Breiling, R. E. (2014). Annual business turbine aircraft accident review. (Annual

Report). Boca Raton, FL: Robert E. Breiling Associates, Inc.

Brown, M. E., Treviño, L. K., & Harrison, D. A. (2005). Ethical leadership: A social

learning perspective for construct development and testing. Organizational

Behavior and Human Decision Processes, 97(2), 117-134.

doi:10.1016/j.obhdp.2005.03.002

Butler, M. C., & Jones, A. P. (1979). Perceived leader behavior, individual

characteristics, and injury occurrence in hazardous work environments. Journal of

Applied Psychology, 64(3), 299.

Byrne, B. M. (2010). Structural equation modeling with AMOS basic concepts,

applications and programming (2nd ed.). New York, NY: Routledge: Taylor &

Francis Group.

Cai, W. (2005). The impact of safety culture on safety performance: A case study of a

construction company. doi:304986285

 98

Chang, K., & Lu, C. (2009). Effects of safety climate on perceptions of safety

performance in container terminal operations. Transport Reviews, 29(1), 1-19.

Cistone, J., Al-Romaithi, S., & O'Leary, K. (2011). A mixed methods approach to analyze

middle eastern airline flight faults. (Unpublished manuscript). College of

Aviation, Embry-Riddle Aeronautical University, Daytona, Florida.

Clarke, S. (2006). The relationship between safety climate and safety performance: A

meta-analytic review. Journal of Occupational Health Psychology, 11(4), 315-

327. doi:10.1037/1076-8998.11.4.315.

Commeiras, N., & Fournier, C. (2001). Critical evaluation of Porter et al.'s organizational

commitment questionnaire: Implications for researchers. The Journal of Personal

Selling & Sales Management, 21(3), 239-245.

Cooper, M. D., & Phillips, R. A. (2004). Exploratory analysis of the safety climate and

safety behavior relationship. Journal of Safety Research, 35(5), 497-512.

doi:10.1016/j.jsr.2004.08.004

Cox, S., & Flin, R. (1998). Safety culture: Philosopher's stone or man of straw? Work &

Stress, 12(3), 189-201. doi:10.1080/02678379808256861

Craig, S. B., & Gustafson, S. B. (1998). Perceived leader integrity scale: An instrument

for assessing employee perceptions of leader integrity. The Leadership Quarterly,

9(2), 127-145. doi:10.1016/S1048-9843(98)90001-7

Cronbach, L.J. (1951) Coefficient alpha and internal structure of tests. Psychometrika,

16(3), 297-334.

Dillman, D. A., Smyth, J. D., & Christian, L. M. (2009). Internet, mail, and mixed-mode

surveys: The tailored design method. Hoboken, N.J: Wiley & Sons.
 99

Ding, L., Velicer, W. F., & Harlow, L. L. (1995). Effects of estimation methods, number

of indicators per factor, and improper solutions on structural equation modeling fit

indices. Structural Equation Modeling, 2, 119–144.

Erickson, J. A. (1997). The relationship between corporate culture & safety performance.

Professional Safety, 42(5), 29-33.

Field, A. (2009). Discovering statistics using SPSS. (3rd) London, U.K.: SAGE.

Fogarty, G. J. (2004). The role of organizational and individual variables in aircraft

maintenance performance. International Journal of Applied Aviation Studies,

4(1), 73-90.

Freiwald, D. R. (2013). The effects of ethical leadership and organizational safety culture

on safety outcomes (Doctoral dissertation). Retrieved from Paper 55.

http://commons.erau.edu/edt/55

Gibbons, A. M., von Thaden, T. L., & Wiegmann, D. A. (2006). Development and initial

validation of a survey for assessing safety culture within commercial flight

operations. The International Journal of Aviation Psychology, 16(2), 215-238.

doi:10.1207/s15327108ijap1602_6

Goodheart, B. J., & Smith, M. O. (2014). Measurable outcomes of safety culture in

aviation - A meta-analytic review. International Journal of Aviation, Aeronautics,

and Aerospace, 1(4), 1-30.

Grimm, L. G., & Yarnold, P. R. (2000). Reading and understanding more multivariate

statistics (1st ed.). Washington, DC: American Psychological Association.

Hair, J. F., Black, W.C., Babin, B.J., Anderson, R.E., & Tatham, R.L. (2006).

Multivariate data analysis (6th ed.). Upper Saddle River, NJ: Prentice Hall.
 100

Hair, J. F., Black, W.C., Babin, B.J., Anderson, R.E., & Tatham, R.L. (2010).

Multivariate data analysis (7th ed.). Upper Saddle River, NJ: Prentice Hall.

Helmreich, R. L., Foushee, H. C., Benson, R., & Russini, W. (1986). Cockpit resource

management: Exploring the attitude-performance linkage. Aviation, Space, and

Environmental Medicine, 57(12,1), 1198-1200.

Helmreich, R. L., & Merritt, A. C. (1998). Culture at work in aviation and medicine:

National, organizational, and professional influences. Aldershot, England:

Ashgate.

Helmreich, R. L., Sherman, P. J., & Merritt, A. C. (1997). National culture and flight

deck automation: Results of a multination survey. The International Journal of

Aviation Psychology, 7(4), 311-329. doi:10.1207/s15327108ijap0704_4

Hu, L., & Bentler, P. (1999). Cutoff criteria for fit indexes in covariance structure

analysis: Conventional criteria versus new alternatives. Structural Equation

Modeling: A Multidisciplinary Journal, 6(1), 1-55.

doi:10.1080/10705519909540118

Itoh, K., Andersen, H. B., & Seki, M. (2004). Track maintenance train operators' attitudes

to job, organisation and management, and their correlation with accident/incident

rate. Cognition, Technology & Work, 6(2), 63-78.

doi:10.1007/s10111-003-0135-x

Jennings, J. (2008). Human factors: Analysis & classification. Professional Safety, 53(6),

44.

JetNet Inc. (2015, November 7). Retrieved from http://www.jetnet.com

 101

Johnson, S. E. (2007). The predictive validity of safety climate. Journal of Safety

Research, 38(5), 511-521. doi:10.1016/j.jsr.2007.07.001

Joreskog, K. G., & Sörbom, D. (1993). Structural equation modeling with SIMPLIS

command language. Hillsdale, NJ: Lawrence Erlbaum Associates.

Kao, L., Stewart, M., & Lee, K. (2009). Using structural equation modeling to predict

cabin safety outcomes among Taiwanese airlines. Transportation Research Part

E, Logistics & Transportation Review, 45(2), 357.

Kapp, E. A., & Parboteeah, K. P. (2008). Ethical climate & safety performance.

Professional Safety, 53(7), 28-31.

Kline, R. B. (2005). In D. A. Kenny (Ed.), Principals and practice of structural equation

modeling (2nd ed.). New York, NY: The Guilford Press.

Kuhn, T. E. (2009). Exploring the impact of FAA recommended procedures and aviation

industry guidelines on business aviation safety (Doctoral dissertation). Retrieved

from ProQuest Dissertations and Theses database. (UMI No. 305179422)

Li, W., Harris, D., & Yu, C. (2008). Routes to failure: Analysis of 41 civil aviation

accidents from the Republic of China using the human factors analysis and

classification system. Accident Analysis & Prevention, 40(2), 426-434.

doi:10.1016/j.aap.2007.07.011

Lyon, J. S. (2007). The missing link: An examination of safety climate and patient clinical

outcomes in a national sample of hospitals. (Doctoral dissertation). Retrieved

from ProQuest Dissertations and Theses database. (304837988)

 102

McFadden, K. L., & Towell, E. R. (1999). Aviation human factors: A framework for the

new millennium. Journal of Air Transport Management, 5(4), 177-184.

doi:10.1016/S0969-6997(99)00011-3

Mearns, K. J., & Flin, R. (1999). Assessing the state of organizational safety--culture or

climate? Current Psychology, 18(1), 5-17. doi:10.1007/s12144-999-1013-3

Meyer, J. P., & Allen, N. J. (1991). A three-component conceptualization of

organizational commitment. Human Resource Management Review, 1(1), 61-89.

doi:10.1016/1053-4822(91)90011-Z

Mitchell, K., Sholy, B., & Stolzer, A. (2007). General aviation aircraft flight operations

quality assurance: Overcoming the obstacles. Aerospace and Electronic Systems

Magazine, IEEE, 22(6), 9-15. doi:10.1109/MAES.2007.384075

Neal, A., & Griffin, M. A. (2006). A study of the lagged relationships among safety

climate, safety motivation, safety behavior, and accidents at the individual and

group levels. Journal of Applied Psychology, 91(4), 946-953. doi:10.1037/0021-

9010.91.4.946

Neal, A., Griffin, M. A., & Hart, P. M. (2000). The impact of organizational climate on

safety climate and individual behavior. Safety Science, 34(1–3), 99-109.

doi:10.1016/S0925-7535(00)00008-4

National Transportation Safety Board. (1989). Brief of accident: Learjet 35A; N888DJ.

(Brief of Accident No. MIA88LA026). Washington, DC: NTSB.

National Transportation Safety Board. (1997). Brief of accident: Challenger 600S;

N999TF. (Brief of Accident No. ATL96LA073). Washington, DC: NTSB.

 103

National Transportation Safety Board. (2015). Aircraft Accident Report: Gulfstream G-

IV; N121JM. (Aircraft Accident Report No. AAR-15/03). Washington, DC:

NTSB.

O’Connor, P., O’Dea, A., Kennedy, Q., & Buttrey, S. E. (2011). Measuring safety

climate in aviation: A review and recommendations for the future. Safety Science,

49(2), 128-138. doi:10.1016/j.ssci.2010.10.001

O'Toole, M. (2002). The relationship between employees' perceptions of safety and

organizational culture. Journal of Safety Research, 33(2), 231-243.

doi:10.1016/S0022-4375(02)00014-2

Porter, L. W., Steers, R. M., Mowday, R. T., & Boulian, P. V. (1974). Organizational

commitment, job satisfaction and turnover among psychiatric technicians. Journal

of Applied Psychology, 59, 603-609.

Rachels, J. (2002). Elements of moral philosophy. New York, NY: McGraw Hill Higher

Education.

Rasmussen, J. (1982). Human errors: A taxonomy for describing human malfunction in

industrial installations. Journal of Occupational Accidents, 4(2–4), 311-333. doi:

10.1016/0376-6349(82)90041-4

Reason, J. T. (1990). Human error. Cambridge England: Cambridge University Press.

Reason, J. T. (1997). Managing the risks of organizational accidents. Brookfield, VT:

Ashgate.

Schein, E. H. (2004). Organizational culture and leadership. San Francisco: Jossey-Bass.

 104

Shappell, S. A., & Wiegmann, D. A. (1997). A human error approach to accident

investigation: The taxonomy of unsafe operations. The International Journal of

Aviation Psychology, 7(4), 269-291. doi:10.1207/s15327108ijap0704_2

Sheridan, J. E. (1992). Organizational culture and employee retention. The Academy of

Management Journal, 35(5), 1036-1056.

Stolzer, A. J., & Goglia, J. J. (2015). Safety management systems in aviation. Burlington,

VT: Ashgate.

Stolzer, A. J., Halford, C. D., & Goglia, J. J. (2011). Implementing safety management

systems in aviation (Human Factors in Flight Operations Ed.). Burlington, VT:

Ashgate.

Thompson, R. C., Hilton, T. F., & Witt, A. L. (1998). Where the safety rubber meets the

shop floor: A confirmatory model of management influence on workplace safety.

Journal of Safety Research, 29(1), 15.

TRAQPak Business Aircraft Activity Report. (2014). Cincinnati, OH: Aviation Research

Group.

Tye, W. (1980). Civil aviation safety. Aircraft Engineering and Aerospace Technology,

52(10), 2-4. doi:10.1108/eb035668

Vincoli, J. (1990). Pilot factor: A problem in aviation safety. Professional Safety, 35(8),

25.

von Thaden, T. L., Wiegmann, D. A., & Shappell, S. A. (2006). Organizational factors in

commercial aviation accidents. The International Journal of Aviation Psychology,

16(3), 239-261.

Vroom, V. H. (1964). Work and motivation. New York: Wiley.

 105

Walton, R. E. (1985). From control to commitment in the workplace. Harvard Business

Review, 63(2), 77.

Westland, J. C. (2010). Lower bounds on sample size in structural equation modeling.

Electronic Commerce Research and Applications, 9(1/6), 476-487.

doi:10.1016/j.elerap.2010.07.003

Wiegmann, D. A., & Shappell, S. A. (2001). Human error analysis of commercial

aviation accidents: Application of the human factors analysis and classification

system (HFACS). Aviation, Space, and Environmental Medicine, 72(11), 1006-

1016.

Wilson, K. A. (2007). Does safety culture predict clinical outcomes? (Doctoral

dissertation). Retrieved from ProQuest Dissertations and Theses database.

(304746476)

Zohar, D. (1980). Safety climate in industrial organizations: Theoretical and applied

implications. Journal of Applied Psychology, 65(1), 96-102. doi:10.1037/0021-

9010.65.1.96

Zohar, D. (2000). A group-level model of safety climate: Testing the effect of group

climate on micro-accidents in manufacturing jobs. Journal of Applied Psychology,

85(4), 587-596.
 106

APPENDIX A

Permission to Conduct Research

 107

Safety Culture & Performance Survey

Consent for Participation in Survey Research

I am 18 years or older and volunteer to participate in a research study conducted by Kevin O’Leary
(Ph.D. Candidate) from Embry-Riddle Aeronautical University. I understand that the study is
designed to gather information about Safety Culture in Fractional Jet Pilots. I will be one of
approximately 300-700 pilots completing this survey.

1. My participation in this project is voluntary. I understand that I will not be paid for my
participation though a donation to the Corporate Angel Network will be made for each completed
survey.

I may withdraw and discontinue participation at any time without penalty. If I decline to participate
or withdraw from the study, no one will be told.

2. I understand that most respondents will find the survey questions interesting and thought-
provoking. If, however, I feel uncomfortable in any way during the survey, I have the right to end the
survey.

3. Participation involves completing an anonymous 93 question online survey. The survey takes an
average of 13 minutes and can be completed on a most devices with an internet connection
including smart phones (landscape view), tablets or computers.

4. I understand that the researcher will not know my identity and I will not be asked to provide any
identifiable data about myself. My confidentiality as a respondent in this survey will remain secure.
Subsequent uses of records and data will be subject to standard data use policies which protect
the anonymity of individuals and institutions.

5. No organization, institution or company (except the principal researcher) will have access to the
raw responses. This precaution will prevent my individual responses from having any negative
repercussions.

6. I understand that this research study has been reviewed and approved by the Institutional
Review Board (IRB) for the use of Human Subjects in Research at the Embry-Riddle Aeronautical
University. For research problems or questions regarding subjects, the Institutional Review Board
may be contacted through:
 108

David C. Ison, Ph.D. Research Chair

Assistant Professor of Aeronautics College of Aeronautics
Embry-Riddle Aeronautical University, Worldwide
Editor, International Journal of Aviation, Aeronautics, and Aerospace Office
(Cell): (503) 507-5697
email: isond46@erau.edu Skype: david.ison73
Website: http://worldwide.erau.edu

7. If requested, I will be given a copy of this consent form.

8. I have read and understand the explanation provided to me. I have had all
my questions answered to my satisfaction, and I voluntarily agree to
participate in this study. My continuation with this survey will serve as
confirmation of my consent to participate in this study.

Thank you very much for your participation in this important study. Principal

Investigator
Kevin O’Leary Ph.D. Candidate
Embry-Riddle Aeronautical University olearyk1@my.erau.edu
617-600-6868
 109

APPENDIX B

Data Collection Device

 110

Safety Culture & Performance Survey

Survey Introduction

* 1. Are you currently a jet pilot at a one of the following U.S. based fractional Aircraft Management
Companies (AMCs)?
(NetJets, Flight Options, Flexjet or Executive AirShare)

Yes

No

Definition:

Aircraft Management Company (AMC) refers to the organization that operates and manages aircraft while maintaining an operating
certificate such as FAR 135 / Charter or FAR 91K / Fractional.
 111

Safety Culture & Performance Survey

Demographic Information

Demographic Information

* 2. What best describes your position within the Aircraft Management Company (AMC)? (Select one,
please)

Pilot with Office / Management responsibilities

Pilot with other responsibilities (Instructor, Check Airman, etc.)

Pilot (Captain / PIC)

Pilot (First Officer / SIC)

* 3. What category of aircraft based on Maximum Takeoff Weight (MTOW) do you primarily fly?

Light Jet (up to 19,999 lbs)

Mid-sized Jet (20,000 - 29,999 lbs)

Super Mid-sized Jet (30,000 - 39,999 lbs)

Large Jet (40,000 - 49,999 lbs)

Long Range (50,000 lbs or greater)

* 4. How many total hours of pilot experience do you have?

0 - 2,499 hours

2,500 - 4,999 hours

5,000 - 7,499 hours

7,500 - 9,999 hours

10,000 hours or more

 112

* 5. How long have you worked for this Aircraft Management Company (AMC)?

0-4 years

5-9 years

10-14 years

15 or more years

* 6. What year were you born?

 113

Safety Culture & Performance Survey

* 7. Safety is a core value in my Aircraft Management Company (AMC).

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 8. Management is more concerned with making money than being safe.

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 9. Management expects pilots to push for on-time performance, even if it means compromising safety.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 10. Management doesn't show much concern for safety until there is an accident or an incident.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 11. Management does not cut corners where safety is concerned.

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 114

Safety Culture & Performance Survey

* 12. Checklists and procedures are easy to understand.

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 13. My Aircraft Management Company's (AMC's) manuals are carefully kept up to date.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 14. My Aircraft Management Company (AMC) is willing to invest money and effort to improve safety.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 15. My Aircraft Management Company (AMC) is committed to equipping aircraft with up-to-date technology.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 16. My Aircraft Management Company (AMC) ensures that maintenance on aircraft is adequately
performed and that aircraft are safe to operate.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 115

Safety Culture & Performance Survey

* 17. Management goes above and beyond regulatory minimums when it comes to issues of flight safety.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 18. Management schedules pilots as much as legally possible; with little concern for pilots' sleep schedule or
fatigue.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 19. Management tries to get around safety requirements whenever they get a chance.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 20. Management views regulation violations very seriously, even when they don't result in any serious
damage.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 116

Safety Culture & Performance Survey

* 21. Chief pilots do not hesitate to contact line pilots to proactively discuss safety issues.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 22. Chief pilots are unavailable when line pilots need help.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 23. As long as there is no accident or incident, chief pilots don't care how flight operations are performed.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 24. Chief pilots have a clear understanding of risks associated with flight operations.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 25. Pilots often report safety concerns to their chief pilot rather than the safety officer (safety department).
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 117

Safety Culture & Performance Survey

* 26. Dispatch consistently emphasizes information or details (e.g., weather requirements, NOTAMs) that
affect flight safety.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 27. Dispatch inappropriately uses the MEL (e.g., use when it would be better to fix equipment).
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 28. Dispatch is responsive to pilots' concerns about safety.

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 29. Dispatch would rather take a chance with safety than cancel a flight.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 118

Safety Culture & Performance Survey

* 30. Instructors/trainers have a clear understanding of risks associated with flight operations.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 31. Safety is consistently emphasized during training at my Aircraft Management Company (AMC).
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 32. Instructors/trainers teach shortcuts and ways to get around safety requirements.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 33. Instructors/trainers prepare pilots for various safety situations, even uncommon or unlikely ones.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 119

Safety Culture & Performance Survey

* 34. The safety reporting system is convenient and easy to use.

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 35. Pilots can report safety discrepancies without fear of negative repercussions.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 36. Pilots are willing to report information regarding marginal performance or unsafe actions of other pilots.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 37. Pilots don't bother reporting near misses or close calls since these events don't cause any real
damage.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 38. Pilots are willing to file reports about unsafe situations, even if the situation was caused by their own
actions.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 120

Safety Culture & Performance Survey

*
* 39. Safety issues raised by pilots are communicated regularly to all other pilots in this Aircraft Management
Company (AMC).
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 40. When a pilot reports a safety problem, it is corrected in a timely manner.

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 41. Pilots are satisfied with the way this Aircraft Management Company (AMC) deals with safety reports.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 42. My Aircraft Management Company (AMC) only keeps track of major safety problems and overlooks
routine ones.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 121

Safety Culture & Performance Survey

* 43. Personnel responsible for safety hold a high status in the Aircraft Management Company (AMC).
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 44. Personnel responsible for safety have the power to make changes.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 45. Personnel responsible for safety have a clear understanding of the risks involved in flying the line.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 46. Safety personnel have little or no authority compared to operations personnel.

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 47. Safety personnel demonstrate a consistent commitment to safety.

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 122

Safety Culture & Performance Survey

* 48. Management shows favoritism to certain pilots.

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 49. Standards of accountability are consistently applied to all pilots in this organization.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 50. When pilots make a mistake or do something wrong, they are dealt with fairly by the Aircraft
Management Company (AMC).
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 51. When an accident or incident happens, management immediately blames the pilot.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 123

15

Safety Culture & Performance Survey

* 52. Pilots are seldom asked for input when Aircraft Management Company (AMC) procedures are
developed or changed.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 53. Pilots are actively involved in identifying and resolving safety concerns.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 54. Pilots who call in sick or fatigued are scrutinized by the chief pilot or other management personnel.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 55. Pilots have little real authority to make decisions that affect the safety of normal flight operations.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 56. Management rarely questions a pilot's decision to delay a flight for a safety issue.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 124

Safety Culture & Performance Survey

* 57. Pilots view the Aircraft Management Company's (AMC's) safety record as their own and take pride in it.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 58. Pilots who don't fly safely quickly develop a negative reputation among other pilots.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 59. Pilots with less seniority are willing to speak up regarding flight safety issues.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 60. Decisions made by senior pilots are difficult to challenge.

Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 61. Pilots don't cut corners or compromise safety regardless of the operational pressures to do so.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 125

Safety Culture & Performance Survey

* 62. I make errors in my job from time to time.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 63. Workload pressures have at times affected the quality of my work.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 64. I have made errors that have been detected by other pilots.
Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree
 126

Safety Culture & Performance Survey

* 65. I will say something if my peers (other pilots) take shortcuts.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 66. I will say something if my supervisor takes shortcuts.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 67. "Gut instincts" can be used in lieu of the publications and manuals.
Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 68. There are better ways of performing a task than those described in the publications and manuals.
Neither disagree nor
Strongly disagree Disagree agree Agree Strongly agree

* 69. There are better ways of performing a task than those described in the company operations manuals.
Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 70. Bending a procedure is not the same as breaking it.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

\
 127

* 71. Shortcuts, in order to get a task done, are still violations of procedures.
Neither agree nor
disagree Agree Strongly agree
Strongly disagree Disagree

* 72. Reporting mistakes helps other people learn from them.

Neither agree nor
disagree Agree Strongly agree
Strongly disagree Disagree

* 73. Personnel should be encouraged to report their mistakes.

Neither agree nor
disagree Agree Strongly agree
Strongly disagree Disagree
 128

Safety Culture & Performance Survey

* 74. I am willing to put in a great deal of effort beyond that normally expected in order to help this Aircraft
Management Company (AMC) be successful.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 75. I talk up this Aircraft Management Company (AMC) to my friends as a great organization to work for.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 76. I would accept almost any type of pilot assignment in order to keep working for this Aircraft
Management Company (AMC).
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 77. I find that my values and the Aircraft Management Company's (AMC's) values are very similar.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 78. I am proud to tell others that I am part of this Aircraft Management Company (AMC).
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 129

* 79. This Aircraft Management Company (AMC) really inspires the best in me in the way of job performance.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 80. I am extremely glad I chose this Aircraft Management Company (AMC) to work for over others I was considering at the
time I joined.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 81. I really care about the fate of this Aircraft Management Company (AMC).
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree

* 82. For me, this is the best of all Aircraft Management Companies (AMCs) for which to work.
Somewhat Neither agree nor
Strongly disagree Disagree disagree disagree Somewhat agree Agree Strongly agree
 130

Safety Culture & Performance Survey

* 83. Company managers conduct their personal lives in an ethical manner.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 84. Company management defines success not just by results but also the way that they are obtained.
Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 85. Company management listens to what employees have to say.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 86. Company management disciplines employees who violate ethical standards.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 87. Company management makes fair and balanced decisions.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree

* 88. Company management can be trusted.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree
 131

* 89. Company management discusses business ethics or values with employees.

Neither agree nor
disagree Agree Strongly agree
Strongly disagree Disagree

* 90. Company management sets an example of how to do things the right way in terms of ethics.
Neither agree nor
disagree Agree Strongly agree
Strongly disagree Disagree

* 91. Company management has the best interests of employees in mind.

Neither agree nor
disagree Agree Strongly agree
Strongly disagree Disagree

* 92. When making decisions, company management asks "what is the right thing to do?"
Neither agree nor
disagree Agree Strongly agree
Strongly disagree Disagree
 132

Safety Culture & Performance Survey

* 93. I am more likely to make judgement errors in abnormal or emergency situations.

Neither agree nor
Strongly disagree Disagree disagree Agree Strongly agree
 133

Safety Culture & Performance Survey

Thank you!

The principal researcher, Kevin O'Leary thanks you for taking the time to complete this survey.
A donation to the Corporate Angel Network will be made for each completed survey.

Thank you very much!

Kevin
O'Leary
617-600-
6868
olearyk1@my.erau.edu
 134

APPENDIX C

Tables

C1 Descriptive Statistics
 135

Table C1

Descriptive Statistics

Skewness Kurtosis
Std. Std. Std.
Item N Min Max Dev Var. Statistic Error Statistic Error
OC7. Safety is a
core value in my
Aircraft 305 1 7 1.58 2.50 -1.59 0.14 2.04 0.28
Management
Company (AMC).
OC8.
Management is
more concerned
305 1 7 1.94 3.75 0.36 0.14 -1.21 0.28
with making
money than being
safe.

OC9.
Management
expects pilots to
push for on-time 305 1 7 1.84 3.37 0.80 0.14 -0.61 0.28
performance, even
if it means
compromising
safety.

OC10.
Management
doesn't show
much concern for 305 1 7 1.79 3.21 0.87 0.14 -0.45 0.28
safety until there
is an accident or
an incident.

OC11.
Management does
not cut corners 305 1 7 1.88 3.52 -0.14 0.14 -1.33 0.28
where safety is
concerned.

OC12. Checklists
and procedures are 305 1 7 1.31 1.72 -1.34 0.14 1.32 0.28
easy to
understand.
 136

OC13. My
Aircraft
Management
Company's 305 2 7 1.02 1.04 -1.68 0.14 3.75 0.28
(AMC's) manuals
are carefully kept
up to date.

OC14. My
Aircraft
Management
305 1 7 1.31 1.71 -1.08 0.14 1.42 0.28
Company (AMC)
is willing to invest
money and effort
to improve safety.

OC15. My
Aircraft
Management
Company (AMC) 305 1 7 1.51 2.27 -0.89 0.14 0.32 0.28
is committed to
equipping aircraft
with up-to-date
technology.

OC16. My
Aircraft
Management
Company (AMC)
ensures that
305 1 7 1.59 2.52 -0.88 0.14 -0.10 0.28
maintenance on
aircraft is
adequately
performed and
that aircraft are
safe to operate.

OC17.
Management goes
above and beyond
305 1 7 1.53 2.35 -0.65 0.14 -0.38 0.28
regulatory
minimums when it
comes to issues of
flight safety.
 137

OC18.
Management
schedules pilots as
much as legally
possible; with 305 1 7 1.80 3.23 -0.66 0.14 -0.74 0.28
little concern for
pilots' sleep
schedule or
fatigue.

OC19.
Management tries
to get around
305 1 7 1.76 3.11 0.64 0.14 -0.65 0.28
safety
requirements
whenever they get
a chance.

OC20.
Management
views regulation
violations very 305 1 7 1.33 1.76 -0.94 0.14 0.56 0.28
seriously, even
when they don't
result in any
serious damage.

OP21. Chief pilots

do not hesitate to
contact line pilots 305 1 7 1.72 2.97 -0.50 0.14 -0.75 0.28
to proactively
discuss safety
issues.

OP22. Chief pilots

are unavailable 305 1 7 1.54 2.36 1.03 0.14 0.28 0.28
when line pilots
need help.

OP23. As long as
there is no
accident or 305 1 7 1.74 3.02 1.00 0.14 -0.14 0.28
incident, chief
pilots don't care
how flight
 138

operations are
performed.

OP24. Chief pilots

have a clear
understanding of
305 1 7 1.43 2.06 -1.15 0.14 0.85 0.28
risks associated
with flight
operations.

OP25. Pilots often

report safety
concerns to their
305 1 7 1.65 2.71 -0.12 0.14 -0.95 0.28
chief pilot rather
than the safety
officer (safety
department).

OP26. Dispatch
consistently
emphasizes
information or
details (e.g., 305 1 7 1.83 3.34 -0.27 0.14 -1.11 0.28
weather
requirements,
NOTAMs) that
affect flight
safety.

OP27. Dispatch
inappropriately
uses the MEL 305 1 7 1.84 3.38 0.07 0.14 -1.23 0.28
(e.g., use when it
would be better to
fix equipment).

OP28. Dispatch is
responsive to 305 1 7 1.47 2.16 -1.01 0.14 0.43 0.28
pilots' concerns
about safety.

OP29. Dispatch
would rather take
305 1 7 1.73 2.99 0.69 0.14 -0.64 0.28
a chance with
safety than cancel
a flight.
 139

OP30.
Instructors/trainers
have a clear
understanding of 305 2 7 1.15 1.33 -1.24 0.14 1.51 0.28
risks associated
with flight
operations.

OP31. Safety is
consistently
emphasized
305 2 7 1.14 1.30 -1.31 0.14 1.78 0.28
during training at
my Aircraft
Management
Company (AMC).

OP32.
Instructors/trainers
teach shortcuts 305 1 7 1.01 1.02 1.95 0.14 5.60 0.28
and ways to get
around safety
requirements.

OP33.
Instructors/trainers
prepare pilots for
305 1 7 1.29 1.67 -1.20 0.14 1.30 0.28
various safety
situations, even
uncommon or
unlikely ones.

FS34. The safety

reporting system 305 1 7 1.37 1.88 -1.25 0.14 1.20 0.28
is convenient and
easy to use.
FS35. Pilots can
report safety
discrepancies
305 1 7 1.38 1.90 -1.66 0.14 2.68 0.28
without fear of
negative
repercussions.
 140

FS36. Pilots are

willing to report
information
regarding
305 1 7 1.58 2.51 -0.20 0.14 -0.96 0.28
marginal
performance or
unsafe actions of
other pilots.

FS37. Pilots don't

bother reporting
near misses or
305 1 7 1.49 2.21 0.57 0.14 -0.61 0.28
close calls since
these events don't
cause any real
damage.

FS38. Pilots are

willing to file
reports about
unsafe situations, 305 1 7 1.16 1.35 -1.09 0.14 1.29 0.28
even if the
situation was
caused by their
own actions.

FS39. Safety
issues raised by
pilots are
communicated
305 1 7 1.80 3.25 -0.64 0.14 -0.76 0.28
regularly to all
other pilots in this
Aircraft
Management
Company (AMC).

FS40. When a
pilot reports a
305 1 7 1.53 2.36 -0.36 0.14 -0.60 0.28
safety problem, it
is corrected in a
timely manner.
 141

FS41. Pilots are

satisfied with the
way this Aircraft
Management 305 1 7 1.67 2.78 -0.36 0.14 -0.86 0.28
Company (AMC)
deals with safety
reports.

FS42. My Aircraft
Management
Company (AMC)
only keeps track 305 1 7 1.52 2.32 0.65 0.14 -0.38 0.28
of major safety
problems and
overlooks routine
ones.

FS43. Personnel
responsible for
safety hold a high
305 1 7 1.49 2.21 -0.59 0.14 -0.19 0.28
status in the
Aircraft
Management
Company (AMC).

FS44. Personnel
responsible for
305 1 7 1.57 2.47 -0.42 0.14 -0.70 0.28
safety have the
power to make
changes.

FS45. Personnel
responsible for
safety have a clear 305 1 7 1.62 2.63 -0.83 0.14 -0.11 0.28
understanding of
the risks involved
in flying the line.
FS46. Safety
personnel have
little or no
authority 305 1 7 1.67 2.79 0.10 0.14 -0.94 0.28
compared to
operations
personnel.
 142

FS47. Safety
personnel
demonstrate a
305 1 7 1.41 1.99 -0.92 0.14 0.47 0.28
consistent
commitment to
safety.

IS48.
Management 305 1 7 1.69 2.84 -1.00 0.14 0.10 0.28
shows favoritism
to certain pilots.

IS49. Standards of
accountability are
consistently 305 1 7 1.96 3.84 -0.10 0.14 -1.35 0.28
applied to all
pilots in this
organization.

IS50. When pilots

make a mistake or
do something
wrong, they are 305 1 7 1.71 2.91 -0.48 0.14 -0.77 0.28
dealt with fairly
by the Aircraft
Management
Company (AMC).

IS51. When an
accident or
incident happens, 305 1 7 1.70 2.89 0.12 0.14 -0.84 0.28
management
immediately
blames the pilot.

IS52. Pilots are

seldom asked for
input when
Aircraft
305 1 7 1.82 3.31 -0.31 0.14 -1.15 0.28
Management
Company (AMC)
procedures are
developed or
changed.
 143

IS53. Pilots are

actively involved
in identifying and 305 1 7 1.70 2.90 -0.39 0.14 -0.97 0.28
resolving safety
concerns.

IS54. Pilots who

call in sick or
fatigued are
scrutinized by the 305 1 7 2.02 4.07 0.32 0.14 -1.26 0.28
chief pilot or other
management
personnel.

IS55. Pilots have

little real authority
to make decisions 305 1 7 1.73 3.01 1.31 0.14 0.55 0.28
that affect the
safety of normal
flight operations.

IS56.
Management
rarely questions a 305 1 7 1.93 3.73 -0.51 0.14 -1.09 0.28
pilot's decision to
delay a flight for a
safety issue.

IS57. Pilots view

the Aircraft
Management
Company's 305 1 7 1.40 1.95 -1.01 0.14 0.45 0.28
(AMC's) safety
record as their
own and take
pride in it.

IS58. Pilots who

don't fly safely
quickly develop a 305 2 7 1.15 1.33 -1.04 0.14 1.20 0.28
negative
reputation among
other pilots
 144

IS59. Pilots with

less seniority are
willing to speak 305 1 7 1.42 2.02 -0.98 0.14 0.29 0.28
up regarding flight
safety issues.

IS60. Decisions
made by senior 305 1 7 1.46 2.13 0.91 0.14 -0.05 0.28
pilots are difficult
to challenge.

IS61. Pilots don't

cut corners or
compromise 305 1 7 1.59 2.53 -0.52 0.14 -0.80 0.28
safety regardless
of the operational
pressures to do so.

ER62. I make
305 1 5 0.55 0.30 -0.53 0.14 3.84 0.28
errors in my job
from time to time.

ER63. Workload
pressures have at
305 1 5 0.87 0.76 -1.34 0.14 2.26 0.28
times affected the
quality of my
work.

ER64. I have
made errors that
305 1 5 0.56 0.31 -0.77 0.14 4.55 0.28
have been
detected by other
pilots.

AT65. I will say

something if my 305 2 5 0.57 0.33 -0.47 0.14 1.86 0.28
peers (other pilots)
take short cuts.
AT66. I will say
something if my
supervisor takes 305 1 5 0.71 0.51 -0.81 0.14 1.57 0.28
shortcuts.
 145

AT67. Gut
instincts can be
used in lieu of the 305 1 5 0.97 0.94 0.40 0.14 -0.52 0.28
publications and
manuals.
AT68. There are
better ways of
performing a task
than those 305 1 5 0.94 0.88 -0.19 0.14 -0.29 0.28
described in the
publications and
manuals.
AT69. There are
better ways of
performing a task
than those
305 1 5 0.98 0.97 -0.19 0.14 -0.49 0.28
described in the
company
operations
manuals.

AT70. Bending a
procedure is not 305 1 5 0.88 0.78 0.38 0.14 -0.45 0.28
the same as
breaking it.

AT71. Shortcuts,
in order to get a
305 1 5 0.81 0.66 -0.87 0.14 1.17 0.28
task done, are still
violations * of
procedures.

AT72. Reporting
mistakes helps 305 2 5 0.59 0.35 -0.62 0.14 0.49 0.28
other people learn
from them.

AT73. Personnel
should be
305 2 5 0.57 0.33 -0.64 0.14 0.04 0.28
encouraged to
report their
mistakes.
 146

PC74. I am
willing to put in a
great deal of effort
beyond that
normally expected
305 1 7 1.24 1.55 -1.33 0.14 2.05 0.28
in order to help
this Aircraft
Management
Company (AMC)
be successful.

PC75. I talk up
this Aircraft
Management
305 1 7 1.72 2.96 -0.64 0.14 -0.50 0.28
Company (AMC)
to my friends as a
great organization
to work for.

PC76. I would
accept almost any
type of pilot
assignment in
305 1 7 1.85 3.43 0.08 0.14 -1.20 0.28
order to keep
working for this
Aircraft
Management
Company (AMC).

PC77. I find that

my values and the
Aircraft
305 1 7 1.77 3.14 -0.47 0.14 -0.84 0.28
Management
Company's
(AMC's) values
are very similar.

PC78. I am proud
to tell others that I
am part of this 305 1 7 1.67 2.78 -0.91 0.14 -0.06 0.28
Aircraft
Management
Company (AMC).
 147

PC79. This
Aircraft
Management
Company (AMC)
305 1 7 1.71 2.91 -0.44 0.14 -0.69 0.28
really inspires the
best in me in the
way of job
performance.

PC80. I am
extremely glad I
chose this Aircraft
Management
305 1 7 1.84 3.40 -0.85 0.14 -0.41 0.28
Company (AMC)
to work for over
others I was
considering at the
time I joined.

PC81. I really care

about the fate of
305 1 7 1.38 1.91 -1.96 0.14 3.72 0.28
this Aircraft
Management
Company (AMC).

PC82. For me, this

is the best of all
Aircraft
305 1 7 1.78 3.18 -1.16 0.14 0.12 0.28
Management
Companies
(AMCs) for which
to work.

EL83. Company
managers conduct
305 1 5 0.93 0.86 -0.34 0.14 0.47 0.28
their personal
lives in an ethical
manner.

EL84. Company
management
defines success
305 1 5 0.99 0.98 -0.29 0.14 -0.41 0.28
not just by results
but also the way
that they are
obtained.
 148

EL85. Company
management
listens to what 305 1 5 1.11 1.24 -0.22 0.14 -0.85 0.28
employees have to
say.

EL86. Company
management
disciplines
305 1 5 0.90 0.81 -0.97 0.14 0.68 0.28
employees who
violate ethical
standards.

EL87. Company
management
305 1 5 1.07 1.15 -0.15 0.14 -0.77 0.28
makes fair and
balanced
decisions.

EL88. Company
management can 305 1 5 1.18 1.40 0.13 0.14 -0.91 0.28
be trusted.
EL89. Company
management
discusses business 305 1 5 1.01 1.03 -0.91 0.14 0.41 0.28
ethics or values
with employees.
EL90. Company
management sets
an example of
how to do things 305 1 5 1.23 1.50 0.02 0.14 -1.12 0.28
the right way in
terms of ethics.
EL91. Company
management has
the best interests 305 1 5 1.12 1.25 0.23 0.14 -0.76 0.28
of employees in
mind.
EL92. When
making decisions,
company
305 1 5 1.09 1.20 0.05 0.14 -0.77 0.28
management asks
"what is the right
thing to do?"
 149

A93. I am more
likely to make
judgement errors
305 1 5 1.01 1.03 0.24 0.14 -0.94 0.28
in abnormal or
emergency
situations.
 150

APPENDIX D

Tables

Table D1

Total Variance Explained for EFA

Extraction Sums of Rotation Sums of

Initial Eigenvalues Squared Loadings Squared Loadings
% of *Cumul % of *Cumul % of *Cumul
Comp Total Var. % Total Var % Total Var %
1 18.886 46.064 46.064 18.886 46.064 46.064 9.320 22.733 22.733
2 2.227 5.431 51.495 2.227 5.431 51.495 8.514 20.766 43.499
3 1.792 4.370 55.865 1.792 4.370 55.865 2.630 6.414 49.913
4 1.585 3.865 59.730 1.585 3.865 59.730 2.260 5.511 55.424
5 1.245 3.036 62.766 1.245 3.036 62.766 2.196 5.355 60.780
6 1.096 2.674 65.440 1.096 2.674 65.440 1.764 4.301 65.081
7 1.033 2.519 67.959 1.033 2.519 67.959 1.180 2.878 67.959
*Cumul % is the Cumulative Percentage
 151

Table D2

Rotated Correlation Matrix for EFA

Components
1 2 3 4 5 6 7
OC8. -.680 -.460
OC9. -.793 -.382
OC10. -.777 -.383
OC11. .481 .441
OC14. .529 .382 .387
OC16. .643 .379 .376
OC17. .618 .423 .331
OC19. -.709 -.428
OC20. .527 .309 .483
OP21. .449 .404
OP22. -.639
OP23. -.661 -.343
OP24. .623 .391
OP27. -.627 -.347
OP28. .708
OP29. -.768
OP31. .475 .482
FS36. .753
FS38. .716
FS40. .452 .462 .420 .330
FS41. .497 .421 .365 .339
FS42. -.531 -.343
FS47. .440 .383 .447
IS48. -.403 -.601
IS49. .380 .596
IS53. .361 .447 .457
ER62. .889
ER64. .897
AT66. .801
AT68. .917
AT69. .891
AT70. .524 -.346
PC74. .687 .396
EL83. .308 .650
EL84. .638
EL85. .409 .718
EL87. .350 .809
 152

EL88. .344 .821

EL90. .333 .831
EL91. .331 .806
EL92. .807
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 13 iterations.
 153

APPENDIX E

Suggested Future CASS Survey Questions

 154

OC8. Management is more concerned with making money than being safe.

OC9. Management expects pilots to push for on-time performance, even if it means
compromising safety.

OC10. Management doesn't show much concern for safety until there is an accident or an
incident.

OC14. My Aircraft Management Company (AMC) is willing to invest money and effort
to improve safety.

OC16. My Aircraft Management Company (AMC) ensures that maintenance on aircraft

is adequately performed and that aircraft are safe to operate.

OC17. Management goes above and beyond regulatory minimums when it comes to
issues of flight safety.

OC19. Management tries to get around safety requirements whenever they get a chance.

OC20. Management views regulation violations very seriously, even when they don't
result in any serious damage.

OP21. Chief pilots do not hesitate to contact line pilots to proactively discuss safety
issues.

OP23. As long as there is no accident or incident, chief pilots don't care how flight
operations are performed.

OP24. Chief pilots have a clear understanding of risks associated with flight operations.

OP27. Dispatch inappropriately uses the MEL (e.g., use when it would be better to fix
equipment).

OP29. Dispatch would rather take a chance with safety than cancel a flight.

FS40. When a pilot reports a safety problem, it is corrected in a timely manner.

FS41. Pilots are satisfied with the way this Aircraft Management Company (AMC) deals
with safety reports.

FS42. My Aircraft Management Company (AMC) only keeps track of major safety
problems and overlooks routine ones.

FS47. Safety personnel demonstrate a consistent commitment to safety.