‘ .9.
L:

:9;

 

‘1‘?
am“
pin.» 2.:
{tin I:
. it...

 

....£§§ 1 , A
. V 42‘...”

 

 

THEE“?

‘Imnlulli‘llllllli:illiilluum L

3 1293 01770 0596

LIBRARY
Michigan State
University

 

U"

 

 

This is to certify that the

thesis entitled

Examining the Effects of Incentive-Based Health Benefits:

A Closer Look at the Effects of Perceived Control and
Fairness on Employee Behaviors and Perceptions

presented by

Deidre Jane Wasson

has been accepted towards fulfillment
of the requirements for

Masters degree in Psychology

 

 

 

( Qajor professor

Date l27//7/77

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

PLACE IN REFURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

J c}; U 5 gafifli
”a” ( 4 _,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EXAMINING THE EFFECTS OF INCENTIVE-BASED HEALTH BENEFITS: A
CLOSER LOOK AT THE EFFECTS OF PERCEIV ED CONTROL AND FAIRNESS

ON EMPLOYEE BEHAVIORS AND PERCEPTIONS

by

Deidre Jane Wasson

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS
Department of Psychology

1998

ABSTRACT

EXAMINING THE EFFECTS OF INCENTIVE-BASED HEALTH BENEFITS: A
CLOSER LOOK AT THE EFFECTS OF PERCEIVED CONTROL AND FAIRNESS
ON EMPLOYEE BEHAVIORS AND PERCEPTIONS

By

Deidre Jane Wasson

A new type of program is on the rise in organizations to compensate for
the escalating costs of health insurance offered to employees as alternative compensation.
In these programs, employees are differentially charged for health insurance based on
their health level as determined by the organization. Reactions to these programs vary
widely among employees, regardless of their individual outcomes, and may be explained
by examining employee perceptions of fairness. Giving employees control over their
own outcomes, as is done in these programs, may increase perceptions of fairness and
lessen the adverse effects of negative outcomes explaining some perceptual and
behavioral variations. Ideally, increased perceptions of fairness would be accompanied
by increased commitment to the program (Korsgaard, Schweiger, & Sapicnza, 1995).
This study investigates the effects of outcome control as it relates to health on perceptions
of fairness by proposing a model that suggests that perceptions of outcome control and

fairness dually influence employees’ commitment to improving their health, which would

subsequently lead to an increase in health-related behaviors. The data required to test the
proposed model were collected in two Midwest hospitals, one employing the incentive-
based health benefits program (N: 243) and the other taking a more traditional benefits
approach (N: 273). Overall the model was supported. Specific findings and

implications of this research are discussed.

To my parents, Terry and Paula Wasson.
They have been a source of strength and love throughout my entire life.
They believed in me when I didn’t and pushed me where I didn’t think I could go.

I will be eternally grateful for all that they have given me.

ACKNOWLEDGMENTS

This research was made possible by the support of the Cook Institute of Grand
Rapids, Michigan and the Outreach Center at Michigan State University.

I wish to acknowledge the substantial amount of time and energy that my chair,
Dr. Daniel R. Ilgen, has invested in this project and in my development as a graduate
student. His attention to detail, knowledge of research, and willingness to provide
detailed feedback has made a major impact upon my approach to this research. I also
wish to sincerely thank my other committee members. Dr. Ann Marie Ryan whose
guidance into the field of health psychology profoundly shaped the way in which I
approached this research effort and Dr. Richard DeShon whose endless patience and
guidance through factor analysis strongly influenced the research results. Combined
together, all of my committee members possessed unique talents and perspectives that
worked incredibly well together to enhance this research effort.

In addition to my thesis committee, I also wish to truly thank my cohort. Without
the support and understanding of KD Delbridge, Danielle Jennings, Lori Sheppard, and
Becca Toney I could not have completed this thesis still smiling. They are appreciated
beyond words and remind me daily why I am so grateful to them and how much they
brighten the basement of Baker Hall.

Finally, I want to thank Rich. Without his infuriatingly perennial and naive good
mood this process would have been almost impossible. Never once, did he pressure me
or question how things were going, he just always knew and supported me with his

enduring strength and love.

TABLE OF CONTENTS

LIST OF TABLES. ............................................................................................................ ix
LIST OF FIGURES ............................................................................................................. x
INTRODUCTION ............................................................................................................... 1
Rewarding Health Behaviors: A specific case ........................................................ 5
Targeted Health Behaviors ............................................................................. 5
Health Quotient .............................................................................................. 5
Reward System ............................................................................................... 6
Exceptions ...................................................................................................... 8
Reactions to the System ........................................................................................... 8
CHAPTER 1
LITERATURE REVIEW .................................................................................................. 10
Incentives ................................................................................................................ 10
Commitment ........................................................................................................... 14
Fairness ................................................................................................................... l7
Distributive Justice ..................................................................................... 17
Procedural Justice ....................................................................................... 21
Perceived Control ................................................................................................... 24
CHAPTER 2
HYPOTHESES ................................................................................................................. 29
The Roles of Outcome, Age, and Accuracy Perceptions on Fairness
Perceptions ............................................................................................................ 29
The Effects of Outcomes on Employee Perceptions of Fairness .............. 29
Effects of Age on Employee Perceptions of Fairness ............................... 30
Effects of Income ...................................................................................... 31
The Relationship between Accuracy and Fairness .................................... 32
Hypotheses Relating to the Relationships between Fairness, Commitment, and
Perceived Control .................................................................................................. 34
The Relationship between Commitment and Behavior ............................. 36
The Relationship between Fairness and Commitment .............................. 36
The Relationship between Perceived Control and Commitment .............. 37
The Relationship between Fairness and Perceived Control ...................... 38

Putting the Pieces Together: Understanding How Incentive Systems Affect the
Relationships between Fairness, Perceived Control, and

Commitment .......................................................................................................... 38
The Effects of the Incentive System on Employee Health Behaviors ...... 39
The Effects of the Incentive System on Commitment .............................. 40
The Effects of the Incentive System on Perceived Control ....................... 41
Two Alternative Models to Explain the Relationship between Incentives
and Fairness ............................................................................................... 42

vi

CHAPTER 3

METHOD .......................................................................................................................... 50
Sample .................................................................................................................... 50
Incentive-based program description ..................................................................... 54
Procedure ................................................................................................................ 55

Procedures followed in the hospital with the incentive system .................. 55
Procedures followed in the hospital with the traditional system ................ 59
Measures ................................................................................................................. 60
Fairness ....................................................................................................... 60
Health—Related Behaviors ........................................................................... 63
Perceived Control ....................................................................................... 67
Commitment ............................................................................................... 68
Perceived Accuracy of Measurement ......................................................... 69
Demographics ............................................................................................. 70

CHAPTER 4

RESULTS .......................................................................................................................... 71
Nature and quality of variables .............................................................................. 71
Relationships Among Variables ............................................................................. 74
Tests of Hypotheses ................................................................................................ 80

CHAPTER 5

DISCUSSION ................................................................................................................... 88

Overview of Results .............................................................................................. 88
The Effects of Outcome, Age, Income, and Perceived Accuracy on
Fairness Perceptions .................................................................................. 89
Hypotheses Relating to the Relationships between Fairness, Commitment,
and Perceived Control ............................................................................... 92

Putting the Pieces Together: Understanding How Incentive Systems
Affect the Relationships between Fairness, Perceived Control, and

Commitment .............................................................................................. 95
Limitations .......................................................................................................... 102
Future Directions ................................................................................................. 105

APPENDICES ................................................................................................................. 108
APPENDD( A: Letters sent to sample of employees at both hospitals ............. 109
APPENDD( B: Consent form used for hospital with incentive-based benefits
system .................................................................................................................. 112
APPENDD( C: Consent form used in traditional health benefits hospital ........ 113
APPENDD( D: Introduction of survey for the hospital using an incentive-based
benefits program .................................................................................................. 114
APPENDD( E: Introduction of survey for hospital using traditional
Benefits ................................................................................................................ 115
APPENDIX F: Items adapted to measure fairness perceptions in the incentive-
based benefits system and the traditional system unless indicated ..................... l 16

vii

APPENDIX G: Factor solutions for fairness items ............................................ 119
APPENDIX H: Items created or adapted to measure health-related behaviors in
the incentive-based benefits system and the traditional benefits system presented
by each health dimension .................................................................................... 120
APPENDIX 1: Factor solution for health behavior self-report items ................. 124
APPENDD( J: Items created to measure perceptions of control over outcomes
and health behaviors in the incentive-based benefits system and the traditional
system unless indicated ....................................................................................... 130
APPENDD( K: Items created to measure perceptions of commitment to change
health-related behaviors in the incentive-based benefits system and the traditional

benefits system .................................................................................................... 132

APPENDD( L: Eight items created to measure perceptions of accuracy in the

incentive-based benefits system .......................................................................... 133
BIBLIOGRAPHY ........................................................................................................... 134

viii

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

LIST OF TABLES

Demographic Characteristics of Sample with Incentive-based
Benefits ........................................................................... 52

Demographic Characteristics of Sample with Traditional Health
Benefits ........................................................................... 53

Intercorrelations between Self-report Dimensions of Health
Behaviors ........................................................................... 65

Intercorrelations between Self-reports of Health Behaviors from the
Surveys with Health Behaviors Measured as Part of the Health
Assessment in the Incentive Hospital .......................................... 66

Reliability Data for Constructed Scales ........................................ 72

Means, Standard Deviations, and Intercorrelations for Entire Data

Set ................................................................................... 76
Intercorrelations between Variables in Incentive Group ..................... 78
Intercorrelations between Variables in Traditional Group .................. 79

Data Illustrating the Differences between Means of the Incentive-based
and Traditional Samples: Independent Samples and t-test for Equality of
Means ............................................................................... 85

LIST OF FIGURES

Figure 1. Model representation of the relationship between fairness, perceived
control, commitment, and health-related
behaviors ........................................................................... 35

Figure 2. Model depicting a positive relationship between incentive-based health
benefits system and fairness perceptions ....................................... 48

Figure 3. Model depicting a negative relationship between incentive-based health
benefits system and fairness perceptions ....................................... 49

INTRODUCTION

For decades organizations have offered medical benefits to employees as a form
of alternative compensation. Recently, however, medical insurance has become
increasingly expensive for everyone, including organizations. In fact, in 1987 United
States organizations spent approximately $95 billion in health care costs, about two to
five thousand dollars per employee (Colosi, 1988) and that amount is rising rapidly every
year due to the increasing costs of health care. Organizations are bearing the brunt of
these costs, as health insurance has become an accepted and expected form of
compensation for employment. Due to the rising costs of health care and the expectation
that employers will provide health care insurance to all employees, organizations have
attempted to hold down these escalating costs through a variety of means.

The first solution generally utilized to control the cost of health care is to require
employees to pay more for their own health care. In some cases organizations have
decreased the coverage employees receive, restricted the amount employers will pay for
certain procedures, or instituted co-pay programs where employees share in the costs of
health care. All these methods cut the costs of health care to the organization, but they do
so by turning the costs either in dollars or in increased health risk over to the employees.

The second solution to aid organizations that are dealing with health care costs is
to attempt to control the behaviors of employees so that the health care they receive is
less costly. There are two specific methods that have been implemented to control
employee health-related behaviors. The first is to restrict what health care employees
may seek by requiring that they get approval from the insurance provider before any

health care is sought or obtained. The second method to control the behaviors of

employees has been to require they be a part of an HMO. HMOs require that employees
receive health care from a specific set of physicians who are employed by the insurance
provider and who actively strive to keep health care costs at a minimum.

A third solution to reduce the rising costs of health care is becoming more
popular. A few organizations have tried to reduce costs by attempting to raise overall
employee health. This alternative operates on the notion that healthy employees use less
health services, therefore cost the organization less. Organizations are realizing what
health professionals and researchers have long known, that to save money and increase
health, the single most important health factor is prevention and that most behavioral or
psychological determinants of health are modifiable (Baum, Krantz, & Gatchel, 1997).
For instance, the most important modifiable health risk to Americans is cigarette smoking
which is linked to over 350,000 deaths a year from heart disease, cancers, and lung
disease (Matarazzo, 1988). In addition, the single leading cause of death in the United
States has been cardiovascular disease for the past decade (Higgins & Luepker, 1998:
NHLBI, 1994). Research suggests that the emergence of cardiovascular disease in people
is greatly determined by lifestyle, specifically, poor diet, lack of exercise, and smoking
(Krantz, Grunberg, & Baum, 1985). Finally, obesity is also linked to a variety of life
threatening illnesses such as hypertension, diabetes, and cancer (Sjostrom, 1993;
Stunkard & Wadden, 1993). While each of these important health issues have both
genetic and environmental factors, researchers support the notion that the effects of a
genetic predisposition can be mitigated to some degree by the performance of healthy

behaviors (Brownwell & Wadden, 1992).

Hoping to improve employee overall health, some organizations encourage
employees to lead healthier lifestyles by creating Employee Assistance Programs (EAP)
which provide employees with counseling and health services or in-house gyms for
employees to use free of charge and health programs such as smoking cessation and
weight loss for all employees to voluntarily attend (Weiss, Fielding, & Baum, 1991). In
addition, a few organizations provide healthy individuals with monetary incentives and
disincentives to the unhealthy by requiring them to contribute more than their healthy
counterparts to the cost of their health insurance. These programs were created to cut
health care costs immediately and encourage a healthier work force for the long term.
This approach combines principles from the previous solutions to reduce the costs of
health care immediately by requiring the most costly, unhealthy, employees to pay more
for their health care while also providing a long term reduction in health care costs by
increasing overall employee health. One of these incentive-based programs will be
investigated in this study.

This system of differentially providing incentives to employees based on their
health to reduce costs and increase employee health presents a novel situation to
researchers. The new system operates on the same premise that automobile insurance
policies have used from the start, that of making the most dangerous, costly persons
covered pay more for their insurance. However, applying the auto insurance analogy to
health insurance differs in one fundamental way. Drivers with records of frequent
reckless driving are, for the most part, responsible for their behaviors. On the other hand,
an individual with high blood pressure has a higher risk of developing cardiovascular

disease (Somes, Harshfield, Alpert, et al., 1995), but it is less clear that the "fault" for the

hypertension lies as completely with his or her behavior. In the hypertension example,
blame is not as clearly discernible because controversy still surrounds the degree to
which an individual can control his or her blood pressure (Smith, Turner, Ford, et al.,
1987). For instance, a child of one parent with high blood pressure has a 50% greater
chance of developing hypertension than a child of two parents with normal blood
pressure. That risk rises to a 95% chance that a child of two parents with hypertension
will have high blood pressure as an adult (Taylor, 1995). The same is true for other
important aspects of health (Baum, Krantz, & Gatchel, 1997; Braunwald, 1994).

Individuals can be genetically or biologically predisposed to be obese (Ezzell,
1995; Faust, 1980; Hakas, Gajiwala, Maffei, et al., 1995; Keesey, 1993), have high
cholesterol (Kannel, Castelli, & Gordon, 1979; Stoney, Matthews, McDonald, &
Johnson, 1988), to smoke (Silverstein, Kelley, Swan, & Kozlowski, 1982), and to drink
alcohol (Dawson, Harford, & Grant, 1992; Morse & Flavin, 1992; Pickens & Svikis,
1991). While we know that these health factors are the result, in most cases, of an
interaction between any genetic or biological predisposition and the environment
(Brownell & Wadden, 1992), the question that arises is exactly how much can the
environment reverse what we are genetically predisposed to be or do. Thus, an interesting
issue that evolves from this type of program is whether or not employees perceive a
health benefits system that attaches incentives to the level of health as fair.

In this study, we are concerned with the effects of an incentive-based health
benefits program on employees, specifically, employee fairness perceptions. Two
hospitals were surveyed, one employing an incentive-based health benefits program and

the other using a traditional health benefits program. With these two samples, we will

investigate the perceptions of fairness, control over health, employee commitment to a
healthy lifestyle, and health-related behaviors. Since a number of features of the
incentive system itself influence the nature of the hypotheses regarding reactions to it, we
will first give a brief description of the program and then review the research relevant to
this study.

Rewarding Health Behaviors: A specific cafl

This study will look closely at an incentive-based health benefits program, which
was created to encourage positive health behaviors among employees. This program
assumes that employees have direct control over their health behaviors and awards
financial incentives for healthy behaviors and overall good health and disincentives to the
unhealthy.

_T_argeted Health BMrs. To determine employee health level or health risk,
this program evaluates all employees and their dependent spouses on eight health-related
dimensions. The following dimensions are listed in order of decreasing importance when
used for assessing health risk: blood pressure, cholesterol (HDL) level, body fat/muscle
tissue ratio, cardio-vascular fitness level, tobacco usage, motor vehicle safety, alcohol
consumption and nutrition.

Health Quotient. Employees’ and any dependent spouses’ health levels are
evaluated on the eight aforementioned health dimensions once a year in the summer
during a two-week assessment period. Four of the health dimensions are assessed with
objective means. In particular, blood pressure is measured using a pressure cuff,
cholesterol level is measured with a blood test, body fat/muscle ratio is determined with a

caliper test, and cardio-vascular fitness is assessed by heart-rate during a timed walk.

The other four health dimensions are measured with self-reports of tobacco and alcohol
use, motor vehicle safety, and nutrition behaviors.

Approximately one month following the health assessment employees receive
their and their spouse’s “health quotients” which are their composite scores (or averaged
composite scores if the employee has a dependent spouse) based on all the health
measures. Health quotient scores range from -25 to +25. The composite score is
calculated by weighting the scores within each health dimension and summing them. All
of the objectively measured health dimensions are given more weight than the
subjectively reported dimensions. This is an important distinction since the objectively
measured health dimensions may be more difficult to change than those that are
subjectively measured. For example, a change in nutrition habits can be made over the
time span of a few weeks or months, but those changes may not affect cholesterol level
and blood pressure for a long time.

In addition, when employees and dependents are notified of their health quotients
they are given a description of their scores on each of the individual dimensions, and are
provided with individualized suggestions directed at improving their overall health.
Employees are also provided discounts to athletic clubs and on the purchase of fitness
equipment to encourage more healthy lifestyles. Within hospital facilities, employees are
encouraged to participate in health programs offered to improve nutritional habits, stop
smoking, and improve cardio-vascular fitness.

Regan! System. When employees receive notification of their and their
dependent spouse’s health quotients from the organization they are also given the

incentive (or disincentive) amount they will receive with every paycheck. The health

quotient, which ranges from -25 to +25, directly translates into the monetary incentive or
disincentive each employee will receive throughout the coming year. For example, an
individual with a health quotient of +15 will receive an extra $15 per bi-monthly pay
check totaling $30 per month. On the other hand, a health quotient of -15 would mean
that a deduction of $15 will be taken from the employee’s bi-monthly paycheck. If, in
addition, the employee receiving the $15 deduction had a dependent spouse who was
covered under the same health insurance policy and this spouse had a health quotient of
+15, the employee would receive no deductions nor incentives as their two health
quotients would be averaged to determine the incentive amount. The health quotient
scores each individual receives will be in effect for one full year following the annual
health assessment. Not until the next health assessment will the individual be able to
improve his or her health quotient and modify his or her incentives (or disincentives).

At this point, it might be useful to discuss what is meant by an incentive or
disincentive. Technically, we could refer to the monetary outcomes in this study as either
rewards/punishments or incentives/disincentives. We have chosen the later. While the
monetary gains or losses are assigned based on an individual’s health level and are kept
for an entire year, we believe that they serve as an incentive for the individual to improve
his or her health further over the coming year to either reduce the amount of disincentive
the person received or increase the amount of incentive the person received. For the
remainder of this paper, the terms incentive and disincentive will be employed to describe
the monetary gains or losses each employee receives associated with his or her health

quotient.

Exceptions. Each employee participating in the health benefits program (nearly
100% of the employees) has a health quotient; however there are a few exceptions made
for employees who are physically unable to participate or who have an uncontrollable
health condition. For instance, a physician’s excuse for the fitness walk may be accepted
for those physically incapable for various reasons. In the cases when individuals are
exempt from completing a part of the assessment, they are automatically assigned a score
of zero (no monetary gain or loss) for that dimension which is then summed with their
scores on the other dimensions. The system also provides an option for those who choose
not to participate in the health benefits system. Those who choose not to will be assigned
a health quotient of -25, a pay deduction of fifty dollars per month for the entire year.
Be_actions to the System

Understanding the structure and nature of this incentive-based benefits system is
of importance as we investigate the implications these types of programs have for
employees. The system of differentially awarding employees based on their overall
health level will be investigated in this study. In particular, we are interested in the
effects these types of systems have on employee fairness perceptions and employee
perceived control over the system and the effects that these two have on employee health-
related actions. In an attempt to understand the relationships between these variables, a
model is proposed where fairness and perceived control dually affect employee
commitment, which influences an individual’s behaviors. We believe that individual
perceptions of fairness and perceived control are both determined, in part, by the benefits

system the organization employs.

In the next few pages, research providing the theoretical foundation for the
present study will be presented. First, a brief review of the research explaining the
effects of incentives on employee behaviors will be provided. Following that discussion,
a review of some commitment research will be presented. Subsequently, research on
both fairness and perceived control will be presented. Throughout these literature
reviews, research from the health psychology domain will be included due to the nature
of our samples and the health benefits programs involved in this study. Following the
review of this literature, we will discuss the hypotheses and models proposed for

investigation.

Chapter 1

LITERATURE REVIEW

Incentives

For over a half a century researchers have studied the effects of various incentives
on employee reactions to their work environments such as performance, job satisfaction,
and organization or job commitment (Hull, 1943; Lewin, 1938; Skinner, 1969; F.W.
Taylor, 1947; Thorndike, 1911; Tolman, 1932). The use of incentives is common in
situations where the goal is to influence employee behaviors. The discussion presented
on incentives will be rudimentary; however, for more comprehensive reviews on the use
of incentives to influence employee behaviors refer to Campbell and Pritchard (1976),
DeLeo (1972), Guzzo (1979), and Lawler (1971).

Incentives utilized in the work place can be financial or non-financial. Examples
of various financial incentives range from the common (wage increases for good
performance) to the unique and creative (free theater tickets). Non-financial incentives
are frequently an economical option for organizations. Examples are additional time off
for good attendance, promotions, job changes, or, simply, praise from a superior.
Regardless of the type of incentive employed, if the system is designed carefully and
clearly with employee concerns considered, incentive systems are often successful in
their aim to influence employee behaviors (Eisenberger & Cameron, 1996).

The theoretical foundation for the use of incentive systems can be found in two of
psychology’s most prominent theories: expectancy and reinforcement theories. Classical
reinforcement theory (Skinner, 1969) operates on the assumption that a rewarded

behavior is more likely to be repeated than an unrewarded behavior. Incentive systems

10

are based on this theory, in that if you provide an incentive for a behavior it is more likely
that the behavior will be demonstrated. This theory also applies to undesirable behaviors.
For instance, unwanted behaviors can be discouraged through either punishment or
reinforcement of the desired alternative behavior. Expectancy theory (Vroom, 1964) is
similar to reinforcement theory except it predicts that individuals will consider the
relative value of the promised outcome(s) and evaluate the likelihood of achieving the
behaviors necessary for the outcome when deciding how to act. For incentive systems
this theory suggests that individuals will evaluate the promised outcome for desirability
and then decide whether or not those outcomes are worth the effort required to attain the
outcomes. Expectancy theory, then, would explain why, in some cases, incentive
systems do not work for certain individuals.

Beyond expectancy and reinforcement theories, there is another explanation for
the effectiveness of incentive systems. When incentives are made salient before
behaviors are exhibited they provide a goal for individuals to strive for, given the goal is
desirable. Once the desired behaviors have been attempted, individuals receive feedback
about their performance through the receipt or withholding of the promised incentive. In
the workplace, research has found that incentives are effective when they provide
feedback to employees about their job-related behaviors in a positive way and provide
employees with goals to strive toward (Locke & Latham, 1990; Pritchard, Jones, Roth,
Stuebing, & Ekeberg, 1988).

In the workplace, incentives are often tied to behaviors relating to employee
performance. Meta-analyses by Guzzo, Jette, and Katzell (1985) have shown that the use

of incentives has an average effect size of .57 standard deviations on employee

11

performance. However, in all situations, the strength of the incentive depends on the
circumstances at the time of implementation, the behavior to be influenced, and the
methods used (Eisenberger & Cameron, 1996). Little is known in the field of industrial
and organizational psychology about the effects of incentive systems on employee health
level and the subsequent effects on other work-related behaviors. There does exist,
however, a large body of research in the health psychology domain about the effects of
various incentive systems on health-related behaviors. We’ll talk now about a few of
those studies, focusing on those that examined smoking and alcohol use, two health
behaviors that are measured in the incentive-based health benefits program involved in
this study.

Given that smoking is the single most modifiable health risk factor for disease
(Baum et al., 1997) it is no surprise that quite a bit of research has been directed at
incentives that may help encourage people to quit. One of the first studies to investigate
the effects of incentives on smokers was run by Bernstein and Glasgow (1979). They
provided monetary incentives to people provided that they quit smoking. Each day the
researchers met with their participants and, if a blood test showed no nicotine in their
systems, they were given a monetary sum. The incentive system showed good success at
the early stages, but as time went on and the blood tests became more erratic the
monetary incentive became less powerful. They concluded that monetary incentives
were a good method to get people to quit smoking, but not for long term maintenance.
Most other research endeavors directed at encouraging smoking cessation have been
directed at biological techniques and have met with greater success (Baum et al., 1997;

Silverstein et al, 1982).

12

Another leading cause of death and illness is alcohol, and research has also been
directed at reducing alcohol consumption among alcoholics. Two specific studies seem
directly applicable to our investigation as they investigated the effects of an incentive
system that utilized both self-given rewards and punishments (Bigelow, Libson, &
Lawrence, 1973; Griffiths, Bigelow, & Liebson, 1978). They placed alcoholics into five
different groups. The two groups that interest us involve the self-reward group and the
self-reward/self-punishment group. Individuals within these groups were given
suggestions about how and when to reward or punish themselves, but were given the
opportunity to decide for themselves when and what incentive methods to actually
employ. These two groups experienced almost identical reductions in alcohol
consumption, reported more sobriety than the three other groups, and maintained their
soberity for the longest period. However, as in the case of smoking cessation methods,
the aversive biological programs seem to be the most effective (Dawson et al., 1992;
Pickens & Svikis, 1991).

Overall, research suggests that an incentive-based benefits system may be a
powerful tool when attempting to encourage employees to behave in a more healthy
manner. However, employee commitment to improving or maintaining healthy behaviors
must be present and issues of the fairness of the system must be addressed before any
conclusions can be drawn concerning the effectiveness of and reactions to these types of

programs.

13

Commitment

Before employees will modify or maintain their health-related behaviors to
improve or maintain their overall health, research suggests that they must be committed
to improving or maintaining their health. There are several motivation/commitment
theories in health psychology that explain when and why people choose to (or not to)
modify their health-related behaviors. We will discuss two of the most popular and
empirically supported.

The first theory is called the Health Belief Mode, or HBM for short, and was
created by Janz and Becker (1984). They proposed the model to explain why people do
or do not engage in healthy behaviors. There are four specific antecedents that they
believe determine motivation and commitment to performing healthy behaviors. The first
determinant is perceived susceptibility and relates to whether or not an individual
believes that he or she is likely to develop any health side effects due to his or her
behaviors. Next, and related to the first dimension, is perceived seriousness and relates to
whether or not an individual believes that any side effects would be serious or life-
threatening. Once the individual perceives that some outcomes of his or her behaviors
are likely and serious, he or she has to believe that changing his or her health behaviors
would decrease the likelihood that he or she would develop the side effect or illness.
Finally, the individual has to believe that the benefits associated with changing his or her
health behaviors would outweigh the sacrifice of giving up his or her present health
behaviors. As an example we can consider a smoker. For the individual to quit smoking
he or she has to believe that smoking is related to lung cancer and that if he or she did

develop lung cancer the consequences would be quite serious and possibly life-

14

threatening. Before the smoker would actually commit to quitting he or she would have
to believe that if he or she quit now the risks associated with smoking would be greatly
diminished and that the decreased risk was worth any side effects he or she might
experience after quitting. We’ll talk more about this theory when we discuss individuals’
perceptions of control over their health, one of our proposed antecedents of commitment.

Another theory that is very popular in health psychology for explaining when and
why people initially commit to changing their health-related behaviors is the Theory of
Reasoned Action (TRA) developed by Ajzen & Fishbein (1980). This theory was
created as a general theory to explain the relationships between beliefs, attitudes,
intentions, and behaviors. Similar to the HBM model (J anz & Becker, 1984), according
to TRA people are under control of their behavior, and they base their behaviors on
rational decision making processes. The most immediate influence on behavior is the
intention to engage in the behavior. Two types of beliefs influence this commitment.
First, commitment is influenced by an individual’s attitude towards the behavior, whether
he or she sees the behavior as favorable or unfavorable. The second influence on
commitment is a more social factor called a subjective norm. Subjective norms refer to
the perceived social pressure to perform, or not perform, the behavior. Relatively few
large scale interventions have been based on TRA (Curry & Emmons, 1994).

In addition to discussing research linking motivation and commitment to health
behavior change, it is also useful to look at other areas of commitment research. In
general, commitment research focuses on improving work-related performance. For the
sake of the present study, we will assume that processes influencing the display of health

behaviors are similar to those influencing task/job performance. Both entail the

15

performance of certain behaviors relevant to the goal. Logically, then we can extend the
assumption to conclude that employee reactions to a goal of improving health will be
similar to one of improving any other aspect of performance. Before a discussion of the
goal commitment research is begun, a cursory review of general goal setting research will
be discussed.

Research studying the effects of goals on employee performance in a variety of
domains has been quite prolific (Locke & Latham, 1990; O’Leary Kelly, Martocchio, &
Frink, 1994). This research has repeatedly found that difficult, specific goals often lead
to increased performance on simple tasks (Austin & Vancouver, 1992; Locke, Shaw,
Saari, & Latham, 1981). However, employee commitment to the goal must be present
before a behavioral change is implemented and maintained (Locke, 1968).

Goal commitment has recently received increased recognition in the literature as
an important factor in the relationship between goals and performance. Hollenbeck and
Klein (1987) have defined goal commitment based on the earlier work of Campion and
Lord (1982) and Locke, Shaw, Saari, and Latham (1981). Goal commitment, according
to Hollenbeck and Klein (1987), is “the extension of effort, over time, toward the
accomplishment of an original goal and emphasizes an unwillingness to abandon or lower
the goa ” (p. 213). In this study, employees are given a goal of improving or maintaining
their health both explicitly and implicitly through the incentive system.

In 1968, Locke was among the first to propose the moderating effects of goal
commitment on the relationship between goals and performance by proposing that
performance is only affected by goals when employees are committed to the goal.

However, little research has been conducted to test this relationship (Hollenbeck & Klein,

16

1987). Regardless of the limited research conducted, it is very important to consider the
effects of commitment on behaviors. As illustrated by Locke and Latham (1990), given
the same goal, “it is not necessarily the case that all subjects are equally committed to
their personal goals” (p. 216).

In conclusion, we will explore whether employees under each type of benefits
system are committed to maintaining or improving their health levels. In addition, we
will investigate whether the incentive-based group is more committed to improving their
health-related outcomes than the traditional group. We believe that this will be the case
as they are subjected to norms that encourage healthy behaviors, are encouraged to
believe that they have control over their health, are given more control over their benefits
costs, and are encouraged to take an active role in determining their own individual
outcomes.

Fairness

Fairness is an individual’s assessment of the appropriateness of outcomes
received, the processes by which the outcomes were distributed, and the interpersonal
treatment and information the individual received concerning the decision, process, or
outcome of interest (Greenberg, 1994). Following the direction of this definition, the
research relating to the distribution of outcomes, commonly termed distributive justice,
will be reviewed first followed by a review of the literature on procedural justice, or
perceived fairness of the process by which a decision is made.

Distributive Justice. The underlying assumption of distributive justice is that
peoples’ perceptions about the fairness of what they obtain (their outcomes) are strongly

influenced by how they believe those outcomes are distributed across people. In

17

particular, those who invest more should receive more, where investments are broadly
defined as putting in more effort, having more education, training or seniority, or any of a
number of possible inputs. An individual’s perception of the fairness of the distribution
of outcomes is commonly explained through equity theory (Adams, 1965; Homans,
1961; Walster, Berscheid, & Walster, 1973). Equity theory operates on the premise that
people believe that they should receive rewards consistent with their inputs relative to a
referent comparison (Adams, 1965; Cohen, 1987), commonly someone in the same or a
similar job.

An example illustrating the application of equity theory involves examining
employee reactions to compensation. According to the theory, if an employee is working
harder at a job than a peer is in an equivalent job, that employee will expect to receive
higher outcomes, perhaps in the form of a bonus or praise for his or her hard work.
However, if employees perceive that they are receiving too little in accordance with their
outputs, these employees may compensate through a variety of behaviors (Greenberg,
1990a). Some notable compensatory behaviors are lowered job performance to the level
of referent others and stealing company time, resources, or money (Greenberg, 1993b).
On the other hand, if an employee perceives that he or she is earning more than peers
who are working just as hard or harder, he or she may raise performance to maintain a
feeling of equity or, more likely, change the person or people to whom he or she is
making comparisons. The goal, in any case, is to maintain a feeling of equity through
behavioral or cognitive modification.

Perceptions of equity or distributive justice in personnel selection arise from a

combination of hiring expectations and the outcome of the hiring decision (Adams, 1965;

18

Gilliland, 1993). If employees feel that they were not hired because they were
unqualified then no feelings of inequity or unfairness should arise. However, if
employees feel that they were equally qualified for the job, but unhired for another
reason, then feelings of injustice may arise resulting, possibly, in negative opinions of the
organization. If employees feel unjustly hired, that they were unqualified, but hired for
an external reason, we would expect that feelings of injustice would also arise resulting
in, perhaps, increased performance, dissatisfaction, higher turnover intentions or lower
organizational commitment (Gilliland, 1993). However, research suggests that this is not
the case. Employees who receive positive or favorable outcomes may experience some
initial feelings of injustice, but they quickly resolve these feelings by altering their beliefs
of why they were hired. These studies suggest that employees who receive negative
outcomes are much more likely to perceive distributive injustice than those receiving
positive outcomes.

The key to this paradox is the outcome one receives. The outcomes related to a
procedure or decision are critical to the formation of fairness perceptions (Brockner et al.,
1994; Greenberg, 1987, 1990a, 1993b; Leventhal, 1976; Shapiro & Buttner, 1988). If
an outcome is considered negative, it is far more likely that the individual will respond
with feelings of inequity or unfairness than if the outcome is positive (e. g. Greenberg,
1987). Perceptions of fairness are manifested by the general tendency for people to
perceive as agreeable those outcomes that best serve their interests.

If outcomes are positive, individuals are likely to consider the distribution fair
according to egocentric or self-serving biases (Greenberg, 1983; 1994). Self-servin g

bias is a distributive justice term that explains how individuals often assume that they

19

deserve higher outcomes than referent others. Gilliland (1994), in his study on the
distributive justice of a selection system, found that individuals who received the job
were more likely to have positive impressions of the fairness of the selection system
regardless of the job-relatedness of the process. Those who did not receive the job, but
expected to, experienced increased negative perceptions of distributive justice when the
system was obviously job-related and the decision to not hire was evident. In this study,
a negative outcome, not getting the job, but expecting to, predicted negative distributive
justice perceptions. The outcome influences an individual’s reaction to a decision or
process regardless of whether one is considering distributive or procedural justice
perceptions (Shapiro & Buttner, 1988). However, the relationship between procedural
justice and outcomes is quite complex with the treatment and information an individual
receives when confronted with a potentially negative outcome mitigating some feelings
of injustice (Brockner, Wiesenfeld, Stephan, & Hurley, 1997). The issues concerning
procedural justice and outcome favorability will be discussed further when we discuss the
nature of procedural justice in greater detail.

The distribution of outcomes is of interest in the present study because health-
bascd outcomes are distributed differentially, which could cause feelings of inequity
when employees feel that they are not receiving what they deserve. Employees who
believe that their health quotients do not reflect their overall health may feel that they
have been inequitably treated. However, there is another aspect of justice that may
further facilitate our understanding of employee reactions to this incentive system.

Researchers have also considered the effects of the treatment and information individuals

20

receive concerning the distribution of outcomes on fairness perceptions, employee
behaviors, and attitudes (Greenberg, 1982; Greenberg et al., 1991).

Procedural Justice. According to justice theorists (e.g., Leventhal, 1973), the way
in which outcomes are distributed across people is only one of two major sources of
fairness perceptions; the other source stems from a broader set of conditions that guide
the way in which people are treated when outcomes are provided to them and the process
by which a decision is made. This is termed procedural justice or procedural fairness.

The information people have about the way rewards are distributed and the
interpersonal style of those administering the outcomes are but two of the factors that
combine to influence perceptions of fairness regarding procedures (Greenberg, 1994).
These process-oriented sources of fairness incorporate factors involved in arriving at and
carrying out decisions related to outcome distribution as opposed to factors related to
whom the outcome itself is distributed (Thibault & Walker, 1978). Procedural justice is
often referred to as the social determinant of justice (Bies, 1987; Greenberg, 1990a, b, c,
1993a, 1993b; Greenberg, Bies, & Eskew, 1991; Greenberg & McCarty, 1990).

These social factors not only are of interest as they affect general perceptions of
the fairness of situations, but also because they may impact independently on the degree
of positive or negative affect associated with outcomes. Peoples’ acceptance of negative
outcomes is affected by their beliefs about the fairness of the information they receive
about the outcomes (informational justice) and the interpersonal conditions that
surrounded the distribution of the outcomes (interactional or interpersonal justice) (e. g.,

Greenberg, 1983). This point is elaborated below.

21

Informational justice is related to how the rewards, decisions, or procedures are
determined (Thibaut & Walker, 1975). Informational justice is what information an
employee is given concerning how any outcomes (decisions, procedures, or rewards) that
are salient to an individual are determined, distributed, or decided. The information an
individual is given concerning an outcome has strong impacts on how a procedure or
event is perceived. For example, Landy, Barnes-Farrell, and Cleveland (1980) found that
employees’ perceptions of the performance evaluation process were stronger
determinants of their fairness perceptions than the actual evaluation, whether positive or
negative. These findings suggest that instilling positive employee perceptions about
procedures may lead to increased perceptions of fairness.

Procedures are perceived as fair when resource distributions are consistent across
persons over time, free from bias, based on accurate information, correctable,
representative of all recipients’ concerns, and based on prevailing moral and ethical
standards (Greenberg, 1987; Leventhal 1976, 1980). Resource distributions that are
inconsistent with Leventhal’s guidelines for fair procedures are found unfair
(Barrett-Howard & Tyler, 1986; Greenberg, 1987; Sheppard & Lewicki, 1987).

Using fair procedures does not determine fairness perceptions alone. These
perceptions are also likely to be affected by the perceived value of the outcome to form
an overall perception of fairness. Greenberg (1987) found that in high and medium
monetary award conditions all participants found the process fair. However, in low
award conditions the procedure determined fairness perceptions. Thus in cases of
unfavorable outcomes, perceptions of fairness may be improved by utilizing fair

procedures. Using fair procedures alone is not enough to explain perceptions of

22

procedural justice. How individuals are treated, interactional justice, when presented
with a novel situation or unfavorable outcomes may also play a role in perceptions of
fairness.

As mentioned earlier, interactional justice may also affect employee perceptions
of procedural justice. Interactional justice relates to the treatment individuals receive
concerning a decision or new procedure. Negative outcomes were more readily accepted
when employees felt that the decision-makers were sensitive to general employee views
(Tyler, 1988), the decision-makers were honest (Bies, 1986), and employees affected by
the decision or procedure were treated in a courteous and civil manner (Bies & Moag,
1986).

Greenberg (1990) found that employees who received pay cuts had higher
perceptions of fairness when the situation was thoroughly explained to them with
sensitivity and understanding than when the decision was simply presented without
consideration. Other research also suggests that explanations for negative outcomes
mitigate people’s reactions to those negative outcomes (Cropanzo & Folger, 1989; Folger
& Martin, 1986; Shapiro & Buttner, 1988) These studies once again showed that using
honest explanations presented with sensitivity greatly reduced adverse reactions to
negative outcomes including theft, lower performance, decreased commitment and
lowered trust in the supervisor or decision makers (Greenberg, 1993b; Konovsky &
Cropanzo, 1991). The preceding reactions to injustice present another aspect of justice
research, the effects that fairness perceptions have on employee behavior.

Acceptance of negative outcomes is also facilitated by procedural justice attempts

which include thoroughly explaining with sensitivity how and why a decision or

23

procedure was implemented (Bies & Moag, 1986; Brockner et al, 1994). Greenberg
(1994) investigated the acceptance of a work site smoking ban and found that heavy
smokers, those experiencing the most negative outcome, were strongly affected by
procedural justice actions, informational and interactional attempts by the organization to
facilitate acceptance. Whereas the nonsmokers, unaffected by the procedure, considered
the smoking ban as fair and were thus unaffected by attempts to increase procedural
justice perceptions. Thus, negative outcomes do influence fairness perceptions, but low
fairness perceptions can be mitigated by attempts to increase perceptions of procedural
justice. In addition, procedural justice attempts will have a greater impact on acceptance
of negative outcomes when outcomes are more severe rather than less (Brockner et al,
1994; Shapiro et al, 1994). Another relevant finding suggests that feelings of resentment
are triggered by the receipt of negative outcomes from a decision or procedure to
accomplish a goal that could have been equally accomplished through a decision or
procedure with less severe outcomes (Fol ger, 1986).

Overall, it is important that employees perceive aspects of their work environment
as fair as the consequences of unfairness can be costly to organizations. There are a
number of guidelines and research findings to suggest how to develop a fair system and
to distribute outcomes so that they are perceived fair.
Perceived Control

When we discussed research suggesting that commitment had to be present before
behaviors were exhibited we mentioned two theories in conjunction with the health
psychology domain. The two theories both detailed different antecedents of commitment

to changing behaviors. The second theory we discussed was the Theory of Reasoned

24

Action (T RA) and was developed by Ajzen and Fishbein (1980). This theory, as we
discussed earlier, proposes two requirements for commitment, individual and norm
beliefs in favor of the behavioral change, not specifically health behavior change. We do
believe that these are requirements for commitment, but we think the first theory we
discussed in the commitment section holds some especially interesting suggestions for
our present research.

The Health Belief Mode, or HBM for short, was created by J anz and Becker
(1984). As we described earlier, there are four requirements for commitment to changing
health behaviors. The first two relate to whether or not the individual believes he or she
will develop some serious side effects or illness due to some behavior and that any illness
would be quite serious. The last requirement is that the benefits of changing relevant
behaviors outweighs any inconvenience. However, it is the third requirement in which
we are most interested. The third requirement for commitment is that the individual has
to perceive that changing his or her health-related behavior would actually decrease the
risk of becoming ill or developing side effects. This factor is called perceived control and
in our study we believe that employees must perceive that they have control over their
health-related outcomes before they will commit to changing their behaviors.

This notion of perceived control, while still relatively new to the justice and
industrial/organizational psychology domains, has also been researched in conjunction
with attribution theory (W einer, 1985) in both health and educational psychology
research (Baum et al., 1997). Attribution theory has been used to explain individuals
reactions to their performance, whether good or bad. Attribution theory suggests that

individuals attribute their successes and failures based on three general properties: locus,

25

stability, and controllability. Locus of attribution suggests that individuals decide
whether their performance or health was due to something within themselves or
something external to themselves, possibly within the environment (Heider, 1958). The
second property that determines the attribution of causes is stability and refers to whether
a cause is considered stable, i.e. constant, or unstable, fluctuating and not a permanent
state (W einer et al. 1971). The last property and the one of most interest in this research
is controllability and suggests that certain conditions are subject to volitional control
while others are not (Weiner, 1979).

The research surrounding controllability suggests that when something is
considered under a person’s control he or she is more likely to be motivated to try to
control that behavior. The motivation to control those behaviors stems from an
expectation that changing the behavior is possible and will lead to some positive (or
negative) consequences for their actions (for a review see Weiner, 1985). These findings
suggest several key points for the present research. If individuals perceive that they have
control over their health levels they will likely be more motivated to try to change their
health-related behaviors (i.e., more committed to change). In addition, if individuals
perceive that they have control over their health levels and that any changes made will be
detectable by the measurement methods they will be more likely to change their
behaviors. When direct control over outcomes is perceived, as is possible in the present
study, commitment to the program or decision should be evident (Lind & Tyler, 1988).

Placing control over outcomes into the hands of employees has some interesting
implications for research surrounding employee perceptions of fairness in the workplace

related to health insurance programs. As discussed in the auto insurance example

26

presented in the introduction, the health insurance considered in the present research is
based on the assumption that individuals have control over their health. But individual
control over many aspects of health may be limited. Research suggests that some portion
of individual health is determined by, in part, genetics, not by the choices individuals
make concerning their behaviors (Baum et al., 1997; Braunwald, 1994). Individuals
within programs such as these may feel that even when they institute the changes
suggested following their health appraisals they do not change their health (i.e.,
cholesterol, blood pressure, and body fat ratio) dramatically enough to affect their
monetary outcomes. Proponents of these programs may counter this argument with the
notion that there are five other aspects upon which an individual ’3 health index is based
that are directly under an individual’s control: nutrition, motor vehicle safety, alcohol and
tobacco usage, and fitness level. There are two counter arguments to that belief. First,
research does suggest that both alcohol and tobacco use and addiction have some genetic
and biological predisposing factors (Dawson et al., 1992; Morse & Flavin, 1992;
Silverstein et al., 1982). Second, in the system to be studied here these five dimensions
in conjunction were assigned less than half of the importance given to blood pressure,
cholesterol, or body fat ratio when calculating the health quotient.

Returning to our discussion on fairness or justice perceptions, we know
procedures or processes are considered fair when outcome distributions are consistent
across persons over time, free from bias, based on accurate information, correctable,
representative of all recipients’ concerns, and based on prevailing moral and ethical
standards (Greenberg, 1987; Leventhal 1976, 1980). If, in turn, individuals do not

perceive that they have control over their outcomes they may also not perceive that the

27

procedures utilized prior to and during the distribution of outcomes are fair, that they are
not consistent over time, not free from bias, not based on accurate information, and not
correctable. We have argued here that many of these phenomena also are likely to
operate with respect to behaviors related to displaying healthy behaviors. The basis for
this position is that there is a strong expectation in the United States that employers will
provide health insurance for their employees. Therefore, when a health insurance system
is offered to employees that differs in a number of ways from “typical” employer systems
and affects such an important outcome as money, that new system is likely to be judged
in terms of fairness, especially when the outcomes are negative. Furthermore, when
components of that system attach rewards to the levels of employee health, beliefs about
personal control over health levels are likely to also affect perceptions of the fairness of
the system. For this and other reasons, there was reason to believe that the degree to
which employer health insurance systems are perceived as fair will relate to the way

employees respond to health insurance systems. Specific hypotheses will follow.

28

Chapter 2

HYPOTHESES

The hypotheses and models that will be proposed in this study center around the
issues of fairness, perceived control, commitment, and behavioral change in the context
of an incentive system. We are very interested in the specific effects that the incentive
system has on employee perceptions of fairness. To address these issues we will examine
fairness from three distinct perspectives. First, we will examine the effects of four
general variables on fairness perceptions. Following the presentation of those four
propositions, we will present a framework for understanding the interrelationships
between fairness, perceived control, and commitment as we believe that these three
employee perceptions are critical when considering behavioral change. Finally, we will
expand the discussion of fairness, perceived control, and commitment to include the
effects of the incentive system itself.

The Roles of Outcome, Age. Incomeband Accuracy Perceptions on Fairness Perceptions.

As we believe that fairness perceptions are very important in understanding
employee reactions to benefits systems, we will begin by proposing several general
variables that are expected to influence employee perceptions of fairness in this study.
Specifically, we propose to investigate the effects of outcome, age, income, and accuracy
perceptions on employee perceptions of fairness.

The Effects of Outcomes on Employee Perceptions of Fairness. Outcomes have
long been accepted as an important aspect in the formation of fairness perceptions.

Reviewing the literature discussed previously, we have evidence that individuals have a

29

tendency to possess self-serving biases (Greenberg, 1987) which lead them to perceive
themselves as deserving the most positive outcomes. In line with theories of self-serving,
or egocentric biases, Greenberg (1994), when investigating the effects of a work site
smoking ban, found that individuals receiving negative outcomes were much more likely
to perceive the process, situation, or outcomes as unfair than someone who receives more
positive outcomes.

The outcomes in this study are especially relevant to justice perceptions. Each
employee within the incentive-based system will receive an outcome. That outcome, as
aforementioned, will vary from negative to positive twenty-five dollars. Based on
previous research, it is expected that each individual’s outcome will influence his or her
perceptions of the fairness of the system.

H1: Within the incentive system, employees’ perceptions of the fairness of

their benefits system will be positively affected by the value of their

individual outcomes.

Effects of Age on Employee Perceptions of Fairness. Because of the nature of
the incentive program, an employee’s age may influence his or her perceptions of the
fairness of the benefits system. Research suggests that employees who are older may be
less convinced of their ability to change their overall health (Baum et al., 1997; Taylor,
1995). With that in mind, expectancy theory (Vroom, 1964) would predict that
employees who believe that they cannot change their health level should not expect that
they would get the desired outcomes, more money, and would thus not attempt to change
their health related behaviors. For older individuals this suggests that the system, which

operates on the assumption that people can control their health, is not fair.

30

In addition, fairness research has repeatedly supported Leventhal’s (1976)
guidelines for fair procedures which specify that resource distributions must be free from
bias if they are to be considered fair (Greenberg, 1987). A system that discriminates
against those that believe they have little control over their health levels would be
considered unfair.

H2: Perceptions of control will mediate the relationship between age and

fairness perceptions such that older employees within the incentive-based

system will believe that they have less control over their health levels than
younger employees and subsequently perceive incentive-based benefits
programs as less fair.

Effects of Income. In the situation under investigation, monetary outcomes are
not individually assigned according to income level so that all participants experience
equivalent levels of incentives or disincentives depending on their health quotient score.
This aspect of the incentive program leads to a situation where even though some
employees are receiving the same outcomes, they are likely to perceive the personal value
of the outcome differently, depending on their personal financial state. For example,
losing $25.00 per pay period for one earning $100,000 per year is likely to be seen as less
serious than for someone earning $25,000 per year.

In addition, research does support the finding that individuals in lower
socioeconomic groups are less educated about their health and may not perceive that the
consequences of their health-related behaviors are as serious as they could be (Baum et
al., 1997). Subsequently, individuals from these groups smoke more, consume more

alcohol, are more obese, and exercise less than higher socioeconomic groups (NHLBI,

31

1994). If larger portions of lower socioeconomic groups are receiving disincentives or
pay deductions, the incentive-based program may be viewed as unfair.

H3: Employee income level will moderate the relationship between health-

related outcomes and employees’ perceptions of the fairness of their benefits

system such that the positive correlation between health-related outcomes
and fairness perceptions will be stronger within lower income groups than
within higher income groups for employees within the incentive-based
system.

The Relationship between Accuracy and Fairness. The accuracy of the health
measurement methods employed by the incentive organization to determine health levels
can influence an individual’s perceptions of fairness. An employee’s perception that the
measurement techniques are inaccurate is hypothesized to decrease that employee’s
perception of fairness. The literature reviewed on fairness was extensive, but there are
common threads that run through all fairness research. Leventhal ( 1976) laid the
foundation for much of the fairness research when he identified the characteristics of
procedures that are considered fair. Leventhal’s guidelines have received much empirical
support (Barrett-Howard & Tyler, 1986; Greenberg, 1987; Sheppard & Lewicki, 1987).
Included in those characteristics are resource distributions that are based on accurate
information. In this case, health levels that are determined through methods which are
perceived inaccurate would not lead to resource distributions that were based on accurate
information, thus violating one of Leventhal’s guidelines for fair procedures.

Measurement accuracy and reliability is of top concern for many medical

professionals (Taylor, 1985). There are two areas of debate that are of interest in our

32

study. The first surrounds the accuracy of objective indicators of health. More
specifically, in the incentive-based benefits system blood tests used to measure
cholesterol levels may be perceived as inaccurate based on research suggesting that they
are notoriously unreliable (Lenfant, 1986; NHLBI, 1994). Similar issues surround
objective measures of body fat/muscle tissue ratio. Specifically, caliper tests used to
determine fat/muscle tissue ratio are considered among the most unreliable measures
available (NHLBI, 1994). A second set of issues surrounds the use of subjective or self-
report measures of health behaviors (Taylor, 1985). People are likely to distort their
health behaviors when given a self-report survey. For exarnple, individuals asked to
report their daily health-related behaviors were very inaccurate and, in almost all cases,
underreported the amount of food and alcohol they consumed and overreported the
amount of exercise they engaged in, even when they were not being judged or penalized
for their reports (Jose & Anderson, 1990; Naditch, 1984).

Our incentive sample is one in which many of the subjects are probably very
aware of these kinds of issues as they are closely tied to the medical field. Suspicious
about the use of inaccurate measurement methods and faulty self-report surveys are likely
to decrease an individual’s perceptions of the fairness of the benefits system.

H4: Employees, within the incentive-based benefits system, who believe that

the measures used to determine their health levels were more accurate are

expected to report higher perceptions of the fairness of their health benefits

system than those who believe the measurements were less accurate.

33

Hypotheses Relating to the Relationships between Fairness Commitment and Perceived

 

$9119.]

At this point we will present a model for examining the interrelationships between
fairness perceptions, perceived control, and commitment. The model will be broken
down into smaller hypotheses that address the specific paths specified in the model itself.

An overall representation of the model can be found in Figure l.

34

 

 

Fairness

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Perceptions +
T Commitment + Health-related
+ to Maintain or % Behaviors
Improve Health
+
Perceived
Control
Figure 1.

Model representation of the relationship between fairness, perceived control,

commitment and health-related behaviors.

 

35

The Relationship between Commitment and Behavior. From the literature

reviewed on commitment (e.g., Hollenbeck & Klein, 1987) and its effects on behaviors,
we have evidence that not all employees are committed the same to a goal and that the
level of commitment does, in fact, influence employee behaviors (Ajzen & Fishbein,
1980; Locke, 1968). In fact, employee performance is affected when employees are
committed to the goal. For our study, this suggests that employees who report higher
commitment to maintaining or improving their health will also report healthier behaviors
and vice versa.

H5: Employees who are more committed to maintaining or improving their

health will report performing more healthy behaviors than those less

committed.

The Relationship between Fairness and Commitment. Equity theory (Adams,
1965), a central theory in justice research, suggests that if an individual believes that if he
or she put in a certain amount of effort he or she would receive certain predetermined
outcomes, he or she is more likely to be committed to providing the proscribed inputs.

For our study, this suggests that individuals who believe that their benefits system
is fair are more likely to be committed to maintaining or improving their health in the
future. On the other hand, individuals who perceive that the system is not fair because
they will not receive a more favorable outcome with the proscribed inputs are unlikely to
be committed to maintaining or improving their health.

H6: Employees who report more positive perceptions of fairness will report

higher levels of commitment to maintain or improve their health-related

behaviors.

36

The Relationship between Perceived Control and Commitment. As described in
the literature review, perceived control is an individual’s belief that he or she has control
over his or her outcomes. In this study, we are interested in employees’ beliefs that they
can control their health levels and, in the case of the incentive group, the monetary
incentives received. Much of the literature reviewed previously relates to this hypothesis.

Health psychology has long been interested in why or why not individuals commit
to improving their health behaviors. One model of intentions and behavior, the Health
Belief Mode (J anz & Becker, 1984) proposed that, among other requirements, individuals
must perceive that by changing their behaviors they have the ability to reduce the risk of
developing serious side effects of illnesses. We propose that the same is true in our
sample, that employees have to believe that modifying their health-related behaviors
would actually improve their health before they commit to changing those behaviors.

Other support for this hypothesis comes from both fairness and motivation
research. Support exists for the assumption that strong commitment should be present
when employees perceive that they have direct control over their outcomes (Lind &
Tyler, 1988). In addition, equity theory (Adams, 1965) and expectancy theory (Vroom,
1964) both suggest that employees who perceive that they have control over their
outcomes through behavioral change are more motivated to attain those outcomes and
devote more of their efforts to attainment. If control is not perceived, then commitment
to modifying behaviors is not likely.

H7: Employees who perceive that they have more control over their health-

related outcomes will report higher commitment to maintain or improve

their health-related behaviors.

37

The Relationship betweerfirimess and Perceived Control. Support for the
relationship between employee fairness perceptions and perceived control can be found
by reexarnining literature reviewed on equity theory (Adams, 1965) and procedural
justice (Greenberg, 1987).

Equity theory would suggest that if an employee’s behavior results in the desired
outcome then he or she should consider the procedure fair. On the other hand, if an
employee puts in a lot of effort and does not receive the desired outcomes, the system is
likely to be viewed as faulty because he or she has not been equitably rewarded for his or
her inputs.

Procedural justice research also suggests that something is considered fair if it is
under an individual’s control. If individuals perceive that they have control over their
health the system would be believed to operate on a moral and ethical standard that is
consistent across individuals (guidelines for fair procedures by Leventhal, 1976) and
should, theoretically, be considered fair. Believing that one can change one's outcomes,
improve one’s health, should accompany positive perceptions of fairness, whether
distributive or procedural justice.

H8: Employees who perceive that they have more control over their health-

related outcomes will report more favorable perceptions of the fairness of

their benefits program than those who perceive less control.
Putting the Pieces Together: Understanding How Incentive Systems Affect the
Relationships between Fairness, Perceived Control. and Commitment

The following hypotheses are directed at understanding how the incentive-based

health benefits system involved in this study would affect the predicted relationship

38

between perceived control, fairness perceptions, and commitment. Of specific interest
are the effects of the incentive system on fairness perceptions.

To begin with, the effects of the incentive system on employee health behaviors,
commitment, and perceived control will be proposed. Then, two alternative models of
the effects of the incentive system on fairness perceptions will be proposed. The present
literature does not clearly predict which of these would best fit the unique incentive
system involved in this research.

The Effects of the Incentive System on Employee Health Behaviors. As
previously discussed, incentive systems encourage employees to perform specific
behaviors. Several theories explain why incentive systems have these effects.

Classical reinforcement theory (Skinner, 1969) operates on the assumption that if
you reward a behavior it is more likely to be repeated. This suggests that there are two
potential incentives inherent in the system under investigation that may encourage
employee behaviors. Being healthy is an incentive to many and the addition of the
monetary incentives is further encouragement to employees to improve or maintain their
health.

Expectancy theory (Vroom, 1964) is similar to reinforcement theory and would
also predict that if an individual expects to receive certain desirable outcomes in return
for some behavior he or she is more likely to exhibit that behavior.

In addition, employees who are covered under the incentive system receive
explicit feedback about their health and how to improve or maintain it. This fact should
increase the differences between the two benefit type groups, such that those who are

covered under the incentive-based program have both specific directions and the financial

39

incentive to improve their health. Research supports this hypothesis. Specifically, we
discussed a study in the introduction which showed that individuals giving self-rewards
and self-punishments reduced their alcohol consumption significantly more than their
control peers or those who solely received information on the adverse side effects of
alcohol consumption (Bigelow et al., 1973; Griffiths et al., 1978).

The conclusion that can be drawn from this body of research is that the incentive-
based system is set up to reinforce (both negatively and positively) employee health
behaviors and those incentives should positively influence employee health-related
behaviors.

H9: Employees with an incentive-based benefits system will report

performing more healthy behaviors than those under a standard health

benefits system.

The Effects of the Incentive System on Commitment. Once again, the
individuals who are covered under the incentive-based health benefits system are
expected to report higher commitment to maintaining or improving their health than those
(under a more traditional system. The logic behind this hypothesis is similar to that of
hypothesis 9 which rests on the theories of reinforcement and expectancy, but support can
also be found in the goal commitment, goal setting, and behavioral change research
domains.

Briefly, goal commitment provides evidence for the proposition that difficult,
specific goals best influence performance (Austin & Vancouver, 1996; Locke & Latham,
1990). In the present situation we have a goal that is either perceived or actually difficult

(staying or becoming healthy), specific directions to improve, and a task that is

40

cognitively simple. These characteristics of the system lead us to expect that employee
health behaviors will be higher among the incentive-based group.

In addition, we discussed the Theory of Reasoned Action (Ajzen & Fishbein,
1980) which proposed a model to understand why individuals do or do not behave in
certain ways. They proposed two antecedents to commitment or intention to change
behaviors. Both relate to our study. The first antecedent requires that the individual have
a desire to change his or her behavior due to a personal belief that the new behavior is
more favorable than the present, for any variety of reasons. The second antecedent
requires that the individual perceive that it is socially preferred to perform the new
behavior than the present behavior. We believe that the incentive group will have a
stronger sense of societal pressure due to propaganda surrounding the program. In
addition, we believe that the information and education involved in conjunction with the
incentive-based program will increase a personal desire to behave more healthy.

In conclusion, the research suggests that employees within the incentive-based
benefits system should report higher levels of commitment due to the decision/goal
setting process and the nature of the program itself present in the incentive-based system.

H10: Employees with an incentive-based benefits system will report higher

levels of commitment to maintain or improve their health-related behaviors

than those under a traditional benefits system

The Effects of the Incentive System on Perceived Control. The presence of the
incentive system is expected to have positive effects on employee perceptions of control

over their health. Literature in support of this hypothesis is not known to the authors, but

41

support can be logically deduced by examining the nature of the system and considering
the HBM model of behavior (J anz & Becker, 1984).

The incentive-based health benefits system is based on the notion that employees
have control over their health. According to the Health Behavior Mode theory of
behavior, control is a requirement for commitment and subsequent behavioral change.
To perceive control, individuals have to perceive that a change in behavior would reduce
the risk of serious illness or other health-related side effects. To encourage employees to
believe that they have control, classes and programs are offered to direct employees in
methods to improve their health and specific directions are given to each employee about
how to improve their health when the assessment results are distributed. These
employees receive more, theoretically, information about their health and improvement
methods than those covered under a more traditional plan where health is not regularly
discussed. Thus, it is logical to hypothesize that,

H11: Employees with an incentive-based benefits system will report more

perceived control over their health-related outcomes than those under a

standard health benefits system.

Two Alternative Models to Explain the Relationship between Incentives and
Fairness. Understanding and hypothesizing the effects of an incentive-based system on
employee perceptions of fairness is not quite as easy as the previous hypotheses. The
literature in our domain is ambiguous concerning what we can expect in terms of fairness
perceptions when confronted with a system such as the incentive-based system in our
study. Although it is expected that the incentive system will impact perceptions of

fairness, the impact may be either positive or negative. If the general expectations are

42

that being paid more if healthy are positive, we expect the contingent benefits to be more
favorably viewed than the standard one. If, on the other hand the contingency is
perceived as inconsistent with the expected norms for the distribution of health benefits,
the relationship should be negative. Thus, at this time we cannot propose a specific
hypothesis describing fairness reactions under the effects of an incentive system because
both a positive and negative relationship appear equally likely to explain how the
incentive-based system influences employee perceptions of fairness.

In the following paragraphs we will present literature and descriptions of the
incentive program that suggest both that the program may be positively related to fairness
perceptions and negatively related to fairness perceptions.

The incentive-based health benefits system has many of the characteristics that
are frequently cited as fair procedures. The review of the literature that we provided
leads the reader to believe that the system under investigation should be considered as
fair and should further increase employee commitment to maintain and improve his or
her health.

Both procedural and distributive justice research suggest that employees covered
under the incentive-based system should report more positive perceptions of fairness than
those employed under the traditional system. Leventhal (1976) provides the guidelines
for fair procedures that suggest that resource distributions that are consistent, free from
bias, based on accurate information, and based on prevailing moral and ethical standards
will be considered fair. All that we have discussed concerning the design of the
incentive-based system under investigation suggests that these requirements were upheld.

While it is potentially true that these are all present in the traditional benefits system, we

43

believe that it is possible that these issues are more salient in the incentive-based sample
as their benefits system is outside the norm and probably subject to closer inspection.

As was discussed in the literature review of faimess/justice theories, there are two
distinct types of procedural justice: informational and interactional justice. This system
may produce high reports of both from employees within the incentive-based system.
Employees within the incentive-based system spend a lot more time discussing and
learning about their benefits system than those within a traditional system. Also, specific
departments are set up throughout the organization that cater to employee needs and
concerns about the incentive system. Both of these aspects of the incentive-based system
should lead employees to perceive their system as more fair than those under a traditional
system.

Finally, distributive justice is often represented by equity theory (Adams, 1965)
which would suggest that those who put more into the system, by taking better care of
themselves, should get the most out of the system. The incentive-based system operates
on this, and all the other aforementioned premises and thus should be considered more
fair. This literature would lead us to hypothesize that employees covered under the
incentive-based system will report higher perceptions of the fairness of their benefits
coverage than those who are covered under the traditional program.

While hypothesizing that employees under the incentive-based system should
report higher perceptions of fairness than those under the traditional program is correct
according to much of the fairness research, there is also some evidence to suggest that

this relationship may be negative.

44

As you may remember, Leventhal’s (1976) guidelines state that a system is fair if
it is moral and ethical, consistent across persons over time, based on accurate
information, and free from bias. These guidelines are for all practical purposes included
in the theoretical design of the system, but employee perceptions of these guidelines are
not necessarily present.

We believe that employees may not believe that this system is moral and ethical
for a variety of reasons. First of all, this is not a traditional method for providing
employee benefits. In some cases, employees under this system have to pay for some
portion of their health insurance and may believe that health insurance is one of their
rights as an employee. Second, employees may not believe that they have control over
their health. There are aspects of health that may or may not be considered under an
individual’s control (Baum et al., 1997; Braunwald, 1994). For instance, research
suggests that obesity (Ezzell, 1995; Faust, 1980; Hakas et al., 1995; Keesey, 1993),
blood pressure (Taylor, 1995), cholesterol (Kannel et al., 1979; Stoney et al., 1988),
alcoholism (Dawson et al., 1992; Morse & Flavin, 1992), and cigarette smoking
(Silverstein et al., 1982) are partly hereditary and not wholly under the control of the
individual. Thirdly, some people may consider this system an invasion of privacy and
health an area of personal life that an organization should not be involved.

In addition, employees may not consider the system consistent across persons
over time. There are two reasons for these inconsistencies that we may want to consider.
First, the health evaluations are done on different days of the week. Each day may have
different climate and weather conditions, different assessment individuals, or different

measurement instruments that may influence the evaluation of health each person

45

receives. In addition, some employees are given special considerations for health
problems. Some employees may feel that these are undeserved or that they have similar
problems and should also be given the same allowances.

Also, employees may perceive that the health evaluations that they receive are not
based on accurate information. As we mentioned earlier, there may be concerns among
employees about the measurement techniques. For instance, some employees may feel
that their health evaluations were inaccurate due to assessor error, faulty measurement
equipment, inappropriate techniques, or simply a bad day.

Finally, employees may not perceive that the incentive-based system is free from
bias. We believe that it is possible that employees may feel that the system is biased
against older individuals, those with chronic health problems, different ethnic groups, et
cetera. It is possible that employees may believe that older people have a more difficult
time changing their health. Also, individuals with special circumstances may not be as
able to take the necessary steps to improve their health.

These issues could continue endlessly, but the point is simple. Even though the
incentive system under investigation is designed to optimize the traditional fairness
guidelines, those relevant characteristics of the system may not be interpreted as intended
leading employees to believe that the system is not fair. This literature would lead us to
hypothesize that employees covered under the incentive-based system will report lower
perceptions of the fairness of their benefits coverage than those who are covered under
the traditional program.

With literature and research existing in support of both a positive and a negative

relationship between fairness perceptions and the incentive-based system, we can’t be

46

specifically propose either hypothesis. Instead, we will propose an exploratory
hypothesis directed at uncovering the relationship between the incentive system and
fairness perceptions in our sample.

H12: Employees within the incentive system will differ in their perceptions
of fairness than employees within the traditional benefits system. The direction of
the relationship is unspecified, with either a positive or negative relationship
plausible.

The negative and positive relationships are depicted in Figures 2 and 3.

In the following section, the method employed to test these proposed hypotheses

will be described including the method used to collect subjects and data as well as the

measures adapted and developed.

47

 

Fairness

 

 

 

 

 

 

 

Perceptions
+
+
+1
Incentive + Perceived + Commitment Health-
System ——’ Control —> Related
Behaviors

 

 

 

 

 

 

 

 

 

Note. Traditional System is coded as l and Incentive System is coded as 2.

Figure 2.

Model depicting a positive relationship between incentive-based health benefits gstem

and fairness mrceptions.

48

 

Fairness

 

 

 

 

 

 

 

Perceptions
+
+T
Incentive + Perceived + Commitment Health-
System —’ Control . Related
Behaviors

 

 

 

 

 

 

 

 

 

Note. Traditional System is coded as 1 and Incentive System is coded as 2.

Figure 3.

Model depicting a negative relationship between incentive-based

health benefits system and fairness perceptions.

49

Chapter 3
METHOD
mm

In this study, data were collected from a sample of employees from a hospital in
the mid-west using an incentive-based health benefits system (N=243) and another
sample of hospital employees in the same city that provided employees with traditional
health insurance (N=273). Approximately 5000 people were employed by the hospital
using the incentive-based program at any given time. From a list of current employees a
sample of 800 employees were randomly selected. A check of personnel records
indicated that only 524 of these were full time employees eligible for participation in the
health care system. Therefore, this latter sample was considered the initial sample. Of
those in the sample, 243 provided useable data. Table 1 illustrates the similarities
between the sample and the entire hospital. These data suggest that the sample is
reasonably representative of the entire hospital workforce. The sample is composed of a
slightly higher percentage of upper-level employees who are likely to be more highly
paid.

Within the hospital utilizing a traditional benefits system 500 employees were
randomly selected from a work force of approximately 2800 employees. Two hundred
seventy-three responded. Table 2 illustrates the demographic makeup of those
respondents. No data were available about the demographics of the entire hospital.
Since this hospital and the incentive one were located in the same large city and in the
same general area they drew from the same labor force. There is no reason to suspect

that their willingness to volunteer would not be similar in both hospitals. Therefore, it

50

seemed reasonable to conclude the sample in this hospital represented reasonably well,
the work force from which it was drawn just as was the case in the incentive hospital. In
both samples, any respondents who were not employed in full time positions were
eliminated from the study. Both samples contained high percentages of female

respondents.

51

Table 1.

Demographic Characteristics of Sample with Incentive-based Benefits.

 

 

 

 

 

 

 

 

 

 

 

 

 

Demographic Characteristics Sample Entire Hospital Population

Female 75% 76%

Ethnic/Race 95% 94%

(% White)

Mean Age 41.3 Data Not Available

Job Type:
Management 14% 9%
Professional 1 6% 1 2%
Registered Nurse (RN) 16% 19%
Technical 18% 16%
Office/Clerical 26% 3 1 %
Maintenance 1 1 % 13%

 

 

52

 

Table 2.

Demographic Characteristics of Sample with Traditional Health Benefits.

 

 

 

 

 

 

 

 

 

 

 

 

Demographic Characteristics Sample
Female 74.5%
Ethnic/Race 95.5%
(% White)

Mean Age 39.76

Job Type:
Management 14.0%
Professional 14.8%
Registered Nurse (RN) 16.0%
Technical 17.7%
Office/Clerical 25.9%
Maintenance 10.7%

 

 

53

 

Incentive-based Program Description

As was mentioned earlier, in the hospital using the incentive-based health benefits
program, a health assessment evaluated health level by evaluating employees and their
dependent spouses on eight health-related dimensions. These were: blood pressure,
cholesterol ratio, body fat/muscle tissue ratio, cardio-vascular fitness level, tobacco
usage, motor vehicle safety, alcohol consumption, and nutrition once a year during a two-
week assessment period. The first four of these eight health dimensions were measured
with physiological measures or a timed walk, and the last four were assessed by self-
report.

The assessment occurs once a year during a two-week period. During that period
all employees and their dependent spouses covered under the incentive-based health
benefits system are asked to schedule an appointment for their assessment. During that
assessment each of the eight health dimensions are assessed through either self-report or
objective measures. (For a detailed description of the assessment procedure, refer to the
“Health Quotient” section in the introduction.)

Employees were provided feedback on their overall health level and suggestions
for improving their health status. In that feedback, employees were notified how much
money they would be charged or awarded based on their health and their spouses’ health
levels. Employees were also encouraged to participate in health programs within the
hospital, such as smoking cessation and weight loss, and were provided discounts on
health equipment and club memberships. Employees and spouses were provided with the
methods to try to improve or maintain their health.

The program was actually introduced in 1993 when employees were informed of

54

the new health benefits system that would be officially started in 1994. At this time, they
were encouraged to go through the annual health assessment so that they could, if they
chose, work on improving their health levels before the actual start of the system.
Employees who did go through the health assessment that first year were given their
health quotient scores as if they would in the coming year and were informed of the
amount of their incentive (or disincentive). Then, in May of 1994, all employees went
through their first official health assessment. The health quotients calculated at the time
were applied to their 1995 salaries. We surveyed the employees during the summer of
and fall of 1996. At this time, the employees were well into their second year of the new
benefits program and were either going through or had gone through the assessments for
1997.

Procedure

The procedure for data collection differed somewhat, depending on the hospital.
We will describe the procedures separately for each hospital sample.

Procedures Followed in the Hospital with the Incentive System. As was
mentioned in the description of the sample, 524 persons were randomly identified from
the list of full time employees in the hospital - the employees eligible to participate in the
health care program. From these individuals, two sets of data were collected. One set
was collected from a survey instrument that was designed to tap employee beliefs and
expectations as well as self-report measures of health and demographic characteristics.
The exact nature of these variables will be described later. The second set of data was
obtained from company records. These included the employees’ health assessment

scores and a number of other personnel records. These too will be described later.

55

Procedurally, the important point is that data were collected from two sources, the
employee and organizational records, and the data had to be collected in a manner that
these two forms of data could be paired for each individual, after the participants gave us
explicit permission to do this. The variables for which it was done under the subheading
of measures in this Method section. Let us turn to the procedure for collecting survey
data.

Two methods for collecting survey data were used in this sample. The first
involved an on-site administration and the second method involved an inter-office
mailing. All 524 employees received an interoffice letter from the researchers describing
the nature of the study and inviting their participation (See Appendix A). This letter
explained that the researchers would be on-site at the hospital to distribute the surveys
during a twenty-four period approximately one month later. All employees were
requested to attend if possible and their supervisors were notified by human resources
within the hospital that all employees should be excused from work for one hour if at all
possible to complete the survey. Within this same mailing, a postcard was included that
allowed employees to indicate whether they would be able to come to the on-site survey
location on a specific date or would like to have the survey mailed to them. In a separate
mailing, the CEO of the hospital wrote a letter requesting that all employees invited to
participate in the study do so for the good of the hospital (See Appendix A).

The procedure for administering the survey on-site focused on insuring employee
privacy while still insuring that the researchers could link these survey data to
organizational records. Each participant received an envelope containing the survey and

other relevant information. Within the envelope, respondents were given two letters, the

56

survey, a consent form, and a large envelope in which to seal the completed survey. The
first letter was from the CEO of the hospital encouraging employee participation and
explaining the benefit of this survey to the organization (See Appendix A). The second
letter explained the purpose of the survey and that the researchers were from Michigan
State University and are independent from the hospital. (See Appendix A) In addition,
this letter stressed that participation was voluntary and explained the necessity of the
consent form. To gain access to employee records for employees covered under the
incentive-based system we asked all participants to complete a consent form specifying
which information within their employee records we would utilize. (See Appendix B.)

A number was printed on the survey and the consent form to serve two purposes.
First of all, we needed to be able to link up survey responses to hospital data on employee
health, specifically the health quotient. At the same time, we also needed to link up
survey responses to other employee data such as absenteeism and other demographic
information. The number also allowed us to avoid requiring participants to print their
name on the survey itself. Confidentiality was strongly stressed; employees were
assured that no one in the organization would be able to identify their responses. In
addition, the organization required that we complete a legal confidentiality agreement
between the insurance provider and the ourselves that insures that all data would be used
responsibly and would be securely kept in both a locked cabinet and stored without any
identifying names attached.

Employees were given as long as necessary to complete the survey. Once

completed, the survey was to be placed in the envelope with the consent form, and given

57

to the researcher present. At this time, the researcher made certain that the survey was
completed and the consent form signed.

At the scheduled on-site survey date, turnout was lower than expected. Only 83
employees providing complete data. Therefore, an alternative survey method was
implemented which involved a mailing of the survey to all employees in the sample who
did not complete the survey during the on-site administration or who previously
requested a survey be mailed to them.

All those who had not yet completed a survey were contacted through interoffice
mail. Included in the envelope were two letters, a survey, and a stamped return envelope.
The first letter from the CEO was the same as that during the on-site survey
administration. (See Appendix A.) The second letter was slightly different and
explained what this mailing was and why it had been done in response to the low on-site
turnout. The surveys were numbered as described for the on-site administration. Since
this was a mailing, it was possible for the researchers to assign participants numbers prior
to the mailing so that we would know who had responded and who had not. In addition,
in cases where the consent form was returned unsigned, the researchers had a way to
identify who that survey belonged to and contact them in the hopes of obtaining a signed
form. Once again, participation was voluntary.

Employees were encouraged to complete the survey as soon as possible and mail
it in the provided envelope to the researchers at Michigan State University. All questions

were directed to the researchers and a telephone number was provided. (See Appendix

D.)

58

Procedures Followed in the Hospital with the Traditional System. The survey
administration in the hospital with the traditional benefits program differed slightly from
that of the hospital with the incentive-based program for two reasons. An on-site survey
administration was not conducted. Also, we did not need to collect any data from
personnel records as there was no health assessment done in conjunction with this
benefits program. The data collection method will now be described.

Five hundred employees were randomly selected from employee records. All
employees in the sample were full-time employees covered by the health benefits
program. All five hundred employees were sent a package at their residences containing
a letter from the CEO of the hospital, a letter from the researchers, a consent form, a
survey, and a stamped return envelope. The letters were almost identical to those sent to
the first hospital. (See Appendix A.) The survey itself was slightly different due to the
fact that these employees were not covered by the incentive-based health benefits
program. The consent form was a formality that allowed us to simply use the collected
survey data and provided the researchers with the opportunity to explain that the survey
was voluntary and that confidentiality would be guaranteed since no personnel record
information would be accessed. (See Appendix C.)

The respondents were instructed to complete the survey in a timely manner. Once
completed, the surveys were to be placed in the pre-addressed stamped envelope along
with their completed consent form. Any questions were directed to the researchers and a

telephone number was provided. (See Appendix E.)

59

Measures

Unless explicitly stated, all variables described below were assessed on the
employee surveys. In the following paragraphs each of the measures utilized in this
study will be discussed in relation to the construct they were used to assess. First,
fairness measures will be discussed followed by descriptions of the measures used to
assess health behaviors, perceived control, commitment, accuracy, and various
demographic variables.

Eai_r_n_§§_s_. Employee perceptions of procedural and distributive justice were
measured using adaptations from previous research based on work by Leventhal (1976,
1980) and Greenberg (1994). From this research, two broad aspects of fairness are
evident and applicable to this study, procedural and distributive justice. All of the
survey items used to assess fairness perceptions utilized a 5—point Likert-type scale
anchored at “Strongly Disagre_e” and “Strongly Aggee” for responses.

The distributive justice scale (3 items) was adapted from Greenberg (1994). An
example item is, “The cost of the health cge system is fair to staff members”. Bias and
consistency are more specific subdimensions within distributive justice. Six items were
also adapted from Greenberg’s work to assess consistency and bias. An example item
assessing consistency is, “When it comes to health benefits, all employees ape treated the
am”. An example item from the three-item scale measuring bias is, “The health
benefits system is bizflagainst those with uncontrollable health problems. (Refer to
Appendix F for a listing of all items.)

Procedural justice was assessed through two types of questions adapted from

Leventhal’s guidelines (1976, 1980). A total of 16 items assessed interpersonal and

60

informational justice, two aspects of procedural justice. Interpersonal justice was
assessed with seven items relating to personal treatment or social sensitivity. An example
item is, “When I have a question or need to talk to someone about health care benefits,_I
arm treated with respect and dignity”. The other aspect of procedural justice,
informational justice, assessed the employee’s ability to question his or her health
quotient score (5 items) and how much justification an individual receives concerning the
nature, functioning, and rationale behind the use of the benefits program (4 items). An
example item dealing with the chance to change or correct problems is, “When I disagree
with something regarding the health care system, there is an appaal process that is fai_r”.
(See Appendix F for a complete listing of all survey items.)

Overall, there are 25 items assessing fairness perceptions, 9 items for distributive
justice and 16 for procedural justice. To verify that the created items actually measured
the aspect of justice they were written for and that the items could be factored into
distributive and procedural justice, we were interested in investigating the factor structure
of the set of items.

To investigate the factor structure of these 25 items, exploratory factor analyses
with varimax rotation were conducted in both the incentive and the traditional samples.
The samples were kept distinct at this stage due to concerns that the fairness constructs
would be different in the two samples. Specifying the extraction of two factors, each
sample created two different factor structures. (Refer to Appendix G for the Varimax
rotated factor loadings of the 25 items on the two factors for each sample.) The

emergence of two different factor structures was surprising. We expected 1 or 2 factors

61

based on the literature from which the items were chosen and expected the factor
structures to be similar within the two samples.

Due to the nature of the hypotheses, desire to compare responses across samples,
and our specific interest in the incentive system itself, we decided to apply the factor
structure of the incentive system to the traditional system. To this end, we first dropped
those items that loaded on both or neither factor in the factor analysis on the incentive
sample. The decisions to drop items were based on item loadings. For instance, an item
that loaded highly or similarly on both of the factors was dropped as was an item that
loaded below .4 on both factors. In addition, at this time, we removed the two items that
were present in the incentive data set, but not the traditional survey. These two actions
left us with 8 items on the first factor (items 2, 3, 4, 5, 6, 8, 10, and 20) and 7 on the
second (items 1, 12, 13, 14, 15, 16, and 19). (Refer to the table in Appendix G labeled
“Secondary two factor solution” for the incentive group). Upon inspection of the content
of these items we identified the nature of the factors themselves to be distributive justice
for the first and procedural justice for the second. The items themselves can be found in
Appendix F.

Once the two factor solution was finalized, we applied the same factor structure to
the traditional data set. Specifically, we deleted eight items from the 23-item pool and
were left with 15 items, eight of which we assigned to the first factor (distributive justice)
and seven of which we assigned to the second factor (procedural justice).

Internal consistency reliability estimates of the distributive justice dimension
yielded alphas of .88 for the entire sample, .86 for the incentive sample, and .88 for the

traditional sample. Internal consistency reliability estimates were also calculated for the

62

procedural justice dimension and yielded alphas of .92 for the entire sample, .84 for the
incentive sample, and .93 for the traditional sample.

The scales themselves were highly correlated in the entire sample (p.71, p<.01),
the incentive sample (556, p<.01), and the traditional sample (gr-:74, p<.01).

flaalth-Related Behayiors. Data on participant health were gathered through both
subjective and, in the case of the incentive system hospital, objective measures. Self-
report items used to assess exercise/fitness and nutrition were adapted from Ribisl (1994).
Items measuring stress, tobacco use, alcohol use, and motor vehicle safety were
constructed for this study. The items were about specific behaviors such as, “1%
puc_kle my seatbelt while I am driving or riding in a motor vehicle”. Responses were
indicated using a 5-point Likert-type scale with anchors at “Strongly Agree” to “Siquly
Disagree”. (Refer to Appendix H for a listing of all health measures.)

Overall, sixty—six items assessed six self-report health behavioral dimensions for
both samples. While there were six different aspects of health represented in the items
we believed that the dimensions would be highly related. For instance, individuals who
exercise a great deal are probably more likely to eat a balanced diet and avoid tobacco
products. Investigating the factor structure of these items would identify if, in fact, this
was the case and the different dimensions of health were interrelated and could be
considered as one indicator of overall health-related behaviors for the analyses.

An exploratory factor analysis with varimax rotation was conducted. Following
the Kaiser normalization criterion guideline of selecting factors with eigen values above
one, there were 13 factors present in the 66 items. (Refer to Appendix I for the varimax

rotated factor loadings of these 66 items on the 13 factors.) The emergence of thirteen

63

factors was surprising because only six specific dimensions of health were assessed.
Further investigation of the eigen values of the 13 factors facilitated an interpretation of
these results. The eigen values of those thirteen factors were 13.81, 8.42, 5.62, 3.27,
2.54, 1.89, 1.74, 1.62, 1.55, 1.29, 1.24, 1.12, and 1.06 respectively. The first factor
accounted for 21% of the total variance and the second factor accounted for 13%. The
ratio of the first factor to the second was 1.53. While this is below a ratio of 2, the most
significant break occurred at the second factor with a reduction of 5 points.

Further evidence in support of using a single overall factor as an indicator of general
health-related behaviors can be gained by investigating the dimensions more closely. All
of the dimensions are correlated significantly with the overall health-related behavior
composite (Table 3) and with the objective measures of health (Table 4). Thus, for the
purposes of our analyses and hypotheses only one factor will be utilized in all analyses.
For the one-factor solution, internal consistency reliability estimates yielded an alpha of
.93 for the entire sample.

Additional data were also available for the hospital using the incentive-based
system. Data on cholesterol, blood pressure, cardio-vascular fitness, and body fat ratio
had been collected by the organization over a two-year period. These data were used to
validate the self-report measures that were available for both sites. The correlation
between the composite subjective and objective measures was g = .37 (p<.05). (Refer to

Table 4.)

64

Table 3.

Intercorrelations between Self-report Dimensions of Health Behfliors

 

Dimension Mean SD 1 2 3 4 5 6
1. Alcohol 3.46 .70 fl

2. Exercise 3.13 .58 .02 £4

3. Nutrition 3.57 .68 -.05 .40* fl

4. MotorVehicle 4.16 .53 .32 -.02 .16* &

Safety
5. Stress 3.41 .54 -.01 .14* .12 .08 p32
6. Tobacco 3.50 .93 -.05 -. 15* .01 .16* -.20* ;4_5

Note. The correlations between the specific dimension and the overall composite of self-
report health behaviors are provided in the diagonal and are underlined. All correlations
marked with an “*” are significant at p<.05

65

Ha 230$:me 2w

2*:

.mo.vg

~23 macaw—9:00 .SVQ “a “56$:me one .12.... 5:» 20:22.50 .3332: toner-:8 mo >255 So 2038

\n—0mo_o ”505 “an: HGOEmwOmmd 530: deummmOr— 05 50¢ wOuSmNUE QmOr—u GOOBuQD chfi—afiboo 05 0.3 mfiomumfiboo GOES-SUSS .QHOZ

 

 

 

 

 

1.5. 3.- 2. mo. co. mo. no. ._. *Nm. * *mm. 5305 :Eo>O
8.- 00.. 3.- 5. 3.- oo.- oo. mo. vo. mmobm
"Sum?- **mN. “1.3”. **mN. :3. film. :5. 33%... mo. ommohoxm
8.- 8.- S. 8.- 8.- a? we. 8. 3. a: 8088.
*2.-. 5.2.- *m_.- 00.. *3r 8.. .34 *3. iom. behav-
* *vm. mo. * * mm. mo. *2. 2. mo. :3. *2. cog-5:2
.43.- N_.- *3 .- cor LA.- No.- cor 5.- 3.3. 33 3:83.
ocmm “mm Esme-2m £055 83 mum-am coat“: Z 83 mcommnom ”Emma

O: .Losu boom U85 -0630 0839—. £255 3:82 .«o 353% Bram

338.25 338.3%

8.5982 “guise-mt. 5.8:

 

.3: mom 33:85 05 E 308333.

 

 

5:3: 2: Mo BE 3 pecan-32 macaw—Lem 5:3: .53 mag-Sm 2: Bot EBB-Eom .2133 «o flamed-tom 50253 macaw—obooeoufi

.v 2an

66

Perceived Control. Employee perceptions of control over health-related outcomes
were assessed through a series of statements developed for this study. As mentioned
previously, each individual employee’s overall health was evaluated by the organization
with the incentive-based benefits system using eight health-related dimensions: blood
pressure, cholesterol, body fat, fitness, tobacco use, motor vehicle safety, alcohol
consumption and nutrition. These eight were the behaviors or aspects of an employee’s
health that the incentive system desired to impact in a positive way. It was reasoned that
a necessary condition for this to occur was that the employee should believe he or she
could influence these aspects of his or her health. Therefore, a measure was developed
for each health dimension of the extent to which the employee believed that he or she had
control over their health.

Employees were asked to rate the control they had over each of the health factors
individually. The item stem read, “How possible do you feel it is to change your health
level on each of the following health dimensions within the next yeafi” Following the
stem each of the eight health dimensions from the health assessment were listed. For
each health factor respondents were asked to rate the extent to which they believed they
could change their behavior on that dimension on a five-point Likert-type scale of
“Nppp” to “Extremely Possible”. (See Appendix J for a complete listing of items.) The
estimate of internal consistency reliability yielded an alpha of .89 over the entire sample

from both hospitals.

67

Commitment. Employee commitment to change health behaviors was measured
in two ways: actual behaviors that implied an interest in changing behavior and
employees’ self-reports of their willingness to change their behaviors. However, actual
behaviors were available only from the hospital using incentive systems. Therefore, the
“objective” behaviors were used to only draw inferences about the validity of the self-
report measures.

Data were collected on the survey to measure employee willingness to change
their behaviors. Eight items relating to each of the eight health dimensions were
developed for this study to assess employee willingness to change. The item stem read,
“How willing are you to work on improving or maintaining this aspect of your health to a
healthy level?” Below the stem were listed each of the eight health dimensions. For each
health factor respondents were asked to rate the extent to which they believed they were
willing to improve or maintain their behavior on that dimension on a five-point Likert-
type scale of “Not at all” to “Extremely Willing”. (Refer to Appendix K for a complete
listing of the items.) The eight items were combined for an overall indicator of
commitment. The internal consistency reliability of the combined scale of these items
was .90 for the entire sample.

Beyond survey responses, data were collected from employee records in the
hospital utilizing an incentive-based system to provide objective data of employee health,
which was collected during the annual health assessment. Records detailing employee
health levels were available for the two years prior to the survey collection date. Other
data, which might indicate commitment, were also available in employee records. For

instance, the organization offered decreased costs of membership to local health clubs,

68

fitness equipment, and weight loss programs. In addition, the organization had a specific
department that promoted health-related programs within the hospital. Examples of these
programs were weight loss, exercise, and smoking cessation. Each instance of
participation in these organizationally sponsored events or discounts was documented in
employee records. For our purposes, each instance of use was totaled for an overall
indication of commitment. To evaluate the validity of the self-report items a correlation
was calculated between the two types of measures. The correlation was significant
(p.18, p<.05), but fairly small. This is possibly due to the relatively few instances of
participation in any of the recorded programs with the average number of programs
attended being .23 with a standard deviation of .55. The low base rate could be due to
involvement in fitness and health programs outside the organization or in programs not
documented by the organization. Thus, the objective data from the hospital with the
incentive plan provided some support for the validity of the self-report measures of
health-related activities, but convergent validity was low. No further analyses will be
conducted utilizing the objective commitment data due to the low base rate of
participation creating severe range restriction.

Perceived Accuracy of Mpasurement. Within the incentive sample, employees’
perceptions of the accuracy of the measures used to establish their Health Quotients were
assessed. Eight items assessed employee perceptions of the accuracy of measurement
techniques used by the organization to determine their health level on each of eight
dimensions. An example item is, “The blood test to measure Cholesterol for the Health
Index isfiaccura_t§_’_’. (Refer to Appendix L for a complete listing of the items.) These

statements were repeated for each of the eight health dimensions with the appropriate

69

health factors and measurement techniques substituted. Participants responded to the
item using a 5-point Likert-type scale from “Strongly Aggee” to “Strongly Disagree”.
The eight items were averaged for an overall measure. The internal consistency
reliability for the overall accuracy measure was alpha = .82.

Demographics. Age, gender, race, education, and job level data were available
through employee records (in the hospital using the incentive-based program) and
through items on the survey. In addition, an item measured annual household income
with six response ranges available to choose from anchored at “Below $ 14,999” to

“$85,000 or Above”.

70

Chapter 4

RESULTS

Natureé and Oglity of Variables

The method section detailed the content and reliability estimates of the five
measures: fairness perceptions, health behaviors, perceived control, commitment, and
measurement accuracy. An overview of these results are provided in Table 5. In this
table, estimates of internal consistency reliability for each of the variables are provided
for the entire sample, the incentive sample, and the traditional sample. Reliability
estimates for the subsamples were provided as some of the hypotheses were proposed for
only one of the samples. In the following paragraphs, descriptive information for the

variables will be provided.

71

Table 5.

Reliability Dy for Constructed Scalg

 

Scale & Grog) Mean N SD # of Items Alpha
Distributive Justice

Overall 3.38 506 1 .02 8 .88
Incentive 3.08 232 .87 8 .86
Traditional 3.70 270 1 .07 8 .88
Procedural Justice

Overall 3.68 504 .95 7 .92
Incentive 3.32 230 .67 7 .84
Traditional 3 .99 270 1 .03 7 .93
Perceived Control

Overall 3.61 507 .89

Incentive 3.63 237 .89 8 .86
Traditional 3.60 267 .89 8 .86
Commitment

Overall 3.67 506 .89

Incentive 3.74 236 .91 8 .88
Traditional 3.60 267 .86 8 .86
Accuracy

Incentive 2.95 237 .80 8 .81
Health Behaviors

Overall 3.54 514 .32 66 .93
Incentive 3.45 243 .27 66

Traditional 3.61 271 .33 66

72

First of all, out of the twenty-five items measuring fairness, two factors were
found using 15 items in the pool. These fifteen items were broken into two fairness
measures, procedural and distributive justice. The mean response on the distributive
justice scale (8 items) for each of the three groups are as follows: entire sample, x=3.38,
sd=1.02; incentive group, x=3.08, sd=.87; and the traditional group, x=3.70, sd=1.07.
The mean response on the procedural justice scale for each of the three groups are as
follows: entire sample, x =3.68, sd=.95; incentive group, x=3.32, sd=.67; and the
traditional group, x=3.99, sd=1.03. The mean responses on both scales were higher in
the traditional group.

The second variable was assessed with a composite of eight items measuring
employee perceptions of control over their health. The mean response on a five-point
scale for the entire sample was 3.61 with a standard deviation of .89, for the incentive
group the mean was 3.63 with a standard deviation of .89, and for the traditional group,
the mean was 3.60 with a standard deviation of .89. Mean perceptions of control were
similar in all three groups.

Third, eight items assessed employee commitment to maintaining or improving
health-related behaviors. These items were averaged for a composite indicator of
commitment. The mean commitment scale response was 3.67 with a standard deviation
of .89 for the entire sample, the mean was 3.74 with a standard deviation of .91 for the
incentive group, and the traditional group, on average, responded 3.60 with a standard
deviation of .86. The incentive group had a slightly higher mean commitment response

in comparison to the traditional group.

73

The fourth variable was assessed with a composite of eight items measuring
employee beliefs about the accuracy of the measures used to assess their health levels on
the eight dimensions. These data were collected only in the incentive group. The mean
response of the incentive group on this scale was 2.95 on a five-point scale with a
standard deviation of .80.

Finally, sixty-six items were combined into one overall measure of health
behaviors. The mean responses on the health behaviors scale was 3.54 with a standard
deviation of .32 for the entire sample, for the incentive group the mean was 3.45 with a
standard deviation of .27, and for the traditional group the mean was 3.61 with a standard
deviation of .33. The traditional group reported, on average, more positive health-related
behaviors with a greater standard deviation.

Morfirips Among Variables.

In this section, three tables detailing the relationships between the variables
discussed above and various other demographic indices will be provided. The first table
(Table 6) provides intercorrelations among the benefits system type or hospital, perceived
control, fairness perceptions, commitment, health-related behaviors, and accuracy
perceptions. Within this table data are provided from the entire sample (N=516), the
incentive hospital (N=243), and the traditional benefits hospital (N=273) individually.
The second table (Table 7) contains an intercorrelation matrix for the incentive group
only and includes additional demographic information and both objective commitment
and health-behavior data in addition to the key variables of interest in the study. The
final table (Table 8) provides an intercorrelation matrix for the traditional benefits

hospital only and includes relevant demographic variables in addition to the key variables

74

of interest. The data provided in these tables will be utilized to evaluate the majority of

the hypotheses proposed in the following section.

75

.mnom

 

ow. .mlo- from. Nb.- wo. mor hm. mvfi mvm 5R03N~
3w. ml **Nm. **wm. *2. ow. mmN hmm >03:oo<.Z
ww. **ON. “1.3. inwm. a. inn 0mm “ECU-O—
DQEmSH
a 30m. *2. no. de Omm .095 .m
352.
.lwaaclemh hm. mod . NMN .ammfl .w
cw. ow. m©.m 5mm 35:00 .5
a: mom 95:8:—
.mnom
mlo. "1.2. ".311. no. ".342. screw. mm. tum-m 3m :53: .0
moi *oo. mo. 33%. wet ow. 59m cow “ECU .m
853
MW. .315. *oo. **mm. mo. wo.m vow .095 .v
853
MWle mo. .1an. NOA wm.m wow .ume .m
m4 m0.- ow. Ed New 35:00 .N
829nm

- om. mm; 2m Ecocom;
295“. _m o. .._.u=_m_

S _ _ o_ a m N. c m w m N _ Om :32 Z 032:8»

 

 

 

 

com Slam Scam e8 macaw—88895 use. .3334an wSucflm 2802

.0 03,3-

76

 

.mnom

 

 

 

 

ow. “rm *NM. 8.. mo. mm. 36 EN 5302.2
cm. mo. nor Ion. 0w. 006 SN 3253500 .3
comma
MN 33‘. no. mo; mmfi Ohm 605.2
0053.
a 5. R: 09m OR .35 .3
0% mm. 09m SN 35:00 .3
3: :0: 10:05:05.;
.mnum
mm. min mvw :23: .N_
ow. mad 5mm >08:00<.:
a. vnfi 0mm 0:253:80 .2
0002:.
no. mm.m 0mm .005 .m
803:.
mm. mod mmm .35 .w
aw. mo.m 5mm 35:00 .h
3: m3.— 95:3:—
2 2 2 E m_ Om $02 Z 03th>

.9380 0 came

77

E00030
00.: N:. *2. 2.. m:. imm. Nor 3.0a. *2. N:.- :0.- **mm.- m0.- 8.0 mm. mm 5302.2
.23:
00.: 00. 00. 50.- m0. m0.- N0. *2. N0.- m0.- Nor m0. R. San mvm :::a0:-:.~:
00.: N0: Imm- iwm. **m:. 0:. :0.- w0.- 00. :0. m:.- 0w. mad bmm >08=00<.::
00.: 3.0m. **w:. iwm. m0.- m0.- 0:. 00. >0. :0.- :0. Eum 0mm £8500 .0:

00:63.

8.: in. it. 8.- 2. 3.- 2. we. 2.. 3. 0mm ca .855
00:35—-

8: :8. :8. :3 3%.. S- 8.- 8.- 5. men as .02: .0
3.5000

8.: 2.. mo.- :. 8. mo- 8- a. S: RN 8385.: s

8.: $3.. 1%.- :8..- 8- 8.- w: 03. Gm 8:830 .0

8.: it..- :.- :2. 2.x.- 8: gm 0mm 2:85 .m

8.: 2. 8.- 8.- 6.: mnm :0 25%: .v

8.: 5. S.- 00 8.: SN 83: .m

8.: :2.- avd 2.3 g cm... .N

00.: 3». mn.: Nvm 50:00 .:

 

m: N: :: 0: 0 m N. 0 m v m N : Gm 4:03: Z 0538:»

 

 

0:80 03800:: E m0Emtm> :003000 m:o::m:0b00:0:::

.h 030-:-

78

 

 

8.: :8. 8.- 8.- 8. 8.- 8.- .8.- ..2. :2. mm. 8.0 EN 0.1.2
8.: 8. 8.- 3.2.. 8. 8.- 8. 8. 8.- 8. 85. SN .8858 .8
00303—-
80 :8. 8. 8. .11.- 2. 3.8. 8. 8.: 8.8. 88 .85 .w
00:00::-
8._ 8. 8.- 8. :. :8. 2. 8.: 88 88 .35 s
35:00
8.: 8. *2- 8. 8.- 2.- 8. 88 SN .28 .0
8.: 1.8.- 8.- 8.- 8.- 8.: :2. 88 .260 .m
8.: 8. 8. 8. 8.: mm :8 8: .v
8.: 8.- 3.:- 8. o: 88 83: .m
8.: 8.- 8.8 0:. 8m 00... .N
8.: 8.. 8.: 88 08:00 .:
2 a w n o m :- m N _ mm :82 z 289$

 

 

9.80 180068 E E 00:0m:._|0> c0058: 20080080005:

.w 030:.

79

Tests of Hypotheses

In the following section each hypothesis will be restated, the analyses conducted
to test each hypothesis discussed and any related findings provided. Each hypothesis will
be discussed in turn and further explanations of the findings can be found in the
discussion section.

Hypothesis 1. First, it was hypothesized that employees’ perceptions of the
fairness of their benefits system would be positively affected by the values of their
individual outcomes. This hypothesis referred to reactions of employees working under
the incentive conditions as the receipt of some tangible outcome is required; it was
explored only within the incentive hospital (N =243). Correlations were computed
between both distributive and procedural fairness perceptions and the health-related
outcome (the Health Quotient Score). The hypothesis that distributive justice perceptions
would be positively related to health-related outcomes was supported (_r=.33, p<.01).
However, the hypothesis relating procedural justice perceptions to health outcomes was
not supported (13:.13, p>.05). Refer to Table 7 for mean, SD, and intercorrelation data.

Hypothesis 2. Hypothesis two proposed that perceptions of control would
mediate the relationship between age and fairness perceptions such that older employees
would believe that they had less control over their health levels than younger employees
and would subsequently perceive their benefits programs as less fair. To test this
hypothesis we first selected the incentive sample (N=243), and computed correlations
between age and control (§=-.O8, p>.05), distributive justice (r:-.OZ, p>.05), and

procedural justice (p.08, p>.05) individually. No correlation was significant and thus

80

further analyses investigating this mediation hypothesis were not conducted in the
incentive sample. Refer to Table 7.

Next, we ran the same analyses in the traditional sample (N=273). We computed
correlations between age and control (F-.O3, p>.05), distributive justice (p.32, p<.01)
and procedural justice (_r=.33, p<.01). Refer to Table 8. These data do not suggest the
support of the mediation hypothesis, but do support a direct positive relationship between
age and fairness in the traditional sample.

Hypothesis 3. The third hypothesis proposed that employee income level would
moderate the relationship between health-related outcomes and employee perceptions of
the fairness of their benefits system for employees within the incentive-based system.
The relationship between outcomes and perceived fairness was expected to be stronger
for those with lower incomes. To examine this hypothesis, only the incentive sample was
used as monetary outcome was a critical variable in the hypothesis and was only
employed in the incentive group.

Hierarchical regression analyses were used to test the moderated hypothesis. In
the first step income and health-related outcome (HQ) were individually entered into the
regression equation and significantly predicted distributive justice perceptions (R2
change=. 140, p<.01). Then, the interaction term was entered and did not lead to a
significant increase in the R2 (R2 change=.002, p>.05). Individuals who receive negative
outcomes perceive the incentive-based system as less distributively fair than those
receiving higher outcomes regardless of their incomes.

Hierarchical regression analyses were also used to test this moderated hypothesis

involving procedural justice perceptions, our second aspect of fairness. In the first step,

81

income and health-related outcome (HQ) were individually entered into the regression
equation and did not significantly predict procedural justice perceptions (R2 change=.020,
p>.05). The interaction term was then added into the regression equation and was also a
non-significant predictor (R2 change=.003, p>.05). Outcomes (HQ) nor income levels
proved important in procedural justice perceptions for the incentive group.

Hypothesis 4. In the fourth hypothesis, employees within the incentive-based
benefits system, who believed that the measures used to determine their health levels
were more accurate, were expected to report higher perceptions of the fairness of the
health benefits system than those who believed the measurements were less accurate. To
test this hypothesis, employees within the incentive group were selected (N=243) because
accuracy data were only available for that sample. Correlations between perceptions of
the accuracy of the health benefits system and both procedural and distributive justice
perceptions were computed. Employee perceptions of the accuracy of the measures used
to determine their health levels significantly correlated with their perceptions of both the
distributive fairness (p.38, p<.01) and procedural fairness (p.32, p<.01) of the
incentive-based benefits system. Refer to Tables 6 or 7.

Hypothesis 5. The fifth hypothesis proposed that employees who were more
committed to maintaining or improving their health would report performing more
healthy behaviors than those less committed. This hypothesis was tested with a
correlation between commitment and self-report health behaviors for the both samples
combined (N=516). The hypothesis was supported (tr-.15, p<.01), suggesting that
employees, in these samples, who report being more committed to improving or

maintaining their health levels do report performing more healthy behaviors. Refer to

82

Table 6. We can also correlate the health quotients of employees within the incentive
system and their self-reports of commitment ($.06, p<.05) which again suggests no
support for this hypothesis. Refer to Table 7.

Hypothesis 6. The sixth hypothesis proposed that employees who reported more
positive perceptions of fairness would report higher levels of commitment to maintain or
improve their health-related behaviors than those who have lower perceptions of fairness.
The hypothesis was tested on the entire sample (N=516) through correlations between
distributive (p.07, p>.05) and procedural justice (F- 14, p<.Ol) perceptions and self-
reported commitment. The data suggest that employees who believe the health benefits
system to be more procedurally fair report higher levels of commitment to improve or
maintain their health-related behaviors while there is no significant relationship between
commitment and distributive justice. Refer to Table 6.

Hypothesis 7. Hypothesis seven proposed that employees who perceived that
they had more control over their health-related outcomes would report higher
commitment to maintain or improve their health-related behaviors than those who
perceived less control. Again, this hypothesis was tested on the entire sample (N=516)
through a correlation between perceived control and self—report commitment. The data
show that employees who perceived higher levels of control over their health also
reported higher levels of commitment to improve or maintain their health—related
behaviors (_r=.64, p<.01). Refer to Table 6.

Hypothesis 8. The eighth hypothesis suggested that employees who perceived
that they had more control over their health-related outcomes would report more

favorable perceptions of the fairness of their benefits system than those who perceived

83

less control. Correlations between distributive and procedural justice perceptions and
self-reports of perceived control were conducted for the entire sample (N=516). The
results suggest that individuals perceptions of distributive justice and control are not
related (_1_=.08, p>.05). However, the results do suggest that individuals within these
samples who believe that their benefits system is procedurally fair do report stronger
beliefs in their ability to control their health levels (p.09, p<.05). Refer to Table 6.
Hypothesis 9. The ninth hypothesis stated that employees within an incentive-
based benefits system would report performing more healthy behaviors than those under
a standard health benefits system. This hypothesis was tested using a directional t-test to
investigate mean differences between the two groups. With equal variances assumed, the
t-test was significant (t=-6.02, 512 df, p<.01), but in the wrong direction. Employees
employed under the traditional benefits system reported more healthy behaviors than
those under the incentive-based system (Refer to Table 9). Possible explanations of this

finding are addressed in the discussion section.

84

Table 9.

Data Illustrating the Differences between Means of the Incentive-based

 

and Traditional Samples: Indepgndent Samples and t-test for Equality of Mew

 

 

Sig. Sig
Variables t df (2-tailed) ( 1 Jailed) Difference
Distributive -6.58 500 .000 .000 -.58
Justice
Procedural -8.45 498 .000 .000 -.67
Justice
Commit. 1.73 501 .084 .042 .14
Perceived .60 502 .550 . 275 .00
Control
Self-Report -6.02 5 12 .000 .000 -. 16
Health
Behaviors
Income -1.06 502 .289 .144 -. 12
Gender- 3.10 510 .002 .001 -.11
Race -.62 509 .538 .269 .00
Education -.44 509 .662 .33 l .00
Age -l.51 513 .133 .067 -l.53
Job Type 5.08 510 .000 .000 .68

85

Hypothesis 10. Hypothesis ten proposed that employees within an incentive-
based benefits system would report higher levels of commitment to maintain or improve
health-related behaviors than those under a traditional benefits system. This hypothesis
was tested with a directional t-test conducted on the differences between mean self-report
commitment levels between the two samples (incentive N=243; traditional N=273). The
hypothesis was supported (t=1.73, 501 df, p<.05). The incentive group reported higher
commitment to maintaining or improving their health-related behaviors than those within
the traditional group. Refer to Table 9.

Hypothesis 11. The eleventh hypothesis proposed that employees under an
incentive-based benefits system would report more perceived control over their health-
related outcomes than those under a standard health benefits system. This hypothesis was
tested with a directional t-test conducted on the differences between mean perceived
control between the two samples (incentive N=243; traditional N=273). This hypothesis
was not supported (1:60, 502 gr, p>.05). Refer to Table 9.

Hypothesis 12: The last hypothesis proposed that employees within the two
different hospitals would report, on average, significantly different mean fairness
perceptions, however directionality was not hypothesized. It was theoretically feasible to
propose that employees under the incentive-based system would report both higher and
lower perceptions of the fairness of their benefits system than those who were covered
under the traditional program. These two models are represented in Figures 2 and 3.

To test this hypothesis and establish the directionality of the relationship, two
independent samples t-tests were conducted. These data are provided in Table 9. The

results suggest that there are significant differences between both mean distributive

86

justice perceptions (t=-6.58, c_l_f=500, p<.01) and procedural justice (t=-8.45, df=498,
p<.01) perceptions between the two benefits programs. In addition to these analyses,
correlations between fairness perceptions and benefits system type were calculated. Both
correlations were significant (distributive justice [=-.28, p<.01; procedural justice g=-.35,
p<.01) and suggested that employees within the incentive-based benefits program
reported lower perceptions of fairness than those within the traditional benefits program

supporting the model depicted in Figure 3.

87

Chapter 5

DISCUSSION

The main purpose of this study was to investigate the interrelationships among
employee perceptions of control, fairness perceptions, and commitment as they relate to
health behaviors under conditions where monetary incentives for health behaviors do or
do not exist. These questions were addressed through the examination of two differing
benefits systems, one offering an incentive health benefits program and the other utilizing
a more traditional health benefits approach. Many of the proposed hypotheses were
significant and suggest, overall, that both perceptions of control and fairness influence
employees reports of commitment to their overall health regardless of the type of benefits
system. The conclusion of this paper reviews and discusses our findings, addresses some
limitations present in the study itself, and, in light of these findings and limitations,
suggests some implications and overall conclusions.
Overview of Results

When first presented, the hypotheses were discussed in three different sections.
The first included hypotheses directed at understanding the effects of four general
variables on fairness perceptions: distributive and procedural justice. The second set
included hypotheses that were directed at creating a framework for understanding the
interrelationships between fairness, perceived control, and commitment as they are
critical in understanding behavioral change. The final set of hypotheses involved the
consideration of the effects of the different health benefits systems on the proposed

relationships between fairness, perceived control, commitment, and health—related

88

behaviors. These three sections will again be used as a framework for discussing the
overall findings of this study.

The Effects of Outcome. Age, Income, and Perceived Accuracy on Fairness
Perceptions. Four hypotheses individually investigated the effects of outcome, age,
income, and accuracy perceptions on employee fairness perceptions with mixed results.
The first hypothesis posited that employees’ perceptions of the fairness of their benefits
system would be positively related to the values of their individual outcomes. This
hypothesis was based on work by Greenberg (1987, 1994) and others who found that
individual’s fairness perceptions are affected by self-serving biases. This work would
suggest that individuals perceive as most fair those outcomes that benefit themselves the
most. As hypothesized, individuals within the incentive hospital receiving more positive
outcomes did perceive their benefits system as more distributively fair (gr-.33; p<.01).
However, the relationship between procedural justice and outcome value was not
significant (1:13, p>.05). These findings suggest that within the incentive group,
employee perceptions of the distributive justice of their outcomes increases with the
favorability of the outcome. Employees who are receiving the more positive outcomes
perceive the benefits system as more distributively fair. Procedural justice perceptions,
however, are unaffected by outcome value.

The second hypothesis proposed that perceptions of control would mediate the
relationship between age and fairness perceptions. The proposed relationship between
age and perceptions of control was supported by work in the health psychology domain
where research suggests that older people are often less convinced of their ability to

change their overall health (Baum et al., 1997; Taylor, 1995). With this in mind, we

89

turned to expectancy theory (Vroom, 1964) and proposed that employees who don’t
perceive that their actions will affect their outcomes won’t change their health-related
behaviors.The mediation hypothesis was not supported. Within the incentive sample, the
correlations between age and control (§=-.08; p>.05), age and distributive fairness (r:-
.02; p>.05), nor age and procedural justice (1:08; p>.05) were significant so any further
investigations into the mediation hypothesis were not conducted. Within the traditional
sample, however, the data were slightly more interesting. The same correlations were
calculated with age and control not significant (rs-.03, p>.05) and both fairness
correlations with age significant (distributive p.32, p<.01; procedural p.33, p<.01).

These data are quite interesting. While neither sample provided support for the
mediation hypothesis, the results between the two samples are very different. Within the
incentive sample, fairness perceptions and age were unrelated. However, within the
traditional sample, the data illustrate that as employees get older, they perceive the
benefits system as more distributively and procedurally fair. This difference could be due
to the nature of the benefits programs under investigation. Perhaps in the incentive
sample employees do not have to be old to perceive the program as unfair, perhaps you
just have to be unhealthy. This explanation is partly addressed in the first hypothesis
were a significant correlation was found between health-related outcome and fairness
perceptions for the incentive group. However, further investigation into this hypothesis
by including the traditional group cannot be performed due to the lack of objective
ratings of healthiness in that sample.

The third hypothesis proposed that employee income level would moderate the

relationship between health-related outcomes and employees’ perceptions of the fairness

90

of their benefits system such that positive correlations between health-related outcomes
and both distributive and procedural fairness perceptions would be stronger within lower
income groups than within higher income groups. This hypothesis was based on the fact
that the impact of a financial amount of money added or subtracted per pay period would
be greater for those with lower incomes than those with higher ones. Research within
health psychology supports this position. Individuals from lower socioeconomic groups
are, on average, less healthy than those in higher socioeconomic groups in the general
population (NHLBI, 1994) and in our sample (correlation between income and HQ is
equal to .16, p<.05). With this in mind, one of the fairness guidelines proposed by
Leventhal (1976) which suggests that fair procedures should be perceived as free from
bias may be violated. If larger portions of lower socioeconomic groups are receiving
disincentives or pay deductions the incentive-based program may be viewed as unfair.

This hypothesis also was not supported. The lack of direct support for the
moderation hypothesis could stem from the fact that both our incentive and traditional
samples were composed of a higher percentage of upper level employees who were likely
to be earning higher incomes. However, within the incentive group, income was
positively correlated with both distributive (1:.27, p<.01) and procedural fairness (_r=.20,
p<.01) perceptions as expected.

The final hypothesis in this section addressed the relationship between
perceptions of accuracy and fairness. Specifically, it was hypothesized that employees
within the incentive-based benefits system who believed that the measures used to
determine their health levels were more accurate were expected to report higher

perceptions of the fairness of the health benefits system than those who believed the

91

measurements less accurate. This hypothesis was based on the foundations for fair
procedures as laid out by Leventhal (1976). According to his work and supported by
empirical findings (e.g., Sheppard & Lewicki, 1987), an individual must perceive that
resource distributions are based on accurate information if they are to be considered fair.
In our study, we hypothesized that individuals may differ in their beliefs in the accuracy
of the measures used to assess their health levels and that any perceptions of inaccurate
assessment would lower perceptions of the distributive and procedural fairness of the
incentive system.

Results supported the hypothesis and suggested that employees within the
incentive system who perceived the measurement methods utilized to determine their
overall health as more accurate also perceived their benefits system as more distributively
(p.38, p<.01) and procedurally (p.32, p<.01) fair. In general, these data suggest that
employees who perceive that the system employed to determine the allocation of health
benefits costs is accurate are more likely to perceive that same system as fair. Expanding
beyond the scope of this study, this finding supports research in support of Leventhal’s
(1976) guidelines indicating that the accuracy of methods used to determine the
distributions of resources is critical in determining perceptions of fairness.

Hypotheses Relating to the Relationships between Fairness Commitment and

 

Perceived Control. There are four hypotheses relating to the relationships between
fairness perceptions, commitment, and perceived control. These four hypotheses set up a
framework to investigate our main interest in this study, the interrelationships among

fairness, control, and commitment in predicting behavioral change. All four hypotheses

92

were supported, providing support for the framework we laid for the relationship between
fairness, control, and commitment.

The first hypothesis in this set is actually the fifth we proposed. In the fifth
hypothesis, employees who were more committed to maintaining or improving their
health were expected to report more healthy behaviors than those less committed. The
foundation for this hypothesis originates in two prominent research fields. First, goal
commitment research suggests all employees are not equally committed to goals and that
commitment levels do influence employee behaviors (Hollenbeck & Klein, 1987; Locke,
1968). In addition, theories by J anz & Becker (HBM, 1984) and Ajzen & Fishbein
(TRA, 1980) both attempt to understand when and why people behave and propose
commitment as a direct antecedent to behavior. In our study, this hypothesis was
supported with a significant correlation in the expected direction (1:. 15, p<.01). These
data suggest that employees who report higher commitment levels to health improvement
or maintenance will also report healthier behaviors.

The sixth hypothesis and the second in this set proposed that employees who
reported more positive perceptions of fairness would also report higher levels of
commitment to maintain or improve their health-related behaviors. Equity theory
(Adams, 1965) provides support for this hypothesis. According to equity theory
employees who believe that with the input of certain behaviors certain predetermined
outcomes will be received are more likely to perform the required input behaviors,
provided the outcome is meaningful to the individual. For our study, employees who
believed that they could improve their health by exhibiting certain healthier behaviors

and receive a monetary outcome were expected to be more likely to report commitment

93

to maintain or improve their health. The results for this hypothesis were noteworthy.

The hypothesis was supported with a significant correlation between procedural
fairness perceptions and self-reports of commitment (1:. 14, p<.01) in the expected
direction, but unsupported by a non-significant correlation between distributive justice
and commitment self-reports (p.07, p>.05). Employees who perceive that their benefits
program is procedurally just are more likely to report being committed to improving or
maintaining their health-related behaviors than those who perceive their benefits program
as less procedurally fair.

The third part of the framework we are proposing is detailed in the seventh
hypothesis. Specifically, we believed that employees who perceived that they had more
control over their health-related outcomes would report higher commitment to maintain
or improve their health-related behaviors than those who perceived less control. Both
fairness and motivation domains provide strong support for this hypothesis. First of all,
research suggests that strong commitment should be present when employees believe
they have control over their outcomes (Lind & Tyler, 1988). Also, support exists in
equity theory (Adams, 1965) and expectancy theory (Vroom, 1964) which both suggest
that employees who perceive control over their outcomes are more motivated to attain
those outcomes and devote more of their efforts to attainment. In addition, the Health
Belief Mode theory (J anz & Becker, 1984) describes four antecedents to commitment to
change health behaviors. One of those antecedents is perceived control. The
hypothesized relationship between perceptions of control and commitment was

significant and in the predicted direction (_r=.64, p<.01).

94

The final part of the framework is described in the eighth hypothesis which
proposed that employees who perceived that they had more control over their health-
related outcomes would report more favorable perceptions of the fairness of their benefits
system than those who perceived less control. Support for this relationship exists in
equity theory (Adams, 1965) and procedural justice (Greenberg, 1987) which both
suggest that perceptions of equity or fairness occur when an outcome is considered, in
part, under the control of the individual. Our data provide mixed support for this finding
with a significant correlation in the predicted direction between control and procedural
justice (p.09, p<.05) suggesting that employees who believe that they have control over
their outcomes are more likely to perceive the process that determines those outcomes as
fair and a non-significant correlation between distributive justice and control (r_=.08,
p>.05).

Putting the Pieces Together: Understanding How Incentive Systems Affect the
Rglationsrhips between firimess. Perceived ContrcL and Commitment. The remaining
five hypotheses were directed at understanding how the incentive-based health benefits
system would affect the framework specified in the previous set of hypotheses
concerning the relationships between fairness, perceived control, and commitment. Two
of our four hypotheses were unsupported.

The first hypothesis within this set is actually the ninth hypothesis that was
proposed. Specifically, we proposed that employees within the incentive-based benefits
system would report performing more healthy behaviors than those under a standard
health benefits system. The rationale behind this hypothesis is fairly intuitive on the

surface, we assumed that the incentive-based system is working and that employees

95

under that system would report performing more healthy behaviors. Research also
supports this hypothesis in such classic theories as Skinner’s (1969) reinforcement theory
or Vroom’s (1964) expectancy theory. Surprisingly, this hypothesis was not supported.
In fact, the directional t-test conducted to investigate any mean differences between
groups was significant (t=-16.28, 469 g, p<.01), but in the opposite direction. These
data imply that employees within the traditional benefits system actually report
performing more healthy behaviors than those under the incentive system.

There are several plausible explanations for this phenomenon. One obvious way
to interpret this finding is to conclude the self-report measures accurately reflect the
levels of health, and it was higher under standard conditions. Yet, this alternative is
extremely unlikely. To reach this conclusion one must assume that (a) the incentive
system lowered health, (b) the standard system raised the health level, (c) there were
differences in levels of health between the two, or (d) the perceptions of actual individual
health level were lower (or more realistic) under the incentive system. Several factors
point to the last as the most likely conclusion.

First, explanations “a” and “b” are unlikely because there is nothing we know
about either system that would predict this pattern of results. The incentive system, for
example, had added classes on nutrition, smoking cessation, and other health-related
programs. There is no evidence that such activities decrease overall health. Likewise,
while the control hospital had some programs to encourage health they had no more than
the incentive one.

With respect to the third explanation, we did measure a number of demographic

variables and none of the comparisons between hospitals was consistent with an

96

interpretation that the traditional hospital would have higher reports of health. In
addition, as we discussed earlier, our two samples were very similar. Specifically, all of
our subjects come from the same type of organization, a hospital. Also, the two hospitals
are located in the same city and are likely drawing from the same applicant pool. Finally,
our two samples had reasonably similar demographics. These similarities provide us no
reason to suspect that the two samples would have different health levels or health-related
behaviors.

Finally, we have to consider the fourth alternative as an explanation of why the
traditional hospital reported more positive health-related behaviors than the incentive
group. There are at least three reasons perceptions of health may have been lower under
the incentive system. First of all, employees in the incentive hospital were lead to be
more critical when assessing their own health levels due to the nature of the program. As
the incentive or disincentive amounts were determined by an annual health assessment,
employees were lead to realistically view their health-related behaviors since they were
tied to an outcome and would, in part, determine their incentive levels the coming year.
Secondly, and related to the first, employees under the incentive system had more
accurate indicators of their health levels due to the annual health assessments which were
probably much lower than what would normally be assumed. Finally, the employees at
the incentive hospital knew that we had access to their health records from the
organization thus creating increased pressure to report realistic and accurate levels of
health and health-related behaviors.

For these reasons, we believe that the incentive-based benefits program may have

altered employees’ perceptions of their health and health-related behaviors, not health

97

levels themselves. Specifically, we are suggesting Beta change, or differences in the
way the two groups view the scaling of health-related behaviors. Unfortunately, we
cannot investigate the presence of Beta change because we cannot address issues of
possible confounds. For instance, we believe that issues such as incentive employees’
perceptions of the accuracy of the subjective health data could also influence health
behavior ratings in our surveys. This would not indicate Beta change and we would have
to control for accuracy perceptions in both groups before Beta change analyses were
conducted as the effects of accuracy and Beta change would look identical in any
analyses. However, we cannot control for accuracy perceptions in both the incentive and
traditional samples as we do not have these measures for the second group.

Overall, we do believe that, for various reasons, this data addressing this
hypothesis do not necessarily suggest that the employees within the incentive system
were less healthy, just that they viewed health in a different, perhaps more accurate,
fashion than the traditional group. In conclusion, it is important to note that while we are
suggesting that the groups were not reporting health-behaviors in the same fashion, the
distribution of people’s self-reports of health-related behaviors should not be affected by
the mean shift in health-related behaviors between the two groups. The next hypothesis
in this section and the tenth proposed in the study further examines this phenomenon by
examining the relationship between employee commitment and benefits system type.

The tenth hypothesis in this study forms the second hypothesis in this section
investigating the role of benefits system type on the fairness, control, and commitment
relationship. Hypothesis ten proposed that employees within the incentive-based benefits

system would report higher levels of commitment to maintain or improve health-related

98

behaviors. The Theory of Reasoned Action by Ajzen & Fishbein (1980) suggests that
there are two antecedents to commitment to change behaviors, the influence of individual
beliefs in favor of the new behavior and the influence of social norms that favor the new
behavior. We believed that both of these would be stronger in the incentive group due to
the increased health education efforts employed by the benefits department.

The hypothesis was supported when examined with a directional t-test (1:1.73,
501 df, p<.05). The two groups did differ in their self-reported commitment levels and in
the predicted direction. This finding provides further reason to interpret the findings of
hypothesis 9 cautiously. While the analyses from hypothesis 9 suggested that employees
within the traditional benefits system were performing more healthy behaviors, findings
concerning hypothesis 10 suggest that employees within the traditional benefits system
are less committed to improving or maintaining their health levels. Perhaps, as we
proposed earlier, employees within the incentive system are more committed to
improving their health, but just have more accurate perceptions of how far they have to
go to do that.

The eleventh hypothesis, the third in this section, continues to progress through
the framework. Specifically, we hypothesized that employees within the incentive-based
health benefits system would report more perceived control over their health-related
outcomes than those in the more traditional benefits system. This hypothesis stemmed
from the design of the incentive-based program which placed a good deal of importance
on establishing easily accessible health-related classes, facilities, and information.

This hypothesis was not supported (1:598, 502 _c_lf, p>.05). There was no

significant difference between the perceptions of control over health-related outcomes in

99

the two groups; the incentive system did not affect individuals’ beliefs in their ability to
change their health levels. We believe that the lack of support for this hypothesis is due
to the length of time the incentive system had been in place. As we discussed in the
method section, the incentive system had been in effect for almost two years prior to our
survey and it is possible that any initial increase in employees’ beliefs that they could
change their health due to the program was probably greatly reduced as employees began
to see that changing their health levels was not as easy as it first seemed.

Up to this point, we have discussed three of the four variables that were
incorporated into our framework for understanding employee willingness to change
behaviors. The previous three hypotheses were directed at the potential effects an
incentive system could have had on employee health behaviors, commitment, and
perceived control with mixed support for our hypotheses. The remaining aspect of the
framework is employee perceptions of fairness. This aspect of the framework was the
most problematic from a theoretical and practical point of view. Primarily, we were not
sure who would perceive their benefits system as more fair. Thus, we proposed to do an
exploratory analysis of the differential effects of the benefits system type on fairness
perceptions. We proposed that employees within the incentive-based benefits system
would perceive their benefits as more or less fair than those within the traditional benefits
system. Consideration of the design of the incentive program and prior research led us to
believe that either could be the case. The data suggest that employees within the
incentive-based benefits system did perceive their benefits as less fair. To establish that
there was a significant difference between the mean perceptions of fairness between the

two groups we conducted an independent samples t-test. The t-test proved significant

100

suggesting that our two samples did differ on their mean fairness perceptions (distributive
t=-6.45, _d_f=500, p<.01; procedural _t=-8.45, gl_f=498, p<.01). In addition to the t-test, we
conducted a correlation between benefits system type and distributive and procedural
fairness perceptions, both correlations were significant (distributive _r_=.28, p<.01;
procedural p.35, p<.01). These data suggest that employees within the incentive-based
benefits system perceived their benefits system as less fair than those within the
traditional benefits system.

Overall, our investigation into the effects of the different benefits systems on our
perceived control/faimess framework for predicting employee change behavior proved
only marginally useful. Two hypotheses were supported. We found that employees
within the incentive-based benefits system perceived their benefits system as less fair
than employees under the more traditional benefits system and that employees under the
incentive system reported higher commitment to maintaining or improving their health
levels than those under the traditional system. One hypothesis was significant, but in the
wrong direction such that employees within the traditional benefits system reported
performing more healthy behaviors than those within the incentive-based benefits system.
Finally, the hypothesized relationship between benefits system type and perceived control
was not supported. These results imply that perceived control perceptions and
commitment levels remain fairly unchanged under incentive-type programs, but that
perceptions of fairness may be influenced.

In the following section, we will discuss several limitations that were present in
the samples and the design of the study. Following that discussion, overall conclusions

and implications will be drawn.

101

Limitations

Throughout the previous section dedicated to reviewing the results concerning the
proposed hypotheses, three limitations proved important when considering the
conclusions drawn from this study and will be discussed.

The first limitation is related to the nature of the sample. Throughout this paper,
we have returned to the same issue underlying our data: are the samples representative of
both hospitals? We have proposed and continue to argue that they are. To prove this
point, we first argued that the demographic make-up of our incentive sample was
representative of the entire hospital employing the incentive-based program. While we
would have liked to have been able to say the same about the traditional sample, we did
not have demographic information on the entire hospital. But, we did have the
demographics on the incentive hospital and could assume that if the two samples were
similar, the traditional hospital was probably adequately represented by the traditional
sample due fact that both hospitals were located in the same city and selected employees
from the same labor market. Thus, we compared the demographic make-up of our
traditional sample to that of the incentive hospital and found that they were quite similar.
We used this as support that the traditional sample represented the entire hospital.

The second limitation also concerns the nature of the samples. Specifically, it
must be stated that these data are based on a quasi control group design (Cook &
Campbell, 1990). In our study, these two hospitals were quite similar in a number of
respects (i.e., location, demographics, salary ranges, and function), they were also quite
dissimilar. For instance, the hospital sizes differed and the culture and climate of the two

hospitals may have differed. Such designs and the issues we just presented allow for

102

other variable effects. While these possible effects have been considered, we realize that
others may exist. We recognize this and stress that the reader should also.

A final limitation we will discuss concerns not the nature of the sample, but that
of the data themselves. Specifically, the majority of our data were self-report, especially
those data involved in any analyses involving both samples. To investigate the
appropriateness of utilizing these self—report data, we found significant correlations
between the overall measures of both subjective and objective health (p.34; p<.05).
This analysis is far from conclusive, it does, however, suggest some preliminary support
for the validity of our self-report data. But, there are still implications that the use of
self-report data could have had for our study.

First of all, we expected that responses on several of our variables (e. g.,
commitment, health-related behaviors, and perceived control) would be heavily slanted to
the positive. We saw evidence that this was the case in our traditional group’s mean
health-related behavior responses. While we discussed why the self-report effects
probably did not affect the incentive group’s health behavior self-reports, we can fully
expect that the group’s reports of commitment and perceived control were influenced.
This limitation could explain the lack of any differences between the mean levels of the
two groups on perceived control and commitment.

Another limitation stems from the amount and nature of our items. Upon
inspection of the items in the appendices it is clear that some of our items are quite
repetitive, especially those measuring commitment and perceived control. It is very
likely that response bias could have played a role in these measures with participants

indicating the same or similar responses for all aspects of health. This could also explain

103

why we did not have mean differences between the groups on commitment nor perceived
control.

In sum, the use of self-report items presents a severe limitation in these data and
was probably the cause of the lack of support for the hypotheses presented concerning the
effects of the benefits system. While we did have access to objective commitment and
health behavior data in the incentive group, we did not have the same data in the
traditional group and we did not have adequate commitment data due to an extremely low
participation rate. In the future, studies should be directed at investigating these types of
phenomenon with more objective measures of perceived control, commitment, and

health-related behaviors.

104

Future Directions

In the search for ways to reduce health care costs, employers have turned to
attempting to influence the behaviors of their employees in ways that lead to using less
costly health care. By far the most frequently used method of control is to attempt to
influence what health care services employees use. Yet, another method is to try to
encourage healthy behavior in order to decrease the need for health care services. Such
influence attempts are rare, but the present study looked at one of these, an incentive-
based benefits system.

We argued a two-step model to explain employee perceptual and behavioral
reactions to these types of systems. The first step proposed a model linking perceptions
of control and fairness to commitment to improve or maintain health levels, and finally,
to actual health behaviors. The second step incorporated the effects of the incentive
system. More or less, our first step was supported while the second was not. There are
several reasons for our failure to provide the second step.

The first reason for our failure could stem from the program itself. Perhaps the
program didn’t influence employee behaviors as expected or maybe the program had
been in effect for too long prior to our involvement and attitudes and behaviors had
returned to normal. Future research should investigate these types of incentive programs
from the start and chart their progress and evolution to adequately explain and predict
employee attitudinal and behavioral reactions.

The second reason also relates to the program itself. Specifically, the incentive
program we investigated focused on influencing health-related behaviors, which are

influenced by many other factors outside of work. For example, everyone has a vested

105

interest in health issues with even the “unhealthy” being concerned with their health
levels. So, we investigated an incentive program that attempted to encourage behaviors
that were, probably, already important to the employees. With this in mind, few
individual health improvements or changes were probably possible with most pe0ple
doing as much as they could to maintain or improve their health levels before the
incentive-based program was introduced. This could explain why the second step was
not successful. Both groups were probably performing the same health-related behaviors,
both believing they could change their health the same amounts, and both committed
similarly to maintaining or improving their health levels regardless of the type of health
benefits system their organization employed. Future research should investigate
incentive programs more closely tied to work-related behaviors that are not a central life
issue and that can be modified with behavioral change.

Finally, and unrelated to the programs themselves, were the use of self-report
measures. We discussed this issue in the limitation section, but the use of these types of
measures could explain why our second step was not successful. Both groups were
equally likely to self-report that they were committed to improving or maintaining their
health and performing healthy behaviors. We had access to objective data in the
incentive organization, but not in the traditional group thus analyses involving both
groups and any objective data were impossible. To avoid this situation in future research,
measures of objective health behaviors would be ideal.

Moving away from the topic of why the second step failed, there is another issue
that our data could not address, but which should be addressed in any future research

endeavors. Specifically, while we say that our model relating perceptions of control and

106

fairness to commitment and health-related behaviors was supported, our data cannot
speak to the directionality of the relationships we hypothesized. It is plausible that the
relationships move in the opposite directions from those we proposed. Specifically, the
model we proposed could be reversed to propose that employees who are living healthy
lifestyles are more likely to report being committed to maintaining those behaviors and
thus perceive an incentive program focusing on the importance of good health as fair
because they do have control over their health outcomes. All of the analyses reported
would support this relationship, however it is unsupported in the literature. Future
research should investigate the exact nature of the relationships.

Our overall conclusions concerning these data and limitations are that employees
do base their commitment and subsequent health-related behaviors on perceptions of
fairness and control and that these types of incentive-based benefits programs may have
unintended effects on the fairness perceptions covered under these types of programs.
Specifically, while perceptions of control remain unchanged between the two groups,
commitment reports were positively affected and fairness perceptions were negatively

affected by the incentive system.

107

APPENDICES

108

APPENDIX A

LETTERS SENT TO SAMPLE OF EMPLOYEES AT BOTH HOSPITALS
Employee name and address]
Dear [Employee name]:

The Cook Institute, a non-profit research unit located in Grand Rapids and Michigan
State University are jointly sponsoring a research study designed to better understand
people’s health related behaviors. Their focus is on the role of the workplace in affecting
the health of employees and their families. They are interested in reactions of employees
to their medical benefits, the work climate as it relates to health, and the nature of their
behaviors as they relate to health. We have agreed to be one of the work sites
participating in this research because we are interested in many of the same things.

The study involves randomly drawing a sample of names from those persons currently
employed at [hospital name]. You were one of those persons randomly selected.
Although selected as part of the sample, the decision whether to participate is yours and
yours alone. Participation is voluntary. In this same mailing, you will receive more
information about the study from Professor Ilgen of Michigan State University that will
help you to make this decision. We urge you to consider this carefully and hope that you
will be willing to participate. We feel that the information that will be gained will be
valuable to us at [hospital name] and to the wide audience of employers as the latter tries
to meet the health needs of their employees.

More information and your copy of the survey is enclosed. Please read it carefully. If
you should have any questions about your involvement in the study, please contact
Deidre Wasson at [phone number].

Sincerely yours,

109

[Employee’s Address]

Dear [Employee’s name]:

As mentioned by Mr. Dillingham, we are conducting a study of factors related to the
health behaviors of employees. I do not have to tell you that health and health care are
extremely important to each of us. Health-related policies and practices that affect all of
us are being instituted every day. Yet, often these policies and practices are instituted
with little clear information about their impact on the health and welfare of each of us.
Our research is aimed at gathering such information.

Since the work place plays a major role in providing medical benefits and in affecting a
number of behaviors related to our health, the Cook Institute for Research and Education
of Grand Rapids along with Michigan State University is conducting a study of work and
health and well-being. Employment settings may encourage or discourage people to
engage in healthy behavior. Attempts to affect health through mandatory or voluntary
means lead to such things as rules for smoke free work spaces or the choices of foods in
the cafeteria. We are looking into a number of factors related to health at work in hopes
of learning what works well and what does not. It is our hope that by learning more
about health-related practices at work, health benefit systems, and peoples’ health-related
behaviors, better informed policy and practice decisions can be made at the national
level, by employers, and by you and me.

Part of the study involves administering a survey to employees. The survey takes about
15 minutes to complete. We have randomly selected a number of [hospital name]
employees and are asking them to volunteer to complete a survey anonymously. You
have been selected as part of that sample. Although we can offer no direct payment for
this, we hope they will agree with us that having data on which to make better informed
decisions regarding health is critical to all of us and that you are willing to provide such
data. For our part, we will provide you with a letter informing you of some of the
conclusions we have reached from the data when the analyses are complete.

Data collected from [hospital name] will be reported in one of two ways. The primary
way will be to combine it with other organizations where employees were asked the same
questions. A report will be prepared from this to be submitted to scientific and
professional journals. A second report will be prepared only for [hospital name]. Here
all the data will be reported in aggregate form in a way that will be focused toward
identifying things that are working well and those that might be improved concerning
peoples’ health and the nature of their work, work setting, and health benefits. This
information will be shared with participating organizations and their employees.

Participation in the study simply involves filling out the enclosed survey booklet. Once

you have completed the survey, please place the booklet into the self-addressed stamped
envelope so that it may be returned directly to us.

110

Thank you in advance for your participation.

Sincerely yours,

111

APPENDIX B

CONSENT FORM USED FOR HOSPITAL WITH INCENTIVE-BASED BENEFITS
SYSTEM

Survey No.

 

Consent Form
IRB 96-150
Spring 1996

I understand that I have been randomly chosen to participate in the study conducted by
researchers from Michigan State University in association with [hospital name].

My participation involves the completion of a questionnaire about the Health Quotient
Program and other conditions related to work and health. Responses to the questionnaire
will be kept strictly confidential. Any publication of the data will be aggregated with no
names or individual identification. My participation in this study is voluntary. IfI
choose to fill out the questionnaire and find items in it that I do not wish to complete, I
may leave them blank.

As has been explained to me, the questionnaire will be matched to personnel records on
file with my employer. I give the Michigan State University researchers permission to
access my Health Quotient data and my employment records that might be related to the
health care system. In this way, health behaviors can be related to attitudes and opinions.

Signature: Date:

 

 

Print Name:

 

112

APPENDIX C

CONSENT FORM USED IN TRADITIONAL HEALTH BENEFITS HOSPITAL

Survey No.

 

Consent Form
IRB 96-150
Fall 1996

I understand that I have been randomly chosen to participate in the study conducted by
researchers from Michigan State University in association with the Cook Institute.

My participation involves the completion of a questionnaire about conditions related to
work and health. Responses to the questionnaire will be kept strictly confidential. Any
publication of the data will be aggregated with no names or individual identification. My
participation in this study is voluntary. If I choose to fill out the questionnaire and find
items in it that I do not wish to complete, I may leave them blank.

Signature: Date:

 

 

Print Name:

 

113

APPENDIX D

INTRODUCTION OF SURVEY FOR THE HOSPITAL USING AN INCENTIVE-
BASED BENEFITS PROGRAM

This questionnaire is being administered to a sample of present and former employees of
the [hospital name] organization. A primary objective of the survey is to better
understand the reactions of employees to the health benefits program and to assess the
nature of the work climate with respect to health. A secondary concern is general levels
of satisfaction and opinions about working at [hospital name]. For the most part, these
feelings and reactions will be studied with respect to possible health programs. For
scientific purposes, the data collected will be combined with data collected at other sites
outside of the [hospital name] system in order to gain a better understanding of health
care systems from a broader perspective than just one company.

As a former employee of [hospital name] your perspective is unique and valuable to both
[hospital name] itself and the research team from Michigan State University.

Your participation in the study is voluntary and all data you provide will be kept strictly
confidential. If you should decide not to complete any or all of the questionnaire, that is
your right. The number on the top of the questionnaire is used to be able to pair
questionnaire opinion data with Health Quotient data and also personnel records data. As
a protection of confidentiality, responses to the questionnaire will never appear in any
form matched with an individual’s name. Filling out the questionnaire will represent your
willingness to be in the study and allow us to create a record with no name, but with both
survey data and employment record data.

To complete the questionnaire, please turn the page and follow the instructions. If you
have any questions, feel free to ask us. For those of filling this out at the worksite, we
will be there to respond to your concerns. For those of you doing this by mail, contact us
if you have any questions.

Thank you in advance for your participation.

114

APPENDIX E

INTRODUCTION OF SURVEY FOR HOSPITAL USING TRADITIONAL BENEFITS

This questionnaire is being administered to a sample of present employees of the
[hospital] organization. A primary objective of the survey is to better understand the
reactions of employees to their health benefits program and to assess the nature of the
work climate with respect to health. A secondary concern is general levels of satisfaction
and opinions about working at [hospital]. The data will be combined with data collected
at other sites outside of the [hospital] system in order to gain a better understanding of
health care systems from a broader perspective than just one employer. As an employee
of [hospital your perspective is unique and valuable. When combined with many other
employees’ views these are useful to both [hospital] itself and the research team from
Michigan State University.

Your participation in the study is voluntary. All data are anonymous. Do not put your
name on the questionnaire. Numbers do appear on return envelopes to be used only for
second mailings if necessary. If you should decide not to complete any or all of the
questionnaire, that is your right. As a protection of confidentiality, responses to the
questionnaire will never appear in any form matched with an individual's name. Filling
out the questionnaire will represent your voluntary participation in the study.

Please turn the page and follow the instructions. Thank you in advance for your
participation.

115

ITEMS ADAPTED TO MEASURE FAIRNESS PERCEPTIONS IN THE INCENTIVE-

APPENDIX F

BASED BENEFITS SYSTEM AND THE TRADITIONAL SYSTEM UNLESS

 

 

 

 

 

 

 

 

 

INDICATED
Item Incentive system items Traditional system Specific Reverse
# items Fairness Coding
Domain
My employer provides a Same Procedural
1 reasonable explanation for Justice
the way our health benefits
system is designed
Staff members who are Same Distributive
2 greater health risks should Justice
have to pay more for their
health insurance
It is not fair for staff Same Distributive
members who live healthy Justice
3 lifestyles to pay the same
for company health
insurance as those who
don’t
4 Our health benefits system Same Distributive
is a fair one Justice
Since the health of Same Distributive
dependents affects the cost Justice
of health insurance, it is
5 most fair to consider
dependents health and
health behaviors along with
that of staff members
The administration of the The administration Distributive
6 [program name] treats all of the health Justice
employees fairly benefits system
treats all employees
fairly
Some groups of staff Some groups of Dropped R
7 members are unfairly staff members are
treated by the [program unfairly treated by
name] the health benefits
system

 

 

 

 

 

116

 

 

 

 

 

 

 

 

 

 

 

 

When it comes to health Same Distributive
8 benefits, all staff members Justice
are treated the same
The [program name] is The health benefit Dropped
biased against those with system is biased
9 uncontrollable health against those with
problems uncontrollable
health problems
The [program name] favors The health benefits Distributive
10 some groups of staff system favors some Justice
members groups of staff
members
Staff members who are Staff members who Dropped
sometimes less than are sometimes less
11 perfectly honest take than perfectly
advantage of the [program honest take
name] advantage of the
health benefits
system
When I have a question or Same Procedural
need to talk to someone Justice
12 about health care benefits, I
am treated with respect and
dignity
Any testing done by the Not applicable Dropped
[assessment center name]
N/A related to the [program
name] is done in a way that
respects my privacy
I like the way that they I like the way that Procedural
treat me when I have to they treat me when Justice
13 deal with anything I have to deal with
regarding the [program anything regarding
name] the health benefits
system
When I disagree with When I disagree Procedural
14 something regarding the with something Justice
[program name], there is an regarding benefits
appeal process that is fair there is an appeal
If I suspect that there has Not applicable Dropped
been a mistake regarding
N/A my Health Quotient, the
[benefits office] is very
willing to address it with
me
I feel free to raise questions I feel free to raise Procedural

 

 

 

 

117

 

 

 

 

 

 

 

 

 

 

 

 

about the [program name] questions about the Justice
15 with my supervisor or the health benefits
benefits office system with my
supervisor or the
benefits office
They listen to you here if They listen to you Procedural
16 you have concerns about hear if you have Justice
the [program name] concerns about
health benefits
17 The health care system at Same Dropped
this organization is a fair
one
18 My health care system is Same Dropped
administered fairly
19 Staff working with health Same Procedural
care treat me fairly Justice
20 Staff members at [hospital Same Distributive
name] get what they Justice
deserve from the health
care system
21 The cost of the health care Same Dropped
system is fair to staff
members
I am satisfied with the Same Dropped
22 amount this health care
system requires me to pay
for health care
I feel that the costs to me Same Dropped
23 of the health care system

 

are unfair

 

 

 

118

 

 

APPENDIX G

FACTOR SOLUTIONS FOR FAIRNESS ITEMS

 

 

 

Primary 2-factor solution Secondary 2-factor solution
Incentive Sample Traditional Sample Incentive Sample
(N=243) (N=273) (N=243)

Items Factor] Factor2 Factor] Factor2 Factorl Factor 2
4 .757 .301 .637 .299 .712 .368
20 .749 .358 .733 .201 .675 .407
2 .734 -.010 -.174 .746 .706 .1 14
17 .707 .443 .746 .227 Dropped

6 .694 .369 .560 .405 .632 .465
5 .666 .006 -.094 .696 .592 .180
3 .664 -.106 -.181 .787 .659 .018
18 .645 .510 .769 .152 Dropped

2 l .639 .414 .693 .259 Dropped

8 .604 .236 .536 .369 .520 .317
22 .555 .412 .632 .036 Dropped

10 .545 . 132 .504 .462 .434 .220
23 .479 .319 .592 .170 Dropped

9 .412 .209 .400 .410 Dropped

7 .316 .189 .526 .425 Dropped

16 .265 .710 .719 -.113 .225 .710
13 .309 .702 .725 -.120 .282 .671
15 .288 .698 .614 -.025 .232 .699
12 -.032 .675 .637 -. 158 -.025 .609
14 .294 .582 .413 .000 .268 .510
l .337 .576 .644 -.063 .298 .595
19 .349 .500 .751 -.026 .280 .485
1 l -.020 . 133 .205 .317 Dropped

119

APPENDIX H

ITEMS CREATED OR ADAPTED TO MEASURE HEALTH-RELATED

BEHAVIORS IN THE INCENTIVE-BASED BENEFITS SYSTEM AND THE

TRADITIONAL BENEFITS SYSTEM PRESENTED BY EACH HEALTH

 

 

 

 

 

DIMENSION
Health Item Symbol Health Behavior Reverse
Measured Coded
I consider myself a very relaxed STl Stress
person, under little stress.
I am almost always tense and I ST2 Stress R
find it difficult to relax.
I am effective in managing my ST3 Stress
stress.
Generally, I am satisfied with my ST4 Stress
life.

 

My relationships with my family STS Stress
and friends are strong.

 

I get along well with my family ST6 Stress

 

 

 

 

 

 

 

 

 

 

and friends.

I regularly experience significant 8T7 Stress R
tension from my job.

I rarely seem to have enough time ST8 Stress

to get all my work done.

My co-workers are often pushed to 8T9 Stress R
the limit due to the amount of

work they have.

Staff members here are under a lot ST10 Stress R
of pressure.

Staff members here worry because STll Stress R
of their jobs.

I would smoke cigarettes or cigars TOBl Tobacco Use R
while working if it were allowed.

I feel that smoking is a nice way to TOB2 Tobacco Use R
take a break from work.

I like to smoke on my work TOB3 Tobacco Use R
breaks.

I often see other staff members TOB4 Tobacco Use R

smoking while at work.

 

 

 

 

 

120

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I think smoking is a bad habit. TOBS Tobacco Use
I am proud to be a non-smoker. TOB6 Tobacco Use
I would like a lenient smoking TOB7 Tobacco Use
policy, one that allows smoking at

work.

I feel that it is not acceptable to TOB8 Tobacco Use
smoke at this workplace.

For the past year I have not TOB9 Tobacco Use
smoked a single cigarette.

During the past year, I have not TOB10 Tobacco Use
used any tobacco products.

I use tobacco products at least TOBll Tobacco Use
once a day.

I use tobacco products once in a T0812 Tobacco Use
while, but I do not consider my

use a habit.

I regularly used tobacco products TOBl3 Tobacco Use
in the past, but I no longer do.

I drink at least one alcoholic ALCl Alcohol Use
beverage at least once a week.

I enjoy an alcoholic drink after ALC2 Alcohol Use
work or on the weekend with

friends.

An alcoholic beverage after work ALC3 Alcohol Use
helps me to relax.

I rareny drink alcohol. ALC4 Alcohol Use

I drink alcoholic beverages almost ALCS Alcohol Use
everyday.

I will have an alcoholic beverage ALC6 Alcohol Use
every once in a while.

I eat healthy snacks such as carrot NUTl Nutrition/Diet
sticks, low-fat yogurt, or fruit

when I’m hungrybetween meals.

I think I am a good role model for NUTZ Nutrition/Diet
making nutritious food choices.

Honestly, I have unhealthy eating NUT3 Nutrition/Diet
habits.

I make an effort to include NUT4 Nutrition/Diet
vegetables, salads, or fruit in my

meals at work.

I am concerned about the amount NUTS Nutrition/Diet
of cholesterol in the foods I eat.

I regularly choose high fat foods NUT6 Nutrition/Diet

for lunch because I am just too
busy.

 

 

 

121

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I rarely, if ever, eat high fat foods. NUT7 Nutrition/Diet

Often, while I am at work, I eat NUTS Nutrition/Diet

potato chips, or candy bars for a

quick snack.

I make an effort to include whole NUT9 Nutrition/Diet

grains, fruits, vegetables, and

lowfat dairy in each of my meals.

I really don’t care about what I eat. NUT10 Nutrition/Diet

I just grab whatever I can for lunch NUTll Nutrition/Diet

while I’m at work.

I don’t have time to worry about NUT12 Nutrition/Diet

eating foods from the 5 basic food

groups.

I always buckle my seatbelt while SAFl Motor Vehicle

I am driving or riding in a motor Safety

vehicle.

I rarely buckle my seatbelt while SAF2 Motor Vehicle

in a motor vehicle. Safety

I ride a motorcycle. SAF 3 Motor Vehicle
Safety

When I ride (or if I rode) a SAF4 Motor Vehicle

motorcycle, I always wear a Safety

helmet.

During the past 5 years, I have SAFS Motor Vehicle

almost always had one or more Safety

points on my drivirirecord.

I rarely have even one point on my SAF6 Motor Vehicle

driving record. Safety

I never drive in a motor vehicle SAF‘7 Motor Vehicle

when I have had too much to Safety

drink.

I always make sure that we have a SAF8 Motor Vehicle

designated driver when I know Safety

that I will be drinking away from

home.

I am capable to drive home after SAF9 Motor Vehicle

less than 2 drinks. Safety

I never drive home if I have had SAF10 Motor Vehicle

even one alcoholic beverage. Safety

I belong to a fitness or health club. EXERl Fitness/Exercise

I always find time to exercise EXER2 Fitness/Exercise

before or after work.

I would consider myself a “health EXER3 Fitness/Exercise

nut” because I like to exercise so
much.

 

 

 

122

 

 

 

 

 

 

 

 

 

 

 

 

I am actively working to improve EXER4 Fitness/Exercise

my physical fitness.

I participate in sports to keep EXERS Fitness/Exercise
physically active.

I walk for exercise during lunch or EXER6 Fitness/Exercise

other breaks.

I exercise, other than walking, EXER7 Fitness/Exercise

during normal working hours.

I think that people who exercise EXER8 Fitness/Exercise R
are a bit “crazy”.

I feel that exercise is not very EXER9 Fitness/Exercise R
important.

Honestly, I think exercise is a EXERl Fitness/Exercise R
waste of my time. 0

I really think the benefits of EXERl Fitness/Exercise R
exercise are overrated. 1

 

 

 

 

123

 

 

APPENDIX I

FACTOR SOLUTION FOR HEALTH BEHAVIOR SELj-REPORT ITEMS

 

Factor] Factor2 Factor3 Factor4 FactorS
EXERlO .959 —.002 -.00] .005 .006
SAF2 .936 .004 .005 .006 —.003
EXER] l .919 -.002 .00] .008 .006
SAF3 .916 -.000 .004 .003 .00]
EXER9 .899 -.005 -.00] .008 .106
ALC7 .880 .105 -.007 .005 -.004
EXER8 .877 -.004 -.002 .007 .122
SAFS .866 .004 .007 .004 -.003
NUT10 .860 .004 .112 .002 .001
TOB12 .846 .122 -.006 .104 -.102
NUT6 .759 .005 . 122 .005 .007
NUT] 1 .758 .005 .132 -.004 .008
NUT12 .712 .004 .155 -.001 .007
ALC3 .710 .009 .000 .007 -.001
NUTS .651 .005 .l 14 .004 .008
TOB 10 -.001 .868 .003 .005 .004
TOB9 -.002 .867 -.000 .006 .009
TOB3 .008 .861 .005 .1 17 .004
TOB] 1 .004 .842 .003 .006 .008
TOB2 .007 .810 .006 .146 .008
TOB] .010 .743 .005 .233 -.000
TOB7 .007 .500 .002 .277 .002
NUT2 .007 .005 .788 .158 .245
NUT] .004 .003 .766 .006 .213
NUT9 .005 .002 .763 -.001 .206
NUT4 .002 .003 .697 .003 . 1 19
NUT7 -.000 —.001 .635 .005 .287
NUTS .005 .005 .63 1 -.009 .165
SAF 1 .009 .008 .470 .004 .010
ST10 .010 .139 -.001 .815 .007
ST9 .006 .008 -.001 .801 .007
ST8 . 180 .003 .005 .724 . 106
ST7 .009 .163 .005 .720 -.003
STll .000 .16] .000 .715 .112
ST] .005 . 144 .172 .563 .008
EXER3 .007 .008 .345 . 102 .742
EXERS .002 .002 . 194 .003 .730
EXER2 -.002 .007 .377 . 123 .703
EXER] .109 .003 .009 .001 .636
EXER7 . 108 .002 . 150 .006 .635

124

Factor] Farctor2 Factor3 Factor4 FactorS

EXER4 -.001 .007 .477 .002 .578
EXER6 -.001 .003 .269 .125 .530
SAF7 .009 .009 .133 -.005 .160
SAF8 .004 .003 .129 .002 .237
SAF9 .162 -.001 .181 .006 .009
SAFll 106 .105 .223 .004 .198
SAF4 .172 .009 .210 .109 -.112
ALC2 .271 .003 .004 .147 .003
ALC] .432 .006 .001 .007 -.000
SAF10 .004 .009 .198 .006 . 146
ALC8 -.3 17 .005 .127 .236 .009
STS .008 .183 .156 .237 .108
ST6 .009 .184 .141 .229 .107
ST4 .002 .23] .194 .309 .003
ST3 -.001 .160 .184 .396 .135
ALCS .005 -.000 -.003 .009 .009
ALC4 .005 -.004 .009 .008 . 146
TOB13 .003 .141 .004 .127 .009
TOBS .004 .150 -.000 .003 -.001
TOB8 .000 .296 .009 .216 .002
TOB6 .003 .446 .010 . 128 .003
NUT3 .31 l .006 .244 .134 .142
ST2 .010 .143 .220 .472 -.004
SAF6 .007 .006 .204 .000 .010
TOB4 .200 .200 -.000 .179 .00]
ALC6 -.003 .159 -.001 .00] .128

125

 

Factor6 Factor7 Fggor8 Fzgtor9 FactorIO
EXERIO .007 .002 .000 .00] .000
SAF2 .006 .00] -.000 .003 -.002
EXER] 1 .009 -.000 .003 .000 -.002
SAF3 .003 .002 -.001 .001 .000
EXER9 .005 .001 -.001 -.002 .004
ALC7 -.001 .248 .007 .006 .003
EXER8 .006 .003 .003 .002 -.003
SAFS 010 -.001 -.004 -.003 .00]
NUT10 .005 -.001 .003 .002 .001
TOB 12 -.004 .1 10 -.000 .007 .003
NUT6 .005 .000 .002 .003 .006
NUT] 1 .001 .005 .007 .003 .003
NUT12 .005 .002 .006 .010 .004
ALC3 -.002 .501 .010 .002 .002
NUT8 .005 .003 .005 -.005 .003
TOB10 .010 -.002 -.002 -.004 .004
TOB9 .009 .000 -.007 -.003 .002
TOB3 .001 .006 . 146 .005 .137
TOB] ] -.002 .003 .009 .001 .002
TOB2 .003 .007 . 176 .005 . 148
TOB] .004 .003 . 172 .106 .174
TOB7 -.002 .009 .138 -.001 .400
NUT2 .009 .001 .003 .002 -.003
NUT] .005 .002 -.000 .009 .005
NUT 9 .003 .005 .126 -.003 .004
NUT4 .002 .1 13 .179 —.005 -.006
NUT7 . 150 .002 -.008 .004 .1 10
NUTS . 198 .003 .008 .007 -.000
SAF] .310 -.002 .158 -.005 .008
ST 10 .001 .009 .146 .003 .005
ST9 .001 .007 .009 .136 .005
ST8 .002 . .009 -.002 -.003 .006
ST7 .000 .001 .109 .008 .00]
ST]] .002 .118 .192 .004 .140
ST] -.001 .001 .238 .008 .006
EXER3 .008 .002 .002 .002 .008
EXERS .008 -.002 .010 .124 .000
EXER2 -.00] .003 .005 .001 .003
EXER] . 149 -.003 .003 .005 .005
EXER7 .154 . 139 -.002 .008 -.006
EXER4 . l 13 -.003 .126 .007 .002
EXER6 .143 . 133 .006 .003 -.104

126

Factor6 F4actor7 Factor8 Factor9 fiactorlO

SAF7 .815 -.000 .007 .008 —.001
SAF8 .785 .001 .005 -.004 -.003
SAF9 .751 . 158 .009 .004 .007
SAFll .492 .238 -.006 -.l76 .005
SAF4 .348 -.010 -.176 .256 .l 17
ALC2 .005 .851 .004 .006 .003
ALC] -.000 .746 .006 .006 .005
SAF10 .336 .609 -.006 .009 -.105
ALC8 .110 .538 -.010 .154 .009
ST5 .004 -.001 .823 .005 .136
ST6 .007 . -.001 .806 .010 . 154
ST4 .008 .009 .640 .108 .002
ST3 .002 .000 .444 .009 .006
ALC5 -.000 .010 .002 .801 .004
ALC4 .165 .106 .005 .714 .160
TOB13 -. 168 -.001 .179 .574 —.009
TOB5 .002 -.003 . 146 .010 .740
TOB8 .005 .005 .006 .004 .570
TOB6 .001 .007 .006 -.002 .519
NUT3 . 147 -.004 .007 .005 .005
ST2 . 135 -.003 .259 .131 .009
SAF6 .192 -.003 -.007 .134 .157
TOB4 .008 .002 -.004 . 133 .245
ALC6 .27] .303 .177 -.213 .295

127

 

 

 

Factor 1 l fictorl 2 Factor] 3

EXERIO .003 .00] -.002
SAF2 -.004 .007 .000
EXER] l -.001 .001 -.002
SAF3 -.005 .003 -.003
EXER9 -.00] -.001 -.005
ALC7 -. 160 .005 .007
EXER8 .002 -.002 .001
SAFS -.006 -.003 .008
NUT 10 .204. 002 .001
TOB 12 -.005 .008 .005
NUT6 .398 -.002 -.004
NUT] 1 .430 .004 -.004
NUT12 .435 .006 .003
ALC3 -.138 .002 .003
NUT8 .487 -.002 -.006
TOB10 .002 .177 .159
TOB9 .004 .178 .166
TOB3 -.002 -.008 -. 129
TOB] 1 .004 .008 .1 17
TOB2 -.00] -.008 -. 145
TOB] .003 -. 120 -.244
TOB7 .001 -.185 -.l 13
NUT2 .002 005 -.001
NUT] .004 .1 18 -.00]
NUT9 .007 .007 .002
NUT4 .009 -.007 .009
NUT7 .007 008 .001
NUTS -.001 .003 .005
SAF] .115 -.]80 .008
ST10 .004 -.105 -.001
ST9 .004 .001 .009
ST8 .001 .1 17 .005
ST7 .006 .192 -.008
ST] 1 -.003 -. 155 -.007
ST] -.002 .49] -.010
EXER3 -.001 -.00] .005
EXERS -.001 .009 -.002
EXER2 .001 -.002 .001
EXER] .002 -.002 -. 126
EXER7 .1 1 1 .008 188
EXER4 -.007 -.008 -.001

 

 

Factor] 1 Factor12 Factor] 3
EXER6 . 174 -.006 . 168
SAF7 .009 .003 .126
SAF8 .004 .00] .144
SAF9 .003 .001 -.010
SAF 1 1 -.00] .144 -.009
SAF4 -.009 -.184 .172
ALC2 -.008 .003 .003
ALCl -.127 .009 .004
SAF10 .235 -.134 -.161
ALC8 .223 -.010 -.001
ST5 -.003 .000 -.002
ST6 -.007 -.001 -.002
ST4 .169 .201 .003
ST3 .010 .403 -.003
ALC5 -.003 .003 .001
ALC4 .001 .007 -. 105
TOB13 .010 .000 .183
TOBS .007 .129 .003
TOB8 -.003 -.219 .005
TOB6 .00] .163 .154
NUT3 .652 .001 .007
ST2 -.004 .505 -.007
SAF6 .005 -.004 .676
TOB4 .003 .326 -.407
ALC6 -. 145 .199 .305

129

APPENDIX J
ITEMS CREATED TO MEASURE PERCEPT IONS OF CONTROL OVER
OUTCOMES AND HEALTH BEHAVIORS IN THE INCENTIVE-BASED BENEFITS

SYSTEM AND THE TRADITIONAL SYSTEM UNLESS INDICATED

1. Perceived control over health.

Consider each of the dimensions below individually and answer the following
items based on how POSSIBLE you feel it is to CHANGE your health level on each
health dimension within the next year. Please keep in mind, this statement is not
referring to your Health Quotient Score, just your general opinion about how easily you
could change your health level on each dimension.

1. Cholesterol level
2. Blood pressure

3. Body fat

4. Fitness level

5. Motor vehicle safety
6. Tobacco use

7. Alcohol use

8. Nutritional habits

Responses were indicated on a five-point Likert-type scale anchored at Not at all
Possible to Very Possible.

2. Perceived control over outcomes*

Assume you changed your behaviors to try to improve your health. How likely is
it that your overall Health Quotient Score would improve?

1. Cholesterol Level
2. Blood Pressure

3. Body Fat

4. Fitness Level

5. Motor Vehicle Safety
6. Tobacco Use

7. Alcohol Use

8. Nutritional Habits

Responses were indicated on a five-point Likert-type scale anchored at Extremely
Unlikely to Extremely Likely.

130

*Denotes that these data were not collected within the traditional benefits system.

131

APPENDIX K
ITEMS CREATED TO MEASURE PERCEPTIONS OF COMMITMENT TO
CHANGE HEALTH-RELATED BEHAVIORS IN THE INCENTIVE-BASED
BENEFITS SYSTEM AND THE TRADITIONAL BENEFITS SYSTEM

Please consider how Willing you are to work on improving your health on each of
the eight health dimensions to a level within a healthy range.

1. Cholesterol level

2. Blood pressure

3. Body fat

4. Fitness level

5. Motor vehicle safety

6. Tobacco use

7. Alcohol use

8. Nutritional habits

Responses were indicated for each item on a Likert-type five point scale with anchors at
Not at all_Willing and Extremely Willing.

132

Appendix L.
EIGHT ITEMS CREATED TO MEASURE PERCEPTIONS OF ACCURACY IN THE

INCENTIVE-BASED BENEFITS SYSTEM

For each of the following statements, please rate how much you agree with each
statement according to the scale described below.

I. The blood test to measure C/HDL for the Health Quotient is accurate.

2. The pressure cufi‘ test to measure blood pressure for the Health Quotient is

accurate.

3. The caliper test used to measure body fat for the Health Quotient is accurate.

4. The fitness walk test to measure fitness for the Health Quotient is accurate.

5. The self report measure to estimate motor vehicle safety for the Health
Quotient is accurate.

6. The self report used to measure tobacco use for the Health Quotient is

accurate.

7. The self report used to measure alcohol consumption for the Health Quotient

is accurate.

8. The self report used to measure nutritional behavior for the Health Quotient is

accurate.

Responses were indicated on a five-point Likert-type scale anchored at Strongly Disagree
and.Strongly Disagree.

133

BIBLIOGRAPHY

134

Bibliography

Abramson, L. Y., Seligman, M. E. P., and Teasdale, J. (1978). Learned

helplessness in humans: Critique and reformulation. Journal of Abnormal Psychology,
87, 49-74.

Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.),
Aflmces in Experimental Social Psychology (Vol. 4, pp. 267-299). New York:
Academic Press.

Ajzen, I., & Fishbein, M. (1986). Understanding attitudes and predicting social
behavior. Englewood Cliffs, NJ: Prentice-Hall.

Arvey, R. D. (1972). Task performance as a function of perceived effort—

performance and performance-reward contingencies. Organizational Behavior and
Huraan Performance, 8, 423-443.

Atkinson, J. W. (1958). Towards experimental analysis of human motivation in
terms of motives, expectancies, and incentives. In J. W. Atkinson (Ed.), Motives in
fantasy, action, and society (pp. 288-305).

Atkinson, J. W. & Reitman, W. R. (1956). Performance as a function of motive

strength and expectancy of goal attainment. Journal of Abnormal and Social Psychology,
_5_3, 361-366.

Austin, J. T. & Vancouver, J. B. (1996). Goal constructs in psychology:
Structure, Process, and Content. Psychological Bulletin, 120(3), 338-375.

Bandura, A.(1986). The explanatory and predictive scope of self-efficacy theory.

Special issue: Self-efficacy theory in contemporary psychology. Journal of Social and
Clinical Psychology. 4(3), 359-373.

Barrett-Howard, E., & Tyler, T. R. (1986). Procedural justice as a criterion in
allocation decisions. Journal of Personaligurnd Social Psychology, 50, 296-304.

Berger, B. G. (1984). Running toward psychological well-being: Special
considerations for the female client. In M. L. Sachs and G. Buffone (Eds), Running as t
herapy: An integrated approach. Lincoln: University of Nebraska Press.

Bernstein, D. A. & Glasgow, R. E. (1979). Smoking. In 0. F. Pomerleau and J.
P. Brady (Eds.), Behavioral medicine: Theory and practice. Baltimore: Williams &
Wilkins.

Berry, R. E. & Boland, J. P. (1977). The economic cost of alcohol abuse. New

York: Free Press.

135

Bies, R. J. (1986, August). Identifying principles of interactional arstice: The
case of corporate recruiting. Symposium conducted at the annual meeting of the
Academy of Management, Chicago, IL.

Bies, R. J. (1987). The predicament of injustice: The management of moral
outrage. In L.L. Cummings & B. M. Staw (Eds.), Research in organizational behavior
(Vol. 9, pp. 289-319). Greenwich, CT: JAI Press.

Bies, R. J ., & Moag, J. S. (1986). Interactional justice: Communication criteria
of fairness. In R. J. Lewicki, B. H. Sheppard, & M. Bazemab (Eds.), Research on
Negotiation in Organizations (Vol. 1, pp. 43-55). Greenwich, CT: JAI Press.

Borland, R., Owen, N ., Hill, D., & Schofield, P. (1991). Predicting attempts and
sustained cessation of smoking after the introduction of workplace smoking bans. Health

 

Psychology, 10, 336-342.

Bourgeois, L. J ., III, & Eisenhardt, K. M. (1988). Strategy decision processes in a

high velocity environment: Four cases in the microcomputer industry. Management
Science. 34. 816-835.

Braunwalk, E. (1994). Myocardial ischemia, infarction, and failure: An odyssey.

Cardioscience. 5, 139-144.

Broaunwald, E. (Ed.) (1994). Principle of internal medicine, 12th ed. New
York: McGraw-Hill.

Brockner, J ., Wiesenfeld, B., Stephan, J ., & Hurley, R. (1997). The effects on
layoff survivors of their fellows survivors’ reactions. Journal of Applied Social

Psychology, 27(10), 835-863.

Campion, M. A., & Lord, R. G. (1982). A control systems conceptualization of

the goal setting and changing process. Organizational Behavior and Human
Performance. 30, 265-287.

Cassileth, B. (1969). Reinforcement management: An approach to motivating
Army trainees (HumRRO Technical Report 69-17, AD672484). Fort Knox, KY:
HumRRO Division 2.

Cohen, R. L. (1987). Power and injustice in intergroup relations. In H. W.
Bierhoff, R. L. Cohen, & J. Greenberg (Eds.), Justice in Social Relations (pp. 65-84).

New York: Plenum Press.

Colosi, M. (1988, April). Do employees have the right to smoke? Personnel
Journal, pp. 72-79.

136

Cropanzo, R. & Folger, R. (1989). Referent cognitions and task decision
autonomy: Beyond equity theory. Journal of Applied Psychology, 74, 293-299.

Daltroy, L. H. & Liang, M. H. (1991). Advances in patient education in
rheumatic disease. (Review). Annals of the Rheumatic Diseases 50(suppl.3), 415—417.

 

Date], W. (1970). Contingency management in basic combat training. Current

Trends in Army Medical Service Psychology (PP. 106-120). Washington DC:
Department of the Army, Office of the Surgeon General.

Date], W., & Legters, L. (1971). The psychology of the Army recruit. Journal of
Biological Psychology, 12(2), 24-40.

Dawson, D. A., Harford, T. C., and Grant, B. F. (1992). Family history as a

predictor of alcohol dependence. Alcoholism: Clinical and Experimental Research. 16,
572-575.

DeLeo, P. J. (1972). An errmirical investigation of expectancy-valence concepfi

and relationships. Unpublished doctoral dissertation, Purdue University, West Lafayette,
IN.

Deutsch, M. (1957). Trust and suspicion. Conflict Resolutiona 2, 265-279.

Dimsdale, J. E., Alpert, B. S., and Schneiderman, N. (1986). Exercise as a
modulator of cardiovascular reactivity. In K. A. Matthews, S. M. Weiss, T. Detre, et al.
(Eds.), findbook of stress. reactivity. ml cardiovascular disease. New York: Wiley.

Dockstader, S. L., Nebeker, D. M., Nocella, J ., & Shumate, E. C. (1980).
Incentive management training: Use of behavioral principles for productivity
enhancement (NPRDC-TR-80-29). San Diego, CA: Navy Personnel Research and
Development Center.

Evans, M. R. (1970). The effects of supervisory behavior on the path-goal
relationship. Organizational Behavior and Human Performance. 5. 277-298.

Ezzell, C. (1995). Fat times for obesity research: Tons of new information, but
how does it all fit together. J oumai of NIH Research. 7, 39-43.

Faust, I. M. (1980). Nutrition and the fat cell. International Journal of Obesity. 4,
314-32].

Folger, R. (1986). Rethinking equity theory: A referent cognitions model. In
H. W. Bierhoff, R. L. Cohen, & J. Greenberg (Eds.), Justice in social relations_ (pp.
145-162). New York: Plenum Press.

137

Folger, R., & Martin, C. (1986). Relative deprivation and referent cognitions:

Distributive and procedural justice effects. Journal of Experimental Social Psychology,
2_2_, 531-546.

Folkins, C. H. & Sime, W. E. (1981). Physical fitness training and mental health.
American Psychologist, 36, 373-389.

Frankenhaeuser, M. (1983). The sympathetic-adrenal and pituitary—adrenal
response to challenge: Comparison between the sexes. In T. M. Dembroski, T. H.
Schmidt, and G. Blumchen (Eds.), Biobehavioral bases of coronal heart disease. Basel:
Karger.

Galbraith, J ., & Cummings, L. L. (1967). An empirical investigation of the
motivational determinants of task performance: Interactive effects between

instrumentality-valence and motivation-ability. Organizationaflehaviorard Human
Performance. 2, 237-257.

Gavin, J. F. (1970). Ability, effort, and role perception as antecedents of job
performance. APA Experimental Publication System. 5(190A).

Gilliland, S. W. (1993). The perceived fairness of selection systems : An
organizational justice perspective. Academy of Ma_nagement Review. 18. 694-734.

Gilliland, S. W. (1994). Effects of procedural and distributive justice on
reactions to a selection system. Journal of Applied Psychology, 79, 5, 691-70]

Graen, G. (1969). Instrumentality theory of work motivation: Some experimental
results and suggested modifications. Journal of Applied Psychology, 53, 1-25.

Greenberg, J. (1982). Approaching equity and avoiding inequity in groups and
organizations. In J. Greenberg & R. L. Cohen (Eds.), Equity and justice in social
behavior (pp. 389-435). New York: Academic Press.

Greenberg, J. (1983). Overcoming egocentric bias in perceived fairness through

self-awareness. Social Psychology Quarterly, 46, 152-156.

Greenberg, J. (1987). A taxonomy of organizational justice theories. Academy of
Management Review, 12, 9-22.

 

Greenberg, J. (1990a). Employee theft as a reaction to underpayment inequity:
The hidden costs of pay cuts. Journal of Applied Psychology. 75, 561-568.

Greenberg, J. (1990b). Looking fair vs. being fair: Managing impressions of

organizational justice. In B. M. Staw & L. L. Cummings (Eds.), Research in
organizational behavior (Vol. 12, pp. 111-157). Greenwich, CT: JAI Press.

138

Greenberg , J. (1990c). Organizational justice: Yesterday, today, and tomorrow.
J cum] of Marggement. 16, 399-432.

Greenberg, J. (1993a). The social side of fairness: Interpersonal and
informational classes of organizational justice. In R. Cropanzo (Ed.), Justice in the

workplace: Approaching fairness in human resource management (pp. 79-103).
Hillsdale, NJ: Erlbaum.

 

Greenberg, J. (1993b). Stealing in the name of justice: Informational and

interpersonal moderators of theft reactions to underpayment inequity. Organizational
Beaavior and Human Decision Processes. 54, 81-103.

Greenberg, J. (1994). Using socially fair treatment to promote acceptance of a
work site smoking ban. Journal of Applied Psychology, 79, 2, 288-297.

Greenberg, J ., Bies, R. J. & Eskew, D. E. (1991). Establishing fairness in the
eye of the beholder: Managing impressions of organizational justice. In R. Giacalone &
P. Rosenfeld (Eds.), Applied impression management: How image making affects
managerial decisions (pp. 111-132). Newbury Park, CA: Sage.

Greenberg, J. & McCarty, C. (1990). The interpersonal aspects of procedural
justice: A new perspective on pay fairness. Labor Law Journal. 41. 580-586

Griffiths, R. R., Bigelow, G. E., and Liebson, I. (1978). The relationship of social
factors to ethanol self-administration alcoholics. In P. E. Nathan, G. A. Marlatt, and T.

Y. Loeberg (Eds.), Alcoholism: New directions in behaviorairesearch and treatment.
New York: Plenum.

Guzzo, R. A. (1979). Types of rewards, cognitions, and work motivation.
Academy of Marggement Review. 4(1), 75-86.

Guzzo, R. A., Jette, R. D., & Katzell, R. A. (1985). The effects of

psychologically based on intervention programs on worker productivity: A meta-
analysis. Personnel Psychology, Vol 38(2), 275-291.

Hackman, J. R., & Porter, L. W. (1968). Expectancy theory predictions of work
effectiveness. Organizational Belgvior and Human Performance. 3. 417-426.

Hakas, J. L., Gajiwala, K. S., Maffei, M., et al. (1995). Weight reducing effects
of the plasma protein encoded by the obese gene. ScienceL269, 543-546.

Heider, F. (1958). The psychology of integpersonal relations. New York:
Wiley.

Hirsch, J. & Knittle, J. L. (1970). Cellularity of obese and nonobese human
adipose tissue. Federal Proceedings, 29, 1516-152].

139

Hollenbeck, J ., & Klein, H. (1987). Goal commitment and the goal setting
process: Problems, prospects, and proposals for future research. Journal of Applied
Ps cho o 72, 212-220.

Homans, G. C. (1961). Social behavior: Its elementaryforms. New York:
Harcourt, Brace, and World.

Hulin, C. (1991). Adaptation, persistence, and commitment in organiZations. In
M. D. Dunnette & L. M. Hough (Eds.), Handbook of Industrial and Organizational
Psychology(2nd ed., Vol. 2, pp. 445-505). Palo Alto, CA: Consulting Psychologists
Press.

Hymowitz, N. (1991). Long-term smoking intervention at the worksite: Effects
of quit-smoking groups and an “enriched milieu” on smoking cessation in adult white-
collar employees. Health Psychology, 10(5), 366-369.

J anz, N. & Becker, M. (1984). The health belief model: A decade later. Health
Education Quarterly, 2, 1—47.

Jorgenson, D. O., Dunnette, M. D., & Pritchard, R. D. (1973). Effects of the
manipulation of a performance-reward contingency on behavior in a simulated work

setting. Journal of Applied Psychology, 57, 271-280.

Kane, R. A. (1991). Psychological and social issues for older people with cancer.
(Review). Cancer 68, 2514-2518.

 

Kannel, W. B., Castelli, W. P., & Gordon, T. (1979). Cholesterol in the
prediction of atherosclerotic disease: New perspectives based on the Framingham study.
Annals of Internal Medicine. 90. 85-91.

Konovsky, M. A. & Cropanzo, R. (1991). Perceived fairness of employee drug
testing as a predictor of employee attitudes and job performance. Journal of Applied

Psychology, 76, 698-707.

Korman, A. K., Glickman, A. S., & Frey, R. L. (1981). More is not better: Two
failures of incentive theory. Jouma] of Applied Psychology, 66, 255-259.

Korsgaard, M. A., Schweiger, D. M., & Sapienza, H. J. (1995). Building
commitment, attachment, and trust in strategic decision making teams: The role of
procedural justice. _Aaademy of Management Journal, 38, 60-84.

Krantz, D. S., Grunberg, N. E., & Baum, A. (1985). Health psychology. Annual
Review of Psychology, 36, 349-383.

140

Krantz, D. S. & Schulz, R. (1980). A model of life crises, control, and health
outcomes: Cardiac rehabilitation and relocation of the elderly. In A. Baum and J. E.
Singer (Eds.), Advances in environmental psychology, vol. 2. Hillsdale, NJ: Erlbaum.

Landy, F. J ., Barnes-Farrell, J. L., & Cleveland, J. N. (1980). Perceived fairness
and accuracy of performance and accuracy of performance evaluation: A follow-up.
Journal of Applied Psychology, 65, 355-356.

Langer, E. J. (1975). The illusion of control. Journal of Personality and Social
Psychology, 32, 311-328.

Lawler, E. E. III. (1964). Managerslob performance and their attitudes toward
their pay. Unpublished doctoral dissertation, University of California, Berkley.

Lawler, E. E. III., & Porter, L. W. (1967). Antecedent attitudes of effective
managerial job performance. Organizational Behaviognd Human Performance. 2, 122-
142.

Lerner, D. J. & Kannel, W. B. (1986). Patterns of coronary heart disease

mortality in the sexes: A 26—year follow-up of the Framingham population. American
_Haart Journai 11], 383-390.

Leventhal, G. S. (1976). The distribution of rewards and resources in groups and
organizations. In L. Berkowitz & E. Walster (Eds.), Advances in experimental social
psychology (Vol. 9, pp. 91-131). San Diego, CA: Academic Press.

Leventhal, G. S. (1980). What should be done with equity theory? In K. J.
Gergen, M. S. Greenberg, & R. H. Willis (Eds.), Social exchange: Advances in theory
and research (pp. 27-55). New York: Plenum Press.

Lewin, K. (1938). The conceptual representation and the measurem_ent of
psychological forces. Durham, NC: Duke University Press.

Levin-Epstein, S. (1986). Where's the smoke: Problems and policies concerning
smoking in the workplace: Spacial report. Rockville, MD: Bureau of National Affairs.

Lind, E. A., & Tyler, T. R. (1988). The social psychology of procedural justice.
New York: Plenum Press.

Locke, E. A. (1968). Toward a theory of task motivation and incentives.
Organizational Behavior and Human Performance. 3. 157-189.

Locke, E. A., & Latham, G. P. (1990). A theory of goisetting and task
performance. Engelwood Cliffs, NJ: Prentice-Hall.

141

Locke, E. A., Shaw, K. N., Saari, L. M., & Latham, G. P. (1981). Goal setting
and task performance: 1969-1980. Psychological Bulletin, 90, 125-152.

McFarlin, D. B. & Sweeney, P. D. (1992). Distributive and procedural justice as
predictors of satisfaction with personal and organizational outcomes. Academy of
Management Journal, 35(3), 626-637.

 

Mahoney, M. J. & Amkoff, D. B. (1979). Self-management. In 0. F. Pomerleau

and J. P. Brady (Eds.), Behavioral__medicine: Theory and practice. Baltimore: Williams
& Wilkins.

Mahoney, M. J ., Moura, N. G. M., & Wade, T. C. (1973). The relative efficacy of
self-reward, self-punishment, and self-monitoring techniques for weight loss. Journal of
Consulting and Clinical Psychology, 40, 404-407.

Markoff, R. A., Ryan, P., & Young, T. (1982). Endorphins and mood changes in
long-distance running. Medical Science Sports Exercise. 14, 11-15.

Mento, A. J ., Steel, R. P., & Karren, R. J. (1987). A meta-analytic study of the

effects of goal setting on task performance: 1966-1984. Organizational Behavior and
Human Decision Processes, 39, 52-83.

Morse, R. M. & Flavin, D. K. (1992). The definition of alcoholism. The Joint
Committee of the National Council on Alcoholism and Drug Dependence and the
American Society of Addiction Medicine to Study the Definition and Criteria for the

Diagnosis of Alcoholism. Journal of the American Medical Association. 268(8), 1012-
1014.

National Heart, Lung, and Blood Institute (1994). Report on the task force on
research in epidemiology and prevention of cardio-vascular diseafia.

Opsahl, R. L., & Dunnette, M. D. (1966). The role of financial compensation in
industrial motivation. Psychological Bulletin, 66, 94-118.

Parker, S. D., Brewer, M. B., & Spencer, J. R. (1980). Natural disaster, perceived
control, and attiributions to fate. Personality and Social Psychology Bulletin, 6, 454-459.

Pickens, R. W. & Svikis, D. S. (1991). Genetic influences in human substance
abuse. Journal of Addictive Diseases. 10(1-2), 205-213.

Porter, L. W., & Lawler, E. E. III. (1968). Managerial attitudes and performance.
Homewood, IL: Dorsey Press.

Pritchard, R. D., DeLeo, P. J ., & von Bergen, C., Jr. (1974). Incentive motivation
techniques evaluation in Air Force technical training (AF-HRL-TR-74-24, AD-AOOS

142

302). Lowry AFB, CO: Technical Training Division, Air Force Human Resources
Laboratory.

Pritchard, R. D., Jones, S. D., Roth, P. L., Stuebing, K. K., & Ekeberg, S. E.
(1988). Effects of group feedback, goal setting, and incentives on organizational
productivity. Journal of Applied Psychology, 73, 2, 337-358.

Pritchard, R. D., & Sanders, M. S. (1973). The influence of valence,

instrumentality, and expectancy on effort and performance. J ouma] of Applied
Psychology. 57. 55-60.

Rodin, J. (1986). Health, control, and aging. In M. M. Baltes and P. B. Baltes
(Eds.), The psychology of control and aging. Hillsdale, NJ: Erlbaum, 139-165.

Rosenstock, I. M. (1966). Why people use health services. Milbaak Memorial
Fund Quarterly, 44, 94-127.

Rotter, J. B. (1966). Generalized expectancies for internal versus external control
of reinforcement. Psychological Monographs, 80, 1-28.

Seligman, M. E. P. (1975). Helplessness: On depression, development, and
death. San Francisco: W. H. Freeman.

Shapiro, D. L., & Buttner, E. H. (1988, August). Adequate explanations: What
are they, and do they enhance procedural justice under severe outcome circumstances?
Paper presented at the annual meeting of the Academy of Management, Anaheim, CA.

Shapiro, D. L., Buttner, E. H., & Barry, B. (1994). Explanations for rejection
decisions: What factors enhance their perceived adequacy and moderate their

enhancement of justice perceptions? Organizational Behavior and Human Decision
Processes.

Sheppard, B. H., & Lewicki, R. J. (1987). Toward general principles of
managerial fairness. Social Justice Research 1 161-175.

 

Silverstein, B., Kelley, E., Swean, J ., and Kozlowski, L. (1982). Physiological

predisopositon toward becoming a cigarette smoker: Experimental evidence for sex
differences. Addictive Behavion 7. 83-86.

Sjostrom, L. (1993). Impacts of body weight, body composition, and adipose
tissue distribution on mobility and mortality. In A. J. Stunkard and T. A. Wadden (Eds.),
Obesity: Theory and Therapy. New York, Raven Press, 13-42.

Skinner, B. F. (1969). Contingencies of reinforcement: A theoretical analysis.
New York: Appleton-Century-Crofts.

143

 

Smith, T. W., Turner, C. W., Ford, M. H., et al. (1987). Blood pressure reactivity
in adult male twins. Health Psychology, 6, 209-220.

Somes, G. W., Harshfield, G. A., Alpert, B. S., et al. (1995). Genetic influences
on ambulatory blood pressure patterns: The Medical College of Virginia Twin Study.
American Journal of Hypertension, 8, 474-478.

Stone, D. L., & Vine, P. L. (1989). Some procedural determinants of attitudes
toward drug testing. Paper presented at the Fourth Annual Conference of the Society for
Industrial/Organizational Psychology, Boston.

Stoney, C. M., Davis, M. C., & Matthews, K. A. (1987). Sex differences in
physiological responses to stress and coronary heart disease: A causal link?

Psychophysiology, 24, 127-131.

Stunkard, A. J., & Wadden, T. (1993). Obesity: Theory and Therapy. 2nd ed.
New York: Raven Press.

Taylor, F. W.(l947). Scientific Management. New York: Harper.

Taylor, S. E. (1995). Health Psychology. New York: McGraw-Hill.

Taylor, S. E., Helgeson, V. 8., Need, G. M., & Skokan, Z. Q. (1991). Self—
generated feelings of central and adjustment to physical illness. Journal of Social Issues,

42(4), 91-109.

Thibault, J ., & Walker, J. (1975). Procedural justice: A psychological analysis.
Hillsdale, NJ: Erlbaum.

Thibault, J., & Walker, J ., (1978). A theory of procedure. _C_alifomia Law
Review 66, 541-566.

 

 

Thorndike, E. L. (1911). Animal Intelligence. New York: MacMillan.

Tolman, E. C. (1932). Purposive behavior in animals and men. New York:
Appleton-Century.

Tyler, T. R. (1988). What is procedural justice? Law and Society Review. 22,
301-335.

. Tyler, T. R., Rasinski, K. A., & Spodick, N. (1985). Influence of voice on
satisfaction with leaders: Exploring the meaning of process control. Journal of

Personality and Social Psychology. 48, 72-81.

Vroom, V. H. (1964). Work_and motivation. New York: Wiley.

144

Wallston, K. A. & Wallston, B. S. (1978). Development of the multidimensional
health locus of control (MHLC) scales. Health Education Monographs. 6, 160-170.

Walster, E., Berscheid, E., & Walster, G. W. (1973). New directions in equity
research. Journal of Personality and Social Psychology, 25, 151-176.

Weiner, B. (1979). A theory of motivation for some classroom experiences.
Journal of Educational Psychology. 71. 3-25.

Weiner, B. (1985). An attributional theory of achievement motivation and
emotion. Psychological Review. 92. 548-573

Weiner, B., Russell, D., & Lerman, D. (1979). Affective consequences of causal
ascriptions. In J. H. Harvey, W. J. Ickes, & R. F. Kidd (Eds.), New directions in
attribution reseaih (pp. 59-88). Hillsdale, NJ: Erlbaum.

Weiss, S., Fielding, J ., and Baum, A. (Eds) (1991). Perspectives in behavioral
medicine: Health at woglg. Hillsdale, NJ: Erlbaum.

145

"lllfiltllll”1711111111“