Hannah“... ..

iii-60.): I
.\ . T ..1§|‘6
1.3111... .

 

t 31.‘ ’5.
1"... 158v
ti ‘1. vi

‘1‘?! A
u! Isiah}:

 

IGAN STATE UNIVERSITY LIBRARIES

11111111111111111 1 * 11 1 1111

l

 

 

 

 

 

 

 

 

l M l

H “1‘
3 1293 01111 4122

This is to certify that the

dissertation entitled

Psychological Adaptability and Exercise Adherence:
The Influence of Personality
on Exercise Habits

presented by

Kay Ellen Ketzenberger

 

has been accepted towards fulfillment
of the requirements for

Ph.D Counseling Psychology

 

 

degree in

"1’5”":

Major professor

 

Date February 9, 1995

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

LIBRARY
Michigan fitate
University

 

 

 

PLACE II RETURN BOXto romanthb chockomtrom your record.

TO AVOID FINES mum on or baton date duo.

DATE DUE DAT- . PATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An Nfltmdlvo Action/EM Opponunlty Itltltulon
m1

PSYCHOLOGICAL ADAPTABILITY AND EXERCISE ADHERENCE:
THE INFLUENCE OF PERSONALITY

ON EXERCISE HABITS

BY

Kay Ellen Ketzenberger

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Counseling, Educational Psychology
and Special Education

1995

ABSTRACT
PSYCHOLOGICAL ADAPTABILITY AND EXERCISE ADHERENCE:

THE INFLUENCE OF PERSONALITY
ON EXERCISE HABITS

BY
Kay Ellen Ketzenberger

Studies of exercise adherence have been guided primarily by
atheoretical or predominately cognitive/behavioristic
theoretical perspectives. This study employed a
psychodynamic framework to examine exercise habits, thus
contributing to a theoretical expansion of research
approaches to the study of exercise adherence.
Psychological adaptability was hypothesized to impact
exercise persistence; psychological adaptability was
operationalized by personality variables of low
compulsiveness, low impulsiveness, low perceptual rigidity,
and high tolerance for ambiguity. Assessment instruments
included the Obsessive Compulsive Scale (OCS), the Barratt
Impulsiveness Scale-10th Revision (BIS-10), the Breskin
Rigidity Test (BRT), the Tolerance for Ambiguity Scale
(ATZO), and a demographic questionnaire. Demographic data
and data on other factors shown to influence exercise
adherence were also collected: age, gender, exercise
enjoyment, partner support, exercise format, exercise type,
and intensity. Subjects were volunteers from private health
clubs, the university community, and a community club for

recovering alcoholics (N=4OS). The subject pool consisted

of 257 women and 148 men, with a mean age of 40 years,
ranging from 18 to 84 years. Subjects were categorized into
four exercise groups based on their report of how often they
exercised per week, and how long they had been exercising.
Groups consisted of super-exercisers, general exercisers,
intermittent exercisers, and non-exercisers. Correlation
analysis identified the relationships among the
psychological variables, in particular an interesting lack
of relation between compulsiveness and impulsiveness. The
correlation of the psychological variables and other factors
with exercise group membership was examined through the use
of discriminant analysis. Exercise enjoyment was the most
powerful predictor of exercise group membership, with
impulsiveness contributing a statistically significant but
practically minor addition to predictive accuracy. None of
the other variables were statistically significant. The
implications of these findings is discussed, as well as

directions for future research.

Copyright by
KAY ELLEN KETZENBERGER

1995

ACKNOWLEDGEMENTS

Writing this dissertation has been one of the more
complex, difficult, perplexing, engrossing, and exhausting
experiences of my life. Throughout the process of dealing
with both the fascination and the paralysis involved in
completing a dissertation, I was fortunate enough to receive
the support of the following people.

I wish to acknowledge the impact of several early
mentors. My undergraduate academic advisor, Jeri Falk, M.S.,
played a pivotal role in encouraging me to complete my
undergraduate studies and continue with graduate work.
Sally Grenard-Moore, Ph.D and Melba J. T. Vasquez, Ph.D.
were both wonderful models of counseling psychologists and
sparked my interest in pursuing the doctoral degree in this
field; both their influence and encouragement to pursue my
academic career were invaluable, and I will be forever
thankful to know them. I would especially like to thank
Sally Grenard-Moore for her unwavering support and humor
throughout all the steps involved in this graduate program.

Of course, I would like to take this opportunity to
formally thank the members of my dissertation committee,
lDrs. Linda Forrest (Chair), John Powell, Deborah Feltz, and

lWartha Ewing. All have provided valuable and experienced

feedback throughout the process of this project. I want to
particularly thank Dr. Linda Forrest, who has been my
academic advisor throughout my tenure at Michigan State
University. Her commitment to student development (both
professional and personal) is, I believe, unparalleled among
the faculty at this University, and I appreciate her
availability and willingness to see this project through
even while on sabbatical. Her support and guidance
throughout my graduate experience have been treasured, and I
am grateful to have had the opportunity to work with her.

I would also like to thank two good friends, Ann Hower
and Karyn Boatwright, who were always ready to either laugh
or commiserate, and generally shared both the best and the
worst of the process.

Lastly, I want to acknowledge the tremendous support of
my family. My parents have been a well-spring of
encouragement, and have provided both the emotional and
financial foundation from which I've operated throughout my
graduate program. I’m grateful to have had them as parents.
Thanks also to my brother Steve and sister-in-law Ellen, who
helped me maintain perspective and whose home provided a
welcome respite from the University. Finally, I want to
thank my brother Patrick, for the simple fact of his
existence and that he and his family are part of my life. I

missed you.

vi

TABLE OF CONTENTS

LIST OF TABLES ......................................... ix
CHAPTER 1: INTRODUCTION ............................... 1
Research Question ....................... 6
Hypotheses .............................. 15
Significance of the Study ............... 19
A Caveat ................................ 20
CHAPTER 2: LITERATURE REVIEW .......................... 22
Adaptability,
Impulsivity/Reflectivity/Compulsivity
and Rigidity ............................ 22
Personal Factors Influencing Exercise
Adherence ............................... 32
Situational Factors ..................... 43
Miscellaneous Factors ................... 45
Summary ................................. 46
CHAPTER 3: METHODS .................................... 47
Subjects ................................ 47
Data Collection ......................... 48
Design .................................. 50
Measures ................................ 51
Hypotheses .............................. 59
Data Analysis ........................... 62

vii

CHAPTER 4:

CHAPTER 5:

APPENDICES:

A:

B
C
D:
E

REFERENCES

RESULTS ....................................
Tests of Hypotheses .....................
Post-Hoc Analysis .......................

Statistical Assumptions .................

Summary .................................

DISCUSSION .................................
Sampling and Measurement Considerations..
Discussion of Results ...................
Implications for Practice ...............
Limitations .............................
Directions for Future Research ..........

Summary .................................

Variable Codes ............................
Invitation to Participate Poster ..........
Consent Form ..............................
Demographic Questionnaire .................

Assessment Instruments

E.1 - Barratt Impulsiveness Scale (BISlO)..
E.2 - Obsessive Compulsive Scale (OCS)
E.3 - Ambiguity Tolerance Scale (AT-20)
E.4 - Breskin Rigidity Test (BRT) .........

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

viii

85

86

97

99

100

106

108

109

110

111

113

114

115

LIST OF TABLES

TABLE 3.1
Research Survey Distribution and Return Rate ..

TABLE 3.2

Means and Standard Deviations for Enjoyment,
Impulsiveness, and Intensity for the Total
Sample and for Exercisers Only .............

TABLE 3.3
Exercise Group Membership .....................

TABLE 4.1
Correlation Matrix for Personality Variables ..

TABLE 4.2
Means and Standard Deviations for Personality
Factors (with MANOVA Results) ..............

TABLE 4.3
Discriminant Function Results for Analysis
including OCS, BISlO, BRT and AT20 .........

TABLE 4.4
Standardized Discriminant Function Coefficients
for Function 1 .............................

TABLE 4.5
Classification Results for Analysis 1a
with OCS, BISlO, BRT and AT20 as Predictors

TABLE 4.6
Classification Results for Analysis lb
with BISlO as Predictor ....................

TABLE 4.7
Classification Results for Analysis 1c
with ENJOY as Predictor ....................

TABLE 4.8

Classification Results for Analysis 2
with ENJOY as Predictor for All Exercisers .

ix

.......... 61

.......... 63

TABLE 4.9
Classification Results for Non-exercisers vs.

All exercisers with ENJOY ............................ 76
TABLE 4.10
Classification Results for Non-exercisers vs.
Super-exercisers with ENJOY and BISlO ................ 77
TABLE 4.11
Summary Statistics for Discriminant Analyses ............ 79
TABLE 4.12
Means and Standard Deviations for ENJOY across
Exercise Groups ...................................... 80
TABLE 5.1
Comparison of Norm vs. Current Subjects
on the Psychological Variables ....................... 89

Chapter 1: INTRODUCTION

There is widespread agreement that physical exercise
results in considerable health benefits. Numerous studies
offer evidence that exercise contributes to improved
physical health (Cooper, Pollack, Marin, White, Linnerud, &
Jackson, 1976; Haskell, 1984), as well as to improved
psychological functioning, particularly with anxiety and
depressive states (Folkins & Sime, 1981; Morgan, 1977a;
Morgan, 1979; Morgan, Roberts, Brand & Feinerman, 1970).
Physical activity has also been found to be strongly
associated with healthier psychological functioning in
adolescent women (Covey & Feltz, 1991). Thus, not only does
physical activity improve physical and emotional functioning
in adults, but it is also related to healthy psychological
development in adolescence. Unfortunately, "Even among
those . . . enrolled in structured exercise programs, both
for prevention/health enhancement and for rehabilitation
following . . . coronary heart disease (CHD), adherence is
disappointedly low: Roughly half of the participants will
have dropped out by 3 to 6 months (noncardiac subjects)"
(Martin, et al, 1984, p. 795). The literature is in strong
agreement that, on average, 50% of exercise participants
drop out within the first six months (Dishman, Sallis, &
Orenstein, 1985; Taylor, Buskirk, & Remington, 1973; Ward &
Morgan, 1984), and some ". . . data show that the largest
percent dropout in an exercise program occurs during the

first 12 weeks." (Pollock, 1988, p. 262). People

2
undertaking solitary, individual exercise regimens are even
less likely to maintain adherence (Massie & Shepard, 1971).

The problem of exercise adherence has become a major
interest to researchers and practitioners in a variety of
fields, most notably medicine and exercise science.

Interest continues to grow as the relationship between
regular physical exercise and improved physical and mental
health becomes clearer, and as ever greater numbers of
people undertake exercise programs (Tuson & Sinyor, 1993).
Given this cultural trend towards greater physical activity,
all those professions committed to healthy development and
improved quality of life should make correspondingly greater
efforts toward understanding the factors that impact
exercise adherence.

The majority of the research has stemmed from medical
efforts to treat and prevent coronary heart disease; most of
the remainder of the research has been conducted under the
auspices of university exercise science and kinesiology
departments, generally in the service of identifying
specific factors associated with adherence in normal
populations. Unfortunately, "The use of psychological

variables in adherence research has been characterized by

the lack of accompanying models . . ." (Sonstroem, 1988, p.
125-126). Dishman (1982), cited in Sonstroem (1988),
pointed out that ". . . adherence research has evolved from

applied and pragmatic questions and has been

characteristically atheoretical in nature" (p. 125).

3

More recently, researchers have generated an increasing
number of investigations grounded in one or another
psychological theory, primarily based on cognitive and
social-cognitive (Bandura, 1977, 1986) theories. This
psychologically based research foundation includes
investigations of the influence of self—efficacy on exercise
(Garcia & King, 1991; McAuley, 1992; McAuley & Jacobson,
1991; McAuley & Gill, 1983), the role of outcome
expectancies (Desharnais, Bouillon, & Godin, 1986; Maddux,
Norton, & Stoltenberg, 1986; Rodgers & Brawley, 1991), the
effects of self-schema and other cognitive activity on
exercise behavior (Kendzierski, 1988, 1993) commitment to
physical activity (Corbin, Nielsen, Borsdorf, & Laurie,
1987; Deeter, 1988), and the influence of affect on the
experience of exercise (Gauvin & Rejeski, 1993; Hardy &
Rejeski, 1989; Kendzierski & DeCarlo, 1991; McAuley &
Courneya, 1994).

With the exception of the last area mentioned, that of
affect and its influence, the recent theoretically-based
research has focused almost exclusively on cognitive
factors. In their metapsychological discussion of the
exercise psychology research base, Fahlberg, Fahlberg and
Gates (1992) note that

Although the cognitive behavioral approaches used in

exercise psychology differ from classical behaviorism,

these approaches, when viewed from a metapsychological

perspective, still represent only the behavioristic

4

force of psychology. . . .the current models in

exercise psychology, including for example the

psychological model for physical activity participation
., the psychobiological screening model . . ., the

health belief model . . ., the theory of reasoned

action . . ., locus of control . . ., and the various

"self" theories (e.g., self-esteem . . . self-efficacy

perceived competence . . .) represent only the
behavioristic force in psychology. That is, they do
not stem from psychoanalytic, humanistic/existential,
or transpersonal psychology. . . . Although
psychodynamic and phenomenological approaches to
exercise behavior have been published and presented

(e.g., Berger, 1980; Berger & Mackenzie, 1980;

Fahlberg, 1990a; Kostrubala, 1981; Perry & Sacks, 1981;

Sours, 1981), they remain a rarity (pp. 174-175).

These authors suggest that research stemming from other
psychological approaches (e.g., humanistic/existential,
psychodynamic, transpersonal) would contribute ". . . a new
psychological understanding of exercise behavior through an
expanded perspective that helps to overcome the reductionism
of prevailing approaches" (p. 175). Keeping in mind that
the early research in most fields of inquiry are descriptive
and tend to focus on the observable, "It is not being
suggested that behavioristic approaches are wrong but rather
that they are partial." (p. 177).

The recent interest in the influence of affect on

5
exercise behavior represents an expansion towards a more
existential/psychodynamic approach; that is, the meaning
and/or experience of exercise is the focus. The theoretical
base of the study to be described here is more closely
related to a psychodynamic orientation, in that ".
certain aspects of behavior are viewed as a result of
processes that occur in the mind in the form of conflicting
conscious and unconscious forces (May & Yalom, 1989)"
(p. 175).

Although psychology certainly holds no exclusive claim
to expanding and/or generating relevant theory, the
profession does have considerable expertise in applying
theory to behavior. Counseling psychology in particular
should become more involved in the study of exercise
adherence, for two major reasons.

First, counseling psychology has traditionally focused
on normal populations, been committed to developing
theoretical foundations relevant to practice, and
facilitating healthy development. Since a primary purpose
of exercise research is to help develop interventions for
facilitating adherence, counseling psychology's expertise in
theory development, and the development of interventions,
would be a valuable addition.

Second, there are clearly psychological factors that
impact adherence, and the investigation of such factors
should include the perspective of those educated in

psychological theory. The field of exercise adherence can

6

only benefit from including the perspectives and skills of
psychologists in a multidisciplinary team effort to better
understand this area, as the larger field of sport
psychology has benefitted from such a team approach.
Research Qgestion

This study addresses the question of whether a
relationship exists between an exerciser’s level of
psychological adaptability and their likelihood of adhering
to exercise. Although the behavioral observation that long-
term adherers are more often achievement—oriented, hard-
driving, Type A personalities is commonplace (R. K. Dishman,
August, 1992, personal communication), it is descriptive
only and lacks explanatory power. The focus on a
psychological factor such as adaptability is an effort to
deepen our understanding of such observations by explaining
them as behavioral manifestations of psychological
operations and thus identify the psychological bases of
observable behaviors. Hardy and Rejeski’s (1989) work on
the relation between exercise—induced affect and level of
perceived exertion supports this line of investigation.
They

argue that physiological cues are most ambiguous under

conditions of moderate physical strain. The

theoretical and practical intrigue of this finding is

that attraction to the level of physical exertion

accompanying an effective aerobic stimulus may have

strong social psychological foundations. (p. 313)

7
The rationale for investigating the impact of
psychological adaptability rests upon a basic premise about
the effects of exercise. This premise and its implications
represent a theoretical foundation for linking psychological

adaptability with exercise initiation and adherence.

Basic Premise: Exercise has two important effects:
a) it causes fatigue and physiological changes
b) it induces unfamiliar, ambiguous stimuli
(Hardy & Rejeski, 1989), results in
heightened physical perceptions and fosters
increased self-awareness (Berger & Mackenzie,
1980; Kostrubala, 1976).

Supporting the unfamiliarity of exercise stimuli is
Hardy and Rejeski’s (1989) work on in-task exercise affect,
which found that subjects' recall of their affect during a
moderate intensity exercise session was uncorrelated with
the affect reported during the task itself. Conversely,
posttask recall of in-task affect for easy and difficult
workloads were significantly correlated. They concluded
that "the absence of any relationship between the FS [affect
reported during the exercise task] and recall of affect at
the 60% workload suggests there is something unusual about
moderate work intensity [and suggested that] this ambiguity
of physiological cues at moderate work intensity requires
additional research" (p. 313).

The changes induced by exercise require significant

8
adjustments on the part of the exerciser. Perhaps most
obvious is the need to accomodate an increased sense of
fatigue and lower energy level so often experienced by the
beginning exerciser, either by modifying the exercise
regimen or by making adjustments to one’s daily schedule.
Though less obvious, it is even more important to integrate
internally the unfamiliar, often ambiguous bodily sensations
that accompany exercise and lead to heightened physical and
emotional self-awareness.

For example, several researchers have noted the
relationship between running and increases in self-awareness
(cited in Berger, 1984): "Running provides a woman with an
outstanding opportunity to increase awareness of her
feelings and behavior and encourages her to examine the
underlying psychodynamics" (p. 187). Whether this increased
awareness is experienced as positive, negative, or
ambiguous, it represents a change in self-perception that
must somehow be integrated into the self-concept to avoid
debilitating cognitive dissonance and eventual dropout from
exercise. "Changes in the body as a result of fitness
training might reasonably be expected to alter one’s body
image, which is highly correlated with and might be expected
to radiate to self-concept (Zion, 1965) . . ." (Folkins &
Sime, 1981, p. 380). The exerciser’s level of psychological
adaptability is postulated as a direct influence on the
ability to successfully adjust to and integrate these

changes, and thus acts as a direct influence on exercise

behavior.

Psychological adaptability, then, is the variable of
interest in this study. Adaptability is defined as "the
quality of being . . . able to change without difficulty so
as to conform to new or changed circumstances" (Webster’s
New Twentieth Century Dictionary, 1983, p. 21). Rephrasing
this definition into psychological terms, psychological
adaptability will be defined herein as the ability to
recognize and respond constructively to conditions that
challenge some element of one’s current psychological
reality, and therefore adjust to new situations and
circumstances. If the basic premise of this study is
accepted, that the initiation of exercise induces
unfamiliar, ambiguous physical and emotional experiences
that the exerciser must adjust to, then some minimum level
of psychological adaptability should be necessary to the
continuation of physical exercise.

Unfortunately, there appears to be no established
measure for directly assessing psychological adaptability.
Therefore, it will be operationalized as a continuum of
impulsiveness/reflectivity/compulsiveness, rigidity, and
tolerance of ambiguity. The rationale for operationalizing

adaptability in these terms is given below.

10
Psychological Adaptability and a Continuum of
Impulsiveness/Reflectivitv/Compulsiveness

Viewing psychological adaptability in terms of
impulsiveness, reflectivity and compulsiveness is based on
Miller’s (1988) integrative model of cognitive processing,
which hypothesizes an impulsivity-reflectivity-compulsivity
continuum of cognitive style. Composed of ".
regularities and variations in the basic ego functions of
perceiving, thinking, remembering, and judging" (Messer &
Schacht, 1986), cognitive styles vary in ". . . having a
narrow, concentrated field of attention rather than a broad,
diffuse one [and in] being relatively more or less tolerant
of ambiguous stimuli or unfamiliar experiences . . ."
(Miller, 1988, p. 64), and serve to shape a person's
"general mode of functioning" and "more general aspects of
the personality" (p. 65), thus representing a broadly
influential personality trait.

Most supportive of the rationale behind this proposal
is Miller's (1988) conclusion that ". . . impulsivity and
compulsivity both teeter at the periphery of a more stable,
mature central hub of reflectivity, a core of ego-autonomous
integrative processing . . ." (p. 70). Though Miller does
not use the term, psychological adaptability seems
essentially synonymous with ego—autonomous integrative
processing, or reflectivity. According to this model, both
impulsive and compulsive cognitive styles interfere with

effective integrative processing by impairing reflectivity,

11

and should therefore also interfere with attempts to
persevere in any activity that requires such integrative
processing, such as an exercise program.

Impulsives are relatively deficient in the ability to
". . . modulate attention, affect, thought, and behavior

and explains why, under conditions of social frustration

or ambiguity, behavior regresses to the use of more
primitive response strategies. . ." (p. 66). Shapiro (1965)
points out that the impulsive's subjective experience of
impulse and whim is different from the non-impulsive
person’s experience, that

the experience of urge or impulse is not, from this

point of View, a detached perception of an actual

seizure that overrides what one wants to do, but rather

is a distortion and attenuation of the normal

experience of wanting in which the sense of active

intention and deliberateness is especially impaired

and, usually, even further disavowed for defensive

reasons (p. 137).
This lack of active intention and deliberateness ".
follows from a general deficiency of active organizing and
integrative mental functions" (Shapiro, 1965, p. 138-139);
together with the impulsive’s deficiencies in internal
modulation (Miller, 1988), an impulsive cognitive style
should drastically impair the ability to tolerate ambiguity
and integrate new experience.

Though impulsive and compulsive cognitive styles are

12
usually thought of as radically different, the core process
appears to be the same. Both impulsive and compulsive
people have substantial difficulty tolerating much affective
variety or fluctuations because they lack adequate
reflectivity, which ". . . expresses itself as the tendency
to reflect on the validity of problem solving under
conditions where several possible alternatives are available
and there is some uncertainty over which is the most
appropriate" (Miller, 1988, p. 68). Messer and Schacht
(1986) write that ". . . reflective individuals appear
superior to impulsives across a range of cognitive and
perceptual tasks. . ." (p. 157), but also note the
likelihood of overlap between very high reflectiveness and
compulsiveness:

It is important to keep in mind that extreme

reflectivity (expressed in obsessive, brooding,

ruminative, or perseverative styles) can be as
maladaptive in its own way as is impulsivity. With the
exception of this extreme, however, reflectivity
generally offers an adaptive advantage over

impulsivity. (p. 157)

At this other extreme, compulsive personality styles
entail ". . . a special restriction of attention which is
always narrowly concentrated and sharply focused, never
relaxed or just diffusely aware. This results in a severe
limitation in the overall mobility and range of attention in

daily activities. . ." (Miller, 1988, p. 66), and are

l3
". . . characterized by a 'superreflectivity’ that is really
a pseudo reflectivity. Personal events are scanned and
probed so deliberately and concentratedly that, instead of
being allowed to normally 'sink in' to memory and
consciousness, they are actually prevented from being
integrated . . ." (p. 69). The compulsive personality’s
difficulty in understanding and integrating new and
different experiences has also been noted by Shapiro (1965):
Elements on the periphery of attention, the new or the
surprising, that which can only be apprehended
impressionistcially — all these are only potentially
distracting and disruptive to the obsessive-compulsive,
and they are avoided exactly by the intensity and the
fixed narrowness of his preoccupation. . . ." (p. 30)
This lack of reflective problem solving with ambiguous
and unclear stimuli seems highly applicable to the
physical/affective experiences engendered by physical
exercise; robust reflectivity should represent the level of
psychological adaptability most supportive of the ability to
tolerate ambiguity and adapt psychologically to the physical

and emotional changes resulting from physical exercise.

Psychological Adaptability. RiQiditY and Tolerance of
Ambiguity

The proposed inverse relationship between psychological
adaptability and rigidity can be supported by two lines of

reasoning. First, it is intuitively logical that the

l4
ability to adapt to changes that impact the physiological
and cognitive/emotional spheres requires a certain degree of
cognitive and emotional flexibility. Flexibility and
rigidity have been treated throughout the literature as
diametric opposites on a bipolar continuum (i.e., Breskin,
1968; Rubenowitz, 1963); "Rigidity and flexibility are two
traits lying on the extremes of the same bipolar continum
(sic) varying only in degree and not in kind" (Chhaya, 1985,
p. 14). Since no psychometric measures were found for
flexibility, and a good measure for rigidity does exist
(Breskin, 1968), the alternative is to assess adaptability
through this inverse factor.

Second, cognitive rigidity has been noted as one of the
primary hallmarks of the compulsive cognitive style
(Shapiro, 1965); this connection has since been carried to
the point of operationalizing compulsiveness in terms of
rigidity (Primavera, Hochman, & Reynolds, 1975). Rigidity
has been found to correlate highly with factors that also
correlate with compulsiveness:

The finding that non—verbal rigidity was positively

related to the need for Order and negatively related to

the need for Change seems to be consistent with

[findings that] the BRT [Breskin Rigidity Test] may

measure what has been termed ’obsessive-compulsive

intellectual rigidity’. An active search for order and
an active avoidance of change would appear to be

characteristic of the behavior of many obsessive-

15

compulsive personality types. (Simon, Primavera,

Klein, & Cristal, 1972, p. 130).

Also, rigidity and intolerance of ambiguity have been
found to be strongly associated (Frenkel—Brunswik, 1948b;
Thurstone, 1944), an important correlation considering Hardy
and Rejeski's (1989) findings that moderately intense
exercise is frequently experienced as generating ambiguous
affects. Thus, the ability to tolerate ambiguity should
impact adaptability for two reasons: 1) directly, as the
ability to tolerate the ambiguous physiological cues and
affects induced by exercise, and 2) indirectly, through its
inverse connection with rigidity.

W

This study sought to examine the relationships among
the personality variables of compulsiveness, impulsiveness,
rigidity, and tolerance of ambiguity. Specific hypotheses
regarding their relationships are specified in the Methods
chapter. The study also hypothesized that a positive
relationship exists between psychological adaptability and
exercise adherence.

Psychological adaptability is operationalized as low
impulsiveness, low compulsiveness, low rigidity, and high
tolerance of ambiguity; a measure for each of these traits
will be used.

The utility of this model would lie in its ability to
predict the likelihood of adherence for individuals;

however, whether these personality traits are systematically

16
related to exercise behavior at all needs to be established
first. If so, these traits should be able to distinguish
different categories of exercisers, with the categories
based on frequency and duration of exercise. For the
purposes of this study, four adherence categories have been
defined, including super—exercisers, general exercisers,
intermittent exercisers, and non-exercisers; these
categories will be fully described in the Methods chapter.

It is possible that a subject must cut back or stop
exercising due to injury, which precludes a choice for
continued exercise. Therefore, a question on the
demographic questionnaire asks subjects why they stopped or
cut back exercising; those who indicate that injury
precluded continuous exercise but have resumed exercising
will be classified as though no interruption occurred; those
who note injury as the reason for exercise cessation and are
no longer exercising will not be classified and not included
in the exercise analysis.

Since impulsive, compulsive, rigid, and intolerant
personality styles are hypothesized as impairing the ability
to adapt to an activity that generates heightened self-
perceptions requiring integration into the self—concept, the
model's strongest prediction is that general exercisers
should score significantly lower on these traits than people
who exercise only intermittently.

As it stands, this model may account less well for

super-exercisers, and so predicting how they will score on

17
these variables is more difficult. The only clear
prediction is that super-exercisers will score significantly
lower on impulsiveness than either general or intermittent
exercisers; they may also score low on compulsiveness and
rigidity, and higher on tolerance of ambiguity as the model
would currently predict.

The prediction that low compulsiveness is associated
with super-adherence runs counter to the common-sense
intuition that high compulsiveness should be positively
related to high adherence, particularly at the very high
frequencies of super-exercisers. The existence of exercise
addiction would seem to support this view of a linear
relationship between compulsiveness and adherence. A
modification of the theory could, however, reverse this
prediction and account for high compulsiveness in super-
exercisers.

Basically, the modification suggests that exercise is a
different psychological experience for super-exercisers than
for general and intermittent exercisers. Supporting this
possibility is the reality of exercise addiction. The
extreme compulsiveness associated with "exercise addiction"
serves a qualitatively different function than normal (or
even relatively high) levels of compulsiveness in normal
populations, with much in common with other addictions,
including withdrawal symptoms upon the cessation of
exercise. Addictive exercise, then, fulfills a different

need for the exercise addict than it does for a normal, non-

18
addicted exerciser. For the addict, exercise is a
ritualized behavior serving the same anxiety-suppressing
function as other compulsive behaviors (such as excessive
handwashing and obsessive checking). If exercise addicts
comprise a significant portion of the super-exerciser group,
this group should evidence very high scores on
compulsiveness, significantly higher than any of the other
exercise groups. They should also score in the rigid and
intolerant directions. It is intuitively reasonable that
high compulsiveness would support the ability to maintain
very high frequency exercise.

The possibility that exercise is a different
psychological experience for super-exercisers suggests a
modification to the model of how the personality traits of
interest are related to exercise behavior. At some point,
high compulsiveness stops being a liability and becomes a
behavioral asset for maintaining high-frequency exercise.
Whether it becomes a psychological asset, however, is open
to question, given the possibility of exercise addiction.

According to this modified model, super-exercisers will
score significantly higher on compulsiveness than all other
groups. Given the theoretical connections between
compulsiveness, rigidity, and tolerance of ambiguity, super-
exercisers should also score significantly higher than all
other groups on rigidity and lower on tolerance of
ambiguity. Again, however, this is only a speculative

prediction. It will be interesting to see whether the data

19
support either model.

How the non-exercise group will score on the
personality measures in also unclear, primarily because the
motivations behind people's choices regarding whether or not
to exercise are so diverse. People who choose to remain
sedentary may differ from those who initiate exercise on any
number of variables, and so may well represent quite a
different population than exercisers. Therefore, they may
show wide variability on the personality traits of interest
here. The non-exercise group may serve best as a comparison
group for the various exercise groups.

In summary, this proposal hypothesizes that the
personality traits of impulsiveness, compulsiveness,
rigidity, and tolerance of ambiguity will be able to
distinguish between super-exercisers, general exercisers,
intermittent exercisers, and possibly non-exercisers. This
general hypothesis will be stated more specifically in the
Methods section. The assessment instruments to be used will
also be discussed there.

Significance of the Study

This study offers several possible benefits.
Clinically, the ability to distinguish between intermittent,
general, and super-exercisers at the beginning of an
exercise program would aid in targeting specific
interventions towards the different groups in the service of
group-specific goals, with particular focus on the

intermittent group in hopes of increasing adherence.

20
Theoretically, significant results would support further
investigation from the psychodynamic/existential perspective
discussed by Fahlberg, et a1 (1992), and serve to further
expand the research base of exercise adherence.
A Caveat

A caution is warranted. The model proposed above
addresses only intrapsychic factors, and suggests that overt
exercise behavior is systematically affected by these
internal, individual factors. Attention to internal,
psychological factors is still a relatively new area of
inquiry, one worth exploring.

However, this model is incomplete as a comprehensive
explanation of exercise behavior, in that it ignores the
social context within which exercise decisions and behaviors
occur. Considerable research attention has focused on the
social context of exercise, particularly the difference in
adherence between individual and group exercisers, and found
contextual variables to have significant impact on exercise
persistence. The power of social reality on behavior is not
contested; rather, the internal, intrapsychic factors in
this proposal very likely interact with the social context
so that the behavior of interest is multiply determined and
interactional in nature.

To eventually understand the interaction between social
and intrapsychic variables, however, the intrapsychic must
first become better defined through initial studies.

Intrapsychic factors will likely explain only a small to

21
moderate amount of variance in exercise adherence; once the
contribution of such factors is better understood, the task
will be to better integrate these intrapsychic, individual
influences with the interpersonal, social, and cultural

influences on exercise behavior.

Chapter 2: Literature Review

Psychological adaptability, impulsiveness,
reflectivity, compulsiveness, and rigidity are the central
constructs in this study. Therefore, the first portion of
this chapter will review the current state of knowledge
regarding these factors. The review will then focus on the
exercise adherence literature, with an initial focus on the
personal factors that have been hypothesized to impact
exercise adherence, including psychological and biological
variables. Lastly, the relationships between various
situational factors and exercise adherence will be reviewed,

and several miscellaneous variables will then be covered.

Adaptability, IgpulsivityZReflectivitv/Compulsivity and

Rigidity

 

Psychological Adaptability

Unfortunately, no research has yet sought to
investigate the influence of psychological adaptability,
impulsiveness, reflectivity, compulsiveness, or rigidity on
exercise behavior. A model of psychological adaptation to
trauma and victimization has been proposed in the broader
psychological literature (McCann, Sakheim, & Abrahamson,
1988), where psychological adaptation is defined as
". . . emotional, biological, cognitive, behavioral, and
interpersonal response patterns" (p. 558). Defining
psychological adaptation as response patterns, however,

identifies it in terms of extrapsychic consequences rather

22

23
than an intrapsychic process resulting in observable
behavior. In McCann's et al (1988) view, psychological
adaptation is primarily the result of a variety of cognitive
operations rather than a cause of behavior (though a degree
of reciprocity is acknowledged); the rationale of this
current proposal postulates that the process of adaptability
has effects as well. Thus, McCann’s et a1 (1988) model does
not appear applicable to this proposal, and no other
psychological approach to the process of adaptability was
found in the literature. This proposal, therefore, will
attempt to assess adaptability indirectly through the
measurement of factors either shown to be associated with
adaptiveness or factors that can theoretically be expected

to influence adaptability.

ImpulsivenessZReflectivityZCompulsiveness

Miller’s (1988) model of impulsivity-reflectivity—
compulsivity is based on the construct of reflection—
impulsivity originally introduced by Kagan, Rosman, Day,
Albert, and Phillips (1964), and the cognitive style of
compulsiveness elucidated by Shapiro (1965), and Reed (1968,
1977). Reflection-impulsivity (R-I) has generated a great
deal of research, primarily with children. Miller (1988)
contends that this research base is generalizeable to
adults, having ". . . broad applicability to individuals in
different age ranges" (p. 68).

The reflective style has been found to be the more

24
mature and adaptive style across practically every dimension
studied. Reflective children display significantly more
verbal control of behavior, and use more covert, private
speech than impulsives; the authors concluded that
reflectives were more mature in their use of speech
(Meichenbaum & Goodman, 1969). Schleifer and Douglas (1973)
found that reflectives based their judgments more often on
intentions than on consequences, indicating that reflectives
use more mature moral judgment. Also, Salkind and Nelson
(1980) note that impulsivity is the earlier response pattern
in terms of development, with reflectivity developing only
as a child ages. Thus, reflectivity represents a more
mature and adaptive developmental level.

There is strong evidence that reflectives are
significantly more field independent than impulsives
(Campbell & Douglas, 1972; Schleifer & Douglas, 1973), and
that both reflectivity and field independence contribute
positively to effective problem-solving (Neimark, 1975).
Also, "On a variety of perceptual, conceptual, and
perceptuomotor problem-solving tasks involving response
uncertainty (i.e., when the answer is not immediately
obvious), reflectives consistently perform better than
impulsives, behaving like more mature children" (Messer,
1976, p. 1037). Perhaps this superior problem-solving
ability (compared to impulsives) is related to findings that
reflectives are more efficient information processors

(Swanson & Schumacher, 1986). The only study found that

25
used adult subjects (a college sample), found no relation
between reflection-impulsivity and problem—solving.
However, the authors speculated that their nonsignificant
results could have resulted from the likelihood that the
college population is skewed toward reflectivity and
therefore the sample likely represented a restricted range
of R-I; thus, the separation of the sample into reflective
and impulsive categories was nonrepresentative of the
general population (Wolfe, Egelston, & Powers, 1972).

The relationship between reflectivity and problem—
solving has been called into question, based on findings
that a positive relationship exists between reflectivity and
IQ (Barrett, 1977; Block, Block, & Harrington, 1974). In
response, Messer (1976) reviewed the literature concerned
with reflectivity and IQ, and confirmed a very modest
correlation that accounted for 3% to 11% of the variance in
reflectivity. The correlation between reflectivity and IQ
was most noticeable when the IQ test questions were
multiple-choice and/or nonverbal, and therefore less
dependent on verbal production and more similar in format to
the MFFT. Also, the correlation decreased when the IQ test
was given first, indicating an operative practice effect;
and, the relationship declined with age. Messer and Schact
(1986) state

Apart from the relatively small size of the R-I/IQ

correlation, which appears to pose only a small threat

to the discriminant validity of the R-I construct,the

26
precise meaning of this research remains elusive
because composite IQ scores typically represent a
hodge-podge of abilities, skills, preferences, and
styles. Such composite scores are likely to share some
variance with almost any other cognitive dimension,
simply by virtue of their diversity. (p. 185)
Thus, reflection-impulsivity can be considered a robust
dimension of cognitive style largely unthreatened by IQ.
The positive correlates to reflectivity noted above
should not be taken to mean there are no negative
consequences to high reflectivity. Reflectives have been
shown to manifest greater anxiety than non-reflectives,
lower ego strength, and lower self-satisfaction (Spitzer,
1981). Surprisingly, reflectivity was not associated with
either self—awareness or a tendency to attend to internal
stimuli more than non-reflectives, but was correlated with
an ideal-self discrepancy and anxiety (Jennings, Holstrom &
Karp, 1986);
The ideal-self discrepancy suggests that the anxiety
may be associated with a self-critical psychological
posture, and attendent feelings of low self-
esteem . . . and dissatisfaction with self.
[reflectives show more of a] readiness to adopt some
evaluative distance from oneself, often with critical
self—appraisal and accompanying psychological distress.
(p. 92)

Although these findings are relatively tentative, and

27
considerably outweighed by the positive associations
discussed above, there may be a price to be paid for a
heightened ability for self—evaluation and seeing oneself
clearly.

In terms of adaptability, then, reflectivity appears to
have both advantages and disadvantages. Automatically
assuming that the findings of heightened anxiety and ideal-
self discrepancy are disadvantages, however, is faulty;
after all, a major rationale underlying psychotherapy is
that growth, development, and adjustment are best
facilitated by realistic and clear perceptions about self
and other (Miller, 1986).

The other pole of the impulsivity-reflectivity-
compulsivity continuum (compulsivity) lacks a significant
research base, particularly among non-clinical populations.
Shapiro (1965) described quite well the compulsive cognitive
style, discussed above. Reed (1968) concurred with
Shapiro’s view of the compulsive style as a special
restriction of awareness, but hypothesized also that the
compulsive's mental processing is functionally impaired in
terms of the structure and integration of experience.
Compulsives attempt to compensate for impairment in their
ability to spontaneously organize and synthesize experience,
by over—intellectualizing and analyzing the available data
to the virtual exclusion of all intuitive processing and
understanding (Reed, 1977). This impairment in integration

of experience should also necessarily impair the

28
compulsive’s ability to sustain activity that generates
ambiguous experiences, as in a physical exercise regimen.
Unfortunately, this theory lacks the support of an empirical
research base, due to the dearth of research on

compulsiveness in non-clinical populations.

Rigidity

The concept of rigidity has received considerable
attention in the psychological literature. Interest in
rigidity originated in the study of perception; the term
"rigidity" was used to describe perseverative behavior and
difficulties in shifting from one task to another (Cattell,
1935). Subsequently, the accepted definition of rigidity
became ". . . sluggishness in the variation of a response"
(Werner, 1946, p. 43). A link between personality and
rigidity (as perseverative behavior) was first postulated by
Lewin (1935), who suggested that rigidity was the degree of
fixity and tension between the various psychic systems that
make up the personality.

Of particular importance to the present proposal,
Werner (1946) identified rigidity as a non-adaptive pattern
by distinguishing it from stability. Leach (1967)
summarized Werner as ". . . pointing out that rigidity was
not be confused with stability. In a constantly changing
environment, if behavior is to be stable, response must be
flexible. Differentiation of response is therefore

essential to stability" (p. 13).

29

A link between attitudes and individual differences in
perception was provided by Thurstone's (1944) study of
perceptual rigidity, which found, among other things, that
". . . speed of various perceptual functions, such as speed
of closure, [was] shown to be related to such variables as
tolerance of ambiguity" (Leach, 1967, p. 13). Bruner (1948)
expanded this connection between perception and attitudes to
include the influence of deeper personality factors: "What
emerges in the individual’s perceptual field is a compromise
reflecting his adjustive needs . . ." (p. 167).

With the end of World War II and the blossoming
interest in ethnocentricity and authoritarianism, a great
deal of attention was turned to examining the relationship
between perceptual rigidity and social attitudes and
relationships. The rigid social thinking found in
prejudiced people was associated with rigid problem-solving
(Rokeach, 1948), and prejudiced people markedly
oversimplified both current conscious thought and, over
time, their memories as well (Fisher, 1951). Also,
intolerance of ambiguity was positively associated with the
acceptance of popular misconceptions (Levitt, 1953). More
specific to the rigidity-flexibility continuum, Rubenowitz
(1963) found that flexible people were ". . .more
intraceptive and more able to analyse foreign stimuli and
situations than rigid people" (p. 215). More recently,
rigidity has been found to be negatively associated with

both problem-solving and creativity (Gorman & Breskin,

30
1969), and positivity correlated with field dependence in
women (Breskin & Gorman, 1969).

Prime among these investigations was a study of
prejudiced children, with findings that highly prejudiced
children displayed high perceptual rigidity (Frenkel-
Brunswik, 1948a), leading to the conclusion that perceptual
rigidity was not only correlated with personality traits,
but was also, in essence, a manifestation of the total
personality itself, with its beginnings in a lack of
tolerance for emotional ambivalence (Frenkel-Brunswik,
1948b). Also, highly authoritarian people were found to be
intolerant of ambiguity, and tended ". . . to arrive at
premature closure as to valuative aspects, often at the
neglect of reality" (Frenkel-Brunswik, 1949, p. 115). This
researcher viewed rigidity and intolerance of ambiguity as

always going hand-in-hand, if not exactly synonymous.

Summary

Though the majority of Frenkel—Brunswik’s work focused
primarily on social ambiguity, I am proposing that, by
extension, rigid people are also intolerant of ambiguity in
their own emotional and phenomenological experiences.
Frenkel-Brunswik’s (1948b) conclusion that rigidity stems
from and is a manifestation of intolerance of emotional
ambivalence is highly supportive of the hypothesis that
rigidity negatively influences an exerciser’s ability to

tolerate and adjust to the frequently ambiguous stimuli

31
resulting from exercise.

That a rigid personality style constricts perception
and thereby impairs adaptability appears to be supported by
the literature. Leach (1967) discussed this constriction of
perception, stating that ". . . the rigid individual
selects out and filters stimuli in accordance with his
defensive needs" (p. 17), conceiving of

rigidity as a fundamental restriction on the
individual’s perception of his environment, and
therefore of the use which he could make of that
environment. Earlier workers had thought that
individuals needed this kind of restriction on their
perceptual intake, as a defence against ego-involving

situations which they found threatening. (p. 18)
Surely there is little that has more ego-threatening
potential than ambiguous stimuli regarding one’s body and
self-concept. Also, Werner’s (1946) supposition that
behavioral stability requires cognitive/emotional
flexibility is clearly applicable to exercise adherence;
certainly at least the beginning exerciser's internal
environment is in a continuing state of flux between
physical activity, and rest and recovery, as well as
occasional physical soreness. Thus, if the ". . . rigid
individual would require a greater degree of structure in
his environment than a non-rigid person" (Breskin, 1968, p.
1203), and if this greater structure is as necessary in the

exerciser's internal environment as the external

32
environment, then higher rigidity would make psychological
adaptation to exercise difficult due to its negative
influence on the ability to tolerate ambiguity.

As noted earlier, no research to date has linked the
cognitive variables discussed above to exercise performance
or adherence. This review will now focus on those factors
that have been investigated in relation to exercise

adherence.

Personal Factors Influencing Exercigg Adherence

 

Investigations of exercise adherence have generally
classified the variables of interest into personal and
situational categories. The personal category includes
psychological and biological factors; situational variables
include the influence of social support, exercise intensity,
and convenience.

Psychological Factors

Recent research has employed theoretical models of how
people perceive themselves in attempts to better understand
some of the internal forces affecting exercise performance.
Self-efficacy theory (Bandura, 1977) in particular has been
employed to investigate the influence of self-perception in
a variety of areas, including sport and exercise behaviors.
Also, the effect of self-schema, or how people generalize
about themselves, has also been investigated in regard to
exercise.

Although the impact of self—efficacy has been studied

in a number of arenas (including attitudes toward math and

33
science, behavioral medicine, and sport and motor
performance), "it is only recently that researchers have
begun to examine the potential value of efficacy cognitions
in explaining and predicting exercise behaviors" (McAuley,
1992, p. 110). An early study sought to identify predictors
of adoption and continuation of physical activity (Sallis,
Haskell, Fortmann, Vranizan, Taylor, and Solomon, 1986);
they found that the importance of individual factors varied
depending on the stage of activity (i.e., adoption or
maintenance), as well as across differing levels of
intensity. This study identified self-efficacy as a
significant predictor of exercise adoption, and also
significantly influential in maintaining moderate intensity
exercise once begun.

Another early study examined whether self-efficacy and
outcome expectations could discriminant between adherers and
dropouts in a twice weekly, ll-week exercise program
(Desharnais, et al, 1986). Results indicated that both
self—efficacy and outcome expectations contributed
significantly to the ability to distinguish between adherers
and dropouts, although ". . . self-efficacy [was] a more
central determinant of adherence than expectation of
outcome" (p. 1158). Unfortunately, this study suffered a
number of shortcomings that confound its findings. First,
the adherence criteria did not account for whether subjects
exercised outside the program. Also, the criteria were

appeared arbitrary; adherers were those who attended at

34
least 20 of the 22 total sessions, while missing no more
than one session in any given week. Thus, an exerciser who
attended 19 of the 22 total sessions was classified a
dropout, with no theoretical difference between the 19-
session attender and the 20-session attender. Lastly, the
measurement of self-efficacy was crude, consisting of only
one item on a 7—point Likert scale, and therefore of
questionable value in fully tapping subjects’ beliefs
regarding their ability to complete the program. Although
interesting and supportive of further research, this study’s
results were weakened by methodological flaws.

McAuley and Gill (1983) addressed the measurement issue
in an earlier work that developed a Physical Self-Efficacy
Scale (PSE). Working with collegiate women’s gymnastics,
they found that the PSE was a reliable measure for general
physical self-efficacy. For predicting actual performance,
however, only the task-specific items of the PSE and the
subjects’ individual predictions of performance were
significant predictors. The authors interpreted these
findings as support for Bandura’s (1986) contention that
self-efficacy is highly task specific, and therefore the
most appropriate measurement of self-efficacy is at the
microanalytic level that taps one’s beliefs about the
ability to perform the tasks specific to the overall
behavior, and the ability to overcome the typical obstacles
to performance, rather than a general or global level.

McAuley, Duncan, Wraith, and Lettunich (1991)

35
investigated the effects of a successful exercise experience
on self-efficacy beliefs regarding a variety of future
exercise behaviors, which included walking, biking, and sit—
ups. After successfully completing a submaximal bicycle
ergometer workout, self-efficacy perceptions were
significantly raised for all subjects, and for all three
types of exercise. This study demonstrated strong support
for the reciprocal relationship between self-efficacy and
past performance accomplishments posited by Bandura (1986).

No studies using an appropriate, task—specific measure
of self-efficacy have yet failed to find a statistically
significant effect for self-efficacy, even when modest.
Thus, self-efficacy theory appears a very promising avenue
for further research interest on cognitive mediation of
exercise behaviors.

Another framework of cognitive mediation of behavior
receiving recent attention is self-schemata, originally
introduced by Markus (1977), and defined ". . . as implicit
theories used by individuals to make sense of their own past
behavior and to direct the course of future behavior" (p.
78). In applying schema theory to the exercise domain,
Kendzierski (1988) defines self-schemata in more cognitive
terms, as ". . . cognitive generalizations about the self
derived from past experience and focused on those aspects of
the self that are regarded by the individual as important

." (p. 45). Subjects who were self-schematic for exercise

were found to exercise more frequently, did a variety of

36
exercise activities, and had developed more strategies for
getting themselves to exercise when they did not feel like
it than non-exerciser schematics or aschematics
(Kendzierski, 1988). Another study found that exercise—
schematic subjects were significantly more likely to have
adopted an exercise program than non-schematics
(Kendzierski, 1990). This study also examined the
information processing of a sample of exercise-schematics,
finding that schematics differed from non-schematics in
endorsing more phrases related to exercise as self-
descriptive, took significantly less time to make schema-
congruent judgments, and reported more past exercise
experiences.

In summary, the influence of cognitive mediation on
exercise perceptions and behaviors is undoubtedly
significant and important to pursue in future research.
Clearly, how people think about themselves and their
abilities greatly influences behavior, and interventions
aimed at strengthening those thoughts and beliefs appear
quite possible from these cognitive frameworks.

Other research on psychological factors has focused
primarily on attitudes toward physical activity, and level
of self—motivation. To date, attitudes toward physical
activity have not been found particularly useful in
identifying exercise adherers or dropouts in adult
populations (Dishman & Gettman, 1980; Dishman, Sallis, and

Orenstein, 1985; McCready & Long, 1985). Dishman (1982)

37
concluded that attitudes toward physical activity have been
unable to predict either initiation of or adherence to
exercise, finding that nonadherers possessed attitudes
toward exercise as favorable as regular exercisers.
However, his conclusions were based on studies that used
widely different as well as very broad definitions of
attitudes, calling into question the reliability of the
construct he was reviewing. Also, the Physical Estimation
and Attraction Scales (PEAS), the measure most commonly used
to assess attitudes, was developed on adolescent males and
has only limited generalizeability to adults (Sonstroem,
1978), though efforts are being made to revise the PEAS
scales to develop more valid adult scales (Safrit, Wood, &
Dishman, 1985).

Although the full PEAS scores were not predictive of
adherence, the Attraction Scale of the PEAS has been shown
to influence the initial adoption of an exercise program
(Dishman, 1978; Sonstroem & Kampper, 1980). Though the
research is limited to adolescent males, attitudes have also
been strongly associated with self-esteem (Fox, Corbin, &
Couldry, 1985; Sonstroem, 1976) and lack of maladjustment or
neuroticism (Sonstroem, 1976).

Given these findings, Rejeski and Thompson (1993)
concluded that "In short, in 1982, it was premature to
suggest that attitudes were unrelated to either initiation
of or continued involvement in physical activity" (p. 23).

With the development of a more valid instrument for the

38
measurement of attitudes in adults, ". . . interest in
vigorous physical activity and self-perceived ability at
activity could become components of a future, more complex
model explaining adult activity participation" (Sonstroem,
1988, p. 129).

One of the better researched psychological factors to
date is self-motivation, defined by Dishman and Gettman
(1980) as ". . . a generalized, non-specific tendency to
persist in the absence of extrinsic reinforcement and is
thus largely independent of situational influence" (p. 297).
Self-motivation is thought to reflect a personality trait
that helps determine task persistence of many kinds, and is
believed to have a situationally invariant, trait-like
nature (Dishman & Ickes, 1981). Its relationship with
exercise adherence has been defined as positive and linear;
as self-motivation increases, so does adherence.

Unfortunately, research investigating self-motivation
has provided very mixed results. The Self-Motivation
Inventory (SMI) was developed to measure this construct
(Dishman & Ickes, 1981); the validation studies were
interpreted as supporting the utility of the SMI, since
". . . the mean self-motivation score of the dropouts
was significantly (p < .05) lower than that of the
adherers . . ." (Dishman & Ickes, 1981, p. 427). Factor
analysis supported the authors’ contention that the SMI
measures a unitary trait (Dishman, Ickes & Morgan, 1980).

Another study identified percent fat, self-motivation, and

39
body weight as sufficient for 78.9% accuracy overall in
classifying adherers and dropouts (Dishman & Gettman, 1980).
This was a 39% gain over base rate prediction for dropouts,
but only a 16% gain over base rate for adherers.

The main problem with this initial core of construction
and validation studies is that, though the experimental
groups were found to be significantly different in their SMI
mean scores, the utility of the SMI’s predictive ability is
overstated. Group findings have not led to a practical
ability to predict adherence/dropout for individuals. Ward
and Morgan (1984) used Dishman and Ickes’ (1981) body
weight, percent fat, and SMI equation to predict adherence
in two exercise programs; adherers were predicted well but
dropouts were not. Overall accuracy was not different from
chance for women, and was only 68% for men. Other
researchers and reviewers (Dishman, 1986; Martin & Dubbert,
1985) have also concluded that the SMI is a considerably
better predictor of adherence than of dropout, a finding
that contradicts the Dishman and Gettman (1981) results
mentioned above.

Other studies have obtained results supportive of the
utility of self-motivation. Heiby, Onorato, and Sato (1987)
studied marathoners, finding significant negative
correlations between self-motivation and trait anxiety and
depression, and positive correlations with frequency of
self—reinforcement, current exercise habits, and motivation

to exercise. Pain and Sharpley (1986) used the SMI to

4O
predict the completion of a graduate course in psychological
counseling, finding a significantly positive correlation
between SMI score and course completion. This study was
interesting in that it examined the SMI with a non-exercise
population; its significant results indicates that the
construct underlying the SMI generalizes beyond the exercise
population.

Other studies have failed to find significant
relationships between the SMI and adherence. Weber and
Wertheim (1989) investigated the effects of self-monitoring,
special staff attention, body fat percent and self-
motivation on adherence to individualized exercise programs
at a community gym. Findings were significant for type of
intervention, but not for either fat percent or self—
motivation at any point during the 12 weeks of the study.
This is particularly interesting since the criterion for
adherence was relatively liberal compared to previous work.

Gale, Eckhoff, Mogel, and Rodnick (1984) also failed to
find a relationship between self-motivation and exercise
adherence in their six month study. For both men and women,
self-motivation scores did not distinguish nonadherers and
occasional attenders from adherers; self-motivation did
identify early dropouts from adherers and occasional
attenders for men, but not for women. As mentioned above,
Ward and Morgan’s (1984) study failed to distinguish between
adherers and nonadherers; subjects obtained very similar

scores on the SMI regardless of rate of adherence.

41

This review of the self-motivation literature is
relatively brief; however, it is representative of the
diversity of findings regarding this construct, and serves
to point out the controversial nature of self-motivation.
Part of the problem is definitional: "The exact nature of
the behavioral tendency to persevere and of what is being
measured by Dishman’s questionnaire is not yet clearly
defined" (Knapp, 1988, p. 220). Dishman and Gettman’s
(1981) definition of self—motivation as "a generalized, non-
specific tendency to persist" may be less than adequate,
particularly since motivation is most frequently viewed as a
multidimensional, variable factor (Cassidy & Lynn, 1989),
not the unitary construct tapped by the SMI. Also, "self—
motivation" cannot be considered synonymous with "intrinsic"
motivation, since intrinsic motivation has also been found
to be multidimensional in nature (Reeve & Robinson, 1987).
Also, Dishman and Gettman’s (1981) conclusion that the
absence of extrinsic reinforcement means self-motivation is
independent of situational influences cannot be supported;
the number of intrinsic reinforcers is, for all practical
purposes, infinite, and can easily be situationally
dependent.

In summary, more definitional work is required before
the construct of self-motivation can be sufficiently
understood. The literature is equivocal regarding a
positive relationship between self—motivation and exercise

adherence, with some studies finding a moderate relationship

42

with adherence and others finding little or no correlation.
Unfortunately, self-motivation as measured by the SMI has
very limited utility in predicting adherence or dropout, and
cannot be considered reliable for predicting the exercise
behavior of individual exercisers (Gale, et al, 1984; Weber
& Wertheim, 1989).
Biological Factors

Early evidence pointed to a relationship between
adherence and both body weight and percent body fat
(Dishman, 1981; Dishman & Ickes, 1981; Massie & Shephard,
1971), with higher values of each contributing to lower
adherence. Dishman and Ickes (1981) also found metabolic
capacity to be significantly related to adherence, but
Dishman and Gettman (1980) failed to replicate this finding.
Other research examining body fat as a variable, however,
has not been consistently supportive (Sonstroem, 1988).
Weber and Wertheim (1989) failed to find any correlation
between body fat and adherence, and Ward and Morgan (1984)
were unable to replicate Dishman and Ickes’ (1981) findings,
which had body fat as a significant predictor of adherence
(along with body weight and self—motivation). Morgan’s
(1977b) early review concluded that the balance of the
research at that time pointed to the nonsignificance of
percent body fat in predicting exercise adherence.

Other medical factors, such as systolic blood pressure,
angina, and persistent cough, have been found significantly

related to adherence for patients recovering from myocardial

43
infarction (Oldridge, et al, 1982), but have not been
investigated in non—cardiac subjects.

In summary, no biologic factors have unequivocal or
even strong support in the literature as impacting exercise
adherence in normal populations, which is the population of
interest in this proposal. The research on body weight and
percent body fat has been highly inconsistent (Sonstroem,
1988), and the cardiac factors that impact cardiac

exercisers have not been studied in normal populations.

Situational Factors

The most salient situational correlates with exercise
adherence are social support influences and exercise program
variables (Dishman, 1984; Martin & Dubbert, 1985).
Exercisers with supportive spouses persisted twice as often
as those whose spouses were either negative toward or even
neutral to the exercise program (Andrew, et al, 1981; Andrew
& Parker, 1979; Heinzlemann & Bagley, 1970). There is also
considerable evidence that social support from within the
exercise program exerts great influence. Exercisers in
group programs were twice as likely to adhere than
individual exercisers (Massie & Shephard, 1971), and 90% of
the subjects in another study stated their preference for
exercising either in a group or with another person
(Heinzelmann and Bagley, 1970). There is a strong consensus
that exercising in a group format fosters persistence in the

exercise program (Wankel, 1985; Wankel, Yardley, & Graham,

44
1985).

Other situational variables that have been shown to
vary systematically with adherence are the intensity of the
exercise, and convenience of the exercise setting. Higher
exercise intensity has been associated with lower adherence
(Mann, Garrett, Farhi, Murray, & Billings, 1969; Pollock,
Gettman, Milesis, Bah, Durstine, & Johnson, 1977). Andrew,
et a1 (1981) suggested that higher intensity exercise
contributes to exercise dropout through increased levels of
perceived exertion and fatigue. Also, Pollock, et al (1977)
and Mann, et a1 (1969) found that half their subjects
dropped out of the program due to injury, which was
attributed to the intensity of the exercise. Only one study
found no relationship between intensity and adherence
(Oldridge, 1979); however, the intensity in this study was
not as high as in other studies, and both high and low
intensity exercisers experienced high social support. These
results were interpreted to mean that high social support
can moderate the effect of moderate intensity exercise and
serve to equalize the adherence and dropout rates.

Convenience of the exercise program has also been noted
as an important factor in continuing to exercise (Andrew &
Parker, 1979; Massie & Shephard, 1971; Teraslinna, Partanen,
Koskela, & Oja, 1969). However, the convenience factor has
been assessed primarily through retrospective inquiries into
why people had dropped out of exercise, and the credibility

or accuracy of such hindsight self-reports is questionable.

45
Miscgllgneous Factors

Other variables that have been of occasional interest
to researchers include occupational status, cigarette
smoking, past participation in sport or exercise, age, and
perception of physical competence, among others. Of these,
only occupation and smoking can be confidently associated
with exercise adherence; findings on the other variables are
too mixed to interpret reliably, probably because of
definitional, measurement, and methodological differences
and difficulties.

Adherence to exercise appears to be consistently poorer
in exercisers with blue-collar occupations and for
exercisers who smoke (Fielding, 1982; Oldridge, 1982;
Oldridge, et al, 1983). The occupational influence may
result from worker perception, in that "if the blue collar
worker perceives the exercise intensity to be lower than in
his daily work, he may be likely to question the need for
the extra, "unnecessary" exercise" (Oldridge, et al, 1983,
p. 73). Smokers consistently drop out of exercise programs
more often than non—smokers, according to the same studies,
although Shephard, Corey, and Kavanagh (1981) failed to find
a positive correlation between smoking and adherence. In
general, "Continued smoking after an infarction is an
indicator of poor compliance with health behavior change"
(Oldridge, et al, 1983, p. 73), and may indicate an

insufficient commitment to exercise.

46
Summary

This chapter has reviewed the literature on the
elements of cognitive style thought to be relevant to
continued participation in exercise, particularly the
elements of impulsivity, compulsivity, and rigidity. The
literature supports cognitive style as an influence on
psychological adaptability in a variety of arenas;
therefore, it is hypothesized here that these elements of
cognitive style should also impact one’s ability to adapt to
the physical and emotional changes that result from regular
exercise, because they influence the exerciser’s affective
experience and emotional state.

The exercise adherence literature was also reviewed,
with many inconsistent findings reported. One conclusion to
be drawn from this literature is that the atheoretical
approach to studying adherence is beginning to expand into
more theory—driven approaches to research, and the field is

benefiting from this expansion.

Chapter 3: METHODS

This chapter will specify the methodology and
procedures used in this study. The subject pool will be
described, along with the data collection procedure and
study design. Operational measures will then be reviewed.
Lastly, the research hypotheses will be stated and the data
analysis described. Assessment instruments, the variable
coding scheme, introduction/consent letter, and the
demographic questionnaire are in the appendices.

Sub e ts

A total of 434 surveys were obtained from eight sites.
Of that total, 24 included invalid responses to the validity
items of the Obsessive Compulsive Scale (OCS) and were
therefore excluded completely from the analysis. Two other
surveys were excluded because the subject failed to complete
at least two of the instruments, and three were excluded
because the subjects indicated they had been treated for
cardiac problems. The remaining 405 surveys comprised the
total subject pool.

The sample consisted of 148 men (36.5%) and 257 women
(63.5%). The average age was 40 (SD = 12.4) years, ranging
from 18 to 84; mean age for men was 40.4 (SD = 13.4; range =
53) and for women 39.7 (SD = 11.9; range = 66).

Subjects were recruited from five health/fitness clubs
(four in the mid-Michigan area and one in rural Virginia),
from the Michigan State University community (including

worksite wellness participants, and faculty and graduate

47

48
students), and from a community club for recovering
alcoholics. The alcoholic subsample was recruited in an
effort to include non-university, non-health club subjects.
These sites were collapsed into three: private
health/fitness clubs (which required a membership fee); the
MSU community, including participants in the university
worksite wellness program, graduate students in a summer
course, and individual graduate students and faculty known
to the investigator; and the recovering alcoholic group.
2gp; Collection

The study survey consisted of a demographic
questionnaire and four assessment instruments. Ten graduate
students completed the survey before distribution and were
able to infer the purpose of the measures for compulsiveness
and impulsiveness; consequently, those scales were
counterbalanced in presentation in order to counterbalance
the possible ordering effects.

The health/fitness facilities all agreed to allow a
poster and flyers calling attention to the study to be
posted, and kept packets of test surveys at the front desk.
Health club personnel distributed the packets to those
people who expressed an interest in participating. The MSU
worksite wellness program announced the study in their
monthly newsletter for three months, and mailed out surveys
to members who called and requested one. The wellness
program also advertised in their newsletter for non-

exercisers for two months. Other MSU faculty and graduate

49
students were invited to participate; the study was
described to a graduate substance abuse class and surveys
distributed to those who expressed interest, and nine
surveys were distributed to friends and acquaintences of the
PI, all within the MSU community. Surveys were also
distributed from a table set up at a local club for
recovering alcoholics, with a poster describing the study
and inviting participation. All subjects completed the
survey at their leisure and returned them to the PI through
the mail with the stamped envelope provided. Participation
was fully voluntary and anonymous. The total time required
to complete the survey was roughly 20-25 minutes.

A record of the number of packets distributed through
each source was maintained, and return envelopes were color
coded according to their distribution sites. This permitted
the return rate of distributed surveys to be determined.
Distribgtion of Surveys

A total of 804 surveys were distributed through the
various sites. Table 3.2 summarizes the number issued,

distributed and returned for each site.

50

TABLE 3.1
Research Survey Distribution and Return Rate

 

 

 

 

 

Total Total
Issued Distributed
to to Potential
Sites Subjects Returned
Private
health clubs 725 500 249
University
subjects 388 194 102
Recovering
alcoholics 110 110 54
1223 804 405

 

 

 

The PI issued 725 surveys to the various health club
managements. Health club front desk personnel then
distributed them to members who requested to participate.
Those surveys returned to the PI comprised the sample.

All scoring was performed by the PI, with 10% of the
surveys re-scored as reliability checks. Two errors in
scoring were found, indicating a possible 5% error rate.
Subjects were categorized into exercise groups by the
principal investigator based on the information they
provided about their exercise frequency and duration. The
variable coding scheme is listed in Appendix 1.

Dgsigg

Since the research base on the contribution of

personality factors to exercise adherence is still quite

sparse, this study utilized a correlational design in a

51
natural setting. Although correlational designs do not
permit cause-effect conclusions, they are useful for
predictive questions and ". . .in the pilot stages of
ongoing research, or during a research program at a point
where overall inspection of inter-correlations is called
for" (Gelso, 1979, p. 17). Also, a prime strength of the
correlational field study is that it permits high
generalizeability to the population of interest.
ggagureg
Obsessive—Compulsive Scale (OCS)

Psychometric measures of compulsivity have only
recently been developed. The most direct measure of
compulsivity appears to be the Obsessive—Compulsive Scale
(OCS). Also, some researchers believe the Breskin Rigidity
Scale (BRT) may measure obsessive-compulsive traits
(Primavera, et al, 1975); for the purposes of this study,
however, the BRT will operationalize flexibility by
assessing rigidity, its opposite.

The OCS is a 22-item true-false scale containing 2 non-
scored validity items (Gibb, Bailey, Best, & Lambirth,
1983). Scores can range from 0 - 20, with higher scores
representing greater compulsiveness; the validation sample
of 46 men and 68 women students, and 57 individual therapy
clients, had a mean of 11.32 and SD = 4.13; range was not
reported. Internal consistency was assessed through item-
total correlations, with each item found to be significantly

correlated with total score. Test-retest reliability was .82

52
over three weeks. A measure of concurrent validity was
obtained by having two clinical psychologists rate client
subjects’ degree of obsessive-compulsive traits. These
ratings correlated .79 with the OCS, a reasonably high
association offering the OCS good construct credibility. No
sex differences and no differences between student and
client groups were found; however, the client group included
a wide variety of concerns, and was not meant to represent
clinical examples of obsessive-compulsive traits. There was
no reason to believe that compulsiveness is any more endemic
to a general client population than it would be in a general
student population. This scale appeared to be quite a good
measure of compulsivity.
Breskin Rigidity Test (BRT)

A measure of perceptual rigidity has been developed by
Breskin (1968), based on the perspective that

the rigid person’s requirement for structure
would be manifested in the use of the laws of

Praegnanz, i.e., a more rigid person, given the

opportunity to express a preference between pairs of

figures differing only in "goodness of fit," will tend

to select the "better" fit. (pp. 1203-1204).

Thus, the Breskin Rigidity Test (BRT) uses ". . . a
perceptual rather than a conceptual frame of reference to
measure rigidity" (Breskin, 1968, p. 1203), is not dependent
on verbal or mathematical ability, and has been found to be

independent of intelligence (Breskin & Rich, 1971;

53
Primavera & Simon, 1971).

The BRT was constructed and standardized on a sample of
132 undergraduates (68 men and 64 women), and validated by
comparing the scores of art students to accounting and
secretarial students. Significant differences exist between
males and females, with men scoring in the more rigid
direction; the mean for men was 8.5, with a standard
deviation of 3.6, while for women the mean was 6.7 with a
standard deviation of 3.0. The BRT is a 15-item forced
choice scale that asks subjects to choose between two
figures that differ in their degree of perceptual closure; a
choice for the "better fit" or for higher closure is scored
as a rigid response (Breskin, 1968). Scores can range from
0—15, with higher scores indicating greater rigidity.
Reliability is considered more than adequate, with odd-even
reliability equal to .78 and a .98 KR-20. The BRT takes
only 2-3 minutes to complete.

Low rigidity as measured by the BRT has been found to
be significantly associated with creativity and problem-
solving (Gorman & Breskin, 1969), whereas high rigidity was
inversely related to anxiety at a significant level
(Primavera, et al, 1975), implying that rigidity may be one
manner in which the ego effectively binds anxiety. The
literature supports the BRT as a good measure of rigidity.
Barratt Impulsiveness Scale — 10th Revision

The Barratt Impulsiveness Scale - 10th Revision

(BIS-10) has been used extensively with adults. Also,

54
impulsivity as measured by this scale is well-grounded
within a neuropsychological framework that includes
biological, cognitive, and behavioral components (Barratt,
1985a).

The BIS-10 is a 34-item paper-and-pencil scale scored
on a 4-point scale ranging from "rarely-never",
"occasionally", "often", to "almost always-always". It
contains three subscales, representing the three components
of impulsiveness proposed by Barratt (1985b): motor
impulsiveness (Im), non-planning impulsiveness (Inp), and
cognitive impulsiveness (Ic). The multifactorial nature of
the BIS scale (all revisions) has been made clear in a
number of studies (Barratt, 1965, 1972, 1983; Barratt &
Patton, 1983; Barratt, Patton, Olsson & Zuker, 1981); the
three factors mentioned above have been confirmed, although
the cognitive subtrait is the least psychometrically sound
(Barratt, in press; Luengo, Carrillo-de-la—Pena & Otero,
1991) .

Barratt (1985b) reported reliabilities ranging from .89
to .92 for the three subscales (alpha coefficients),
although Luengo et al. (1991) found somewhat lower
reliabilities ranging from .56 to .65. They also reported a
test-retest reliability of .60 over one year. A
standardization subsample of 412 undergraduates obtained a
mean of 43.24 and standard deviation of 15.11.

The BIS—10 has been used in a wide variety of studies,

and construct validity is well-established. Impulsiveness

55

as measured by the BIS—10 has been positively associated
with a number of perceptual-motor tasks (such as reaction
time and visual tracking), intra-individual variability in
performance, cognitive performance (such as time
estimation), and a number of psychophysiological measures
(Barratt, 1985b). High-impulsive subjects perform poorly at
psychomotor tasks and process information less efficiently
(Barratt, 1987), easily angered people are significantly
more impulsive (Barratt, 1991), and the BIS—10 successfully
differentiated subjects diagnosed with antisocial
personality disorder from subjects diagnosed with thought
disorders, mood disorders, and substance abuse disorders
(Barratt, 1985b). In summary, the BIS-10 is a reliable and
highly valid instrument for measuring impulsiveness.
Tolerance of Ambiguity Scale - AT-20

MacDonald’s (1970) Ambiguity Tolerance Scale (AT-20) is
a revision of the Rydell and Rosen (1966) scale. The AT-20
is a 20 item, true-false, paper—and—pencil measure scored
for high tolerance of ambiguity. The addition of four items
to the Rydell and Rosen (1966) test to form the AT-20
improved the original .64 split-half reliability to a split—
half reliability of .86, representing a sizable increase in
reliability. Test-retest reliability was .63 over six
months, and is consistent with test-retest reliabilities of
.71 and .57 (over 1 and 2 months, respectively) of the
original Rydell and Rosen (1966) scale. The AT—2O has been

shown to be free of social desireability effects as well,

56
with a correlation of only .02 (MacDonald, 1970).

The AT-20 was administered to a validation sample of
undergraduate students; a mean of 10.51 and a standard
deviation of 3.32 was obtained; no sex differences were
observed. AT-20 scores were significantly correlated with
rigidity, dogmatism, and frequency of church attendence. An
interesting result was obtained with a subsample of male
physical education majors, which revealed a significantly
negative (-.30, p < .01) correlation between ambiguity
tolerance and rigidity.

In summary, the AT-20 possesses high reliability and
validity, and should be a useful instrument for the
measurement of ambiguity tolerance.

Dependent Variables

The dependent variables vary according to the analysis
used. To determine whether the exercise group means differ
across the assessment instruments, the assessment scores are
the dependent variables for the MANOVA. For the predictive
analysis utilizing discriminant analysis, the dependent
variables are the exercise categories. Subjects were
classified according to the frequency and duration of
exercise reported on the demographic questionnaires.
Categories were defined as follows:

1) Super-exercisers: people who had exercised very
frequently (5 or more times per week) for 6 months or
more. Those who indicated they had cut back or stopped

exercising in the past were classified as super-

2)

3)

57

exercisers only if they had been exercising for three
years or more, the reason(s) for interruption were
unavoidable (injury, surgery, pregnancy, relocation)
with no motivational elements, i.e., "time
constraints", "work obligations", etc., and exercise
was resumed as soon as possible. If the subject had
exercised for less than three years or if motivation
was part of the reason for the interruption(s), the
exerciser was classified as an intermittent exerciser.
General exercisers: people who exercised an average of
2-4 times per week for 6 months or more. Those who
indicated they had cut back or stopped exercising in
the past were classified as general exercisers only if
they indicated they had exercised for three years or
more, whose reasons for interruption did not include
motivational elements, and exercise was resumed as soon
as possible. Otherwise they were classified as
intermittent exercisers.
Intermittent exercisers: a) people who exercised 2-4
times weekly for 3 months or more and then reported
having stopped or cut back on frequency to once a week
or less, and/or
b) people who initiated an exercise plan and failed

to adhere to it for at least 3 months (the

traditional dropouts), and/or
c) people who exercised once a week or less.

d) those who had been exercising for less than 6

58

months, but had previously dropped out of regular

exercise.
Though this category includes those traditionally
defined as dropouts, it is broader in scope by
including those who exercise steadily for short periods
of time and those who exercise irregularly. This more
inclusive definition will hopefully better represent
real world exercisers.

4) Non-exercisers: people who never initiated an exercise
plan, or had not exercised in the last three years.
People who indicate that they had been exercising

regularly for less than 6 months, but had not previously

exercised and then stopped, were not classified into an
exercise category, since it was uncertain whether they would
continue for the requisite six months; their responses were
included only in the correlational analysis of the
assessment instruments. Subjects who indicated they stopped
exercising or cut back on exercise frequency due to injury
were not classified; their responses were used only in the
correlational analysis. These classification rules should
make the categories as clean and mutually exclusive as
possible.

Covariates
Information on a number of possible confounding

variables was collected. These included age, sex, perceived

enjoyment of exercise, whether the exerciser’s efforts were

supported by a partner, the exercise format (alone, with a

59
partner, or in a group), type of exercise (aerobic,
anaerobic, or mixed), and exercise intensity. Information
on these variables was supplied by the subject on the
demographic questionnaire; age was measured in years, and
degree of enjoyment and exercise intensity were both
recorded on 10-point Likert scales. The specific coding
scheme for each variable is listed in Appendix A.
Hypotheses

The research questions in this study were twofold.
First, the relationships between the personality variables
of impulsiveness, compulsiveness, rigidity, and ambiguity
tolerance were examined. The following are the hypotheses
regarding these relationships:

1) Compulsiveness will be negatively related to both
impulsiveness and tolerance of ambiguity.

2) Compulsiveness and rigidity will be positively related.

3) Impulsiveness and tolerance of ambiguity will be
negatively related.

4) Tolerance of ambiguity and rigidity will be negatively
related.

The remainder of the research question asks whether a
relationship exists between psychological adaptability and
level of exercise adherence. Psychological adaptability is
operationalized as low impulsiveness, low compulsiveness,
low rigidity, and high tolerance of ambiguity. Specifically:
5) general exercisers will score significantly lower on

impulsiveness, compulsiveness, and rigidity, and higher

60
on tolerance of ambiguity than intermittent exercisers,
6) super-exercisers will score significantly higher on
rigidity and compulsiveness than other exercise
groups, and lower on impulsiveness and tolerance of
ambiguity.

By group, the personality measures should fall as follows:

 

COMP IMP RIG TOL-AMB
(OCS) (BISlO) (BRT) (AT20)
Non-exercisers Wide range of all variables
Intermittent Moderate High Moderate Moderate
exercisers
General Low Moderate Low High
exercisers
Super-exer. High Low High Low

 

COMP = compulsiveness IMP = impulsiveness RIG = rigidity
INT-AME = intolerance of ambiguity
7) Compulsiveness, impulsiveness, rigidity, and
intolerance of ambiguity will combine to predict
exercise group membership.
Site Characteristics
Data were collected from several types of sites:
private health clubs, the university community (excluding
undergraduates), and recovering alcoholics. A preliminary
MANOVA was conducted to assess whether the data from the
differing categories of sites could reasonably be combined
into one subject pool, by determining whether there were

differences among the three sites on the predictor

61
variables. A significant overall multivariate effect was
found (F = 3.74, p = .000); followup univariate F-tests
revealed significant effects for: impulsiveness (F = 14.67,
p = .000, eta? = .07), perceived enjoyment (F = 16.97,

p = .000, eta? = .08), and perceived intensity (F = 17.57,

p = .000, eta2 .09). No differences were found for either
age or sex differences.

For enjoyment, the difference among sources is
accounted for by the non-exerciser group, which was
recruited among the university community and the alcoholic
subgroups. Logically, the recruitment of non-exercisers at
the exercise facilities is unlikely. Table 3.1 compares the
means and standard deviations for enjoyment, impulsiveness,
and intensity across the three sources, for the total sample
versus exercisers only (non-exercisers excluded). However,

intensity is irrelevant for non-exercisers, so only

exercisers’ reported intensity can be shown in the table.

TABLE 3.2
Means and Standard Deviations for Enjoyment, Impulsiveness, and Intensity
for the Total Sample and for Exercisers Only

 

 

 

Enjoyment Impulsiveness Intensity
Total Exer.only Total Exer.only Exercisers only

Source X SD X SD X SD X SD X SD
Private
health clubs 2.23 1.64 2.18 1.56 46.71 14.96 46.65 14.84 7.03 1.58
University
community 3.32 2.29 2.38 1.45 40.78 12.82 41.13 13.38 5.83 1.97
Recovering
alcoholics 3.44 2.22 2.22 1.51 53.69 14.57 49.50 13.27 6.05 1.53

 

 

 

 

 

 

 

 

62

As the table shows, the site differences for enjoyment
disappear when non-exercisers are excluded. The results of
a followup ANOVA for enjoyment by source revealed no
significant difference when non-exercisers were excluded
from the analysis (F = .51, p = .601). Since the observed
effect was attributable to membership in one of the exercise
groups, which was one of the factors of interest in this
study, concern for differential results by site was
unwarranted for the enjoyment variable.

There was somewhat more concern for the impulsiveness
and intensity variables. As the table shows, the exercisers
scored lower than the total sample only for the alcoholic
subsample; thus, non-exercisers exerted only moderate
differential influence for impulsiveness. The pattern of
site difference remained the same. For intensity, there was
no non-exercise group with which to compare.

However, the amount of variance accounted for by site
was low (7% for impulsiveness and 9% for intensity). Also,
the practical significance of a difference in means of 1.2
points for intensity is questionable. Therefore, the sites
were aggregated into a total sample.

Data Analysis

Type I alpha was set at .10, reflecting the exploratory
nature of this research. Efforts were made to ensure that
power equaled .80; however, since the effect size was
unknown, it was difficult to determine the number of

subjects necessary. Hypothesizing a small effect size of

63
.30, 140 subjects per group were necessary to meet a .80
power requirement. It was not possible to recruit this
number, as may be seen in Table 3.3.

TABLE 3 .3
Exercise Group Membership

 

 

 

 

 

 

n

Non-exercisers 44

Intermittent

exercisers 132

General

exercisers 92

Super-exercisers 116

Unclassifiable 21
TOTAL 405

 

 

 

 

There were 21 surveys that were not classifiable. The
most frequent reason was that the exerciser had been
exercising less than six months. The Unclassifiable group
was included only in the correlational analysis, which did
not include the exercise data.

The Non-exerciser group, at n = 44, was the most
difficult to recruit, primarily because of the liberal
definition of exercise used in this study. If the effect
size is as small as postulated, this number may reduce the
power of the analysis to .40 for those analyses including

the Non-exercise group. For the exercise groups, the 92

64
General-exercisers limit power to .65 for those analyses
including only the exercisers. If the effect size were to
be slightly higher at .40, power would also increase.
Hypotheses #1 — 4: These hypotheses were tested through
determination of the correlations between the measures of
impulsiveness, compulsiveness, rigidity, and intolerance of
ambiguity.
Hypotheses 5 Q 6: The hypotheses regarding the mean scores
of the exercise groups were tested by a multivariate
analysis of variance (MANOVA).
Hypothesis # 7: Discriminant analysis was used to assess
the degree to which compulsiveness, impulsiveness, rigidity,
and tolerance of ambiguity predicted exercise group
membership, while accounting for the possible influence of
covariates and mediating variables. This technique is
particularly well-suited to assess which variables
differentiate most effectively between groups. Also, it
predicts group membership for individual subjects, which is
then compared to the actual group membership. The
percentage of correct predictions can then be compared with
the percent of group membership expected by chance.
Discriminant analysis can be used to predict group
membership for both dichotomous outcomes and for more than

two groups.

Chapter 4: RESULTS

The purpose of this study was to explore two related
research questions: 1) the relationships between
compulsiveness, impulsiveness, rigidity, and tolerance of
ambiguity, and 2) the degree to which compulsiveness,
impulsiveness, rigidity, and tolerance of ambiguity
influence exercise habits. The chapter will first test the
hypotheses listed in the preceding chapter, with supporting
or disconfirming conclusions drawn. Considerations
regarding the statistical assumptions underlying
discriminant analysis will be discussed, and post-hoc

analyses presented.

Tests of Hypotheses
Hypotheses 1-4:

It was hypothesized that:
1) Compulsiveness will be negatively related to both

impulsiveness and tolerance of ambiguity.

2) Compulsiveness and rigidity will show a positive linear
relationship.
3) Impulsiveness and tolerance of ambiguity will be

negatively related.
4) Tolerance of ambiguity and rigidity will be negatively
related.
Table 4.1 shows the correlation matrix for the
personality variables, with the associated probability

estimates (2—tailed probability).

65

66

TABLE 4.1
Correlation Matrix for Personality Variables

OCS BRT BISlO AT20
OCS 1.00
p:
BRT .324 1.00

p =.000 p =

BISlO -.081 -.267 1.00
p =.120 p =.000 p =
AT20 -.526 -.281 -.072 1.00
p =.ooo p =.ooo p =.17o p =
OCS Obsessive-Compulsive Scale

BRT Breskin Rigidity Test
BISlO = Barratt Impulsiveness Scale - 10th Revision
AT20 = Ambiguity Tolerance Scale

Hypothesis #1 is partially supported; compulsiveness
and tolerance of ambiguity show the strongest relationship
of all, with a negative correlation of r = —.526, p = .000.
However, compulsiveness and impulsiveness failed to show a
significant difference. Hypothesis #2 is supported, with
compulsiveness and rigidity being positively related at
r = .324, p = .000.

Tolerance of ambiguity and impulsiveness were not
significantly related; therefore, Hypothesis #3 cannot be
supported. Tolerance of ambiguity and rigidity, however,
showed a significant negative correlation of r = -.281,

p = .000, which supports Hypothesis #4.

The hypotheses as stated addressed only linear

67

relationships; the possibility of nonlinear relationships

was explored, however, by examining the scatterplots of the

psychological variables against each other. Unfortunately,
no type of curvilinear relationship could be distinguished
between any of the variables. The significant correlations
noted above were clearly linear when scatterplots were
inspected.

Hypotheses #5 and 6:

It was hypothesized that:

5) general exercisers will score significantly lower on
impulsiveness, compulsiveness, and rigidity, and higher
on tolerance of ambiguity than intermittent exercisers,
and

6) super-exercisers will score significantly higher on
compulsiveness and rigidity than the other groups, and
lower on impulsiveness and tolerance of ambiguity.

Groups were hypothesized to score as follows:

 

COMP IMP RIG TOL-AMB
(OCS) (BISlO) (BRT) (AT20)
Non-exercisers Wide range of all variables
Intermittent Moderate High Moderate Moderate
exercisers
General Low Moderate Low High
exercisers
Super-exer. High Low High Low

 

Means and standard deviations of the psychological

variables are presented in Table 4.2:

68

TABLE 4.2
Means and Standard Deviations for Personality Factors
(with MANOVA Results)

 

 

 

 

ocs BISIO BRT AT20
x so x so x so x so

Non-

exer. 11.77 4.18 48.48 14.22 8.33 3.50 8.44 3.72

Inter.

exer. 10.75 3.87 48.24 15.71 7.71 3.48 9.50 3.64

Gen.

exer. 10.67 4.26 47.55 14.81 8.21 3.27 9.71 4.17

Super

exer. 11.04 4.65 42.26' 13.96 H 7.91 3.25 | 9.52 4.17 |

l

 

 

 

 

 

OCS - Obsessive-Compulsive Scale; BRT = Breskin Rigidity Test
31810 a Barratt Impulsiveness Scale-10th Revision;

AT20 a Ambiguity Tolerance Scale; * - p =.01

df (3. 266)

These hypotheses were tested with a multivariate
analysis of variance (MANOVA). The results identified a
significant overall multivariate effect (F = 1.69,

p = .063). Follow-up univariate F-tests were conducted, and
identified a significant difference between the exercise
groups on impulsiveness, between the super-exercise group
and all other groups (F = 3.94, p = .009); no other
significant differences were found. Follow-up univariate
F-tests, though they do not control Type I error in
themselves, are considered "protected" when used to further
analyze MANOVA results (Bray and Maxwell, 1982), since the
MANOVA procedure controls for the inflation of Type I error
inherent in doing several univariate procedures, and a non-
significant MANOVA result is a priori not analyzed further.

Although several more conversative methods exist for those

69
situations involving highly correlated predictors and when
the linear combinations of variables are complex, ". . . if
the purpose of the analysis is to control Type I error rate
for a set of p univariate ANOVAs, then these methods are
generally appropriate" (Bray and Maxwell, p. 343). In view
of the exploratory nature of this study, protected F—tests
are considered logically sufficient to answer the questions
of which variables contributed to the overall significant
MANOVA and whether group differences existed for each
variable. Thus, these data indicate that Hypothesis #5 was
not supported. Hypothesis #6 was partially supported, since
super-exercisers scored significantly lower on impulsiveness
than all others. They did not score higher on either
compulsiveness or rigidity.
Hypothesis #7:

It was hypothesized that compulsiveness, impulsiveness,
rigidity, and intolerance of ambiguity will combine to
predict exercise group membership. Several discriminant
analyses were performed to test this hypothesis; covariates
were included to examine their possible influence.

The first discriminant analysis was employed to predict
group membership in all four groups (the three exercise
groups and the non-exercisers). Membership in the three
exercise groups was then analyzed. Next, an analysis tested
whether non-exercisers could be distinguished from all
exercisers, and from super-exercisers in particular.

Finally, an attempt was made to differentiate intermittent

70
from general exercisers.

Analysis #1: The OCS, BRT, BISlO, and AT20 were used
to predict group membership in the four exercise groups.
The variable selection criteria employed was the
minimization of Wilks’ lambda, which develops the
classification functions so as to minimize the residual or
unexplained variance. Since there were unequal numbers of
subjects in the exercise groups, the probability of being
assigned to each group by chance was specified as the ratio
of each group to the total (proportional prior
probabilities). Table 4.3 presents the three discriminant

functions derived and Table 4.4 the standardized

coefficients for each of the variables entering the equation

for the one significant function:

TABLE 4.3
Discriminant Function Results for Analysis
including OCS, BIS10, BRT and AT20

 

 

 

 

 

Canonical Percent of]
X2 p Correlation Eigenvalue Variance |
Function 1 18.90 .026 .191 .038 71.51
Function 2 5.44 .245 .101 .010 19.29
Function 3 1.76 .185 .070 .005 9.20
TABLE 4.4

Standardized Discriminant Function Coefficients
for Function 1

 

Discriminant Function

 

 

81810 1.023
OCS -.162
BRT .437

 

 

(AT20 did not enter the equation)

71

Of the three discriminant functions derived, only the
first was significant (X2 = 18.9, p = .026). Classification
accuracy was 37.63%; a significance test for the
classification rate may be carried out using either the
proportional chance criterion or the maximum chance
criterion (Huberty & Barton, 1989). The proportional chance
criterion is generally used when individual group accuracy
is considered a higher priority than overall accuracy, and
thus is the more appropriate criterion in this case. Using
the proportional chance criterion, the classification rate
of 37.63% was statistically significant (2 = 3.34,
p = .001). Unfortunately, all but one of the non-exercisers
were misclassified and 88% of the general exercisers were

misclassified. Table 4.5 shows the classification table.

TABLE 4.5
Classification Results for Analysis 1a
with OCS, BIS10, BRT and AT20 as Predictors

 

 

Actual No. of Predicted Group Membership

Group Cases 1 2 3 4
Non- 43 1 23 1 18
exercisers 2.3% 53.5% 2.3% 41.9%
Intermittent 126 1 73 8 44
exercisers .8% 57.9% 6.3% 34.9%
General 90 0 47 11 32
exercisers .0% 52.2% 12.2% 35.3%
Super 116 0 51 7 55
exercisers .0% 45.1% 6.2% 48.7%

 

 

 

 

 

Percent of cases correctly classified: 37.63%
Predicted Group Key: 1 = Non-exercisers; 2 = Intermittent exercisers;
3 = General exercisers; 4 = Super—exercisers

72

As can be seen by the discriminant function
coefficients, and the fact that the BISlO was the only
variable entering the one significant function, the B1810
appeared to carry most of the discriminatory power. Thus, a
second analysis was performed using only the BISlO as a
predictor. The function derived was significant
(X2 = 12.31, p = .006); the canonical correlation was .179,
with an eigenvalue of .033. Classification accuracy was
almost exactly the same as for the first analysis at 38.90%.
Again using the proportional chance criterion, this
classification rate was significantly better than chance
(2 = 17.58, p = .000). However, none of the non-exercisers
or general exercisers were classified correctly. Table 4.6

shows the classification table.

TABLE 4.6
Classification Results for Analysis 1b
with BISlO as Predictor

 

 

Actual No. of Predicted Group Membership

Group Cases 1 2 3 4
Non- 44 0 32 0 12
exercisers .0% 72.7% .0% 27.3%
Intermittent 132 0 91 O 41
exercisers .0% 68.9% .0% 31.1%
General 91 0 56 0 35
exercisers .0% 61.5% .0% 38.5%
Super 116 0 58 0 58
exercisers .0% 50.0% .0% 50.0%

 

 

 

 

 

Percent of cases correctly classified: 38.90%

These first two analyses used only the personality

factors as predictors, and although their predictive

73
accuracy was better than chance, individual group accuracy
was very poor, with few or none of the non-exercisers and
general exercisers correctly classified. Also, as the ratio
of between~groups to within-groups variation, an eigenvalue
of .033 marks a poor function. Squaring the canonical
correlation of .179 reveals that the analysis accounts for
only 3% of the variance. Thus, although the discriminant
function is statistically significant, impulsiveness alone
has little practical utility for predicting exercise group
membership.

Several variables have been identified, however, as
likely covariates or mediating variables, and their possible
contribution to classification was examined next. A
preliminary oneway ANOVA on age, and chi-squares on sex,
support of significant other, exercise format (whether the
exercise was performed alone, with a partner, or in a
group), and exercise type (aerobic, anaerobid, or mixed)
revealed no significant differences among the groups, and
none of these variables contributed to any significant
discriminant functions. Perceived enjoyment (ENJOY) entered
the equation along with impulsiveness (BISlO).

The best model for discriminating among all four groups
included only ENJOY as a predictor. The significant
discriminant function (X2== 177.47, p = .000) produced an
eigenvalue of .594 and a canonical correlation of .610,
accounting for 37% of the variance and resulting in a

classification accuracy of 48.96%, which was significantly

74

better than chance. Table 4.7 shows the classification.

TABLE 4.7
Classification Results for Analysis 1c
with ENJOY as Predictor

 

 

Actual No. of Predicted Group Membership

Group Cases 1 2 3 4
Non- 44 21 22 0 1
exercisers 47.7% 50.0% .0% 2.3%
Intermittent 132 12 92 0 28
exercisers 9.1% 69.7% .0% 21.2%
General 92 5 57 0 30
exercisers 5.4% 62.0% .0% 32.6%
Super 116 2 39 0 75
exercisers 1.7% 33.6% .0% 64.7%

 

 

 

 

 

Percent of cases correctly classified: 48.96%

Although including the B1810 with ENJOY did result in a
second significant discriminant function (X? = 7.58,
p = .0226), this function did not increase classification
accuracy. Instead, accuracy decreased slightly to 47.00%.
Thus, the best model for classifying subjects into the four
exercise groups employed ENJOY as the sole predictor.

Analysis #2: The next analysis attempted to
discriminate between the three exercise groups. This
permitted exercise intensity (INTENS) to be included;
intensity was irrelevant for non-exercisers, and so could
not be used in previous analyses. Enjoyment and intensity,
however, were significantly correlated (r = -.35, p = .000);
the analysis was not improved when both were included, and
ENJOY was a slightly better predictor when tested
separately. Therefore, ENJOY was retained in the model and

INTENS was dropped.

75

Again, the best classificatory model included only
ENJOY as the predictor. The significant function
0? = 32.9, p = .000) exhibited an eigenvalue of .103 and
produced a canonical correlation of .305, accounting for
9.3% of the variance, and correctly classified 52.65% of
subjects, which was significantly better than chance
(2 = 6.34, p = .000). Table 4.8 shows the classification

results for the predictive model including ENJOY by itself.

TABLE 4.8
Classification Results for Analysis 2
with ENJOY as Predictor for All Exercisers

 

 

Actual No. of Predicted Group Membership

Group Cases 2 3 4
Intermittent 132 104 0 28
exercisers 78.8% .0% 21.2%
General 92 62 0 30
exercisers 67.4% .0% 32.6%
Super 116 41 0 75
exercisers 35.3% .0% 64.7%

 

 

 

 

 

Percent of cases correctly classified: 52.65%

Interestingly, the deletion of the non-exercisers from
the analysis resulted in much less of the variance being
explained (37% when all four groups were predicted, only 9%
for the three exercise groups). To investigate this, a
discriminant analysis was conducted to attempt to
distinguish between the non-exercisers and all exercisers.
Again, the best predictor was ENJOY, with a significant
discriminant function (X? = 147.7, p = .0000), an eigenvalue
of .474, and a canonical correlation of .5673, accounting

for 32% of the variance. Classification accuracy was

76
89.06%, significantly better than chance (2 = 4.14,
p = .000). As before, the very unequal group sizes (non-
exercisers = 44, exercisers = 340) were taken into account
through specification of prior probabilities, and the use of
the proportional probabilities criterion with which to
derive the test statistic. Table 4.9 displays the

classification table.

 

 

TABLE 4 . 9
Classification Results for Non-exercisers vs. All exercisers
with ENJOY
Actual No. of Predicted Group Membership
Group cases Non—exer. All exercisers
Non- 44 21 23
exercisers 47.7% 52.3%
All 340 19 321
exercisers 5.6% 94.4%

 

 

 

 

Percent of cases correctly classified: 89.06%

Interestingly, including the BIS10 resulted in no
change in classification accuracy for either individual

groups or the overall subject group.

Analysis #3: As Table 4.2 notes, the only significant
difference in means across exercise groups was for
impulsiveness, with super-exercisers scoring lower than all
other groups. Thus, an analysis was run to examine whether
the two extremes (non-exercisers and super-exercisers) could
be distinguished.

The best model for discriminating between these two
groups was perceived enjoyment and impulsiveness. The

significant function (X2 = 140.18, p = .000) produced an

77
eigenvalue of 1.44, and a canonical correlation of .768,
accounting for 59% of the variance. Table 4.10 presents the
classification table for non-exercisers versus super-

exercisers.

TABLE 4.10
Classification Results for Non-exercisers vs. Super-exercisers
with ENJOY and 81810

 

 

Actual No. of Predicted Group Membership
Group cases Non—exer. Super-exer.
Non- 44 34 10
exercisers 77.3% 22.7%
Super- 116 4 112
exercisers 3.4% 96.6%

 

 

 

 

Percent of cases correctly classified: 91.25%

The percent of cases correctly classified was 91.25%, which
was significantly better than chance (z = 8.65, p = .000).

Deleting impulsiveness from the equation decreases
overall accuracy only slightly, to 89.38%, and by that
criterion it would not seem particularly useful. However,
its deletion loses 9% accuracy in the classification of non-
exercisers, though losing only .8% accuracy for the super-
exercisers; thus, impulsiveness is a useful factor in
maximizing the identification of non—exercisers versus
super-exercisers.

Analysis #4: Keeping in mind the theoretical
distinction between intermittent and general exercisers, the
last discriminant analysis sought to distinguish between
these two exercise groups. ENJOY and BISlO gave the best

results; although the function was significant

78
(X? = 3.61, p = .05), the eigenvalue was very small at .016,
the canonical correlation of .128 accounted for only 1.6% of
the variance, and the classification equation assigned every
subject to the intermittent category when proportional prior
probabilities were specified. Since the group proportions
were not too dissimilar (59% intermittents, 41% general),
the analysis was re—run using equal prior probabilities.
Summary statistics all remained the same, but overall
classification accuracy dropped from 58.93% to 54.02%.
However, assignments were made to each group rather than
assigning all subjects to the largest group; 48.5% of
intermittent exercisers were correctly classified, and 62.0%
of general exercisers were correctly classified. Using the
proportional chance criterion, this classification rate was
not significantly different from chance (2 = 1.21, p = .22).
Thus, intermittent and general exercisers could not be
distinguished using these variables.

Summary of Discriminant Analyses: Given the lack of
influence of the personality variables in distinguishing the
exercise groups, Hypothesis #7 is not supported, with the
exception that impulsiveness contributes to the
discrimination between non-exercisers and super-exercisers.
Perceived enjoyment of exercise, however, proved a
significant predictor between all four groups, between the
three exercise groups, between non—exercisers and all
exercisers, and between non-exercisers and super-exercisers.

General and intermittent exercisers were indistinguishable.

79
Table 4.11 summarizes the pertinent statistics for each

of the discriminant analyses (for the significant functions)

TABLE 4.11
Summary Statistics for Discriminant Analyses

 

 

Canonical Classification I

p Eigenvalue Correlation(CC) CC2 Accuracy |
Analysis
1 v 2 v 3 v 4 .026 .038 .191 3.6% 37.63%
(with OCS, BRT,
BISlO, and AT20)
l v 2 v 3 v 4 .006 .033 .179 2.9% 38.90%
(with BISlO only)
1 v 2 v 3 v 4 .000 .594 .610 37.2% 48.96%
(with ENJOY only)
2 v 3 v 4 .000 .103 .305 9.3% 52.65%
(with ENJOY only)
1 v 2, 3, & 4 .000 .474 .567 32.1% 89.06%
(with ENJOY only)
1 v 4 .000 1.44 .768 58.9% 91.25%
(with ENJOY & 81810)
2 v 3 .05 .016 .128 1.6% 54.02%
(with ENJOY & 31510)

 

 

 

= Non-exercisers; 2 = Intermittent exercisers
= General exercisers; 4 = Super exercisers

1
3
As the table shows, all the functions were
statistically significant, and all but the last analysis
gave statisically significant classification results.
However, only three of the eigenvalues were large enough to
indicate a meaningful ratio of between-group to within-group
variance; also, their associated canonical correlations were
the only ones large enough to explain a reasonable amount of
variance. All three of the most meaningful analyses
included the non-exercise and super-exercise groups, which

were the most different on both the enjoyment and

impulsiveness variables; differentiating between the

80

superexercisers and the intermittent and general exercisers
was the next most useful result, though still accounting for
only 9.3% of the variance. The other three analyses, though
statistically significant, contained little if any practical
utility in distinguishing between exercise groups.
Post-hoc Analysis

Given the finding that perceived enjoyment was the most
discriminating factor found, enjoyment was examined a bit
further. A one-way ANOVA for enjoyment across exercise
groups was significant (F = 71.49, p = .000). Table 4.11
shows the means and standard deviations for ENJOY across the
exercise groups. As the table shows, most of the difference
lies between the non—exerciser and the other three exercise

groups.

TABLE 4 . 12
Means and Standard Deviations for ENJOY across Exercise Groups

 

 

 

 

 

X SD
Non-exercisers 5.84 2.33
Intermittent exercisers 2.75 1.70
General exercisers 2.32 1.42
Super-exercisers 1.64 1.25
Total sample 2.67 1.98

 

Also, there was a slight but significant negative
correlation between enjoyment and tolerance of ambiguity
(r = -.1322, p = .01), though not with any of the other

personality variables.

81

Only one variable showed significant sex differences.
Given that the literature on the Breskin Rigidity Test (BRT)
had identified differences between male and female scores, a
oneway ANOVA was performed to determine if these differences
were present in this sample as well; results indicated that
men did score significantly higher on rigidity than women
(F = 4.32, p = .04); however, the difference in means was
less than one point, and thus there is little practical
significance to this difference.

Spapigtical Assumptions

As with most statistical techniques, the robustness of
discriminant analysis results depends on the standard
assumptions of normality and homogeniety of variance. In
the multivariate case, the assumptions are that the
independent variables come from a multivariate normal
population, and the population covariance matrices are equal
for all predictors.

All variables except ENJOY showed a largely normal
distribution on a scatterplot. ENJOY was not at all normal,
instead exhibiting a strong ceiling effect of the
measurement method (a 10-point Likert scale). Super-
exercisers overwhelmingly indicated that they love to
exercise, and showed the least amount of variance compared
to the other groups. General and intermittent exercisers
showed a slightly higher mean and somewhat more variability;
the non-exerciser group alone showed a faintly normal

distribution, and the most variation.

82

Along with nonnormality, ENJOY also evidenced
heterogeniety of variance across the exercise groups, with
non-exercisers being the most variable. Box’s M-test was
used to test for homogeneity of variance as part of the
discriminant analysis; a significant result indicates that
the groups have significantly different variances. Again,
ENJOY was the only variable that failed the homogeniety of
variance assumption.

Although ENJOY clearly fails to meet the model
assumptions for discriminant analysis, its relevance and
value to distinguishing between exercise groups should not
be summarily dismissed. Box’s M-test tends to result in
small probability values (indicating heterogeniety) when
sample sizes are large, even when the group covariance
matrices are not too different, as seems to be indicated by
Table 4.12. Also, the test is quite sensitive to non-
normality, and will identify the group variances as unequal
when the normality assumption is violated (Norusis/SPSS,
1990). Given the relatively large sample size in this
study, and the non-normal distribution of ENJOY, the unequal
variances for ENJOY is not of primary concern.

The violation of the normality assumption, however, is
still an issue, since "If multivariate normality is
violated, tests of significance and estimated classification
rates may be biased" (Dillon & Goldstein, 1984, pp. 380-
381). More specifically,

" . . . the performance of the linear discriminant

83
function in nonnormal situations can be very
misleading. In particular, though the overall
classification error rates were usually not severely
affected, the individual group error rates were
distorted in that error rates were generally much
larger than the optimal value for one population group,
and much small for the other . . . however, . . the
linear discriminant function is likely to be affected
least when each variable is bounded . . ., in contrast
to the case where the range of each variable is
infinite" (p. 381).
Measurement of enjoyment was a bounded measure, which gives
some support to the judgment that the discriminant analyses
run with ENJOY as a predictor are reasonably robust to the
violation of the normality assumption. Also, comparative
analyses were performed using logistic regression, which is
considered quite robust to violation of assumptions; results
were very similar to those reported above. The reason
logistic regression was not chosen as the primary analytic
technique was that it requires a dichotomous outcome;
analysis of more than two groups was not possible.
Therefore, though ENJOY exhibits a markedly non—normal
distribution and significantly heterogeneous variances, the
results reported above should still be considered
tentatively viable and defensible. For the reasons
discussed above, we should be able to place at least

provisional confidence in these results.

8 4
SUMMARY

This study sought to determine: 1) the relationships
among the personality factors of compulsiveness,
impulsiveness, rigidity, and tolerance of ambiguity, and
2) whether these personality factors influenced exercise
habits. Several of the hypotheses regarding the
relationships between compulsiveness, impulsiveness,
tolerance of ambiguity, and rigidity were supported.
Compulsiveness was positively related to rigidity,
negatively related to tolerance of ambiguity, and not
significantly related to impulsiveness. Rigidity was
negatively related to both impulsiveness and tolerance of
ambiguity, and impulsiveness and tolerance of ambiguity were
not related.

The results indicate that the personality variables
examined do not help identify different exercise groups,
with the exception of impulsiveness, which helped
discriminate between non-exercisers and super-exercisers.
Rather, enjoyment of exercise was highly influential, and
discriminated significantly among exercise groups. A slight
negative correlation between enjoyment and tolerance of

ambiguity was also found.

Chapter 5: DISCUSSION

This chapter will begin with a brief discussion of
sampling and measurement considerations, followed by a
discussion of the results of hypotheses tests. Implications
of the results for practice will be considered, and,
finally, recommendations for future research will be
discussed.

Sgppling and Measurement Considerations

Overall, the sampling and measurement efforts in this
study were acceptable, though not ideal. The sample size
was sufficient to identify relatively subtle effects, and
statistical significance was obtained with small absolute
differences. The aggregation of sites into one sample pool
was also considered acceptable, since the pattern of scores
were the same across the exercise groups.

Self-selection of subjects remains a threat to internal
validity, since those who agreed to participate may differ
in some systematic way from those who chose not to
participate. The voluntary nature of participation,
however, made this threat unavoidable since random selection
was not possible. Since this was primarily a group
comparison field study in a naturalistic setting, however,
lack of randomization should not seriously compromise
generalizeability of the results.

Clearly, the measurement of enjoyment of exercise needs
improvement; the ten-point Likert scale used in this study

resulted in a ceiling effect, with a great many subjects

85

86
scoring at the extreme end of the scale. A measure with
greater variability would most likely have offered a more
psychometrically sound assessment of enjoyment, permitting
much greater confidence in the findings.
Discussion of Rgsultg

Relationships Between Personality Factors

The intercorrelations between the personality variables
offers some interesting information regarding the
relationships among compulsiveness, impulsiveness, rigidity,
and tolerance of ambiguity. The most surprising finding was
the lack of any relationship between compulsiveness and
impulsiveness; a negative relationship was hypothesized but
not found (Hypotheses #1). This finding implies that
compulsiveness and impulsiveness are orthogonal to each
other, that they are tapping different psychological
processes with little or nothing in common.

Thus, the theory that impulsiveness, reflectiveness,
and compulsiveness form a continuum of cognitive style is
not supported by these data, at least if the continuum is
linear. Since this study fails to provide any empirical
support for the continuum theory itself (and the literature
provides no experimental support), there is no foundation
for operationalizing psychological adaptability as a
continuum of impulsiveness, reflectiveness, and
compulsiveness. Therefore, the theoretical basis of this
study has not been empirically supported.

The strong negative correlation between compulsiveness

87
and tolerance of ambiguity does, however, support the view
that compulsiveness is both cognitively and emotionally
limiting. This conclusion is based on the perspective that
high tolerance of ambiguity is a psychologically healthy and
constructive trait, and supports an ability to function well
in unfamiliar and uncertain circumstances.

The positive relationship between compulsiveness and
rigidity (Hypothesis #2) provides only moderate support for
the literature that suggested rigidity and compulsiveness
are essentially synonymous; a correlation of .324 accounts
for only 10% of the variance in compulsiveness. However,
25% of the variance in compulsiveness is accounted for by
tolerance of ambiguity, and thus 35% of the variability in
compulsiveness can be explained by a combination of
tolerance of ambiguity and rigidity. Also, the negative
relationship between tolerance of ambiguity and rigidity
(Hypothesis #4) support the conclusion that these three
factors are integrally related.

The major surprise in these results was the lack of
relationship between impulsiveness and compulsiveness, as
mentioned earlier; however, impulsiveness and tolerance of
ambiguity also failed to show any relationship
(Hypotheses #3), again indicating that impulsiveness is
likely tapping a different psychological mechanism than
either compulsiveness or tolerance of ambiguity. The
negative correlation between impulsiveness and rigidity,

then, is somewhat difficult to explain. The best

88
possibility is that rigidity may be an element underlying
both impulsiveness and compulsiveness. Yet, impulsiveness
and compulsiveness themselves share no variance; rigidity
must operate differently with each. Unfortunately, I could
find no previous work that investigated the relationship
between impulsiveness and rigidity with which to compare
these results.

Another explanation is that the assessment instruments
used in this study were less than acceptable in construct
validity, and did not tap the relevant traits adequately.
Standardized measures are only the best estimates available
for measuring a psychological process, after all, and do not
directly represent that process. However, the strong
reliability and validity data included in the Methods
section suggest that this explanation is not a particularly
good one; all the instruments used appear strongly related
to the theoretical constructs they represent. Also, the
scores on the various assessment instruments in this study
appear reasonably similar to the normative data noted in the
Methods section, indicating that though this subject pool
has greater age variability than the undergraduate samples
in the standardization studies, their performance on the
variables of interest are similar to prior samples. Such
similarity strengthens the view that this sample in not
skewed in some systematic way that would threaten
generalizeability. Table 5.1 compares the means and

standard deviations of the psychological variables obtained

89
in this study to the normative data provided by the

standardization samples:

TABLE 5.1
Comparison of Norm vs. Current Subjects
on the Psychological Variables

 

 

 

Norm Data Current Subjectg
MEAD §2 M292 §2
OCS 11.32 4.13 10.96 4.22
BRT(male) 8.5 3.60 8.44 3.63
(female) 6.7 3.00 7.71 3.22
BISlO 43.24 15.11 45.88 14.76
AT20 10.51 3.32 9.39 3.98

 

 

 

 

Lastly, we should keep in mind that the correlations
are only moderate (.324 with compulsiveness, -.267 with
impulsiveness), and may simply not be large enough to permit
much interpretation.

Exercise Group Results and Influences

Pyschological Variables

Though disappointing, the theory that psychological
adaptability affects exercise habits is clearly not
supported, at least insofar as it was operationalized in
this study. Compulsiveness, rigidity, and tolerance of
ambiguity showed no relationships to frequency and duration
of exercise; impulsiveness showed the only statistically
significant influence on exercise group membership, though
it contributed only slightly to prediction.

There are several possible reasons for these

nonsignificant results. First, the psychological processes

90
of compulsiveness, rigidity, and tolerance of ambiguity may
simply have no connection to exercise choices and habits;
this would imply that Hardy and Rejeski’s (1989) conclusion
that moderate intensity exercise induces ambiguous affect
has no effect on how often or how long people exercise.
This is the most likely explanation, since there were not
even any trends that would encourage further investigation
into these factors. This does not suggest, however, that
affect itself is unrelated to exercise; rather, it is
ambiguity that appears irrelevant.

Second, but less likely, is the explanation suggested
above that the assessment measures were inadequate to the
task of identifying the psychological processes being
examined. Although the possibility cannot be ignored that
other measures may tap other aspects of compulsiveness,
impulsiveness, rigidity, or tolerance of ambiguity that are
more directly relevant to exercise, past empirical studies
have shown the instruments employed in this study to be
reliable and valid measures.

Another possible but unlikely explanation for the lack
of significant results is that the sample size may have been
insufficient to identify very subtle effects. Though the
non-exercise group had only 44 subjects, the smallest group
of exercisers were the 92 general exercisers. The power
analysis discussed in the methods chapter concluded that a
small effect size of .30 had a 40% chance of being

discovered even for the those analyses including the

91
smallest group, and a 65% chance for analyses limited to the
exercise groups (Cohen, 1977). This may not seem very
powerful; however, even slightly higher effect sizes
increase power considerably (effect size of .40: n of 44, =
power of 59%; n of 92 = power of 85%). Though the sample
sizes in this study may not be sufficiently powerful to
identify very small effects, it should be adequate to
identify effects that are of a size to be useful.
Therefore, insufficient power is not a likely explanation
for the nonsignificant results in this study.

Lastly, a methodological explanation for these results
would suggest that the criteria by which exercise groups
were categorized failed to accurately identify intermittent
and general exercisers. Certainly the inability to
distinguish between intermittent exercisers and general
exercisers on any variable was disappointing, especially
given the initial rationale that intermittent and general
exercisers should differ on a variety of factors. The
categorization criteria used in this study resulted in three
exercise groups rather than four: non-exercisers, super-
exercisers, and a confusing mix of intermediate exercisers
whose exercise schedules varied widely but did not comprise
identifiably separate categories. Although subsequent
attempts were made to differentiate intermittent from
general exercisers by using more stringent, exclusive
selection criteria, such attempts were ineffective, and

results remained essentially the same as for the original

92
categorization. Thus, the decision criteria for assignment
to exercise groups may have mixed together several types of
intermittent and general exercisers that might be
distinguishable by other means, resulting in an ambiguous
and indistinct category composed of unidentified subgroups.

This inability to distinguish intermittent exercisers
from regular, moderate frequency exercisers also implies
that many general exercisers, over extended periods of time,
may actually be intermittent exercisers, and that this study
simply failed to identify the interruptions to the exercise
regimen. A meaningful number of general exercisers,
however, indicated that there had been no interruptions to
their exercise programs over many years, and these subjects
also could not be distinguished from the obviously
intermittent group.

The most likely explanation for the lack of significant
results, therefore, is the one offered first: that general
and intermittent exercisers simply do not differ on the
variables investigated here, and the theory that exercise-
induced affect influences these two groups differently is
simply incorrect. Thus, this study does not improve our
ability to prospectively identify those who struggle with
maintaining a regular exercise program, nor contribute
directly to intervention efforts aimed at facilitating

greater adherence among intermittent exercisers.

93
Other Variables

The predictive analyses found no relationships between
group membership and the control variables of age, sex,
support, exercise format, or exercise type. These results
are in conflict with previous work that identified spousal
and social support as significantly related to adherence
(Heinzelmann & Bagley, 1970); previous work has failed to
identify any age, sex, or exercise type effects, so these
results are not in conflict with past work in these areas.

A possible explanation for this discrepancy in findings
is that measurement of support in this study probably
differs from previous work, in that many subjects were
single, and among those who were partnered the overwhelming
majority identified their partners as supportive of their
exercise efforts. Heinzelman and Bagley’s (1970) subjects
were almost all married, and the investigators made the
assessment rather than subject self-report; their method of
assessing support may have generated different results than
in this study.

As stated in the results chapter, perceived enjoyment
of exercise was the major factor that distinguished among
non—exercisers, intermittent and general exercisers, and
super-exercisers. By itself, enjoyment was able to
accurately predict group membership for every analysis
except the analysis between intermittent and general
exercisers. Non-exercisers and super-exercisers were the

most distinct on enjoyment, with intermittent and general

94
exercisers basically indistinguishable in their levels of
exercise enjoyment. Considering how strong a predictor
enjoyment appears to be, more theoretical and research
attention is clearly warranted on the importance of
enjoyment as a motivator. This is particularly true since
the measurement of enjoyment in this study was quite crude
(a 10-point Likert scale); future research employing a more
sophisticated measure of enjoyment could make a much greater
contribution to our understanding of the importance of
exercise enjoyment.

Interest in enjoyment comes primarily from the sport
and participation motivation literatures, with only one
study found from exercise psychology prior to 1989 that
recognized enjoyment as a relevant factor in exercise
involvement (Wankel, 1985). Enjoyment as it impacts
participation in sports has begun to generate research
attention, particularly as a component of a sport commitment
model (Scanlan & Lewthwaite, 1986). Most of the work
carried out to date on sport enjoyment has focused primarily
on the sources of enjoyment, including both intrinsic and
extrinsic factors. Earlier theorizing viewed sport
enjoyment as a consequence of participating in intrinsically
motivating activities (Csikszentmihalyi, 1975; Deci & Ryan,
1985). However, Scanlan and Simons (1992) make a strong
case that ". . . sources of enjoyment in sport span a
broader spectrum than is subsumed under the concept of

intrinsic motivation" (p. 205), pointing out that

95
Csikszentmihalyi’s (1975) list of reasons why soccer,
hockey, and basketball players enjoyed their sport
emphasized extrinsic over intrinsic reasons. Thus, Scanlan
and Simons (1992) define sport enjoyment ". . . as a
positive affective response to the sport experience that
reflects generalized feelings such as pleasure, liking, and
fun" (pp. 202-203), inclusive of both extrinsic and
intrinsic factors, and being composed of cognitive and
affective elements.

As mentioned, most of the work on enjoyment has focused
on enjoyment of sport activities; the sources of sport
enjoyment have begun to be identified, and enjoyment has
been included as one factor contributing to sport commitment
(Scanlan & Lewthwaite, 1986). However, the findings on
sport enjoyment may not generalize completely when the focus
shifts to exercise enjoyment. An important difference is
that most exercise is likely a considerably narrower
experience than most sport activities. Of
Csikszentmihalyi’s (1975) ". . . eight relatively

independent reasons for enjoying sport activities

competition/measuring self against others,
- development of personal skills,

- friendships/companionship,

- activity itself,

- enjoyment of the experience/use of skills,
- measuring self against own ideals,

- prestige/reward/glamour, and

96
- emotional release" (Scanlan & Simons, 1992, p. 205), which
were ranked in order of importance to enjoying the sport
experience, several would seem not to apply to the exercise
experience, including the two top reasons and the prestige
factor. Exercise generally (though certainly not always)
lacks a competitive aspect, which helps define the sport
experience, and also more often lacks the camaraderie and
many of the opportunities for friendship found in team
sports.

Thus, the sources of exercise enjoyment are quite
likely more circumscribed than those of sport enjoyment, and
also seem likely to depend more on intrinsic factors, such
as enjoyment of the experience itself, measuring oneself
against one’s own ideals, and emotional release. The sport
enjoyment construct views the intrinsic, experiential source
of enjoyment as only one among many, and not necessarily
primary; for exercise enjoyment, the intrinsic factors
mentioned above all seem based to some degree on the
exerciser’s sense of "flow", or "the state in which people
are so involved in an activity that nothing else seems to
matter; the experience itself is so enjoyable that people
will do it even at great cost, for the sheer sake or doing
it" (Csikszentmihalyi, 1990, p. 4).

Although extrinsic reasons for engaging in exercise
certainly exist and contribute to exercise enjoyment, it
seems reasonable that the principal foundations of exercise

enjoyment are primarily internal and based in the

97

exerciser’s subjective experience of the exercise activity.
The exploratory work on the impact of affective factors in
exercise supports this line of reasoning. Although
enjoyment, per se, is not directly addressed in the exercise
literature, research has begun to focus on positive versus
negative psychological affect during and after exercise, at
varying workloads, and on developing assessment instruments
for the measurement of exercise related affect and enjoyment
(Gauvin & Rejeski, 1993; Hardy & Rejeski, 1989; Kendzierski
& DeCarlo, 1991; McAuley & Courneya, 1994; Parfitt, Markland
& Holmes, 1994). The results of the current study on the
influence of enjoyment on how often and how long people
exercise strongly supports this research direction.
Igplicationgifor Prgcticg

The more interesting results of this study are the
findings that super-exercisers are significantly less
impulsive than all other exercisers and non-exercisers, and
that no relationship existed between impulsiveness and
compulsiveness. This challenges the common View that super—
exercisers are highly compulsive folks, and that higher
compulsiveness among exercisers is linearly related to
higher rates of exercise. Instead, impulsiveness appears to
have a significant influence on behavior, with exercisers
showing no differences on compulsiveness regardless of rate
of exercise.

This finding suggests that interventions aimed at

helping people return to exercising after interruption may

98
be useful; to paraphrase the non-smoking slogan "Don’t quit
quitting", we might suggest "Don’t quit starting" to
relapse—prone exercisers. Although no interventions to date
specifically address impulsiveness as a cause of exercise
dropout, that may be a constructive direction to explore.

Perhaps the most meaningful result is the recognition
that the exerciser’s perception of enjoyment is a strong
influence on whether that person will continue exercising.
This finding implies that leaders of group exercise programs
should encourage enjoyment of the activity as much as
possible, to help participants interpret anything ambiguous
as enjoyable, and to work at helping exercisers re-interpret
mildly negative reactions as neutral or positive.

Although enjoyment and impulsiveness have shown a good
ability to distinguish between supers and non-exercisers,
research has yet to determine whether that is because of
inherent differences between the two groups, or whether
enjoyment increases as duration increases; the possibility
exists that super-exercisers have learned to enjoy exercise.
This latter explanation, that enjoyment may be a matter of
experience and/or training, appears most likely; Parfitt, et
al (1994) investigated psychological responses to low and
high workloads by highly active and less active exercisers,
and found that high-active exercisers reported significantly
more positive psychological affect in the high workload
condition than less active subjects, with no differences in

the low workload condition; "At a theoretical level this

99

suggests that high-active individuals may have attenuated
their distress schema, whereas the low—active individuals
have not" (p. 184). Enjoyment of exercise may well be
learned; thus, those interested in helping facilitate
exercise adherence would likely do well to encourage
moderate workloads for new exercisers, in an effort to keep
them engaged long enough to learn to enjoy it.
Limitations

There are several notable limitations to this study.
First, all results should be cautiously interpreted given
the subject self-selection discussed earlier. Since there
is no way to determine whether those who chose to
participate in this study differ in any systematic way from
those who did not, the assumption that these results would
generalize to another exercising sample is only tentative.

The violation of statistical assumptions in the
measurement of enjoyment, though moderated as previously
discussed in the Results chapter, poses a threat to the
reliability of these results, and the need for a more valid
assessment of exercise enjoyment is critical. Results need
to be interpreted cautiously, especially since correlational
studies cannot offer causal inferences.

Lastly, the use of predictive discriminant analysis
(PDA) generates a classification rule consisting of those
linear combinations of the predictor variables that result
in the greatest differentiation of the dependent groups.

The ideal method is the generation of an external rule,

100
"where the cases used to generate the classification table
are different from those used to build the classification
rule" (Huberty & Barton, 1989, p. 161). The current study,
however, employed an internal analysis, where the
classification rule was built using the same subjects upon
which the classification table was obtained. The problem
with internal analyses is that they capitalize on sample—
specific variation that may not generalize to other samples,
and therefore the predictive accuracy of the analysis may be
positively biased. Although not ideal, "an internal
analysis may be acceptable if the total number of cases is
large; large is defined by the current authors as a data set
in which the smallest group size is five times the number of
predictor variables" (Huberty, et al, 1989, p. 161). The
smallest group size in this study was the non-exerciser
group, with 44 subjects, and seven predictors were the most
employed in any one analysis. Thus, use of an internal
analysis should not taint the results, though cross—
validation on a different sample would have strengthened
these findings.

Directions for Future Research

 

Further research on the impact of enjoyment on exercise
adherence is clearly indicated, as well as the larger domain
of the affective experience of exercise, of which enjoyment
is one element. Though infrequent, the calls for attention
to the influence of affect on both sport involvement and

exercise participation continue to surface. As noted by

101
Scanlan and Simons (1992), Harter (1981) suggested well over
ten years ago that "we should resurrect ’joy’ as a
legitimate construct and restore affect and emotion to its
rightful place, as central to an understanding of behavior"
(p. 4). Several years later, Wankel (1985) noted that "If
progress is to be made in making exercise a more enjoyable
experience . . . information is needed on how participants’
goals and program factors are related to the quality of the
experience (i.e., enjoyment)" (p. 276). Scanlan and Simons
(1992) continue to note a need to "bring positive affect to
light . . . on [sport] motivation, and add a little ’heart’
to the otherwise highly cognitive treatment of the topic"
(p. 199), and others also continue to note that "One
relatively unexplored factor is the extent to which an
individual enjoys doing the physical activitiy involved in
his or her exercise program" (Kendzierski & DeCarlo, 1991,
p. 50). Certainly the results discussed above support
further investigation in this direction.

A broader approach to the understanding of enjoyment
could also be useful; one promising line of research is the
neurophysiology of pleasure. The exercise field has
suggested the "runner’s high" hypothesis that physical
exertion of a sufficient degree releases endorphins, the
body’s natural narcotics. Another useful and fascinating
direction might be to examine the neurophysiological
pathways by which substance abuse and addiction impacts and,

depending on the substance (i.e., cocaine), impairs the

102

ability to experience pleasure and enjoyment. Doidge (1990)

proposes a rather fascinating connection between

psychoanalytic theory and the neurochemistry responsible for

pleasureable reactions to cocaine, by suggesting that an

analytic theory of appetitive versus consummatory pleasures

is congruent with the neurochemistry associated with cocaine

use .

He summarizes that

Klein’s (1987) theory of appetitive and consummatory
pleasures, briefly stated, asserts that there are at
least two broad, independent pleasure systems. One
regulates appetitive pleasure and in normal situations
acts to boost energy for the pleasureable pursuit of a
goal. A second system of consummatory pleasure
underlines the pleasureable enjoyment of food, sex, and
other interests already obtained. Each system has its
own underlying neurochemical basis. . . . The
appetitive pleasure system, once turned on, gives rise
to a nonsatiating pleasure. This state is a desireable
state of desiring, triggered by a sense of imminent
satisfaction . . . It is quite different from the
yearning kind of desire that is based upon an awareness
of a lack of something. (pp. 146-148)

Doidge (1990) then describes the phenomenology of a

cocaine high and ties the above theoretical model of

pleasure to the neurochemistry of cocaine, and concludes

that

This incentive system [the neurochemical response to

103

cocaine ingestion] has important similarities with

Klein’s appetitive pleasure system. . . . In summary,

the work on cocaine and the work of Klein suggest that

exogenous stimulants have the effect of turning on an
internal appetitive pleasure system for placing the
organism in high gear, increasing energy, and pleasures

of pursuit, and apparently, self-esteem (pp. 152-153)

Investigating the neurochemical correlates of exercise
may be a highly useful research direction, considering that
exercise has been described in much the same terms as above.
Such an increased knowledge of the neurophysiology of
pleasure could only contribute to attempts to understand the
pleasure gained from physical activity and exercise,
especially in view of this study’s findings that enjoyment,
or pleasure, plays a pivotal role in how much and how long
people exercise.

Further research on enjoyment will, of course, require
an assessment instrument with much better psychometric
properties than the 10-point Likert scale used in this
study. The affective realm and its impact on exercise
participation has captured recent research attention, as
evidenced by a recent focus on developing reliable and valid
assessment instruments for several aspects of the affective
experience of exercise. Kendzierski and DeCarlo (1991) have
developed perhaps the most directly applicable scale to the
study of enjoyment, with their Physical Activity Enjoyment

Scale (PACES), an 18-item scale consisting of 7-point

104
bipolar scales assessing a variety of affective experiences
during physical activity. Validation studies during the
scale’s development found a significant negative correlation
between subjects’ PACES scores and scores on a boredom
proneness scale. Thus, those who reported being prone to
boredom were significantly less likely to enjoy physical
activity. The PACES appears able to "provide the necessary
tool for examining the relationship between enjoyment and
exercise adherence, as well as for identifying variables
that affect individuals’ enjoyment of both exercise and
sport" (pp. 59—60).

Two other promising new assessment tools have recently
been developed as well. The Subjective Exercise Experiences
Scale (SEES) is a 12-item scale assessing three factors of
Positive Well-Being, Psychological Distress, and Fatigue
(McAuley & Courneya, 1994). The Exercise-Induced Feeling
Inventory (EFI) is also a 12-item scale, measuring Positive
Engagement, Revitalization, Tranquillity, and Physical
Exhaustion (Gauvin & Rejeski, 1993).

Both scales offer advantages depending on the
particular interests of the potential study. The
availability of psychometrically strong assessment tools is
encouraging, and promises at least the possibility that
affect and enjoyment will be further explored.

Related to improved assessment is the question of how
to address the problem encountered in this study of how to

more accurately classify exercise patterns. A promising

105
approach would be to analyze the data with a statistical
technique capable of utilizing continuous data; the
categorization of subjects into exercise groups resulted in
the loss of considerable information, since the decision
criteria were based on cutoff points of continuous scores
rather than easily identified nominal groups (i.e., gender,
race, etc.). One such statistical approach would be the use
of canonical correlation, which is, in essence, similar to
multiple regression with more than one dependent variable.
Canonical correlation could use the psychological variable
scores to predict exercise frequency and duration, with the
advantage that all the variability inherent in continuous
data would be preserved and used in the statistical
computations. Preservation of all available variability
makes this a more sophisticated statistical analysis;
canonical correlation represents, therefore, the next
logical and appropriate research procedure for this type of
data, and would increase the confidence we could place in
the results.

Lastly, the finding of low impulsiveness in the super-
exerciser group compared to other exercise groups deserves
further study. Are super-exercisers low on impulsiveness
from the beginning, before they ever begin an exercise
program? Would current non-exercisers who score low on
impulsiveness be more likely to persevere once they began an
exercise program? Conversely, might super—exercisers have

learned to moderate those impulsive behaviors that

106
threatened their ability to continue exercising regularly?
If so, impulsiveness could not be considered an unchangeable
trait, but rather an aspect of personality and behavior
amenable to change. Longitudinal studies could be employed
to distinguish unchangeable personality traits from
personality changes associated with regular exercise. In
either case, our understanding of the factors influencing
exercise behavior would be improved, as would our ability to
identify potential dropouts and possibly aid those
interested in maintaining a steady exercise program.
SUMMARY

Enjoyment of exercise was identified as an important
influence on exercise behavior, and one that deserves
further study. Of the psychological factors investigated,
impulsiveness contributed slightly to the ability to
distinguish between non-exercisers and super-exercisers. No
other variables were significantly related to exercise group
membership.

The ideal goal of obtaining results that would
contribute to the development of interventions to promote
greater exercise adherence was not achieved. At the applied
level, the results suggest that enjoyment and impulsiveness
are worth the attention of those attempting to increase
adherence to exercise programs. More generally, this
research should be viewed as supporting the relatively
recent focus on the importance of psychological affect,

particularly enjoyment, on exercise experience and habits.

107

As a field study in a naturalistic setting, these
results should be quite strong in external validity.
Nothing was manipulated, and results should describe the
exercising population well. This study was interested in a
healthy exercising population, and therefore drew its sample
from that group. It will hopefully add to the research base
with non-medical subjects, and should permit credible
generalization to exercisers who voluntarily engage in

personally determined exercise activities.

APPENDICES

108

APPENDIX A

Variable Codes

Variable # Variable Name Description
1 Case Number Consecutively numbered cases

2 Source 1 = Private health clubs

2 = University community

3 = Recovering alcoholics

3 Group 1 = Non-exercisers
2 = Intermittent exercisers
3 = General exercisers
4 = Super exercisers
5 = Unclassifiable
4 Age Age in years
5 Sex 1 = Male
2 = Female
6 ENJOY 1 - 10: 1=Greatly enjoy
10=Greatly dislike
7 EXTYPE 1 = Aerobic exercise
2 = Anaerobic exercise
3 = Mixed
8 SUPPORT 1 = Unattached
2 = Supportive
3 = Neutral
4 = Non-supportive
5 = Not applicable
9 FORMAT 1 = Exercises alone
2 = Exercises with a partner
3 = Exercises in a group
4 = Not applicable
10 INTENSITY 1 - 10: 1=Very easy
10=Very hard
11 OCS 1—20: 1=Low compulsiveness
20=High compulsiveness
12 BRT 1-15: 1=Low rigidity
15=High rigidity
13 BIS10 0-136: 0=Low impulsiveness
136=High impulsiveness
14 AT20 1-20: 1: Low ambig. tolerance

20: High tolerance

109

APPENDIX B

Invitation to Participate Poster

RESEARCH STUDY

EXERCISE
AND

PERSONALITY

You are invited to participate in a study investigating the
influence of personality traits on exercise habits. The
survey takes about 20-30 minutes and is anonymous; a stamped
return envelope is provided. Results upon request. If

interested, please ask at the desk for a questionaire.
THANK YOU!

If you have any questions, please call:

Kay Ketzenberger, Doctoral Candidate
Counseling Psychology Program
Michigan State University

(517) 353—8112

110

APPENDIX C

Consent Form

August 11, 1993

Hello:

I am conducting a research study into the possibility that
exercisers and non-exercisers differ on several personality traits. It
is well-known that many people who begin exercise programs either drop
out within a short time or have great difficulty in maintaining regular
exercise habits. To better understand the relationship between personal
characteristics and exercise, information about the exercise habits of
many people is necessary, including those who do not exercise at all.
Your participation in this study would help greatly in understanding why
some people exercise more regularly than others.

The enclosed questionnaire contains a short demographic survey and
several brief psychological assessments. It should take you no longer
than 20-25 minutes to complete the survey. Be sure to complete both
sides of all double-sided pages. You have the right to refuse to answer
any question; however, the information provided by each answer is
important in gaining a better understanding of exercise. As a
participant, your answers will remain anonymous even from me as the
researcher, because names and addresses are pp; to be put on the
questionnaires.

Your participation in this study is entirely voluntary. You may
choose not to participate or to end your participation at any time.
Completion and return of the questionnaire will indicate you have read
this letter and your participation is voluntary.

The results of this research will be made available to
professionals in the area of sport and exercise psychology. and to the
fitness facilities and gyms who are interested in these results. Also,
you may receive a summary of the results by writing "copy of results
requested" on the back of the return envelope and printing your name and
address below. To insure your anonymity, please do not put this
information anywhere on the questionnaires themselves.

I will be glad to answer any questions you might have. Please
feel free to call me at (517) 353-8112.

Thank you very much for your help!

Sincerely,

Kay E. Ketzenberger
Doctoral Candidate, Counseling Psychology

The return of the completed survey reflects your
informed and voluntary consent to participate in this study.

111

APPENDIX D

Demographic Questionnaire

1. Age 2. Sex 3.0ccupation

 

 

 

4. Where did you obtain this questionnaire?
at my fitness club or gym

on the MSU campus

from a friend or acquaintence
other (specify)

 

 

you ever been treated for heart trouble? yes no
., been prescribed medication, a specific diet
or exercise program by a physician for a heart condition)

m
"m Ii |
m m
- <
010

6. On the following lo-point scale, please mark how much you enjoy
physical exercise:

 

1 2 3 4 5 6 7 8 9 10
great- moderate neutral moderate greatly
I love it enjoyment no comment dislike dislike

Current Exercise Habits

7. Are you currently exercising? yes no
(NOTE: Do not include physician-prescribed physical therapy)

IF YES. cogplete items 8 - 14. then skip to #18:
IF NO. skip to #15.

 

8. On average, how often do you exercise?

every day

5-6 days per week

2-4 days per week
once a week

less than once a week

 

9. How long have you exercised at this rate?

less than 3 months

3-6 months

6 months to 1 year

over 1 year
- how long?

 

10. What type of exercise activity do you usually engage in?
(i.e., running, walking, swimming, weight training, aerobics
classes, racquetball, etc.)

 

11. Support factors:

I am currently single/unattached

My partner is supportive of my exercise efforts
My partner is neutral toward my exercise efforts
My partner is non-supportive of my efforts

12.

13.

14.

112

Most often, I exercise alone
Most often, I exercise with an exercise partner
Most often, I exercise with a group of people

On the following 10-point scale, please mark the average
intensity of your workouts over the past week:

 

l 2 3 4 5 6 7 8 9 10
very fairly moderate fairly very
easy easy hard hard

Within the last 3 years, have you begun to exercise and then
either stopped exercising completely or dropped to once a
week or less?

yes no IF YES: Why?

 

IF YOU ARE NOT CURRENTLY EXERCISING:

15.

16.

17.

Have you ever begun exercising intending to exercise
regularly?

 

 

 

yes no

IF YES: How long did you exercise?
How long has it been since you stopped
exercising?

 

Within the last 3 years, have you exercised regularly (2 or
more times per week) for a period of 3 months or more, and
then stopped exercising entirely?

yes no

 

Within the last 3 years, have you exercised regularly (2 or
more times per week) for a period of 3 months or more, and
then dropped to exercising only infrequently (once a week or
less)?

yes no

 

 

IF YES to either of the two previous questions:
Why?

 

Past Sport Participation

18.

 

Are you now or have you ever participated in intercollegiate
sports?

yes no

 

IF YES: Which sport(s)?

 

113

IXPPEEH)IX.IE.1

PERSONAL EVALUATION - BIS 10

DIRECTIONS: People differ in the ways they act and think in different
situations. This is a test to measure some of the ways in which you act
and think. Read each statement and darken the appropriate circle on the

right side of the page.

Answer quickly and honestly.

w 01 w 04 w t: N b) N t: w h: N h) N +4 H +4 H +4 H +4 H +4 H
e 04 N +4 o (o m ~4 m Ln o 04 w +4 o u: m 41 m Ln e 04 w +4 o

\OIDQOlU'IANNH

HHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH

plan tasks carefully ............................... 0
do things without thinking ......................... 0
make up my mind quickly ............................ 0
am happy-go-lucky .................................. 0
don’t "pay attention" ..............................

have "racing" thoughts ............................. 0
plan trips well ahead of time ...................... 0
am self-controlled .................................

concentrate easily ................................. O
save regularly ..................................... O

"squirm" at plays or lectures ......................
am a careful thinker ..............................
plan for job security .............................

say things without thinking ........................ 0
like to think about complex problems ................ 0
change jobs ........................................
act "on impulse" ...................................
get easily bored when solving thought problems ..... 0
have regular health check ups ...................... 0

act on the spur of the moment ......................
am a steady thinker ................................ 0

change residences ..................................

buy things on impulse .............................. O
can only think about one problem at a time ......... 0
change hobbies .....................................
walk and move fast .................................
solve problems by trial-and-error .................. O
spend or charge more than I earn ................... O
talk fast ..........................................
often have extraneous thoughts when thinking ........ 0

am more interested in the present than the future

am restless at the theater or lectures .............
like puzzles ....................................... 0

am future oriented .................................

C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) 0

Do not spend too much time on any statement.

(3 O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O

O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C) O C)

114

APPENDIX E.2

OCS

DIRECTIONS: Mark T for true and F for false. Do not spend
too much time on an item. There are no right or wrong

answers,

H

11.

12.

l3.
14.
15.
16.

17.
18.

19.
20.
21.
22.

and so your first response is important.

I feel compelled to do things I don’t want to do.
I usually check things that I know I have already
done.

I can walk 30 miles an hour.

I often do things I don’t want to do because I can
not resist doing them.

I seldom keep a daily routine.

I feel compelled to always complete what I am
doing.

I often feel the need to double check what I do.
I’d rather do things the same way all the time.

I seldom have recurring thoughts.

I seldom am compelled to do something I don’t want
to do.

I don’t feel uncomfortable and uneasy when I don’t
do things my usual way.

If I don’t feel like doing something it won’t
bother me not to do it.

I usually never feel the need to be organized.

I am uneasy about keeping a rigid time schedule.
My birthday comes once a year.

I am often compelled to do some things I do not
want to do.

I like to keep a rigid daily routine.

I believe there is a place for everything and
everything in its place.

I seldom check things I know I have already done.
I am not obsessed with details.

I often have recurring thoughts.

I

like to do things differently each time.

115

APPENDIX E.3

THE AT-go SCALE

DIRECTIONS: Please do not spend too much time on the
following items. There are no right or wrong answers and
therefore your first response is important. Mark T for true
and F for false. Be sure to answer every question.

10.

ll.

12.

A problem has little attraction for me if I don’t
think it has a solution.

I am just a little uncomfortable with people
unless I feel that I can understand their
behavior.

There’s a right way and a wrong way to do almost
everything.

I would rather bet 1 to 6 on a long shot than 3 to
1 on a probably winner.

The way to understand complex problems is to be
concerned with their larger aspects instead of
breaking them into smaller pieces.

I get pretty anxious when I’m in a social
situation over which I have no control.

Practically every problem has a solution.

It bothers me when I am unable to follow another
person’s train of thought.

I have always felt that there is a clear
difference between right and wrong.

It bothers me when I don’t know how other people
react to me.

Nothing gets accomplished in this world unless you
stick to some basic rules.

If I were a doctor, I would prefer the
uncertainties of a psychiatrist to the clear and
definite work of someone like a surgeon or X—ray
specialist.

13.

14.

15.

16.

17.

18.

19.

20.

116

Vague and impressionistic pictures really have
little appeal for me.

If I were a scientist, it would bother me that my
work would never be completed (because science
will always make new discoveries).

Before an examination, I feel much less anxious if
I know how many questions there will be.

The best part of working a jigsaw puzzle is
putting in that last piece.

Sometimes I rather enjoy going against the rules
and doing things I’m not supposed to do.

I don’t like to work on a problem unless there is
a possibility of coming out with a clear-cut and
unambiguous answer.

I like to fool around with new ideas, even if they
turn out later to be a total waste of time.

Perfect balance is the essence of all good
composition.

117

Swmnfiozfi new $9 Mama.
cwwnm w Boar on n70 HHDo
szn «0 ~50 unmade ndmn
we: Soon cannon. Hamanmw
«so humane noun woo Huwm
n70 gown on are n80 an
own: Hnms.

‘13

‘11
O

 

>wwmzc HM m . b

wwmwaz HMmH

IVA
IO

 

 

 

 

.. w p.+

 

Hr

Hm.

H+

DIOR

 

> ”T“
v ——L—.

 

_..__J
~——~c

LIST OF REFERENCES

118

REFERENCES

Andrew, G. M., Oldridge, N. B., Parker, J. O., Cunningham,
D. A., Rechnitzer, P. A., Jones, N. L., Buck, C.,
Kavanagh, T., Shephard, R. J., Sutton, J. R., &
McDonald, W. (1981). Reasons for dropout from
programs in post coronary patients. Medicine and
Science in Sports and Exercise, i3, 164-168.

Andrew, G. M. & Parker, J. O. (1979). Factors related to
dropout of post myocardial infarction patients from
exercise programs. Medicine and Science in Sports and
Exercise, ;1, 376-378.

 

Bandura, A. (1977). Self-efficacy: Toward a unifying
theory of behavior change. Psychological Review, pg,
191-215.

Bandura, A. (1986). Social foundations of thought and

 

action: A social cognitive theory. Englewood, Cliffs,
NJ: Prentice Hall.

Barratt, E. S. (1965). Factor analysis of some
psychometric measures of impulsiveness and anxiety.

Psychological Reports, i5, 547-554.

Barratt, E. S. (1972). Anxiety and impulsiveness: Toward
a neuropsychological model. In: C. D. Spielberger
(Ed.), Anxiety: Current trends in theory and research
(Vol. 1, pp. 195-222). New York: Academic Press.

Barratt, E. S. (1983). The biological basis of
impulsiveness: The significance of timing and rhythm

disorders. Personality and Individual Differences, 4,

 

 

387-391.

Barratt, E. S. (1985a). Impulsiveness defined within a
system model of personality. In: C. D. Spielberger &
J. N. Butcher (Eds.), Advances in personality
assessment (Vol. 5, pp. 113-132). New York: Earlbaum.

Barratt, E. S. (1985b). Impulsiveness subtraits: Arousal
and information processing. In: J. T. Spence & C. E.
Itard (Eds.), Motivation. emotional and personality,

 

(pp. 137-146). North Holland: Elsevier.

119

Barratt, E. S. (1987). Impulsiveness and anxiety:
Information processing and elctroencephalograph

topography. Journal of Research in Personality, 22,
453-463.

Barratt, E. S. (1991). Measuring and predicting aggression
within the context of a personality theory. Journal of

Neuropsychiatry, 2, 535-539.

 

 

Barratt, E. S. (in press). Impulsivity: Integrating
cognitive, behavioral, biological and environmental
data. In: W. McCown & M. Shure (Eds.), The impulsive
client: Theory. research. and treatment. Washington,
D.C.: American Psychological Association.

Barratt, E. S. & Patton, J. H. (1983). Impulsivity:

Cognitive, behavioral and psychophysiological
correlates. In: M. Zuckerman (Ed.), Biological bases

of sensation seeking, impulsivity and anxiety (pp. 77-
122). New York: Earlbaum.

Barratt, E. S., Patton, J. H., Olsson, N. G. & Zuker, G.
(1981). Impulsivity and paced tapping. Journal of
Motor Behavior, 22, 286-300.

Berger, B. G. (1984). Running toward psychological well-
being. In: M. L. Sachs & G. W. Buffone (Eds.),
Running as therapy: An integrated approach (pp. 172-

197). Lincoln, Nebraska: University of Nebraska
Press.
Berger, B. G. & Mackenzie, M. M. (1980). A case study of a

woman jogger: A psychodynamic analysis. Journal of
Sport Behavior. 2, 3-16.

Barrett, D. E. (1977). Reflection-impulsivity as a
predictor of children’s academic achievement. Child
Development, 42, 1443-1447.

Block, J., Block, J. H., & Harrington, D. M. (1974). Some
misgivings about the Matching Familiar Figures test as
a measure of reflection—impulsivity. Developmental

Psychology, 22, 611—632.

Bray, J. H. & Maxwell, S. E. (1982). Analyzing and
interpreting significant MANOVAs. Review of
Educational Research, 22, 340-367.

Breskin, S. (1968). Measurement of rigidity, a non-verbal
test. Perceptual and Motor Skills, 21, 1203-1206.

Breskin, S. & Gorman, B. S. (1969). On rigidity and field-
dependence. Perceptual and Motor Skills, 22, 541—542.

120

Breskin, S. & Rich, K. D. (1971). Correlation of non-
verbal rigidity and intelligence in elementary school
children. Perceptual and Motor Skills, 22, 194.

Bruner, J. (1948). Perceptual theory and the Rorschach
test. Journal of Personality, 11, 157-168.

Campbell, 8. B. & Douglas, V. I. (1972). Cognitive styles
and responses to the threat of frustration. Canadian
Journal of Behavioral Science, 4, 30-42.

Cassidy, T. & Lynn, R. (1989). A multifactorial approach
to achievement motivation: The development of a
comprehensive measure. Journal of Occupational
Psychology, 52. 301-312.

Cattell, R. B. (1935). On the measurement of
perseveration. British Journal of Educational
Psychology, 2, 76—92.

Chhaya, S. (1985). Rigidity-flexibility and sex in
relation to Zeigarnik effect. Indian Psychological
Review, 22, 14-16.

Cohen, J. (1977). Statistical power analysis for the
behavioral sciences, Revised Edition. New York:
Academic Press.

Cooper, K. H., Pollack, M., Marin, R., White, S., Linnerud,
A., & Jackson, A. (1976). Physical fitness levels vs
selected coronary risk factors. Journal of the
American Medical Association, 22p, 166-169.

Corbin, C. B., Nielsen, A. B., Borsdorf, L. L. & Laurie, D.
R. (1987). Commitment to physical activity.
International Journal of Sport Psychology, 18, 215—222.

 

Covey, L. A. & Feltz, D. L. (1991). Physical activity and
adolescent female psychological development. Journal
of Youth and Adolescence, 22, 463-474.

Csikszentmihalyi, M. (1975). Beyond boredom and anxiety.
San Francisco: Josey-Bass.

Csikszentmihalyi, M. (1990). Flow: The psychology of
optimal experience. New York: Harper Collins
Publishers, Inc.

Deeter, T. E. (1988). Does attitudinal commitment predict
physical activity participation? Journal of Sport
Behavior, 2;, 177-192.

121

Desharnais, R., Bouillon, J., & Godin, G. (1986). Self-
efficacy and outcome expectations as determinants of
exercise adherence. Psychological Reports, 52, 1155-
1159.

Dillon, W. R. & Goldstein, M. (1984). Multivariate
analysis: Methods and applications. New York: John
Wiley & Sons.

Dishman, R. K. (1978). Aerobic power, estimation of
physical ability, and attraction to physical activity.

Research Quarterly for Exercisg and Sport, 42, 285-292.

Dishman, R. K. (1981). Biologic influences on exercise

adherence. Research Quarterly for Exgrgise and Sporg,
52. 143-159.

Dishman, R. K. (1982). Compliance/adherence in health-
related exercise. Health Psychology, 4, 237-267.

Dishman, R. K. (1984). Motivation and exercise adherence.
In J. M. Silva & R. S. Weinberg (Eds.), Psychological
foundations of sport. Champaign, Illinois: Human
Kinetics Publishers, Inc.

Dishman, R. K. (1986). Exercise compliance: A new view
for public health. The Physician and Sportsmedicine,
14. 127-145.

Dishman, R. K. & Gettman, L. R. (1980). Psychobiologic
influences on exercise adherence. Journal of Sport

Psychology, 2, 295-310.

Dishman, R. K. & Ickes, W. (1981). Self-motivation and
adherence to therapeutic exercise. Journal of
Behavioral Medicine, 4, 421-438.

Dishman, R. K., Ickes, W. & Morgan, W. P. (1980). Self-

motivation and adherence to habitual physical activity.
Journal of Applied Social Psychology, 44, 115-132.

Dishman, R. K., Sallis, J. F., & Orenstein, D. R. (1985).
The determinants of physical activity and exercise.
Public Health Reports, 100, 158-171.

Doidge, N. (1990). Appetitive pleasure states: A
biopsychoanalytic model of the pleasure threshold,
mental representation, and defense. In R. A. Glick &

S. Bone (Eds.). Pleasure beyond the pleasure principle
(pp. 138-173). New Haven: Yale University Press.

122

Fahlberg, L. L., Fahlberg, L. A., & Gates, W. K. (1992).
Exercise and existence: Exercise behavior from an
existential-phenomenological perspective. The Sport
Psychologist, 6, 172-191.

Fielding, J. E. (1982). Effectiveness of employee health

improvement programs. Journal of Occupational
Medicine, 24, 907-916.

Fisher, J. (1951). The memory process and certain psycho-
social attitudes with special reference to the law of
Pragnanz. Journal of Personality, 12, 41-47.

 

Folkins, C. H. & Sime, W. E. (1981). Physical fitness
training and mental health. American Psychologist, 24,
373-389.

Fox, K. R., Corbin, C. B., & Couldry, W. H. (1985). Female

physical estimation and attraction to physical
activity. Journal of Sport Psychology, 1, 125-136.

Franklin, B. A. (1988). Program factors that influence
exercise adherence: Practical adherence skills for the
clinical staff. In R. K. Dishman (Ed.). Exercise

adherence: Its impact on public health (pp. 237-258).
Champaign, IL: Human Kinetics Publishers, Inc.

 

Frenkel-Brunswik, E. (1948a). Tolerance towards ambiguity
as a personality variable. American Psychologist, 2,
268.

Frenkel-Brunswik, E. (1948b). A study of prejudice in

children. Human Relations, 4, 295-306.

Frenkel-Brunswik, E. (1949). Intolerance to ambiguity as
an emotional and perceptual personality variable.
Journal of Personality, 22, 108-143.

Gale, J. B., Eckhoff, W. T., Mogel, S. F., & Rodnick, J. E.
(1984). Factors related to adherence to an exercise
program for healthy adults. Medicine and Science in

Sports and Exercise, 24, 544-549.

Garcia, A. W. & King, A. C. (1991). Predicting long-term
adherence to aerobic exercise: A comparison of two

models. Journal of Sport & Exercise Psychology, 42,
394-410.

Gauvin, L. & Rejeski, W. J. (1993). The Exercise-Induced
Feeling Inventory: Development and initial validation.
Journal of Sport and Exercise Psychology. 15, 403-423.

123

Gibb, G. D., Bailey, J. R., Best, R. H., & Lambirth, T. T.
(1983). The measurement of the obsessive-compulsive

personality. Educational and Psychological
Measurement, 42, 1233-1238.

Gorman, B. S. & Breskin, S. (1969). Non-verbal rigidity,
creativity, and problem solving. Perceptual and Motor
Skills, 22, 715-718.

Hardy, C. J. & Rejeski, W. J. (1989). Not what, but how
one feels: The measurement of affect during exercise.
Journal of Sport & Exercise Psychology, 22, 304-317.

Harter, S. (1981). The development of competence
motivation in the mastery of cognitive and physical
skills: Is there a place for joy? In G. C. Roberts

 

 

and D. M. Landers (Eds.). Psychology of motor behavior
and sport - 1980 (pp. 3-29). Champaign, IL: Human
Kinetics.

Haskell, W. L. (1984). Overview: Health benefits of
exercise. In J. D. Matarazzo, S. M. Weiss, J. A. Herd,
N. E. Miller, & S. M. Weiss (Eds.), Behavioral health:

A handbook of health enhancement and disease
prevention. New York: John Wiley & Sons.

Heiby, E. M., Onorato, V. A., & Sato, R. A. (1987). Cross—
validation of the Self-Motivation Inventory. Journal
of Sport Psychology. 2. 394-399.

Heinzelmann, F. & Bagley, R. W. (1970). Response to
physical activity programs and their effects on health
behavior. Public Health Reports, 22, 905-911.

Huberty, C. J. (1984). Issues in the use and
interpretation of discriminant analysis. Psychological
Bulletin, 22, 156-171.

Huberty, C. J. & Barton, R. M. (1989). Methods, plainly
speaking: An introduction to discriminant analysis.
Measurement and Evaluation in Counseling and
Development, 22, 158-168.

Jennings, P. S., Holmstrom, R. W., and Karp, S. A. (1986).
Personality correlates of reflectivity. Psychological
Reports, 22, 87-94.

Kagan, J., Rosman, B. L., Day, D., Albert, J. & Phillips, W.
(1964). Information processing in the child:
Significance of analytic and reflective attitudes.
Psychological Monographs, 12, (1, Whole No. 578).

124

 

Kendzierski, D. (1988). Self-schemata and exercise. Basic
and Applied Social Psychology, 2, 45-59.

Kendzierski, D. (1990). Exercise self-schemata: Cognitive
and behavioral correlates. Health Psychology, 2, 69-
82.

Kendzierski, D. (1993). Excuses, excuses, excuses: A

cognitive behavioral approach to exercise

implementation. Journal of Sport and Exercise
Psychology , 22, 207-219.

Kendzierski, D. & DeCarlo, K. J. (1991). Physical Activity
Enjoyment Scale: Two validation studies. Journal of

Sport and Exercise Psychology, 22, 50-64.

Knapp, D. N. (1988). Behavioral management techniques and
exercise promotion. In R. K. Dishman (Ed.), Exercise

adherence: Its impact on public health (pp. 203-236).
Champaign, IL: Human Kinetics Books.

Kostrubala, T. (1976). The joy of running. Philadelphia:
J. B. Lippincott.

Leach, P. J. (1967). A critical study of the literature
concerning rigidity. British Journal of Social and

Clinical Psychology, 2, 11-22.

Levitt, E. E. (1953). Studies in intolerance of ambiguity.
I. The decision-location test with trade—school
children. Child Development, 24, 263-268.

Lewin, K. A dynamic theory of personality. New York:
McGraw-Hill.

Luengo, M. A., Carrillo-de-la-Pena, M. T. & Otero, J. M.
(1991). The components of impulsiveness: A comparison
of the I.7 Impulsiveness Questionnaire and the Barratt

Impulsiveness Scale. Personality and Individual
Differences, 22, 657-667.

Maddux, J. E., Norton, L. W., & Stoltenberg, C. D. (1986).
Self-efficacy expectancy, outcome expectancy and
outcome value: Relative effects on behavioral

intentions. Journal of Personality and Social
Psyghology , 22, 783-789.

Mann, G. V., Garrett, H. L., Farhi, A., Murray, H., &
Billings, F. T. Exercise to prevent coronary heart
disease: An experimental study of the effects of
training on risk factors for coronary disease in men.
American Journal of Medicine, 42, 12-27.

125

Markus, H. (1977). Self-schemata and processing
information about the self. Journal of Personalitygand
Social Psychology, 22, 63-78.

 

Martin, J. E. & Dubbert, P. M. (1985). Adherence to
exercise. Exercise and Sport Sciences Review, 22, 137-
167.

Martin, J. E., Dubbert, P. M., Katell, A. D., Thompson, J.
K., Raczynski, J. R., Lake, M., Smith, P. O., Webster,
T. S., & Cohen, R. E. (1984). Behavioral control of
exercise in sedentary adults: Studies 1 through 6.
Journal of Consulting and Clinical Psychology, 52, 795-
811.

Massie, J. F. & Shephard, R. J. (1971). Physiological and
psychological effects of training--a comparison of
individual and gymnasium programs with a
characterization of the exercise "drop-out". Medicine
and Science in Sports, 2, 110-117.

McAuley, E. (1992). Understanding exercise behavior: A
self-efficacy perspective. In G. C. Roberts (Ed.),

Understanding motivation in exercise and sport (pp.

107—127). Champaign, IL: Human Kinetics.
McAuley, E. & Courneya, K. S. (1994). The Subjective
Exercise Experiences Scale (SEES): Development and

preliminary validation. Journal of Sport and Exercise
Psychology, 22, 163-177.

McAuley, E., Courneya, K. S., & Lettunich, J. (1991).
Effects of acute and long-term exercise on self-
efficacy responses in sedentary, middle-aged males and
females. The Gerontologist, 22, 534-542.

McAuley, E., Duncan, T. E., Wraith, S. C., & Lettunich, M.
(1991). Self-efficacy, perceptions of success, and
intrinsic motivation. Journal of Applied Social
Psychology, 24, 139-155.

McAuley, E. & Gill, D. L. (1983). Reliability and validity
of the Physical Self-Efficacy scale in a competitive

 

sport setting. Journal of Sport Psychology, 5, 410-
418.
McAuley, E. & Jacobson, L. B. (1991). Self-efficacy and

exercise participation in sedentary adult female
exercise patterns. American Journal of Health
Promotion, 5, 185-191.

126

McCann, I. L., Sakheim, D. K., & Abrahamson, D. J. (1988).
Trauma and victimization: A model of psychological
adaptation. The Counseling Psychologist, 22, 531-594.

McCready, M. L. & Long, B. C. (1985). Locus of control,
attitudes toward physical activity, and exercise
adherence. Journal of Sport Psychology, 1, 346-359.

Meichenbaum, D. H. & Goodman, J. (1969). Reflection-
impulsivity and verbal control of motor behavior.

Child Development, 42, 785-797.

Messer, S. B. (1976). Reflection-impulsivity: A review.
Psychological Bulletin, 22, 1026-1052.

Messer, S. B. & Schacht, T. E. (1986). A cognitive-dynamic
theory of reflection-impulsivity. In J. Masling (Ed.).
Empirical studies of psychoanalytic theories. Volumg‘2.
New Jersey: Lawrence Erlbaum Associates, Publishers.

Miller, A. (1986). Thou shalt not be aware: Society’s
betrayal of the child. New York: Meridian.

Miller, L. (1988). Neurocognitive aspects of remorse:
Impulsivity--compulsivity--reflectivity. Psychotherapy
Patient, 63-76.

Morgan, W. P. (1977a). Psychological consequences of
vigorous physical activity and sport. In M. Gladys
Scott (Ed.), The Academy Papers. Iowa City: American

 

Academy of Physical Education.

Morgan, W. P. (1977b). Involvement in vigorous physical
activity with special reference to adherences. In L.
I. Gedvillas & M. E. Kneer (Eds.), National Collegg
Physical Education Association Proceedings. Chicago:
University of Illinois-Chicago.

Morgan, W. P. (1979). Anxiety reduction following acute
physical activity. Psychiatric Annals, 2, 36-45.
Morgan, W. P. & Pollock, M. L. (1978). Physical activity
and cardiovascular health: Psychological aspects. In
F. Laundry & Organ (Eds.), Physical activity and human

 

well-being. Miami: Symposium Specialists.

Morgan, W. P., Roberts, J. A., Brand, F. R., & Feinerman,
A. D. (1970). Psychological effects of chronic
physical activity. Medicine and Science in Sports, 2,
213-217.

127

 

 

Neimark, E. D. (1975). Longitudinal development of formal
Operations thought. Genetic Psychology monographs, 21,
171-225.

Norusis, M. S./SPSS, Inc. (1990). SPSS/PC+ Advanced
Statistics 4.0. Chicago: SPSS Inc.

Oldridge, N. B. (1979). Compliance with exercise programs.
In M. L. Pollock and D. H. Schmidt (Eds.), Heart

Disease and Rehabilitation. Boston: Houghton Mifflin.

Oldridge, N. B. (1982). Compliance and exercise in primary
and secondary prevention of coronary heart disease: A
review. Preventive Medicine, 11, 56-70.

Oldridge, N. B., Donner, A. P., Buck, C. W., Jones, N. L.,
Andrew, G. M., Parker, J. O., Cunningham, D. A.,
Kavanagh, T., Rechnitzer, P. A., & Sutton, J. R.
(1983). Predictors of dropout from cardiac exercise
rehabilitation: Ontario exercise-heart collaborative

study. The American Journal of Cardiology, 21, 70-74.

Pain, M. D. & Sharpley, C. F. (1986). Some psychometric
data on predictive validity of Self-Motivation
Inventory. Perceptual and Motor Skills, 22, 294.

Parfitt, G., Markland, D. & Holmes, C. (1994). Responses
to physical exertion in active and inactive males and
females. Journal of Sport & Exercise Psychology , 12,
178-186.

Perkins, K. A. & Epstein, L. H. (1988). Methodology in
exercise adherence research. In R. K. Dishman (Ed.).

Exercise adherence: Its impact on public health (pp.

 

399-416). Champaign, IL: Human Kinetics Publishers,
Inc.

Pollock, M. L. (1988). Prescribing exercise for fitness
and adherence. In R. K. Dishman (Ed.). Exercise

adherence: Its impact on public health (pp. 259-278).
Champaign, IL: Human Kinetics Publishers, Inc.

Pollock, M. L., Gettman, L. R., Milesis, C. A., Bah, M. D.,
Durstine, L., & Johnson, R. B. (1977). Effects of
frequency and duration of training on attrition and

incidence of injury. Medicine and Science in Sports,
2, 31-36.

Primavera, L. H., Hochman, S. H., & Reynolds, W. F. (1975).
Compulsivity, manifest anxiety, and gullibility.
Psychological Reports, 21, 505-506.

128

Primavera, L. H., & Simon, W. E. (1971). Non-verbal
rigidity and academic achievement in high school and

college. Perceptual and Motor Skills, 22, 1178.

Reed, G. F. (1968). Some formal qualities of obsessional
thinking. Psychiatric Clinics, 1, 382-392.

Reed, G. F. (1977). Obsessional personality disorder and
remembering. British Journal of Psychiatry, l30, 177-
183.

Reeve, J., & Robinson, D. T. (1987). Toward a
reconceptualization of intrinsic motivation:
Correlates and factor structure of the Activity-Feeling
Scale. Journal of Social Behavior and Personality, 2,

23-36.

Rejeski, W. J. & Thompson, A. (1993). Historial and
conceptual roots of exercise psychology. In P.
Seraganian (Ed.), Exercise psychology: The influence

 

of physical exercise on psychological processes (pp. 3-
35). New York: John Wiley & Sons, Inc.

Rodgers, W. M. & Brawley, L. R. (1991). The role of
outcome expectancies in participation motivation.

Journal of Sport and Exercise Psychology, 12, 411-427.

Rokeach, M. (1948). Generalised mental rigidity as a
factor in ethnocentrism. Journal of Abnormal and

Social Psychology, 42, 259—278.

Rubenowitz, S. (1963). Emotional flexibility-rigidity as a
comprehensive dimension of mind: An empirical study of

a construct, and its psychological and social
implications. Stockholm: Almqvist & Wiksell.

Safrit, M. J., Wood, T. M. & Dishman, R. K. (1985). The
factorial validity of the physical estimation and
attraction scales for adults. Journal of Sport

Psychology, 1, 166-190.

Salkind, N. J. & Nelson, C. F. (1980). A note on the
developmental nature of reflection-impulsivity.
Developmental Psychology, 12, 237-238.

Sallis, J. F., Haskell, W. L., Fortmann, S. P., Vranizan, M.
S., Taylor, C. B., & Solomon, D. S. (1986).
Predictors of adoption and maintenance of physical
activity in a community sample. Preventive Medicine,
12, 331-341.

129

Scanlon, T. K. & Lewthwaite, R. (1986). Social
psychological aspects of the competitive sport
experience for male youth sport participants: IV.
Predictors of enjoyment. Journal of Sport Psychology,

 

 

2, 25-35.

Scanlon, T. K. & Simons, J. P. (1992). The construct of
sport enjoyment. In G. C. Roberts (Ed.). Motivation
in sport and exercise. Champaign, IL: Human Kinetics
Books.

Schleifer, M. & Douglas, V. I. (1973). Moral judgments,

behaviour and cognitive style in young children.
Canadian Journal of Behavioural Science, 5, 133-144.

Shephard, R. J., Corey, P., & Kavanagh, T. (1981).
Exercise compliance and the prevention of a recurrence
of myocardial infarction. Medicine and Science in

Sports and Exercise, 12, 1-5.

Simon, W. E., Primavera, L. H., Klein, B., & Cristal, R. M.

 

(1972). Personality and nonverbal rigidity: Some
preliminary findings. The Journal of Psychology, 22,
127-132.

Sonstroem, R. J. (1976). The validity of self-perceptions

regarding physical and athletic ability. Medicine and
Science in Sports, 2, 126-132.

Sonstroem, R. J. (1978). Physical estimation and
attraction scales: Rationale and research. Medicine
and Science in Sports, 12, 97-102.

Sonstroem, R. J. (1988). Psychological models. In R. K.
Dishman (Ed.). Exercise adherence: Its impact on
public health (pp. 125-154). Champaign, IL: Human

 

Kinetics Publishers, Inc.

Sonstroem, R. J. & Kampper, K. P. (1980). Prediction of
athletic participation in middle school males.

Research Quarterly for Exercise and Sport, 21, 685-694.

Spitzer, (1981). Reflective process. In M. D. Lynch, A. A.

 

 

Norem-Hebeisen, & K. J. Gergen (Eds.). Self-concept:
Advances in theory and research. Cambridge:
Ballinger.

Swanson, H. & Schumacher, G. (1986). Reflection-

impulsivity and script-action recall. Bulletin of the
Psychonomic Society, 24, 28-30.

130

Taylor, H. L., Buskirk, E. R., & Remington, R. D. (1973).
Exercise in controlled trials of the prevention of
coronary heart disease. Federation Proceedings, 22,
1623-1627.

Teraslinna, P., Partanen, A., Koskela, A., & Oja, P.
(1969). Characteristics affecting willingness of
executives to participate in an activity program aimed
at coronary heart disease prevention. Journal of
Sports Medicine and Physical Fitness, 2, 224-229.

Thurstone, L. L. (1944). A factorial study of perception.
Chiacgo: University of Chicago Press.

Tuson, K. M. & Sinyor, D. (1993). On the affective
benefits of acute aerobic exercise: Taking stock after
twenty years of research. In P. Seraganian (Ed.).
Exercise psychology: The influence of physical

exercise on psychological processes (pp. 80-121). New
York: John Wiley & Sons, Inc.

Wankel, L. M. (1985). Personal and situational factors
affecting exercise involvement: The importance of
enjoyment. Research Quarterly for Exercise and Sport,
22, 275-282.

Wankel, L. M. (1988). Exercise adherence and leisure
activity: Patterns of involvement and interventions to
facilitate regular activity. In R. K. Dishman (Ed.).

 

Exercise adherence: Its impact on public health (pp.
369-396). Champaign, IL: Human Kinetics Publishers,
Inc.

Wankel, L. M., Yardley, J. K., & Graham, J. (1985). The

effects of motivational interventions upon the exercise
adherence of high and low self-motivated adults.
Canadian Journal of Applied Sport Sciences, 12, 147—
156.

Ward, A. & Morgan, W. P. (1984). Adherence patterns of
healthy men and women enrolled in an adult exercise
program. Journal of Cardiac Rehabilitation, 4, 143-
152.

Weber, J., & Wertheim, E. H. (1989). Relationships of
self-monitoring, special attention, body fat percent,
and self-motivation to attendance at a community

gymnasium. Journal of Sport and Exercise Psychology,
11, 105-114.

Webster’s New Twentieth Century Dictionary, Unabridged (2nd
Ed.). (1983). Cleveland: Simon and Schuster.

131

Werner, H. (1946). The concept of rigidity: A critical
evaluation. Psychological Review, 22, 43-52.

Wolfe, R., Egelston, R., & Powers, J. (1972). Conceptual
structure and conceptual tempo. 'Perceptual and Motor
Skills, 22, 331-337.

 

“11111111111111?