EFFECTS OF 24 HOURS OF EXERCISE WITHDRAWAL ON
MOOD STATES OF INDIVIDUALS HIGH AND LOW ON
AMY LYNN HAGAN
A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF
FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF MASTER OF SCIENCE IN EXERCISE AND SPORT SCIENCES
UNIVERSITY OF FLORIDA
Amy L. Hagan
This thesis is dedicated to all those few individuals who believed in my ability
and perseverance when so many people questioned me. Specifically, this thesis is
dedicated to my mom and dad who always listened to my struggles, provided support,
and gave an encouraging word even when they were not sure of what I was doing.
Additionally, this thesis would have never been accomplished without the support from
Sarah Drew and Sue Craven; both of them changed my life forever and helped me
discover the freedom that is available to me. I thank them for always believing in me.
This thesis could have never been completed without the help and guidance of
numerous individuals. I would like to thank Dr. Heather Hausenblas for her knowledge
and guidance in implementing and writing this thesis. She has been so much more than
an advisor-a mentor would be more appropriate. I would also like to thank my other
committee members-Dr. Chris Janelle and Dr. Dan Connaughton-for their feedback and
guidance during this process. The insightful discussions with Danielle Symons Downs
and Aaron Duley also contributed to a more thorough understanding of exercise
dependence and deprivation. Additionally, I would like to thank Living Well and Timm
Lovins for the use of their facility while testing.
Many individuals helped me with the testing procedures and I am grateful.
However, a special acknowledgement is needed for Kathleen Kehrer and MacKenzie
Eugene who assisted in most of the testing.
I would not have stayed sane during the process of writing my thesis if not for the
encouragement, support, and laughter from my officemate, Beth Fallon. We shared many
experiences and I will forever treasure our friendship and good times. Stress-relieving
moments such as "lights out" made the hard times more comfortable. She shared past
and current research and thesis-writing experiences ; and allowed me to use them as a
guide. This was indispensable.
TABLE OF CONTENTS
A C K N O W LE D G M EN T S ............................... ......... ................................................... iv
1 INTRODUCTION AND REVIEW OF LITERATURE ............. ................
E explanations for D dependence ......................................................................................... 2
Psychological Explanations ............................................... ............................. 2
P erson ality traits.......................................... .......................... 2
A effective states ......................................... 7
Physiological Explanations .................................................................................. 8
Beta-endorphin theory of endogenous opioids ................................................. 9
Sym pathetic arousal hypothesis.................................................. ... ................. 10
Psychobiological Explanations ........................................... .......................... 12
S u m m a ry ................... ................................................................ 1 7
D efining Exercise D ependence .................................................... ......... .............. 17
Review of Exercise Deprivation or Withdrawal Literature ........................................ 21
Experimental Studies ............................................. ............ .. 21
Quasi-Experim ental Studies..................................................... ......................... 23
Prospective Studies ............................................... ........ ............ .. 26
C orrelational Studies................... .......... .... .............. ........................ ................ .. 27
G general Sum m ary .... ........................ ........ .............................. .............. 28
Prior R research Lim itations.................................... ....... .................. .............. 29
P u rp o se ............ ..... ............................ .. ........................................... 3 1
Significance of the Study ............................................. ....................... .............. 32
H y p o th e sis.............................................................................................. ............... ..... 3 2
2 M E T H O D ..............................................................................3 5
P articip ants ....................................................................................... ............... 3 5
In stru m e n ts ......... ....................... ................................................................... 3 5
D ro p -O u ts ................................ ................................................................... 4 6
D ata A n a ly sis ......... ....................... ................................................................. 4 7
Hypothesis 1 ......................................................... 47
Hypothesis 2................................................................ 48
H hypothesis 3................................ ............... ..... 48
Assumptions For Statistical Analyses..... .................... .............. 48
3 R E S U L T S .................................................................. 5 1
M manipulation C heck.. ................. ............... ................ .. ............... ........ ................... 5 1
R liability of M measures ................................ ........ ...... ............ .......... .. 51
M missing D ata ........................................................... ........ ...... 54
Prim ary H ypothesis.................................. ................ .. ........... ....... .. .......... .. 54
Secondary H ypothesis.............................. .............. .......................... .................. 58
Third Hypothesis........................................ .............. 61
4 DISCUSSION ............. .. ........................................... 64
F first P u rp o se ........................................................... ........ ...... 6 5
Second Purpose ............................................. .................... ..... 67
Third Purpose ......... ..... ...... ... ....................... 69
L im itatio n s .............................................................................. 7 0
Theoretical Implications ............... ........................... 72
Future D directions and Conclusions ........................................ ..... .............. 73
A IR B A P P R O V A L .................................................................................. ..75
B DRIVE FOR THINNESS SUBSCALE .......................................................... 76
C EXERCISE DEPENDEN CE SCALE ......................................................................77
D EXERCISE MOTIVATION SCALE....................................................................79
E REASONS FOR EXERCISE INVENTORY ................................ .....................82
F LEISURE-TIME ACTIVITY QUESTIONNAIRE ...............................................83
G EXERCISE-INDUCED FEELING INVENTORY (EFI) .......................................84
H PR OFILE OF M O OD STA TES .................................................... .....................85
I F E E L IN G S C A L E .......................................................................... ....................87
J STATE-TRAIT ANXIETY INVENTORY.................................... .....................88
K POSITIVE AND NEGATIVE AFFECT SCHEDULE ............................................89
L RATING OF PERCEIVED EXERTION ...................................... ............... 90
M BRUCE PROTOCOL TEST REPORT ....................................... ............... 91
N PRE PARTICIPATION QUESTIONNAIRE .................................. ...............93
O 24-H O U R H IST O R Y ........................................................................ ..................94
P C O N SE N T F O R M ......... ...... ........... ................. ........................... .....................95
L IST O F R E FE R E N C E S .......................................................................... ....................98
B IO G R A PH ICA L SK ETCH ......... ................. ..........................................................107
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science in Exercise and Sport Sciences
EFFECTS OF 24 HOURS OF EXERCISE WITHDRAWAL ON
MOOD STATES OF INDIVIDUALS HIGH AND LOW ON
Amy Lynn Hagan
Chairman: Dr. Heather Hausenblas
Major Department: Exercise and Sport Sciences
It has been reported that individuals who are dependent on exercise experience
deprivation sensations when exercise is interrupted. The purposes of this study were to
determine the following:
* Do exercise-dependent individuals experience mood alterations after 24 hours of
* Do motives and reasons to exercise differ between individuals high and low on
* Do physiological differences such as body composition and cardiorespiratory fitness
vary between the high and low exercise-dependent groups?
Forty-one undergraduate females, (i.e., 20 low and 21 high exercise-dependence
individuals) completed mood measures pre and post exercise and pre and post quiet rest
using a repeated measures design. Results showed no mood differences between the high
and low exercise-dependent groups after the 24 hours of exercise deprivation. Significant
group differences on reasons and motives to exercise were found, with the high group
displaying more internal and external motivation and exercising for mood, weight
control, attractiveness, enjoyment, and tone more than the low exercise-dependent group.
Finally, the high exercise-dependent group had a lower body composition and a higher
level of cardiorespiratory fitness compared to the low exercise-dependent group.
Implications of results are discussed in addition to future research directions.
INTRODUCTION AND REVIEW OF LITERATURE
Researchers have consistently found that physical activity results in the
improvement of both psychological and physiological health (United States Department
of Health and Human Services [USDHHS], 1996, 2000). Because of the numerous
benefits associated with exercise and the high prevalence of sedentariness in North
America, there is concern by health care professionals and government agencies on how
to increase an individual's activity level (Dishman, 1994; USDHHS, 2000). Facilitating
physical activity adoption and adherence, however, continues to remain a challenge given
that approximately 40% of American adults are sedentary (USDHHS, 2000), and that
50% of sedentary adults beginning an exercise program drop out within six months
At the other extreme of the physical activity continuum are individuals who
exercise excessively. Paradoxically, their extreme exercise behavior may result in
negative psychological and physical consequences. This excessive behavior is termed
exercise dependence and it is gaining attention by researchers (Hausenblas & Symons
Downs, 2001). Consistent with many psychological constructs, exercise dependence
does not have a universal definition (Veale, 1995). It is frequently defined as exercise
that is continued in spite of an injury, placed above all other responsibilities (e.g., job and
family), and withdrawal or deprivation effects are experienced with a decrease or
cessation of exercise (Chan & Grossman, 1988; Szabo, 1995; Veale, 1991).
In this chapter the literature examining exercise dependence and deprivation are
reviewed and the study purposes and hypotheses are explained. Specifically, in the first
section the explanations for the antecedents of exercise dependence are discussed. In the
second section, exercise dependence is defined with a focus on the definitional criterion
of withdrawal. The third section reviews studies that investigated the effects of exercise
deprivation. Finally, in the fourth section, the purposes of the thesis are stated and the
hypotheses are presented.
Explanations for Dependence
There are several processes or explanations for the antecedents of exercise
dependence. Research support for these explanations however, is limited. Explanations
can be classified into psychological, physiological, and psychobiological domains.
Research that examined each of these domains is described below.
The personality trait explanation is based on the belief that pathological
personality characteristics such as obsessive-compulsiveness and narcissism are
associated with exercise-dependent individuals. For example, Davis, Brewer, and
Ratusny (1993) examined relationships among the personality characteristics of
addiction, obsessive-compulsiveness, and exercise dependence with exercise behavior.
They hypothesized that there would be a positive relationship among the aforementioned
personality characteristics and exercise behavior. One hundred and eighty-five male and
female physically active participants (M age = 27.82 years) were recruited through ads
and fliers in local fitness clubs and newsletters. Participants completed the Commitment
to Exercise Scale (Davis et al., 1993), the Addiction Scale (Eysenck & Eysenck, 1991),
the Obsessive-Compulsive Personality Scale (Lazare, Klerman, & Armour, 1966), and
the Drive for Thinness, Body Dissatisfaction, and Bulimia Subscales of the Eating
Disorder Inventory (Garner, Olmsted, & Polivy, 1983). Participants were also
interviewed about their exercise participation over the past 12 months.
Consistent with their predication, Davis and her colleagues (1993) found that, for
males only, obsessive-compulsive personality characteristics were positively related to
exercise frequency. In contrast to their hypothesis, exercise frequency was negatively
related to addictiveness in males. Also, frequency of exercise was negatively related to
addictiveness in females, although nonsignificant (p = .07). No other study findings were
significant. Thus, the hypothesized positive relationships among obsessive-
compulsiveness, addictiveness, exercise dependence, and exercise behavior were not
In a similar study Yates and her colleagues (1992) examined the personality
characteristics between obligatory and nonobligatory male and female runners who ran a
minimum of 15 miles per week. According to their responses on an 18-item author-
developed questionnaire and a semi-structured interview that assessed extreme exercise
attitudes and behaviors, the runners were classified as either obligatory or nonobligatory
exercisers. Ten male and 17 female participants were classified as obligatory runners,
while the nonobligatory group comprised 20 males and 19 females. Participants
completed the Profile of Mood States (POMS; McNair, Lorr, & Droppleman, 1981), the
Eysenck Personality Questionnaire (Eysenck & Eysenck, 1991), the Minnesota
Multiphasic Personality Inventory (Dahlstrom & Welch, 1960), the Bem Sex Role
Inventory (Bem, 1981), the Internal-External Locus of Control Scale (Rotter, 1966), the
Eating Attitudes Test (Garner & Garfinkel, 1979), and the Beck Depression Inventory
(Beck, Ward, Mendelson, Mock, Erbaugh, 1961).
The researchers found that the obligatory runners were more likely to follow a
diet, be preoccupied with their body, run alone, and report more positive changes in self-
concept and control over their lives since beginning to run than the nonobligatory
runners. Also, the obligatory males were twice as likely to have elevated scores on the
Minnesota Multiphasic Personality Inventory compared to the nonobligatory male
runners. No other group differences were found between the obligatory and
nonobligatory male and female runners.
In a controversial study, Yates and her colleagues (1983) argued that male
obligatory runners resembled anorexia nervosa patients on certain personality
characteristics (e.g., introversion, inhibition of anger, high expectations, depression, and
excessive use of denial) and labeled this relationship as the anorexia-analogue hypothesis.
To test their hypothesis they examined the personality characteristics of 60 male
excessive exercisers and compared their responses to traditionally reported idiosyncrasies
of anorexia nervosa patients. Although no objective data were reported, Yates et al.
claimed that running and extreme dieting were both dangerous attempts to establish an
identity; that is, to be recognized as either exercise-dependent or anorexic.
Their article was heavily criticized as having no pertinent data, poor
methodology, no relevance to the majority of runners, an over-reliance on extreme
individuals, and an overstatement of similarities between the groups (Blumenthal,
O'Toole, & Chang, 1984). Succeeding empirical research has failed to find a common
psychopathology between exercise-dependents and eating-disorder patients (e.g.,
Blumenthal et al., 1984; Coen & Ogles, 1993). In short, the anorexia analogue hypothesis
is only speculative and has not been fully supported by subsequent research (Coen &
For example, Coen and Ogles (1993) examined the personality characteristics
believed to be common to anorexics and runners according to the anorexia analogue
hypothesis (e.g., anxiety, perfectionism, and ego identity). The participants were 142
male marathon runners who completed the Obligatory Exercise Questionnaire
(Thompson & Pasman, 1991), the Multidimensional Perfectionism Scale (Frost, Marten,
Lahart, & Rosenblate, 1990), the Trait Subscale of the State-Trait Anxiety Inventory
(Speilberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983), the Ego Identity Scale (Tan,
Kendis, Fine, & Porac, 1977), and demographic running characteristics. A median split
on the Obligatory Exercise Questionnaire was used to classify runners as either
obligatory or nonobligatory.
It was found that the obligatory runners reported more perfectionistic
characteristics compared to the nonobligatory runners. Specifically, obligatory runners
were more concerned about making a mistake, had higher personal standards, more
doubts about their actions, and had a higher need for organization than their
nonobligatory counterparts. The obligatory runners also reported greater trait anxiety
compared to the nonobligatory runners. The authors concluded that the anorexia
analogue hypothesis was partially supported because only some of the personality
characteristics advanced by Yates et al. (1983) were associated with the obligatory
In contrast to the previous study, lannos and Tiggemann (1997) failed to find
personality differences between exercise-dependent and nondependent individuals. They
examined 205 male and female individuals from local gymnasiums. The participants
were divided into three groups according to the number of hours they exercised per week:
light (n = 69, 0-5 hours), medium (n = 107, 5-11 hours), and excessive (n = 29, 11+
hours). The participants completed an author-developed questionnaire which assessed
their level of exercise dependence, the Self-esteem Scale (Rosenberg, 1965), the
Internality Subscale of the Locus of Control Inventory (Levenson, 1981), the Obsessive-
Compulsive Scale (Gibb, Bailey, Best, & Lambrinth, 1983), and the Drive for Thinness
and Bulimia Subscales of the Eating Disorder Inventory. It was found that the excessive
exercisers reported more eating disorder behaviors than the light and medium exercisers,
with females scoring higher than males. Despite this relationship, no significant
differences were found between exercise level and the personality characteristics of self-
esteem, locus of control, and obsessive-compulsiveness. Thus, the authors concluded
that excessive exercise is not related to a pathological personality.
Finally, Estok and Rudy (1986) examined 57 marathon and 38 nonmarathon
female runners (M age = 35.8 years) on physical symptoms (e.g., shin splints and knee or
hip pain), psychosocial symptoms (e.g., anxiety and self-esteem), and addictive behaviors
using a 14-item author-developed scale. It was found that the marathon runners
compared to the nonmarathon runners scored significantly higher on addictive behaviors.
The incidence of injuries among marathon runners was significantly higher than in the
nonmarathon runners. Although there was no difference between the marathon and
nonmarathon groups for psychosocial symptoms, there was a negative association
between addiction and self-esteem. The authors concluded that common running injuries
are more a function of the distance run rather than the level of running addiction.
The affective-states explanation is based on the belief that the improved feeling
states associated with acute exercise bouts may result in exercise dependence. The
rationale for the affective states hypothesis is that acute exercise results in improvements
in levels of depression, anxiety, stress, and mood (Landers & Arent, 2001). It is therefore
suggested that exercise-dependent individuals become "dependent" on these positive
states and the "euphoria" experienced by exercising.
In accordance with this statement, Sachs and Pargman (1979) suggest that
following acute exercise bouts, feelings of relaxation, accomplishment, vigor, and
calmness are achieved. They also state that, "the feeling of well being associated with
exercising is a strong motivator for continued participation in regular physical activity"
(Sachs & Pargman, 1979, p. 145). The belief is that if small doses of exercise produce
positive feeling states, then increased activity would result in more positive mood. For
example, Morgan (1979) states that, "a drug dosage must be increased across time to
create the same quality of sensation, and similarly, exercise duration or intensity (dosage)
must also be increased to maintain an exercise high" (p. 63). These sensations are the
psychological "euphoria" that individuals feel while exercising. It is suggested that this
euphoria can lead to dependence because individuals will exercise at either higher
intensities or longer durations to experience the positive effects (Kagan & Squires, 1985;
Pierce, 1994; Thompson & Blanton, 1987).
The affective-states explanation for exercise dependence resembles tolerance,
which is a component of dependence as defined by the Diagnostic and Statistical Manual-
IV (DSM-IV; American Psychological Association [APA], 1994). According to the
DSM-IV definition for dependence, tolerance is achieved when there is a "need for
greatly increased amounts of the substance to achieve the desired effect or a markedly
diminished effect with continued use of the same amount of the substance" (p. 176).
When individuals experience tolerance they have adapted to and become familiar with
the stimuli or stressor regardless of its negative effects. In regards to exercise, tolerance
can develop despite experiencing negative effects such as sore muscles, achy joints, and
the requirement of time.
Davis (1999) describes the adaptation to tolerance as following:
It has been said that a metaphoric switch seems to be thrown as a result of
prolonged exposure to any reinforcing activity or substance, with the
result that a behavior that was once voluntary becomes compulsive and
moves into the state of addiction. (p. 222)
Research is needed to examine whether tolerance to exercise occurs for exercise-
Physiological explanations center on the premise that chemical processes that
occur in the body during exercise can lead to dependence. These chemicals include
epinephrine, norepinephrine, beta-endorphins, and opioids. Gauvin and Rejeski (1993)
state that changes occur in the metabolic and nervous systems to promote a physiological
desire for exercise. For example, feelings following an acute exercise bout (e.g.,
refreshed, tranquil) are possibly due to the changing chemical processes in the body
(Gauvin & Rejeski, 1993). Furthermore, Scully et al. (1998) state that "mood enhancing
and analgesic properties associated with exercise are influenced by chemicals in the brain
that are akin to opiates" (p. 416). The two existing physiological explanations are
described in the subsequent sections.
Beta-endorphin theory of endogenous opioids
The most common physiological explanations for exercise dependence is the beta-
endorphin theory of endogenous opioids (De Coverley Veale, 1987; Thoren, Floras,
Hoffman, & Seals, 1990). Beta-endorphins are endogenous compounds that play a role
in the transmission of pain impulses by decreasing pain sensitivity. Researchers have
shown that beta-endorphin levels in the blood rise with submaximal exercise due to an
increased need for blood to be transported to the working muscles (Crossman, Jamieson
& Henderson, 1987; Thoren et al., 1990). Based on the physiological processes that
occur in the body when beta-endorphin levels rise, an individual could conceivably
become dependent on these natural chemicals because they mask pain, thereby allowing
individuals to feel better (Crossman et al., 1987). In addition to the analgesic effect,
endogenous compounds can produce addictive behavioral tendencies such as compulsive
exercise (Pierce, Eastman, Tripathi, Olson, & Dewey, 1993). The opioid peptides
produced by the central nervous system however, are impermeable to the blood brain
barrier. Because of the impermeability to the barrier, beta-endorphins cannot solely be
associated with exercise dependence.
The only located study examining the beta-endorphin hypothesis was undertaken
by Pierce and his colleagues (1993). They examined the relationship between beta-
endorphin levels after an acute exercise bout with exercise-dependence scores.
Participants were eight females who engaged in at least three aerobic classes per week.
The Negative Addiction Scale (Hailey & Bailey, 1982) was completed by the participants
and blood samples were taken by venipuncture before they engaged in a high-intensity
aerobic session for 45 minutes. Immediately following the aerobic session blood samples
were reassessed. Although beta-endorphin levels were elevated after the activity bout, a
significant positive correlation between level of exercise dependence and beta-endorphins
failed to appear. In short, despite the intuitive appeal of this hypothesis, research support
Sympathetic arousal hypothesis
Thompson and Blanton (1987) suggest a sympathetic arousal hypothesis to
explain exercise dependence (Figure 1.1). In this hypothesis, dependence is attributed to
exercise behavior and hormonal changes in epinephrine and norepinephrine. Epinephrine
(e.g., adrenaline) and norepinephrine (e.g., noradrenaline) respond to strong emotional
stimuli and prepare the body for challenges from the environment (Powers & Howley,
1997). When the body is repeatedly stressed in the same manner it becomes familiar with
Lowered sympathetic arousal
Figure 1.1 Process of exercise dependence according to the sympathetic arousal
the stressor, thereby becoming more efficient when responding to the stressor. Energy
efficiency is increased causing a change in the dose response, thereby requiring the
individual to exercise more to achieve the same arousal level.
The sympathetic arousal hypothesis is based on two components: physiological
arousal and the opponent process model. The sympathetic arousal component has been
supported with research on depression and the biochemical and psychological effects of
exercise (Doyne, Chambless, & Beutler, 1983; Landers & Arent, 2001; Mandenoff,
Fumeron, Apfelbaum, & Margules, 1982; Morgan, 1985; Morgan, Roberts, Brand, &
Feinerman, 1970). For example, research has found that for depressed individuals,
exercise and drug therapy have similar antidepressant effects due to elevation of
norepinephrine to a healthy level (Morgan, 1985). Furthermore, individuals who were
sedentary prior to experiencing depressive symptoms may alleviate depression by regular
exercise (Doyne et al., 1983; Mandenoff et al., 1982).
The second element of this hypothesis stems from the opponent-process model of
addiction (Solomon, 1980). This model has three phases: 1) affective hedonic contrast,
2) tolerance, and 3) withdrawal. In the first phase, individuals may experience an
affective feeling of pleasure when exercise is performed. Tolerance, the second phase,
then occurs with repetition of the exercise. This causes an increased need for either
higher intensities or longer durations of exercise to experience the same amount of
pleasure or euphoria previously encountered. The last phase, withdrawal, occurs when
the activity is terminated or the effort (i.e., intensity) by the individual does not elicit the
pleasurable euphoria once felt. Withdrawal then produces an aversive reaction that
promotes subsequent exercise in order to diminish the negative effects. No known
studies have examined the existence of the sympathetic arousal hypothesis in explaining
exercise dependence possibly due to its complexity.
A combination of the psychological and physiological domains is referred to as
psychobiological. Sachs and Pargman (1979) report that, based on their interviews with
12 male runners, the vital factors in exercise dependence are a combination of the
psychological and physiological elements of withdrawal. Subsequently, they developed a
two-factor psychobiological model to describe the relationship between running and
dependence. The model is based on the degree of commitment and the amount of
dependence an individual has acquired (Conboy, 1994; Sachs & Pargman, 1984). The
first factor, dependence, is the psychobiologic phenomenon. Commitment, the second
factor, is regarded as the cognitive or intellectual phenomenon (Figure 1.2). The model's
premise is that the addicted runner has high commitment and dependence and when
hih Most amount of
low Least amount of
Figure 1.2 Sachs and Pargman's (1984) hypothesized model of runners'
running deprivation occurs, he or she experiences withdrawal effects. In contrast, a
runner with low commitment and high dependence would not experience withdrawal
effects for a missed run. Research examining this model however, has not consistently
supported the existence of the two factors (Conboy, 1994).
For example, Szabo and his colleagues (1997) attempted to test the model by
examining the relationships among addiction, commitment, and deprivation. Committed
runners were defined as those who run for extrinsic rewards, view running as an
important, but not central, part of their lives, and do not suffer from strong withdrawal
symptoms. In comparison, addicted runners were defined as people who are more likely
to run for intrinsic rewards, view running as the central part of their lives, and experience
strong deprivation sensations when unable to exercise. The authors hypothesized that
deprivation effects from running are related to addiction rather than commitment. That
is, addicted but not committed runners would report withdrawal effects during no run
periods. It was also hypothesized that Type I runners (i.e., the most committed runners)
would have higher mileage per week than Type II runners (i.e., committed runners who
involve themselves in other activities).
To test their hypotheses, Szabo and his colleagues (1997) recruited 100 male and
female participants via an Internet discussion group for runners. The participants
completed the Commitment to Running Scale, the Obligatory Exercise Questionnaire, an
author-developed scale to examine deprivation sensations, running demographics, and
reasons for beginning and maintaining running. The authors found a positive relationship
between running addiction and deprivation sensations, as well as the frequency, duration,
and distance of the run. In addition, participants who started running for reasons other
than health, but continued for health reasons, experienced the most extreme feelings of
deprivation. In comparison, commitment to running was not correlated with the variables
examined. In summary, the study results demonstrated that feelings of deprivation are
components of addiction, not commitment, to running.
Another psychobiological explanation is the general theory of addiction (Jacobs,
1986) which describes the etiology and course of all addictive behaviors. This theory
proposes that there are two predisposing factors which determine if an individual is at-
risk of developing an addiction. The first factor is based on the resting physiological
tension of the body which is between hypo (excessively depressed) and hyper (extreme
excitement). According to this theory for a person to be predisposed to an addictive
behavior his or her resting physiological state must be either chronically hypo or hyper.
The predisposed second factor is psychological in nature and states that certain social and
developmental experiences in childhood (e.g., abandonment, abuse) may result in
addictions. It is proposed that addictive behaviors may allow the individual to escape
from his or her distressing reality and thereby, experience wish-fulfilling fantasies of
being an important person.
Jacobs (1986) proposes that when both of the predisposing factors exist, addictive
behavior can be predicted by charting movement through three sequential stages. The
first is called the Stage of Discovery and occurs when a high level of positive
reinforcement is produced by an experience (which later becomes the addiction).
Because of the reinforcement, the behavior is repeated and the individual continues to
achieve the relief experienced by the first occurrence. The shift from occasional
behaviors to planned, highly motivated, compulsive behaviors (termed overlearning) is
the second stage which is termed Stage of Resistance. This resistance is evident when
outside efforts to discourage the behavior are rejected and the addicted individual
embraces the behavior. Three factors explain the resistance experienced during this time:
1) positive reinforcement from the pleasure experienced from the behavior, 2) extent of
overlearning, and 3) avoidance of the return to the previous resting state. This third factor
is considered to be the key aspect of the general theory of addiction due to the
undesirable resting state.
During the second stage, a reciprocal relationship between positive and negative
reinforcement occurs that prepares the individual to enter the third stage. The final stage,
Exhaustion, is entered when the individual who is emotionally and physically exhausted
experiences a rapid breakdown of the addiction and collapses. At this point, the
individual is most treatable and able to discontinue the addictive behavior. In the general
theory of addiction, the presence of the predisposing conditions and the course of
movement through the three stages is known as the Addictive Personality Syndrome.
Examining the second criterion of the general theory of addiction, Bamber,
Cockerill, and Carroll (2000) tested 291 females recruited from fitness centers and eating
disorder clinics. The Exercise Dependence Questionnaire (Ogden, Veale, & Summers,
1997), the Eating Disorder Examination Self-Report Questionnaire (Fairbum & Beglin,
1994), the General Health Questionnaire (Goldberg & Williams, 1988), the Rosenberg
Self-Esteem Scale, the Eysenck Personality Questionnaire-Revised (Eysenck & Eysenck,
1991), the Body Shape Questionnaire (Cooper, Taylor, Cooper, & Fairburn, 1987), the
Exercise Beliefs Questionnaire (Loumidis & Wells, 1998), and questions assessing
menstrual dysfunction, weight dissatisfaction, weight fluctuation, and physical activity
were administered. Based on responses to the Exercise Dependence Questionnaire and
the Eating Disorder Examination Self-Report Questionnaire, the participants were
divided into one of four groups: (a) exercise-dependent, (b) eating disordered, (c) both
exercise-dependent and eating disordered, or (d) neither exercise-dependent nor eating
disordered (i.e., control group).
Statistical analysis revealed that the exercise-dependent and exercise-
dependent/eating disorder groups experienced a higher percentage of menstrual
amenorrhea and irregular menstrual cycles compared to the eating disorder and control
groups. The exercise-dependent/eating disorder group scored higher on the General
Health Questionnaire compared to the exercise-dependent and control groups, indicating
a lower level of overall health. Additionally, the exercise-dependent and control groups
scored significantly higher on self-esteem compared to the exercise-dependent and
exercise-dependent/eating disorder group. Both the exercise-dependent and exercise-
dependent/eating disorder groups demonstrated significantly more concern with body
shape and weight dissatisfaction than the other groups. For weight fluctuation however,
the exercise-dependent/eating disorder group had the greatest fluctuation compared to the
exercise-dependent and control groups but not the eating disorder group.
According to the Eysenck Personality Questionnaire-Revised, no group
differences existed for psychoticism, social desirability, and empathy. The control group
however demonstrated significantly less neuroticism, addictiveness, and impulsiveness
compared to the other three groups. In terms of exercise, the exercise-dependent and
exercise-dependent/eating disorder groups exercised significantly more than the eating
disorder and control groups. In summary, the exercise-dependence individuals failed to
have the psychological morbidity seen in the eating disorder and the exercise-
dependent/eating disorder groups. The authors concluded that exercise dependence, in
the absence of an eating disorder, may be a rare occurrence and not a pathology.
Equivocal and limited research exists on the hypotheses that have been advanced
to explain exercise dependence. This may, in part, be due to the study designs and
measurements. The psychological explanations of personality traits and affective states
have typically been measured using self-report measures that have inherent biases such as
social desirability and retrospective data collection. The physiological explanations such
as the beta-endorphin theory of endogenous opioids and the sympathetic arousal
hypothesis have been more difficult to examine. This is because physiological tests often
require expensive equipment, trained experts, and are intrusive to the participants. The
psychobiological explanations for dependence, such as the two-factor running model and
the general theory of addiction, are challenging to test because it is difficult to determine
what percentage of dependence can be contributed to psychological versus biological
factors. In short, no single explanation for exercise dependence has been supported. It is
plausible that a combination of psychological and physiological factors contribute to
Defining Exercise Dependence
As with other addictions, a standard definition of exercise dependence does not
exist (Johnson, 1994). Definitions of exercise dependence include behavioral factors
(e.g., exercise frequency), psychological factors (e.g., pathological commitment), and
physiological factors (e.g., tolerance). The definition that has gained the most
recognition was proposed by DeCoverley Veale (1987; Veale, 1995) who recommends a
set of standards for diagnosing exercise dependence based on the DSM-IV criteria for
substance dependence, which includes both a biomedical (e.g., tolerance, withdrawal)
and psychosocial perspective (e.g., interference with social and occupational functioning;
Expanding on Veale's definition, Hausenblas and Symons Downs (2001)
recommend, based on the DSM-IV diagnostic criteria for substance dependence (APA,
1994), that exercise dependence be operationalized as a multidimensional maladaptive
pattern of exercise, leading to clinically significant impairment or distress, as manifested
by three or more of the following: (1) tolerance: which is defined as either a need for
significantly increased amounts of exercise to achieve the desired effect or diminished
effect with continued use of the same amount of exercise; (2) withdrawal: which is
manifested by either withdrawal symptoms for exercise (e.g., anxiety, fatigue) or the
same (or closely related) amount of exercise is taken to relieve or avoid withdrawal
symptoms; (3) intention effects: which represents exercise that is often taken in larger
amounts or over a longer period than was intended; (4) loss of control: which is defined
as a persistent desire or unsuccessful effort to cut down or control exercise; (5) time:
which reflects that a great deal of time is spent in activities necessary to obtain exercise
(e.g., vacations are exercise related); (6) conflict: important social, occupational, or
recreational activities are given up or reduced because of exercise; and (7) continuance:
exercise is continued despite knowledge of having a persistent or recurrent physical or
psychological problem that is likely to have been caused or exacerbated by the exercise
(e.g., continued running despite severe shin splints).
Furthermore, it is difficult to determine whether the individual has physiological
dependence (i.e., evidence of tolerance or withdrawal) or no physiological dependence
(i.e., no evidence of tolerance or withdrawal). Future research is required to establish the
utility of this definition and which of the criteria apply to exercise dependence. Table 1-1
adopts the DSM-IV criteria for substance dependence to define exercise dependence.
Focusing on the second criterion of withdrawal, exercise-dependent individuals
should demonstrate withdrawal effects during periods of no physical activity. When
excessive exercisers discontinue exercising for a short period of time, they often
experience depression, fatigue, anxiety, irritability, restlessness, insomnia, frustration,
and guilt (Morgan, 1979; Scully et al., 1998; Szabo, 1995). Szabo (1995) stated that
withdrawal symptoms are the cardinal feature of exercise dependence. Furthermore, he
suggested that additional research is required to study the magnitude of the effects of
exercise deprivation on dependent individuals. In short, it is important to determine
whether withdrawal exists with exercise dependence as it does with other addictions
Finally, it is important to note that DeCoverley Veale (1987; Veale, 1995)
proposed that a diagnostic hierarchy must occur to validly identify exercise dependence.
He argued that the diagnosis of an eating disorder must first be excluded before a
diagnosis of primary exercise dependence can be made. That is, primary exercise
dependence can be differentiated from an eating disorder by clarifying the ultimate
objective of the exerciser. In primary exercise dependence the physical activity is an end
in itself. In contrast, for secondary exercise dependence the compelling motivation for
physical activity is the control and manipulation of body composition.
Table 1-1: Exercise Dependence Criteria
At least three of the following must be present at one time to be considered exercise-
1. Tolerance: as defined by either of the following:
a) A need for markedly increased amounts of the exercise to achieve the desired
b) Markedly diminished effect with continued use of the same amount of exercise.
2. Withdrawal: as manifested by either of the following:
a) Characteristic withdrawal symptoms including general fatigue, anxiety, irritability,
frustration, and guilt or
b) The same (or closely related) exercise is done to relieve or avoid withdrawal
3. Intention effects: the exercise is often performed at higher intensities or over a
longer period than was intended.
4. Loss of control: there is a persistent desire or unsuccessful efforts to cut down or
5. Time: a great deal of time is spent in preparing to exercise, exercising, or
recovering from its effects.
6. Conflict: important social, occupational, or recreational activities are given up or
reduced because of exercise.
7. Continuance: exercise is continued despite knowledge of having a persistent or
recurrent physical or psychological problem that is likely to have been caused or
exacerbated by exercise.
Review of Exercise Deprivation or Withdrawal Literature
Regardless of the reasons for the deprivation period (e.g., injury, lack of time),
exercise-dependent individuals experience withdrawal symptoms. There is growing
evidence (Crossman et al., 1987; Gauvin & Szabo, 1992; Mondin et al., 1996; Sachs &
Pargman, 1979; Szabo, 1995; Thaxton, 1982) that only 24 hours of deprivation is
required to experience withdrawal symptoms; however, deprivation periods can range
from 24 hours (Conboy, 1994) to one month (Baekeland, 1970). Research designs of
exercise-deprivation investigations include experimental (Thaxton, 1982), quasi-
experimental (Baekeland, 1970; Conboy, 1994; Mondin et al., 1996; Morris, Steinberg,
Sykes, & Salmon, 1990; Wittig, McConell, Costill, & Schurr, 1992), correlational (Estok
& Rudy, 1986; Harris, 1981a; Szabo, Frenkl, & Caputo, 1996), and prospective studies
(Gauvin & Szabo, 1992; Szabo, Frenkl, & Caputo, 1997). Experimental studies are rare
in the exercise-deprivation literature. One possible explanation is that experimental
deprivation studies are difficult to conduct because most exercise-dependent individuals
do not voluntarily stop exercising (Szabo, 1998). Thus, to examine the effects of exercise
deprivation, quasi-experimental and correlational studies have been undertaken to a
greater extent than experimental designs. Prospective studies have also been developed
in which mood is assessed during exercise and no-exercise days. These study designs are
described in greater detail in the subsequent sections.
Thaxton (1982) examined the effects of 24 hours of exercise deprivation with 33
committed runners (24 males and 9 females). Commitment to running was defined as
those who had been averaging five days a week of running for at least one year.
Participants completed the POMS and a galvanic skin response test. The POMS
measures six mood states: tension, fatigue, vigor, anger, confusion, and depression. The
galvanic skin response test is a physiological measure that assesses skin resistance that is
altered by reactions between the autonomic nervous system and a stimulus. The test was
used to measure withdrawal-related tension that was indicated by a high resistance
Using the Solomon four-group design and random selection, half of the subjects
refrained from running (no-run group) while the remaining participants ran for 30
minutes (run group) during a 24-hour period. Half of the run and no-run groups were
pretested using the galvanic skin test. The pretest included counting by seven backwards
from a given number for five minutes. A loud motor was added to increase the amount of
stress. The galvanic skin test was simultaneously used during this process to record skin
resistance. The POMS and galvanic skin response test were administered within a few
hours after the 30-minute run or deprivation period, depending on group assignment. It
was found that the running-deprived group reported greater depression and stress
compared to the no-run group.
In another experimental study, Gauvin and Szabo (1992) employed the experience
sampling method. They examined one week of deprivation using various exercise
modes. The participants (N = 21; 67% male) were randomly assigned into either an
experimental (n = 12) or control group (n = 9). The participants were university students
who engaged in at least 3 bouts of exercise a week with a mean of 7.52 hours of exercise
per week for the past 4 months. The control group continued their regular exercise
routine throughout the five-week study. The experimental group was instructed to stop
exercising from Day 15 to 21 (one week) of the study. Each participant was given a
beeper that sounded four times a day for the length of the study. In response to the beep,
the participant completed a well-being questionnaire that assessed mood states and
physical symptoms (e.g., headache, stomach pain, chest pain, coughing, sore throat, and
stiff/sore muscles). It was found that the experimental group reported more physical
symptoms during the deprivation period than the control group. No group differences in
mood states were evidenced.
Because true experimental deprivation studies are difficult to conduct (Szabo,
1995), there have been a larger number of quasi-experimental studies. Baekeland (1970)
conducted the first quasi-experiment examining the effects of exercise-deprivation on
mood. He encountered great difficulty recruiting habitual exercisers (i.e., individuals
who exercised five to six days a week) who were willing to abstain from exercise for one
month. In fact, the habitual runners refused to participate even when monetary incentives
were provided. The participants he was able to recruit were 14 males who regularly
exercised three to four days per week. During the one-month deprivation period,
participants reported decreased psychological well-being (e.g., increased anxiety,
nocturnal awakening, sexual tension, and anxiety). In short, Baekeland found that: (1)
habitual runners, who ran five to six days a week, refused to interrupt their exercise
program for a one-month period; and (2) regular runners, who ran three to four days a
week, reported withdrawal symptoms during a one-month exercise-deprivation period.
In a more controlled quasi-experimental study, Morris and colleagues (1990)
examined 2-weeks of exercise-deprivation with 40 male runners. For study inclusion,
participants had to run 3 times a week for 10 miles a week for at least 3 months. The
control group was instructed to continue their regular running over the course of the 6-
week study. The experimental group was asked to stop running for Week three and four
during study. The measures used to assess the deprivation effects were the General
Health Questionnaire (Goldberg & Williams, 1988) and the Zung Anxiety and
Depression Scales (Zung, 1974).
It was found that the deprived runners reported increased anxiety, insomnia,
somatic symptoms, social dysfunction, and depression compared to the control group
during the exercise-deprivation period. The amount of time to experience the
psychological changes was dependent on the variable measured. That is, the General
Health Questionnaire subscales of social dysfunction, somatic symptoms, anxiety, and
insomnia were present at the end of the first week of deprivation. In comparison,
depression and anxiety did not manifest until the end of the second week. The authors
suggested that the exercise-deprived individuals were able to cope without running for a
week; but after that time, their coping mechanisms failed and depression and anxiety
In another study using runners, a 4-week reduction in both training intensity and
volume was undertaken (Wittig, McConell, Costill, & Schurr, 1992). Ten male adult
runners who had been training an average of 14.5 years were examined to determine if
exercise reduction, rather than deprivation, had an effect on mood. The participants
trained for four weeks at their normal training load (baseline) and then started a 4-week
period of reduced training. All participants completed a maximal exercise test to assess
their cardiovascular endurance at baseline. The training volume (i.e., mileage per week)
was reduced by 66%, frequency (i.e., number of exercise sessions) by 50%, and intensity
(i.e., effort) at or below 70% of VO2 max. A 5-kilometer run was performed following
baseline training and after the reduced training-regimen. The POMS was administered
before baseline, after baseline, and following the reduced-training period. Increased
negative mood was evident from baseline to following the reduced training. This finding
demonstrated that reduced exercise volume, frequency, and intensity can alter mood; and
that a complete deprivation period may not be required.
Using a measure of dependence, Conboy (1994) examined exercise deprivation
and mood states using Sachs and Pargman's (1984) model of runners' motivation (Figure
1-2). Subjects were recruited from 5 and 10 km running races and local running clubs.
Fifty-one male and 10 female runners completed the POMS, the Commitment to Running
Scale, and a question regarding their frequency of running per week. Using the scores
from the Commitment to Running Scale, a median split was performed to determine the
low and high exercise-dependent groups. The participants did not alter their normal
running schedule for the duration of the study. On at least 10 days when they ran and at
least 2 days when they did not run, they were required to complete the POMS. The
POMS was completed 20-40 min following exercise on run days and on no-run days it
was completed 24 hours after the last run.
It was found that exercise deprivation resulted in increased mood disturbance. In
contrast to the hypothesis however, the high-commitment and high-dependence runners
did not report the greatest withdrawal effect. In fact, individuals scoring low on
commitment and high on dependence reported the greatest mood disturbance on no-run
days. In contrast, high commitment and high dependence participants displayed the least
mood disturbance between run and no-run days.
In another study in which the participants were aware of the length of the
deprivation period, Mondin and his colleagues (1996) examined the effects of three days
of exercise withdrawal for habitual exercisers. Participants were six males and four
females who exercised 6 to 7 days a week for at least 45 min a day. The participants
customarily engaged in cardiovascular exercise that included jogging, swimming, and
cycling. Measurements included the State Anxiety Inventory, POMS, Depression
Adjective Checklist (Lubin, Hornstra, & Dean, 1978), and a questionnaire that examined
sleep, food intake, and general feelings over the past 24 hours. The study was conducted
on five consecutive days (Monday through Friday) and from Tuesday to Thursday the
participants refrained from their regular exercise routine and also limited their lifestyle
exercise (e.g., take the elevator instead of the stairs, park close to buildings). On Monday
and Friday the participants engaged in their regular exercise routine.
Participants completed the questionnaires at the laboratory on Monday and Friday
within 15-20 minutes following their regular exercise. Tuesday through Thursday the
participants returned to the laboratory and completed the questionnaires at the same time
they were administered on Monday and Friday. The authors found that total mood
disturbance increased on Tuesday, peaked on Wednesday, decreased slightly on
Thursday, and returned to baseline on Friday. It was concluded that exercise deprivation
negatively affected mood, tension, and depression in habitual exercisers.
In a prospective study, Szabo and his colleagues (1998) examined runners'
anxiety and mood on running and nonrunning days for 21 consecutive days. Ten female
and 30 male runners were recruited through a local running club. The Commitment to
Running Scale was used to determine the participant's level of devotion to running. The
State Anxiety Inventory and the Exercise-Induced Feeling Inventory (Gauvin & Rejeski,
1993) were used to examine mood and anxiety. The questionnaires were completed
every evening and the participants reported how they felt in the last 24 hours. The
subjects also recorded their running distance and time and stated any major life-event
stresses that occurred.
It was found that the females indicated greater exhaustion during the nonrunning
days than the running days. Also on the nonrunning days, men reported more tranquility
than women. Gender-related differences occurred in regards to running speed and
commitment to running, with males scoring higher than females. When effect sizes were
computed however, only revitalization had a moderate effect for females on running
versus nonrunning days. Although these results support better mood on running
compared to nonrunning days, the differences were minor.
The majority of deprivation studies are correlational (e.g., Harris, 1981b; Szabo et
al., 1996) and require participants to recall how they felt during nonexercise periods.
Therefore, the deprivation period is not manipulated. For example, Szabo and his
colleagues (1996) had participants recall why they began exercising and whether they
experienced withdrawal effects when they discontinued exercise. Participants were
recruited via the Internet through exercise discussion and news groups. In order to be
eligible for study participation, an individual had to be age 18 or older, understand
English fluently, and engage in physical activity at least twice a week. Five exercise
discussion/news groups received a 3-part message every week for 12 weeks. The first
message consisted of the purpose of the study and criteria for participation. The second
message asked for demographic characteristics and motives for beginning to exercise.
The last message included the following questionnaires: the State-Trait Anxiety
Inventory, the Deprivation Sensation Scale (Robbins & Joseph, 1985), and the
Commitment to Physical Activity Scale (Corbin, Nielsen, Borsdorf, & Laurie, 1987).
Bowlers represented the control group because their activity requires minimal physical
exertion, while participants engaging in aerobic exercise, weight training, cross-training,
and fencing represented the experimental group. It was found that participants who
started exercising for health reasons reported less trait anxiety and commitment; however,
they also felt the most deprivation effects compared to those participants who were active
for other reasons.
Finally, Harris (1981b) surveyed 156 females to investigate why they started
running and the psychological and behavioral effects of running. The author-developed
questionnaire assessed physical changes that occurred due to running, alterations in self-
concept, menstrual cycles, and habits such as drinking, smoking, and eating. Participants
who stopped running (90% of those surveyed) felt less energetic, fatter, depressed, and
tense. Because running was discontinued for various reasons however, no information
was provided on the deprivation length or dependence level.
In a comprehensive review of the exercise-deprivation literature, Szabo (1995)
concluded that the majority of studies have found that exercisers report withdrawal
effects during exercise-deprivation periods. The studies reviewed varied markedly on the
number and age of participants, the mode and amount of exercise, the length of
deprivation, and the instruments and methodology used. These inconsistencies among
studies make it difficult to understand the effects of exercise deprivation on psychological
well-being. It has been shown that exercise deprivation leads to negative feelings, but as
to what degree or what point an individual is likely to experience deprivation sensations
is not understood (Szabo, 1998). Withdrawal symptoms may vary from person to person
depending on their level of exercise dependence. In other words, it may take different
individuals a day versus a week to experience the same withdrawal symptoms.
Furthermore, Szabo (1995) commented that recruitment of participants is the greatest
limitation for exercise-deprivation studies because exercise-dependents are not likely to
participate in such a study. Szabo (1995) also noted that most of the studies have
examined runners. Finally, the majority of research has used a single measurement when
gathering data on subjective feelings during deprivation. This creates a limitation
because one measurement may not portray the feelings of an exercise deprived
In summary, Szabo (1998) brought to the forefront the difficulty of studying
exercise deprivation. He stated that the longer the deprivation period, the less likely
exercise-dependent participants are willing to partake in these studies. The research
design of exercise deprivation should not be limited to experimental designs (Szabo,
1997). Thus, more opportunistic (e.g., taking advantage of situations such as bad weather
and not being able to exercise) and descriptive studies (e.g., qualitative research) need to
be underaken. Szabo (1998) concluded that studying exercise-deprivation is not
"hopeless"; however, there are confounds to control and different methodologies that
must be examined.
Prior Research Limitations
Examining the prior literature that exists on exercise-deprivation is critical in
order to design functional future studies. The following deprivation literature limitations
must be taken into account when designing a study. First, the majority of exercise-
deprivation research has examined runners (Chan & Grossman, 1988; Conboy, 1994;
Crossman et al., 1987; Morgan, 1979; Morris et al., 1990; Sachs & Pargman, 1979;
Thaxton, 1982). Exercise dependence, however, is not confined to this activity. Limited
research has been conducted using various sports and activities such as swimming
(Crossman et al., 1987), aerobic exercise, cross-training, and weight lifting (Gauvin &
Szabo, 1992; Szabo, Frenkl, & Caputo, 1996).
Second, the measurements and parameters used to define exercise dependence
have varied across studies. Inclusion criteria have ranged from participants' weekly
exercise frequency (Baekeland, 1970; Carmack & Martens, 1979), duration of weekly
exercise (Gauvin & Szabo, 1992; Thaxton, 1982), mileage per week (Blumenthal et al.,
1984; Chan & Grossman, 1988; Harris, 1981a; Robbins & Joseph, 1985; Wittig et al.,
1992), and psychological inventories (Conboy, 1994; Estok & Rudy, 1986; Szabo et al.,
1998; Thaxton, 1982; Wittig et al., 1992). Third, it is difficult to generalize study
findings due to the varying lengths of exercise-deprivation (Baekeland, 1970; Conboy,
1994; Thaxton, 1982). For example, research has shown that 24-hours of exercise-
deprivation results in withdrawal symptoms; however, deprivation periods have lasted as
long at one month (Baekeland, 1970; Conboy, 1994; Thaxton, 1982).
Fourth, control groups have been either nonexistent or inconsistent. Researchers
have failed to use control groups when examining deprivation (Morris et al., 1990; Wittig
et al., 1992). In studies where control groups were present, the sample has been
inconsistent with the experimental deprivation group. Additionally, recruiting exercise-
dependent participants willing to forgo exercise is difficult (Szabo, 1998). There is also
concern whether advertising for a deprivation study may confound the results because the
participants can mentally prepare for the time off (Bamber, Cockerill, & Carroll, 2000;
Fifth, there has been a lack of physiological assessments of exercise behavior.
This results in an over reliance on self-report measures of exercise behavior. The
majority of studies lack physiological measurements because the tests: (a) are time
consuming, (b) require expensive equipment to conduct, and (c) require experimenter
expertise (Conboy, 1994; Crossman et al., 1987; Mondin et al., 1996; Morris et al., 1990;
Szabo, 1996). Physiological measures are deemed more desired than self-reported
exercise because they are more accurate (American College of Sports Medicine [ACSM],
Finally, measures of mood were often assessed by unstandardized assessment
tools. This could lead to incorrect measurement and interpretation (Baekeland, 1970;
Estok & Rudy, 1986; Gauvin & Szabo, 1992; Wittig et al., 1992). Similarly, instruments
to assess exercise dependence were also prone incorrect measurement and interpretation
(Hausenblas & Symons Downs, 2001).
The first purpose of this study was to examine mood states during a 24-hour
exercise-deprivation period between individuals who were high and low on exercise
dependence symptoms. The second purpose was to examine differences concerning
motives to exercise between the high and low exercise-dependent groups. The third
purpose was to determine if differences between groups existed in regards to the
physiological fitness measurements.
Improving upon previous exercise-deprivation research: (1) a standardized
multidimensional measure of exercise dependence was administered to determine the
high and low dependence groups; (2) standardized measures were used to assess mood
states and exercise; (3) age matched groups were recruited; (4) physiological measures of
exercise behavior were conducted; (5) a manipulation check for the deprivation period
was undertaken; and (6) participants were screened for eating disorders to control for the
confound of primary versus secondary dependence.
Significance of the Study
There are several reasons why this study is of significance to the exercise
psychology field. First, using standardized and valid measures has eliminated many of
the limitations in prior studies. Second, this study examines one of the criteria (i.e.,
withdrawal) for exercise dependence, when modified in accordance to the DSM-IV.
Determining whether deprivation exists and if so, to what extent, will assist in learning
whether this is a criterion of exercise dependence. Finally, examining if differences
occur in motivation to exercise between low and high exercise-dependent individuals
may provide insight to bridging the gap between the two extreme behaviors of exercise.
First, based on previous literature (Szabo, 1995, 1998) it was hypothesized that
individuals high in exercise dependence symptoms would report more mood disturbance
following a 24-hour exercise-deprivation period compared to the low exercise-dependent
group. Specifically, individuals in the high exercise dependence group would indicate a
decrease on the Feeling Scale and the Exercise-Induced Feeling Scale, an increase on the
State Anxiety Inventory, and an increase on the number of negative affect symptoms
following a 24-hour deprivation period compared to the low exercise dependence group.
These results are expected due to the physical and psychological reliance upon exercise
of the high exercise dependence group. Additionally, previous research has supported the
tendency of exercise-dependent individuals to report an increase in negative mood when
exercise deprivation occurs (Baekeland, 1970; Conboy, 1994; Gauvin & Szabo, 1992;
Mondin et al., 1996; Morris et al., 1990; Thaxton, 1982; Wittig et al., 1992).
Second, it was hypothesized that more internal and external motivation to exercise
would be reported by the high exercise-dependent group versus the low dependent group.
In particular, high exercise dependence individuals will score higher on the intrinsic
motivation (IM) subscales of the Exercise Motivation Inventory (Li, 1999; i.e., IM to
know, IM toward accomplishments, and IM to experience stimulation) compared to the
low exercise dependence group. Also the high exercise dependence group will score
higher on the Extrinsic Subscales of Introjection and Identification than the low exercise
dependence group. Elevated levels of intrinsic motivation and self-determination are
likely to be present in individuals with high exercise dependence compared to those with
low levels of exercise dependence because of the discipline necessary to exercise
habitually. Li (1999) demonstrated higher levels of intrinsic motivation and self-
determination in frequent exercisers versus less frequent exercisers.
Third, it was hypothesized that the exercise-dependent group would have a higher
VO2 compared to the low exercise dependence group. Also, the high exercise dependence
group would have a lower percent body fat than the low exercise-dependent group.
Evidence for this hypothesis can be found in the literature on physical activity and
physiological adaptations. That is, fit individuals exhibit a higher estimated VO2 due to
increased efficiency of the cardiovascular system and a lower body composition from
more muscle tissue versus fat tissue than unfit individuals (ACSM, 2000).
Participants were 42 volunteer female students (M = 21.18, SD = 3.54) recruited
from sport and fitness classes at the University of Florida. Because sex differences exist
for mood (APA, 1994), only females were recruited as to not confound the results with
gender reporting biases. Also, participants were restricted to the ages of 18 to 25 because
of the positive relationship between age and mood disturbance (APA, 1994). Sample size
was determined using Potvin's (1996) power tables for repeated measure designs and was
based on an alpha of .05, using an average correlation between repeated measures of .50,
and a meaningful effect size of .70. The sample size of 40 (20 in each group) resulted in a
power that exceeded .85 for the main effects and .75 for the interaction effects of interest
for the analyses conducted.
Drive for Thinness Subscale. The Drive for Thinness Subscale is a 7-item subscale from
the Eating Disorder Inventory-2 (Garner, 1991) that measures the pursuit of thinness
which is the cardinal feature of eating disorders. Statements assessing excessive concern
with dieting, preoccupation with weight, and fear of gaining weight are given with a 6-
point Likert Scale anchored at the extremes with 1 (never) and 6 (always; see Appendix
B). The Drive for Thinness Subscale has demonstrated adequate reliability and validity
(Garner, 1991). The reliability in the current study was good (alpha = .79).
Exercise Dependence Scale. The Exercise Dependence Scale (Hausenblas & Symons
Downs, 2001) is a 30-item scale that assesses exercise dependence symptoms based on
the criteria from the DSM-IV diagnosis for substance dependence (APA, 1994).
Respondents indicate their agreement to the statements using a 6-point Likert Scale
anchored at the extremes with 1 (never) and 6 (always). A high score indicates more
exercise dependence symptoms. The scale also discriminates between exercise
dependent and nondependent individuals. Questions refer to beliefs and behaviors that
have occurred within the last three months (see Appendix C). Examples of items include
"I feel anxious if I cannot exercise," "I organize my life around exercise," and "I often
exercise with more intensity than I intend." Preliminary research has shown the scale to
have adequate reliability and validity (Hausenblas & Symons Downs, 2001).
Exercise Motivation Scale. The Exercise Motivation Scale (Li, 1999) is based on the
self-determination model and contains eight subscales reflecting three types of
motivation: motivation, external motivation, and internal motivation (See Appendix D).
Motivation represents a lack of motivation whereby individuals do not perceive a
connection between their action and the outcome of their action. External motivation is
behavior that is a means to an end. There are four types of extrinsic motivation: (a)
extrinsic regulation which is behavior controlled by external sources such as material
rewards, (b) introjected regulation which represents behavior reinforced through internal
pressures from the individual such as guilt, (c) identified regulation which is behavior
that is internally motivated and performed out of choice, and (d) integrated regulation
which is a fully self-determined form of extrinsic motivation.
Internal motivation represents engaging in a behavior for the pleasure and
satisfaction derived from it. There are three types of internal motivation: (a) internal
motivation to learn which is behavior performed to explore and understand, (b) internal
motivation to accomplish which is behavior engaged in for the pleasure and satisfaction
of accomplishing something, and (c) internal motivation to experience sensations which
represents behavior done to experience a stimulating sensation. Thirty-one statements are
given in which an answer from strongly disagree (1) to strong agree (6) must be selected.
A high score indicates a greater endorsement of motivation. The Exercise Motivation
Scale is a valid and reliable measure of exercise motivation (Li, 1999).
Reasons for Exercise Inventory. Silberstein, Striegel-Moore, Tinko, and Rodin (1988)
developed the Reasons for Exercise Inventory to assess the motives for engaging in
exercise. This inventory consists of 24 statements on a 7-point Likert Scale ranging from
"not at all important" (1) to "extremely important" (7; see Appendix E). The inventory
contains the following seven subscales: Weight Control, Fitness, Mood, Health,
Attractiveness, Enjoyment, and Tone. Higher subscale scores indicate a strong motive to
exercise. Examples of statements include: "To be slim," "To improve flexibility,
coordination," "To cope with stress, anxiety," "To improve my overall health," "To
improve my appearance," "To have fun," and "To improve my overall body shape."
Adequate scale reliability and validity have been established (Silberstein et al., 1998).
Leisure-Time Exercise Questionnaire. The Leisure-Time Exercise Questionnaire is a
three-item scale developed by Godin, Jobin, and Bouillon (1986). Respondents are asked
to recall the number of exercise sessions greater than 20 minutes and the intensity of each
session undertaken within a typical week. The exercise intensities are mild, moderate,
and strenuous (see Appendix F). To determine the metabolic equivalents (METS), the
frequency of exercise is multiplied by the activity intensity. Each intensity level is
appointed a number that is then multiplied by the frequency. The frequency of mild
exercise is multiplied by three, moderate by five, and strenuous by nine. The values for
mild, moderate, and strenuous exercise are added to determine the total exercise index. A
high score represents a greater level of activity. This questionnaire has displayed
adequate reliability and validity (Jacobs, Ainsworth, Hartman, & Leon, 1993).
Exercise-Induced Feeling Inventory (EFI). The EFI is a 12-item questionnaire developed
by Gauvin and Rejeski (1993) that examines feeling states sensitive to acute aerobic
exercise. There are four subscales: Positive Engagement, Revitalization, Tranquility, and
Physical Exhaustion. A response ranging from "do not feel" (1) to "feel very
strongly"(5) is provided by the participants (see Appendix G). Higher subscale scores
indicates the most prominent feeling state, while a low score suggests little or no
occurrence of a particular feeling state. This inventory has demonstrated adequate
psychometric properties (Gauvin & Rejeski, 1993).
Profile of Mood States pomsS). The trait version of the POMS developed by McNair,
Lorr, and Droppleman (1971) was used to assess mood. The 65-item questionnaire
contains the following six subscales: tension, fatigue, vigor, anger, confusion, and
depression (see Appendix H). The participants complete the questionnaire based on
"how you are feeling in general." The scale is anchored at 0 (not at all) to 4 (extremely).
The higher a score, the more prominent the corresponding mood state. The POMS has
demonstrated adequate reliability and validity (McNair et al., 1971).
Feeling Scale. The Feeling Scale, developed by Hardy and Rejeski (1989), is a single
item measure designed to assess current feeling state. Respondents rate their current
feeling state by choosing a number on the 11-point bipolar scale anchored at the extremes
with -5 (very bad) and 5 (very good; see Appendix I). This scale has demonstrated
adequate reliability and validity (Hardy & Rejeski, 1989).
State-Trait Anxiety Inventory (STAI). State anxiety was assessed using Form Y-1 of the
STAI which assesses cognitive aspects of state anxiety (Speilberger, Gorsuch, Lushene,
Vagg, & Jacobs, 1983). The 20-item questionnaire asks respondents to indicate on a 4-
point scale anchored at the extremes with (1) never to (4) always how they currently feel
(see Appendix J). High scores reflect greater state anxiety. Interitem consistency, factor
analytic validity, convergent validity, and divergent validity are adequate (Speilberger et
Positive and Negative Affect Schedule (PANAS). The PANAS is a 20-item affective
measure that assesses positive and negative affect (Watson, Clark, & Tellegen, 1988).
Positive affect (n = 10 items) represents the extent to which a person feels enthusiastic,
active, interested, excited, and alert, for example. High positive affect is characterized by
high energy, full concentration, and pleasurable engagement; whereas low positive affect
includes feelings of sadness and lethargy. Negative affect (n = 10 items) reflects the
degree to which individuals feel distressed, upset, guilty, irritable, and nervous, for
example. Low negative affect represents a state of calmness or serenity, while high
negative affect represents unpleasurable engagement and subjective distress. Participants
indicate their endorsement of the affective items on a 5-point Likert scale anchored at the
extremes with 1 (very slightly or not at all) and 5 (extremely; see Appendix K). The
PANAS has been found to consist of two dominant and relatively independent
dimensions that have adequate psychometric properties (Watson et al., 1988).
Body composition. Body composition was assessed using two methods: Body Mass
Index (BMI) and skinfold measurements. First, weight was assessed using a Health-O-
Meter Scale (Technical Services Inc) in kilograms with a built in measurement stick to
determine height. BMI was determined using height and weight (kg/m2) calculations.
The BMI method has a standard error of 5% (ACSM, 2000).
Second, skinfold measurements, which are based on the principle that the amount
of subcutaneous fat is proportional to the total amount of body fat, were used to estimate
body fat percent. Body composition was estimated using the 3-site (triceps, suprailiac,
and thigh) skinfold method (Lange Calipers; Cambridge Scientific Industries, Cambridge,
MD). The following equation was used based on the three sites to estimate body density:
[1.099421- (.0009929*sum of body composition) + (.0000023* sum of body
composition2) (.0001392 age)] (Jackson & Pollock, 1985). A population specific
formula for white females, age 20-80, was used to convert body density to percent body
fat [((5.03/Body Density) 4.59) 100]; Heyward & Stolarczyk, 1996). Using the
correct techniques and equations, the skinfold method has demonstrated adequate
accuracy (SD = 3.5% body fat; ACSM, 2000).
Heart rate. Heart rate was measured using a Polar Favor heart rate monitor (Polar CIC
Inc., New York) that was strapped around the participant's chest during physical activity.
The ACSM (2000) states that there is a linear relationship between heart rate and VO2
and that heart rate is a valid method to estimate exercise intensity. Heart rate monitors
are an accurate and reliable measure of exercise intensity (ACSM, 2000).
Rating of Perceived Exertion (RPE). Rating of perceived exertion was determined by the
Borg scale (Borg & Noble, 1974). The Borg scale is an interval scale ranging from 6
(very light) to 20 (maximal exertion) and mimics the amount of exertion the heart is
handling (see Appendix L). The larger the number, the higher the heart rate and thus,
more exertion is being given to the task. RPE is a reliable measure to indicate exercise
tolerance of individuals (ACSM, 2000) and has demonstrated adequate psychometric
properties (Borg & Noble, 1974).
Maximal Exercise Test. To estimate oxygen uptake, subjects performed a maximal
graded exercise test using the Bruce protocol (see Appendix M). This test measures the
cardiorespiratory endurance of an individual. The Bruce Protocol is the most common
treadmill procedure and is best suited for younger and active individuals (ACSM, 2000).
An equation to estimate VO2 max that assesses the total amount of time the
participant was able to follow the treadmill protocol was used. The following equation
calculates VO2 in relative terms, accounting for body weight, so comparisons can be
made among the participants: VO2max (ml/kg/min) = 14.8 1.379 (time in minutes) +
0.451 timee) 0.012 timee; Foster, et al., 1984). This equation has been proven to be
reliable and valid (Heyward & Stolarczyk, 1996). The higher the estimated VO2max, the
better the cardiorespiratory fitness level.
Pre-Participation Questionnaire (PPQ). The PPQ is a self-report measures which
identifies medical concerns that would place the participant at increased risk for
complications during the exercise test (see Appendix N). Significant medical conditions
and the risk factors for heart disease, as determined by the ACSM, are formulated in
questions in the PPQ (ACSM, 2000). Participants who indicated "yes" to any question
were asked about the condition. If the condition did not put the individuals at risk, in
conjunction with not interfering with any of the variables that the test is determining, the
subject was allowed to participate in the study.
24-Hour History. The 24-hour history questionnaire determines if the participant is
physically able to perform the maximal exercise test and that their behavior in the past
24-hours would not interfere with the results (e.g., drinking alcohol or caffeine,
smoking). This questionnaire inquires about sleep, food intake, caffeine and alcohol
consumption, and physical activity (see Appendix O). It also asks for a general overall
feeling at the present time including: excellent, very good, good, neither good or bad,
bad, very bad, or terrible. This questionnaire has been shown to be reliable for following
pretest instructions suggested by the ACSM for maximal exercise testing and minimizes
the chances of a poor test (ACSM, 2000).
Manipulation Check A manipulation check was undertaken in an attempt to elicit a true
response from the participants when questioned about refraining from exercise for the
past 24 hours. Two Q-tips with colored tips were placed in labeled baggies for the
participant to perform a "swab" test. After the swab test was completed, the Q-tips were
placed back into the baggie and sealed with tape. This was a bogus test used in an
attempt to determine whether the participants adhered to the deprivation period.
Prescreen. Participants were prescreened using the Drive for Thinness Subscale of the
EDI-2 (Garner, 1991) and the Exercise Dependence Scale (Hausenblas & Symons
Downs, 2001). The scales were administered to sport and fitness classes. Those students
scoring above 14 points on the Drive for Thinness were disqualified from the study
because they are deemed at-risk for an eating disorder (Gamer, 1991). This was done to
prevent a confound in the results by using primary and secondary exercise dependent
individuals. Students scoring in the upper and lower 33% on the Exercise Dependence
Scale were eligible to participate in the study. These individuals were then phoned and
asked if they would be willing to participate in the study. An appointment was made for
the first visit and instructions were given to ensure proper readiness for the maximal
exercise test to be completed. Participants were told that the first appointment consisted
of a fitness assessment and that they needed to wear shorts, a loose fitting top, and
exercise shoes. They were also informed that at least three hours prior to the appointment
they should not eat heavily or consume any type of tobacco, alcohol, or caffeine, and to
not exercise prior to the fitness test.
Visit 1. On the first visit, the participant was informed of the study procedures and
completed the informed consent (see Appendix P) and PPQ. After the PPQ was reviewed
by the experimenter, a series of questionnaires (i.e., Exercise Motivation Scale, Reasons
for Exercise Inventory, Leisure-Time Exercise Questionnaire, POMS) along with
questions pertaining to the participants training schedule were completed. A 24-hour
recall was given to the participant to ensure that the directions given over the phone were
A fitness assessment was then administered at Living Well, the faculty and staff
fitness center for the university. The assessment included height and weight, body
composition measurements (Lange Calipers), as well as the maximal treadmill test.
Weight was measured on a Health-O-Meter Scale in kilograms and height was assessed
in inches using a built-in measurement stick on the scale. Resting blood pressure and
pulse were assessed while the participant was sitting. Blood pressure was measured
using a Labtron Graham-Field Patricia 03-180 Series blood pressure cuff. The resting
pulse was measured for 30 seconds utilizing the radial artery. Body composition was
estimated using the three site measurement of the tricep, suprailiac, and thigh. Skinfolds
were taken on the right side of the body at anatomical locations described by the ACSM
(2000). Each site was measured three times and the average of each site was recorded.
To determine the percent body fat, a population specific formula was used for white
females, age 20-80 (ACSM, 2000). Although the ages in the study ranged from 18-25,
this equation was the most appropriate compared to other equations. In order to not
confound the results, the participants were asked during the assessment if they currently
trained or competed in athletics and any positive responses resulted in dismissal of that
participant. None of the participants met the criteria to be dismissed.
A symptom-limited Bruce protocol was used as the maximal exercise test to
measure estimated VO2max. Two examiners administered this test according to ACSM
guidelines (ACSM, 2000). One examiner was responsible for administering the test,
taking blood pressure, and controlling the grade. The other examiner recorded the heart
rate and blood pressure, controlled the speed, and recorded rating of perceived exertion
measurements. Heart rate was continuously monitored by wearing a heart rate monitor.
The starting treadmill speed was set at 1.7 miles per hour with a 10% grade. Every three
minutes (one stage) the speed was increased and the grade consistently increased by two
percent. At the end of each minute, heart rate was recorded while blood pressure and
RPE were obtained and recorded at the end of every stage. RPE was determined by
holding a large chart in front of the subject and having them verbally express the number
corresponding to how they were feeling. The participant continued the test until
volitional fatigue or they requested to stop. Towards the end of the test, the participant
was verbally encouraged to continue as long as possible to make certain maximum effort
and heart rate were obtained. A cool down was performed to ensure heart rate returned to
pretest level by lowering the grade of the treadmill to zero and reducing the speed to a
normal walking pace.
The examiners were a graduate student trained in exercise testing according to the
ACSM guidelines (ACSM, 2000) and two undergraduate students majoring in exercise
science. The graduate student assumed the role of test administer and the undergraduate
students were trained according to ACSM guidelines and assisted in the testing.
Visit 2. Upon returning for the second visit, the participants were asked to complete a
series of questionnaires assessing mood (i.e., EFI, Feeling Scale, SAI, and PANAS). The
participants were then instructed to choose any cardiovascular exercise machine (i.e.,
Trackmaster TM210 treadmill, Stairmaster 4000PT, Concept II rower, Nordic Track,
Precor EFX546 cross trainer, Stairmaster Freerunner 5400ESS, or NuStep TRS3000
recumbent stepper) and exercise for 30 minutes at a specified heart rate. The specified
heart rate was determined by using the maximum heart rate obtained during the maximal
treadmill test. The heart rate reserve method using the Karvonen formula was utilized to
estimate heart rate at 70% of maximum heart rate. This formula is more accurate than
using the traditional maximum heart rate formula (220 age; ACSM, 2000). A
maximum heart rate of 70% was used to achieve a level both high and low scorers on the
Exercise Dependence Scale could achieve. The low scorers may have struggled to reach
70% while the high scorers would possibly have to refrain from going above 70% of their
estimated VO2 max.
The participants were given the same heart rate monitor used during the
assessment to wear during their cardiovascular session. The watch displaying the pulse
rate was strapped to the equipment so both the participant and the examiner could
monitor the heart rate. Heart rate was recorded every 5 minutes throughout the 30
minutes to ensure the proper effort. The participants were allowed to choose their piece
of exercise equipment. This created a more ecologically valid assessment and reduced
potential anxiety from using either novel or infrequently used equipment (Focht, 2000).
When the 30 minutes was concluded, the participants cooled down on the equipment until
heart rate was within their normal resting range. The same questionnaires that were
completed at the beginning of the session were readministered. The participants were
instructed not to engage in physical activity for the next 24 hours and to limit lifestyle
exercise (e.g., walking/biking to campus).
The participants were also told that upon arrival for the next visit a "swab" test
would be performed that would determine if they exercised during the past 24 hours.
This was a bogus test designed as a manipulation check. It was developed to make the
subject believe that the researchers could determine if she adhered to the deprivation
period. Participants were instructed to return in 24 hours.
Visit 3 (Follow Up). On the last visit, the participants were instructed to perform the
swab test by swiping the inside of each cheek with Q-tips that were colored at one end.
The samples were then put into baggies that were labeled with the participant's subject
number. The series of questionnaires that were completed during the second visit (i.e.,
EFI, PANAS, Feeling Scale, and SAI) were readministered following the swab test. The
participant was then told to sit for 30 minutes in a quiet room where reading and
homework was allowed. The questionnaires were then redistributed and the order of the
questionnaires were counterbalanced to avoid response bias. Upon completion of the
questionnaires, the participants were asked to honestly answer whether they performed
any physical activity within the previous 24 hours. All participants were given a
summary sheet of their fitness assessment results and an e-mail address where they could
contact the experimenter to obtain a copy of the study results.
Over the course of the study, four participants discontinued with the study. Two
of the participants completed the first visit and never returned for the remaining visits
despite numerous phone calls. One of the participants who did not return after the first
visit was contacted by phone and said she did not have time to finish the study. The last
participant who did not finish the study completed the first two visits but had an
unexpected event occur and could not attend the last visit unless the deprivation period
was extended to 48 hours. In order to not confound the results, this participant was
dropped from the study. Of the individuals who discontinued the study, three of the four
were high exercise dependent group while only one represented the low exercise
Prior to examination of the hypotheses the reliability (i.e., internal consistency) of
the measures were undertaken.
Statistical analysis for the first purpose included two separate 2 (group) x 4 (time)
analysis of variance (ANOVA) on the SAI and Feeling Scale with repeated measures on
the time factor. The independent variable was the group (high vs. low dependence) and
the dependent variables were the SAI and the Feeling Scale. Effect sizes were computed
to determine the meaningfulness of the results (Cohen, 1977).
Given the significant correlations between the EFI and PANAS subscales reported
in previous research (Gauvin and Rejeski, 1993), a 2 (group) x 4 (time) multivariate
analysis of variance (MANOVA) with repeated measures on the second factor were used
to examine group differences for the these scales. Group was the independent variable
and the EFI and the PANAS were the dependent variables. Univariate ANOVA with
Bonferonni adjustments were employed for follow-up analysis of the significant
multivariate effects. Effect sizes were computed to determine the meaningfulness of the
results (Cohen, 1977).
For the second purpose of comparing group differences on mood and reasons and
motives to exercise, three separate MANOVAs were conducted on the Reasons for
Exercise Inventory, Exercise Motivation Scale, and POMS. Follow-up univariate
ANOVA was completed to determine the significant multivariate effects. To determine
the meaningfulness of the results, effect sizes were computed.
For the third purpose, one-way ANOVAs were conducted to determine group
differences on the demographic and fitness variables. The independent variable was the
group and the dependent variables were age, height, weight, BMI, body composition,
cardiovascular endurance, and the Leisure-Time Exercise Questionnaire. Bonferonni
adjustments were made when necessary and effect sizes were computed to determine the
meaningfulness of the results.
Assumptions For Statistical Analyses
Before the analyses were conducted, the data were examined to ensure the
appropriate statistical assumptions were met. The ANOVA has four assumptions: 1)
scores in each group are independent, 2) data are on a parametric scale, 3) the population
is normally distributed, and 4) homogeneity of variance exists (Agresti & Finlay, 1997;
Gravetter & Wallnau, 2000). Assumptions 1 and 2 were met because the groups were
independent and the data was on either an interval or a ratio scale. To determine if
assumption 3 was met (the population was normally distributed) examination of the
skewness of the data was undertaken. The homogeneity of variance was checked using
the Box's M test. In the unlikely chance that the test was less than .05, the largest group
variance was examined to ensure the variance was not more than two times the smallest
group variance in order to not have an impact on the F value (Vincent, 1995).
Assumptions for the MANOVA include: 1) random sampling and independent
scores, 2) normal distribution in each group, 3) homogeneity of covariance matrices, and
4) the relationship for the dependent variables are linear. Random sampling is a design
issue and is not violated in this study. To determine normal distribution, univariate
normality was assessed. If normality was not met, the MANOVA is robust to moderate
violations as long as the violation is created by skewness. The assumption of
homoscedasticity is violated by nonnormality of one of the variables or one variable may
have a relationship to the transformation of another variable. Box's M is the statistical
test that is commonly used to check for violations of homoscedasticity. However, if the
test is significant, it will not prove fatal to the analysis. Although unlikely, if
homoscedasticity is violated, Pillai's Trace will be used to interpret multivariate results.
If homoscedasticity is not violated, Wilks' Lambda will be used to interpret the results.
To determine linearity, bivariate scatterplots will be completed with an elliptical shape
indicating the presence of normality (Vincent, 1995).
Repeated measures ANOVA also has four assumptions: 1) observations are
independent, 2) the population distribution is normal, 3) equivalent covariances, and 4)
homogeneity of variance. The first and second assumptions are experimental design
errors and have not been violated. Examining skewness will assess the variance
distribution. The fourth assumption is violated if the effect of the treatment is not
consistent for all the participants. If homogeneity of variance is violated the Mauchly's
Test of Sphericity will be significant and the Greenhouse Geisser adjustment will be used
to adjust the degrees of freedom to obtain a more accurate F statistic (Potvin, 1996).
Assumptions for repeated measures MANOVA include: 1) random sampling and
independent scores, 2) normal distribution in each group, 3) homogeneity of variance-
covariance, and 4) no outliers exist in the data set. The first two assumptions are design
issues and have not been violated. The steps taken when a violation of an assumption
occurs are identical for a MANOVA.
A manipulation check was undertaken to determine if the participants adhered to
the instruction of not exercising for 24 hours. On the final visit participants were asked
to respond either yes or no to the following self-report question: "Have you exercised in
the last 24 hours"? No participants indicated that they had exercised in the previous 24
hours. Thus, it was assumed that the subjects adhered to the deprivation protocol based
on their positive response to the "swab tesf' and their indication that they had not
exercised in the past 24 hours.
Reliability of Measures
Internal consistency scores (i.e., Chronbach's alpha) were determined for each
measure prior to examining the study purposes (Nunnally, 1978). Because the alpha
value is inflated as the number of variables increases, there is no set interpretation as to
what is an acceptable value (George & Mallery, 2001). The general rule of thumb for
reliability interpretations displayed in Table 3.1 was used to interpret the alpha levels of
the study measures (George & Mallery, 2001).
The Exercise-Induced Feeling Inventory Subscales had acceptable to excellent
reliabilities (range = .73 to .93) for all administrations (see Table 3.2). The State Anxiety
Inventory evidenced excellent reliabilities for the four assessments (range = .91 to .94).
Rule of Thumb for Reliability of Measurements Interpretation
Alpha Value I
The reliability scores for the Exercise Motivation Scale Subscales were acceptable to
excellent ranging from .77 to .93 (see Table 3.7). The Reasons for Exercise Inventory
Subscales displayed acceptable to good reliabilities ranging from .69 to .89 (see Table
3.8). The internal consistency scores for the Profile of Mood States (POMS) Subscales
were also adequate with a range of .75 to .92 (see Table 3.9).
Internal Consistency Scores for the Exercise-Induced Feeling Inventory Subscales,
Positive and Negative Affect Scale (PANAS), and the State Anxiety Inventory
Exercise N= 42 Rest N= 42
Pre V Post V Pre V Post V
Exercise-Induced Feeling Inventory
Positive Engagement .86 .87 .87 .86
Revitalization .81 .79 .93 .87
Tranquility .90 .73 .86 .91
Physical Exhaustion .85 .75 .91 .89
Positive Affect .89 .91 .94 .92
Negative Affect .86 .37 .70 .84
State Anxiety Inventory .93 .91 .92 .94
Finally, the Positive Affect Subscale of the PANAS had excellent internal
consistency across all assessments (range = .89 to .94; see Table 3.2). Similarly, the
Negative Affect Subscale of the PANAS had acceptable internal consistencies scores for
the pre-exercise (alpha = .86), pre-rest (alpha = .70), and post-rest assessments (alpha=
.84). In contrast, the post-exercise Negative Affect Subscale was inadequate (alpha =
.37). A factor analysis was attempted to determine if the negative affect subscale factor
structure was related to the inadequate internal consistency. Due to the lack of item
variance (see Table 3.3), however, a factor analysis was unable to be performed. Thus,
the Negative Affect Subscale of the PANAS was eliminated from further analyses.
Implications of the low reliability of the Negative Affect Subscale will be discussed in
the next chapter.
Mean (M) and Standard Deviation (SD) Scores for the Negative Affect Subscale of the
Positive and Negative Affect Schedule (PANAS) for the Post-Exercise Assessment of the
Low and High Exercise Dependent (ED) Groups
Variable Low ED High ED Total
n=20 n=21 N=41
M SD M
1.29 .90 1.24
1.14 .65 1.07
1.14 .48 1.17
1.71 .96 1.61
1.14 .48 1.10
1.57 .93 1.41
1.62 .92 1.49
1.14 .48 1.05
1.24 .54 1.17
1.38 .74 1.34
or not at all, 5 = feeling very strongly.
Each participant was required to complete several self-report questionnaires that
represented 358 data points per person. Seven items from the entire data set were
missing, which represented 1.96% of the total item responses. Because of the small
percentage of missing data, mean replacements were computed for the missing variables
(George & Mallery, 2001).
Eating Disorder Screening
In order to distinguish primary from secondary exercise dependence the Drive for
Thinness Subscale of the Eating Disorder Inventory-2 was administered. No participants
who completed the Drive for Thinness Subscale evidenced scores in the pathological
range of 14 or greater (Garner, 1991). Furthermore, a one-way ANOVA demonstrated no
significant group differences on the Drive for Thinness Subscale between the high (M =
5.35, SD = 4.74) and low (M = 8.47, SD = 4.53) exercise dependent groups [F (1, 36) =
3.11, 2= .09].
The primary purpose of the study was to determine if group differences existed
for mood states following 24 hours of exercise deprivation. First, to examine if baseline
group differences existed on the pre-exercise assessment of Exercise-Induced Feeling
states a one-way ANOVA was conducted. No significant group differences on the
Positive Engagement [F (1, 39) = .12, p= .73], Revitalization [F (1, 39) = .03, p = .87],
Tranquility [F (1, 39) = .13, p = .72], and Physical Exhaustion [F (1, 39) = 1.91, = .17]
Subscales were found. Because there were no baseline group differences, these values
were not covaried in subsequent analyses. Second, the Mauchly's Test of Sphericity was
conducted to examine if the assumption of variance was violated. The test determined
the assumption was not met and therefore, the Greenhouse-Geisser adjustments were
needed. Third, a MANOVA with repeated measures on time (i.e., pre and post exercise,
pre and post quiet rest) was computed on the Exercise-Induced Feeling Inventory
Subscales. Significant main effects for time on the Physical Exhaustion [F (4, 36) = 2.95,
p = .05, w2 = .07], Positive Engagement [F (4, 36) = 15.99, p < .001, w2 = .29],
Revitalization [F (4, 36) = 24.91, p < .001, Y = .39], and Tranquility [F (4, 36) = 9.54, p <
.001, Y = .20] Subscales were evidenced. There was no significant main effect for group
[F (4, 36) = .69, p = .59, Y = .07] or interaction [F (4, 36) = 1.05, p = .40, Y = .36]
To determine where the significance difference occurred for time, post hoc
analysis was performed using the Bonferonni test. Participants reported significantly less
Physical Exhaustion following exercise compared to the pre-exercise and quiet rest
assessments. Positive Engagement and Revitalization were highest after the exercise
session compared to before exercise and before and after the rest period. Finally,
participants reported the most tranquility following quiet rest than before or after exercise
and before quiet rest (1 < .05).
To assess changes in mood an ANOVA with repeated measures on time was
undertaken for the Feeling Scale and the State Anxiety Inventory. Mean and standard
deviation scores for these measures are presented in Table 3.5. First, one-way ANOVAs
were undertaken to determine if baseline group differences existed on the Feeling Scale
and State Anxiety Inventory. Results revealed that there were no group differences on
the baseline assessment of the Feeling Scale [F (1, 39) = .004, p = .95] and the State
Anxiety Inventory [F (1, 39)= .01, p= .92].
Mean (M) and Standard Deviation (SD) Scores for the Exercise-Induced Feeling
Inventory Subscales Across the Assessment Periods for the Low and High Exercise
Dependent (ED) Groups
Positive Revitalization Tranquility Physical
M M SD
SD M SD
Low ED (n= 20)
Pre 6.50 3.14 5.05 2.56 7.15 3.18 3.55 2.68
Post 8.11 3.03 8.20 2.71 7.45 2.93 2.95 1.64
Pre 6.89 3.35 6.11 3.60 8.37 2.48 2.68 2.93
Post 7.11 3.73 5.68 3.76 9.26 2.51 3.63 3.50
High ED (n= 21)
Pre 6.81 2.62 5.19 2.75 6.81 2.77 4.81 3.11
Post 9.10 2.17 9.14 2.22 7.14 2.10 2.86 2.06
Pre 6.45 2.21 5.57 2.96 7.52 2.48 3.90 3.25
Post 6.33 2.15 5.10 2.49 8.48 2.68 4.29 3.20
Total (N= 41)
Pre 6.66 2.85 5.12 2.63 6.98 2.95 4.20 2.94
Post 8.63 2.63 8.68 2.48 7.29 2.51 2.90 1.84
Pre 6.67 2.79 5.82 3.25 7.93 2.48 3.33 3.12
Post 6.70 2.99 5.38 3.13 8.85 2.60 3.98 3.32
Note. A high score indicates prominent feelings while lower scores demonstrate little or no
occurrence of the feeling.
The Feeling Scale was then assessed for the assumption of sphericity using
Mauchly's Test. It was found that homogeneity of variance was present [X2 (5) = 10.35,
p = .07]. Although a repeated measures ANOVA indicated no significant main effect for
group [F (1, 3) = 12.98, p < .001, Y = .51] and interaction [F (1, 3) = 1.27, p= .29, Y =
.03], a significant main effect for time [F (1, 3) = 11.50, p < .001, Y = .23] was found.
Tukey's post hoc test revealed that more positive feelings were reported after the exercise
session than before the exercise session or before and after the rest session (p's < .05).
For the State Anxiety Inventory, Mauchly's Test of Sphericity revealed a lack of
homogeneity of variance [X2 (5) = 14.68, p = .01]. Because the assumption of sphericity
was not met the Greenhouse-Geisser method was used to adjust the degrees of freedom
for an accurate F-test statistic. Analysis revealed no significant main effect for time [F
(1, 2.39) = 2.16, p = .11, Y = .05], main effect for group [F (1, 2.39) = .45, p= .67, Y =
.16], or interaction [F(1, 2.39) = .45, p = .67, Y = .01].
Mean (M) and Standard Deviation (SD) Scores for the Feeling Scale and the State
Anxiety Inventory for the Low and High Exercise Dependent (ED) Groups
Feeling Scale State Anxiety Inventory
M SD M SD
Low ED (n = 20)
Pre 2.55 2.06 33.80 11.06
Post 3.60 1.54 30.15 8.57
Pre 2.84 1.54 30.84 5.83
Post 2.84 1.80 30.84 6.60
High ED (n = 21)
Pre 2.33 1.43 33.55 9.25
Post 3.57 1.40 31.95 7.57
Pre 2.05 1.60 34.14 9.86
Post 2.62 0.92 32.85 10.80
Total (N= 41)
Pre 2.44 1.75
Post 3.59 1.45
Pre 2.43 1.60
Post 2.72 1.40
Note. The Feeling Scale ranges from -5 (very bad) to 5
Inventory a higher score indicates more anxiety.
For the State Anxiety
Mean (M) and Standard Deviation (SD) Scores for the Negative and Positive Affect
Subscales of the PANAS for the Low and High Exercise Dependent (ED) Groups
Positive Affect Negative Affect
M SD M SD
Low ED (n = 20)
Pre 29.70 7.51 11.90 3.49
Post 34.42 8.27 10.70 1.26
Pre 28.45 10.04 10.40 .68
Post 25.95 10.78 10.40 1.19
High ED (n= 21)
Pre 29.86 7.30 13.38 5.55
Post 34.80 6.57 11.29 1.59
Pre 26.48 8.34 12.15 2.54
Post 25.29 6.96 11.86 2.83
Total (N= 41)
Pre 29.78 7.31 12.66 4.66
Post 34.62 7.35 11.00 1.45
Pre 27.44 9.15 11.29 2.05
Post 25.61 8.92 11.15 2.29
The secondary purpose was to compare reasons and motives to exercise and mood
states between the high and low exercise dependence groups. The Exercise Motivation
Scale, Reasons for Exercise Inventory, and POMS was analyzed using a MANOVA to
determine if group differences existed. For the Exercise Motivation Scale, the
MANOVA revealed significant group differences [Wilks' Lambda = .50, F (1, 35) =
4.39, p = .001]. The univariate ANOVA indicated that the high exercise dependence
group scored significantly higher than the low exercise dependent group for Introjected
Regulation [F (1, 42) = 9.90, p = .003, Y = .20], Identified Regulation [F (1, 42) = 18.22,
p < .001, Y = .30], Integrated Regulation [F (1, 42) = 24.91, p < .001, Y = .37], Internal
Motivation to Learn [F (1, 42)= 9.04, p = .004, Y = .18], Internal Motivation to
Accomplish [F (1, 42) = 26.37, p < .001, Y = .39], and Internal Motivation to Experience
Sensation Subscales [F (1, 42) = 30.64, p < .001, Y = .42] (see Table 3.6 for the mean and
standard deviation scores). There were no group differences for the Amotivation [F (1,
42) = 1.64, p = .21, Y = .04] or External Regulation Subscales [F (1, 42) = .25, p = .62, Y
Mean (M), Standard Deviation (SD), and Alpha Level Scores for the Exercise Motivation
Scale for the High and Low Exercise Dependent (ED) Groups
Low ED High ED Total Alpha
n=20 n=21 N=41
Variable M SD M SD M SD
Motivation 4.60 2.09 3.83 1.88 4.16 1.98 .92
External Regulation 8.85 4.49 8.21 3.99 4.18 1.99 .90
Introjected Regulation 11.70 4.03 15.50 3.96 13.77 4.38 .77
Identified Regulation 19.35 2.66 22.33 1.97 20.98 2.73 .78
Integrated Regulation 15.10 3.85 20.08 2.75 17.82 4.11 .88
IM to Learn 13.50 4.17 17.04 3.64 15.43 4.24 .93
IM to Accomplish 15.85 3.45 20.54 2.60 18.41 3.81 .81
IM to Experience 15.50 4.38 21.46 2.69 18.75 4.62 .93
Note. IM = Internal Motivation
A MANOVA examining the Reasons for Exercise Inventory revealed significant
group differences [Wilks' Lambda= .57, F (1, 37)= 3.99, p= .002, Y = .43]. Follow-up
univariate ANOVA demonstrated that the high exercise dependent group scored
significantly higher than the low exercise dependent group for Weight Control [F (1, 43)
= 4.78, p = .03, Y = .10], Mood [F (1, 43)= 13.59, p = .001, Y = .24], Attractive [F (1, 43)
= 14.07, p = .001, Y = .25], Enjoyment [F (1, 43)= 4.78, p < .01, Y = .25], and Tone
Subscales [F (1, 43) = 6.10, p = .02, Y = .12] (see Table 3.8 for mean and standard
deviation scores). Group differences were not found for the Fitness [F (1, 43) = 2.16, p =
.15, Y = .05] and Health Subscales [F (1, 43) = 2.05, p= .16, Y = .05].
Mean (M), Standard Deviation (SD), and Alpha Level Scores for the Reasons for
Exercise Inventory for the Low and High Exercise Dependent (ED) Groups
Low ED High ED Total Alpha
n =21 n = 24 N = 45 Reliability
Variable M SD M SD M SD
Weight Control 13.57 4.69 16.13 3.08 14.93 4.08 .69
Fitness 22.76 3.49 24.29 3.47 23.58 3.53 .83
Mood 15.14 6.13 20.79 4.05 18.16 5.82 .85
Health 22.76 5.00 24.54 3.26 23.71 4.21 .89
Attractive 12.10 4.85 16.58 3.09 14.49 4.56 .88
Enjoyment 8.81 3.54 13.13 4.04 11.11 4.35 .83
Tone 13.05 5.13 16.33 3.76 14.80 4.70 .83
Note. A high score indicates an important reason to exercise.
Finally, a MANOVA was performed for the POMS subscales and results
indicated no mood differences between the groups [Wilks' Lambda = .79, F (1, 37)
1.60, p = .18, = .17]. When examining the mean scores, however, the high exercise
dependent group reported more feelings of depression, anger, confusion, tension, and
fatigue compared to the low group (see Table 3.8). In contrast, the low exercise
dependent group reported more vigor than the high exercise dependent group.
Mean (M), Standard Deviation (SD), and Alpha Level Scores for the Profile of Mood
States (POMS) Subscales for the High and Low Exercise Dependent (ED) Groups
Low ED High ED Total Alpha
n =20 n =21 N = 41 Reliability
Variable M SD M SD M SD
Depression 6.00 5.71 8.13 8.53 7.11 7.32 .91
Anger 4.14 4.23 6.30 7.36 5.27 6.10 .92
Vigor 13.76 6.21 12.00 5.95 12.84 6.07 .86
Confusion 6.00 3.52 7.17 4.54 6.61 4.08 .75
Tension 5.81 4.11 9.17 5.19 7.57 4.95 .84
Fatigue 6.14 3.62 7.70 5.79 6.95 4.88 .83
POMS total 41.86 21.46 50.48 30.92 46.36 26.88
Note. A lower score indicates less prominent feelings.
The final purpose was to determine if group differences existed on the
physiological and psychological demographic characteristics. One-way ANOVA's were
conducted for age, height, weight, BMI, percent body fat, cardiovascular endurance,
Exercise Dependence Scale, and Leisure-Time Exercise Questionnaire scores. No
significant differences were found for age [E (1, 45) = 3.75, p = .06, Y = .08], weight [F
(1, 45) = 3.19, p= .08, Y = .07], height [F (1, 45) = .50, Y = .01], and BMI [F (1, 45) =
2.9, p = .10, Y = .06] (see Table 3.10 for mean and standard deviation scores). In
contrast, significant group differences were evidenced for percent body fat [F (1, 45) =
5.77, = .02, w = .12] and cardiovascular endurance [F (1, 45) = 13.18, p = .001, Y =
.24] with the high exercise dependent group having less body fat and a higher level of
cardiovascular endurance compared to the low exercise dependent group. Significant
differences [F (1, 45) = 284.91, p < .001, Y = .87] occurred between the high and low
groups on the Exercise Dependence Scale with the high group displaying a higher score.
Mean (M) and Standard Deviation (SD) Scores
Dependent (ED) Groups on Physiological Data
Exercise Dependence Scale 38.33 8.14
Age 22.24 3.88
Height (m) 1.54 .04
Weight (kgs) 66.30 15.02
BMI 25.68 5.87
Percent Body Fat 23.72 7.36
Cardiovascular Endurance 30.07 5.74
for the Low and High Exercise
High ED Total
(n = 24) (N = 45)
M SD M SD
115.54 19.5 79.51 41.79
20.25 3.0 21.18 3.54
1.53 .05 1.54 .05
59.96 8.18 62.92 12.15
23.45 2.47 24.49 4.48
19.65 3.59 21.55 5.96
36.38 5.88 33.44 6.58
To assess group differences for the Leisure-Time Exercise Questionnaire, a one-
way ANOVA was performed for mild, moderate, strenuous, and total exercise. It was
found that the high exercise dependent group scored significantly higher than the low
exercise dependent group on the mild [E (1, 43)= 4.17, p= .05, Y = .17], moderate [F (1,
43) = 4.70, p = .04, Y = .09], strenuous [E (1, 43) = 44.06, p < .001, Y = .52], and total
Leisure-Time Exercise Questionnaire scores [F (1, 43) = 31.94, p < .001, Y = .45] (see
Mean (M) and Standard Deviations (SD) Scores for the Leisure-Time Exercise
Questionnaire for the Low and High Exercise Dependent (ED) Groups
Low ED High ED Total
n=21 n= 24 N=45
M SD M SD M SD
Mild 7.05 6.25 11.50 7.90 9.48 7.46
Moderate 10.75 10.67 17.92 11.12 14.66 11.38
Strenuous 12.15 14.39 37.31 10.74 25.88 17.71
Total 29.95 18.87 66.73 23.44 50.01 28.19
Note. A larger score represents a higher level of physical activity.
Although not a specific hypothesis, examining the concurrent validity of the
exercise measure (Leisure-Time Exercise Questionnaire) in relation to the maximal
exercise test results was undertaken (see Table 3.12). The correlation revealed that Total
Leisure-Time Exercise was significantly related to the subcomponents including mild,
moderate, and strenuous exercise. Similarly, cardiovascular endurance was correlated
with strenuous and total leisure-time exercise. The correlation between cardiovascular
exercise and reported leisure-time exercise was highest for strenuous exercise which is
understandable. Thus, self-reported strenuous exercise was linearly associated with a
higher cardiovascular endurance level.
Pearson Correlations for Cardiovascular Endurance and the Leisure-Time Exercise
Cr Mild Moderate Strenuous Total
Cardiovascular endurance 1.0 .21 .21 .46* .42*
LTEQ Mild 1.0 .47* .29 .60*
LTEQ Moderate 1.0 .42* .78*
LTEQ Strenuous 1.0 .85*
* = p < .01 (2-tailed).
Researchers have consistently found that regular physical activity contributes
positively to physical and psychological health (USDHHS, 1996, 2000). For example,
physical inactivity is directly linked to increased morbidity and mortality for at least six
chronic conditions: coronary heart disease, hypertension, obesity, diabetes, osteoporosis,
and mental health disorders (USDHHS, 2000). Furthermore, a lack of regular exercise is
estimated to contribute to 250,000 deaths per year in the United States (Pate et al., 1995).
Despite the health benefits of exercise, facilitating physical activity adoption and
adherence remains a challenge because approximately 40% of American adults are
sedentary (USDHHS, 2000) and 50% of sedentary adults beginning an exercise program
will dropout within 6 months (Dishman, 1994).
Despite the epidemic of sedentariness in North America there are a very small
number of individuals who exercise excessively. Exercise dependence is defined as a
craving for exercise that results in uncontrollable excessive physical activity and
manifests in physiological symptoms, psychological symptoms, or both (Hausenblas &
Symons Downs, 2001). It has been suggested that a cardinal symptom or criterion of
exercise dependence is withdrawal or deprivation effects (Szabo, 1995). Withdrawal
symptoms are manifested by either the characteristic withdrawal symptoms for exercise
(e.g., anxiety, irritability) or the same amount of exercise is engaged in to relieve or avoid
withdrawal symptoms (APA, 1994). The general purpose of this thesis was to examine
the psychological effects of withdrawal in high and low exercise dependent individuals.
This chapter discusses the findings for the first, second, and third purposes as well as
outlines the study limitations and future research directions.
The first purpose of this study was to examine mood states during a 24-hour
exercise deprivation period between individuals who were high and low on exercise
dependence symptoms. It was hypothesized that the high exercise dependent group,
compared to the low exercise dependent group, would report more negative affect, an
increase in state anxiety, and a decrease in positive feeling states after 24-hours of
exercise deprivation. Contrary to the hypothesis, there were no differences between the
high and low exercise dependence groups on the mood measures following the
The lack of group difference in mood between the groups may be due to the
sensitivity of the mood instruments, the length of the deprivation period, or the prescribed
exercise intensity. First, although 24-hours has been reported to elicit withdrawal
symptoms (Conboy, 1994; Modin et al., 1996), a longer time period may be required in
certain populations to elicit deprivation effects. Second, exercise intensity before the 24-
hour deprivation period was prescribed at 70% of the estimated VO2 max. It is plausible
that the high exercise dependent group was exercising at a lower intensity than normal;
and the low exercise dependence group was exercising at a higher intensity than normal.
Thus, self-selected exercise intensities, which elicit a more ecologically valid
environment, may be more likely to produce withdrawal effects (Focht, 2000). Future
research is need to examine withdrawal effects with self-selected versus prescribed
Third, consistent with past researchers, it was found that the PANAS was not
sensitive to mood changes during an acute bout of exercise and quiet rest (Gauvin,
Rejeski, Norris, & Lutes, 1997; Nemanick & Munz, 1994; Rejeski, Gauvin, Hobson, &
Norris, 1995). That is, participants reported high positive affect and low negative affect
at all the assessments which created either a ceiling or floor effect. Furthermore, the
Negative Affect Subscale of the PANAS had inadequate internal consistency and was
dropped from further analysis. To explore the potential "cause" of the poor alpha value,
inter-item consistency, mean, and standard deviation values of each item were examined.
A factor analysis was also attempted to determine if the factor structure of the Negative
Affect Subscale was indeed unidimensional as proposed by Watson, Clark, and Tellegan
(1988). However, because of the lack of variance within each item of the subscale, a
factor analysis could not be performed (Tabachnick & Fidell, 2001).
Other researchers have had similar "problems" with the PANAS (Diener &
Emmons, 1985; Egloff, 1998; Killgore, 2000). For example, Diener and Emmons (1985)
found that affect differed greatly depending on the time frame the questionnaire was
answered (e.g., same day or a week later) and concluded that positive and negative affect
were independent of each other rather than closely correlated. Similarly, Egloff (1998)
found the PANAS to contain two independent subscales while Killgore (2000)
discovered a more appropriate model by splitting the negative affect into two factors.
Thus, the current results coupled with prior researchers' findings, questions the utility of
the PANAS in assessing affect in repeated measures designs.
The second purpose was to examine differences concerning motives to exercise
between the high and low exercise dependent groups. Group differences for motivation
to exercise was found between the low and high exercise dependent groups. Consistent
with the hypotheses, the high exercise dependent individuals were more intrinsically and
extrinsically motivated to exercise compared to the low exercise dependent group
according to the Exercise Motivation Scale. Specifically, for extrinsic motivation, the
high exercise dependent individuals were more likely to exercise due to internal pressures
(e.g., guilt, anxiety) and perform exercise out of choice to achieve personal goals. This is
supported by Summers and colleagues (1982) who found marathon runners participated
in running because of the personal challenges involved. When external sources of
motivation have been internalized it creates pressure and the individual feels the need to
exercise for aesthetic reasons or to prevent feelings of embarrassment if not at a peak
fitness level (Pelletier et al., 1995).
For intrinsic motivation, results revealed that high exercise dependent individuals
were more likely to exercise in order to: 1) explore and understand health or fitness, 2)
experience the pleasure and satisfaction of accomplishing something, and 3) to
experience a stimulating sensation. These results coincide with the motivation necessary
for the high exercise dependent group to continually exercise in the face of obstacles and
is consistent with the findings that exercise dependent individuals report high barrier self-
efficacy (Hausenblas & Symons Downs, 2001). In short, these results are in agreement
with the findings of Li (1999) who found that frequent exercisers displayed higher levels
of intrinsic motivation to learn, intrinsic motivation to experience sensations, integrated
regulation, and identified regulation.
Different reasons for exercising were also apparent between the high and low
exercise dependent groups. Compared to the low exercise dependent group, the high
exercise dependent group was significantly more likely to exercise to control their
weight, to prevent negative mood, to be attractive to the opposite sex, for the enjoyment
of physical activity, and to tone muscles. The finding that the exercise dependent group
reported exercising to prevent negative mood is consistent with the DSM-IV criterion of
withdrawal effects. That is, the exercise dependent individual engages in physical activity
to either prevent negative mood or to promote a positive mood.
It was also found that the fitness and health reasons to exercise were not
significantly different between the high and low exercise dependent groups. This finding
is interesting considering that there were significant fitness level differences between the
two groups. That is, the high exercise dependent group was more fit than the low
exercise dependent group. Fitness and health reasons to exercise may not have differed
between the two groups because the high exercise dependent group already displayed a
high level of fitness and health while the low exercise dependent group had little concern
for health and fitness, possibly because they lack knowledge of the benefits of exercise.
In regards to the mood traits, which was measured by the Profile of Mood States,
it was found that the low exercise dependents reported significantly more tension than the
high exercise dependents. Although statistically nonsignificant the high exercise
dependent group reported more depression, anger, confusion, and fatigue compared to the
low exercise dependent group. These results imply that high exercise dependent
individuals rely on exercise to control or prevent moods such as depression, anger, and
confusion. The high exercise dependent group scored higher on fatigue than the low
exercise dependent group which is consistent with their behavior because the high
exercise dependent group engaged in excessive physical activity.
The third purpose was to determine if group differences existed in regards to the
physiological fitness measurements. After the appropriate computations were completed,
data analysis was undertaken. Consistent with the hypothesis the high exercise
dependent group had significantly higher estimated V02 max than the low exercise
dependence group. As the amount of exercise training increases, VO2 max also
increased. Also, in accordance with the prediction the high exercise dependent group had
significantly lower heart rates than the low exercise group due to adaptations within the
body that occur with training.
Another consequence of regular physical activity is a lower body composition.
As exercise level increases fat levels usually decrease and the amount of muscle may
increase or stay the same (ACSM, 2000). Because of the physiological adaptations,
individuals who are more physically active tend to have lower body composition.
Consistent with the hypothesis it was found that the high exercise dependent group had
lower percent body fat than the low exercise dependent group. In contrast to the
hypothesis no significant group differences were evidenced for BMI. The disparity in
results between the BMI and percentage body fat, both measures of body composition,
may be due to the fact that percentage body fat takes muscle mass into account and is
more accurate. That is, although BMI is a good indicator of body composition in
epidemiological studies, BMI has a high predictive error, especially with physically
active populations (ACSM, 1996). Thus, it is conceivable for regular exercisers to have a
mesomorphic body type, which may result in higher BMI, despite the fact that they have
low percent body fat.
In accordance with a lower body composition, it was also found that the high
exercise dependent individuals engaged in more mild, moderate, and strenuous leisure-
time exercise than the low exercise group (USDHHS, 2000). Taking into account the
various theories explaining the processes of exercise dependence, the difference in the
frequency of strenuous activity for the high exercise dependent group, compared to the
low exercise dependent group, may support the sympathetic arousal hypothesis
(Thompson & Blanton, 1987). The need for an increase in exercise is one of the
components of the hypothesis which can be illustrated by the increasing variance in the
intensity levels of exercise determined by the Leisure-Time Exercise Questionnaire.
Although the current study improved on many aspects of the past literature, it is
not without limitations. Study limitations include: gender, age, self-report measures,
deprivation period, school interference, and equations for physiological calculations.
First, the use of females and an age constraint limits the generalization of the results.
Future researchers are encouraged to examine exercise deprivation in males and
participants of varying ages. Second, self-report measures have an inherent limitation
due to socially desirable responses (Sallis & Owen, 1998). Questionnaires that require
recall also have an inherent bias because of improper recollection of events.
Additionally, the participants may perceive and report aspects of themselves differently
than others may report about them. Thus, future research is needed using both direct and
indirect (i.e., self-report) measures of mood and exercise behavior when examining the
effects of exercise deprivation.
Third, the deprivation period was only 24 hours. Although 24-hours has been
shown to elicit deprivation sensations, it is plausable that a longer deprivation period
could exhibit more drastic changes in mood (Modin et al., 1996). However, when a time
frame longer than 24-hours is used, the likelihood of high exercise dependent individuals
participating is less likely and the dropout rate increases (Szabo, 1995).
Fourth, testing was done during varying times of the school semesters. Thus, the
time throughout the semester when the individual testing (e.g., midterm exams) was
conducted may have effected the participants' mood and exercise behavior. For example,
if a participant completed the study during an exam period, mood and anxiety may have
Fifth, the use of various equations to determine physiological data may result in
measurement errors (ACSM, 2000). For example, there are various equations and
methods to determine body density (ACSM, 1996). The skinfold method of body
composition has a range of error one to two percent higher than hydrostatic weighing, the
"gold standard" to assess body composition. To determine percent body fat, an age-
specific equation was utilized for 18 to 80-year-old females (ACSM, 2000). Some
participants were younger than 18 which may of caused their percent body fat to be
slightly higher or lower than if an equation was used for 16-20 year olds. A maximal
exercise test was performed to evaluate cardiorespiratory fitness and to obtain a maximal
heart rate. Although maximal exercise tests are highly correlated with VO2 max tests, the
former is less accurate than the latter. Gas analyzers were not used in the current study,
however, to determine if maximal effort was reached due to its high cost, need for special
facilities, and scheduling complexities.
Sixth, heart rate range for each participant was used to determine their exercise
intensity of 70% of estimated VO2 max. Although this allows for uniformity, participants
were required to exercise at a prescribed versus self-selected exercise intensity. Recent
research has found that self-selected exercise represents a more ecological valid
environment and is more conducive to illiciting mood states from acute exercise bouts
Finally, although the participants were instructed to not exercise and limit their
lifestyle activity for the 24 hour deprivation period, their physical activity behavior was
not directly monitored. In an attempt to have the participants answer honestly in regards
to their physical activity behavior over the 24 hour deprivation period they performed the
"swab test", which was used as a manipulation check. Also, the last questionnaire
completed inquired if the participant had exercised within the last 24-hours.
A variety of hypotheses have been developed in the exercise dependence
literature to explain the process of dependence. The findings of this study provide
evidence to partially support the affective states hypothesis. The Reasons for Exercise
Inventory revealed that high exercise dependent individuals participate in exercise to
control mood more than low exercise dependent individuals. Whether the high exercise
dependent group exercises to experience positive moods or to avoid negative moods, the
desire to regulate mood is the important aspect. Although not all of the hypotheses to
explain exercise dependence were examined, results from the current study provide
partial support for the affective states explanation.
Future Directions and Conclusions
Despite the increase in research examining exercise dependence and deprivation
in recent years (Hausenblas & Symons Downs, 2001) the antecedents and consequences
of exercise dependence remains unknown. Thus, there are several areas for future
research in this area. First, experimenting with various exercise deprivation time periods
could determine whether there is a threshold of deprivation until withdrawal symptoms
are experienced (APA, 1994; Szabo, 1995). Second, the existence of exercise
dependence across a variety of activities is unknown. For example, depriving an
individual from an activity they enjoy may provoke greater withdrawal effects.
Third, examining the motivation of individuals in different phases of their
commitment and enjoyment to exercise will aid in finding if a point exists where
motivation becomes intrinsic versus extrinsic. Research on changes in motivation to
exercise is crucial to increasing exercise adherence. Reasons why individuals exercise or
are sedentary could further be explored to understand the differences in values and beliefs
that low and high exercise dependent individuals give to exercise.
Fourth, strong conceptual and theoretical frameworks should guide the
development of study hypotheses. For example, the General Theory of Addictions (see
Jacobs, 1986) predicts that people with chronically abnormal arousal states (either high or
low), who also tend to respond to feelings of inferiority and rejection by flight into denial
and fantasy, are at the greatest risk for acquiring a dependency. This theory has been
applied to gambling, eating, and drinking addictions. Research is needed to further
examine its utility for exercise dependence.
Finally, the majority of the research examining the causal mechanisms of exercise
dependence has been correlational (Hausenblas & Symons Downs, 2001). Longitudinal,
ecological momentary assessments, qualitative, and experimental studies are required to
enhance our understanding of the etiological mechanisms of exercise dependence.
In conclusion, the present study has explored differences in low and high exercise
dependent individuals and found that the groups are distinct in certain areas. Motivation
and reasons to exercise, along with physiological data such as heart rate,
cardiorespiratory fitness, and body composition were different between the two groups.
In examining mood though, the groups were similar even after the deprivation period.
Insdlullonml Review Board 9A PsIcholgy Bldg
(jGeesville, FL 32611-2250
Phone: (.52) 392-0433
Fax; (352) 392-9234
hrt','welt ortn uL cdu.'ltrirbfl
TO: Ms. Amy Hagan
FROM: C, Michael Levy, Chair
Unii ersiry of Fiorida
Institutional Review Board
S.UBJECT: Approval of Phrlocol U 2000 154
TITLE: Exercise Deprivation and Mood
I am pleased to advise you thatr he University of Florida Institutional Review Board has recommended
approval of this protocol. Based on its review, the UFIRB determined thai riis research presents no more
than minimal risk to participants. Given your protocol, it is essential that you obtain signed Jucurmenisaion
of informed consent from each participant. Enclosed is the dated, IRB-approved informed consent to be
used when recruiung participants for the research,
If you wish to make any changes to this protocol, including ite need to increase the number of participants
authorized, you must disclose your plans before you implement them so that the Board can assess their
impact on your protocol. In addition, you must report to the Board any unexpected complication s hal
affect your participants.
If you have not completed this protocol by 24-.Feb-20j I, please telephone our office _9.-i0433 ), and we
will discussthe renewal process with you.
It is important that you keep your Deparaneni Chair informed abuuL the 5taltu ofthis research protocol.
cc; Dr. Heather Hausenblas
DRIVE FOR THINNESS SUBSCALE
Instructions. Using the scale provided below, please complete the following
questions as honestly as possible. For each item, decide if the item is true about you
never (1), rarely (2), sometimes (3), often (4), usually (5), or always (6). Please place
your answer in the blank space provided after each statement.
1 2 3 4 5 6
Never Rarely Sometimes Often Usually Always
1. I eat sweets and carbohydrates without feeling nervous.
2. I think about dieting.
3. I feel extremely guilty after overeating.
4. I am terrified of gaining weight.
5. I exaggerate or magnify the importance of weight.
6. I am preoccupied with the desire to be thinner.
7. If I gain a pound, I worry that I will keep gaining.
EXERCISE DEPENDENCE SCALE
Instructions. Using the scale provided below, please complete the following questions
as honestly as possible. The questions refer to current exercise beliefs and behaviors that
have occurred in the past 3 months. Please place your answer in the blank space
provided after each statement.
1 2 3 4 5 6
1. I am unable to reduce how long I exercise.
2. I exercise to avoid feeling stressed.
3. I spend a lot of time exercising.
4. I exercise longer than I intend.
5. I exercise despite recurring physical problems.
6. I continually increase my exercise intensity to achieve the desired
7. I organize my life so that there is always time for exercise.
8. I exercise longer than I expect.
9. I am unable to reduce how often I exercise.
10. I exercise to avoid feeling fatigued.
11. I would rather exercise than spend time with family/friends.
12. I think about exercise when I should be concentrating on school/work.
13. I exercise longer than I plan.
14. I spend a great deal of time in exercise-related activities.
15. I exercise when injured.
16. I exercise to avoid feeling irritable.
17. I exercise to avoid feeling anxious.
18. I spend most of my free time exercising.
19. I continually increase my exercise frequency to achieve the desired
20. I am unable to reduce how intense I exercise.
21. I exercise to avoid feeling restless.
22. I exercise despite persistent physical problems.
23. A great deal of my time is spent exercising.
24. I exercise to avoid feeling tense.
25. I choose to exercise so that I can get out of spending time with family/friends.
26. I exercise more often than I intend.
27. There is no better way to spend my time than to exercise.
28. I continually increase my exercise duration to achieve the desired
29. I exercise more often than I plan.
30. Exercise is my only recreational activity.
EXERCISE MOTIVATION SCALE
WHY ARE YOU CURRENTLY PARTICIPATING IN THIS ACTIVITY?I
Direction: Please read each of the statements listed below and indicate how strongly you
agree or disagree with each statement by circling the appropriate response to the right of
the statement. Use the following response categories:
Strongly disagree Disagree Moderately disagree Moderately agree
(SD) (D) (MD) (MA)
Agree Strongly agree
1. For the pleasure it gives me to experience
positive sensations from the activity.
2. For the satisfaction it gives me to increase
my knowledge about this activity.
3. Because other people believe that it's a
good idea for me to exercise.
4. Because I must exercise to feel good
5. Because I believe that regular exercise is
a good way to enhance my overall
6. Because it is consistent with what I value.
7. I can't understand why I am doing this.
8. Because I feel pressure from others to
9. Because I think that exercise allows me
to feel better about myself.
SD D MD MA A SA
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
WHY ARE YOU CURRENTLY PARTICIPATING IN THIS ACTIVITY?
SD D MD MA A SA
10. For the pleasure I experience while
learning about this activity.
11. For the satisfaction I feel when I get into
the flow of this activity.
12. Because I feel I have to do it.
13. To satisfy people who want me to
14. Because exercising is an important aspect
of how I perceive myself.
15. For the pleasure of understanding this
16. I have no idea.
17. For the pleasure of mastering this activity.
18. Because I think it is a good thing for my
19. For the pleasure I experience when I feel
completely absorbed in the activity.
20. For the satisfaction I feel while I try to
achieve my personal goals during the
course of this activity.
21. Because I would feel guilty if I did not
take the time to do it.
22. Because I value the way exercise allows
me to make changes in my life.
23. It is not clear to me anymore.
24. Because I think exercise contributes to
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
WHY ARE YOU CURRENTLY PARTICIPATING IN THIS ACTIVITY?
25. To comply with expectations of others
26. For the enjoyment that comes from how
good it feels to do the activity.
27. Because I enjoy the feelings of discovering
more about this activity.
28. Because I enjoy the feelings of improving
through participating in this activity.
29. Because I feel that changes that are taking
place through exercise are becoming part
30. For the pleasure I experience while trying
to become the person I want to be.
31. Because I would feel ashamed if I was not
doing anything to improve my current
SD D MD MA A SA
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
REASONS FOR EXERCISE INVENTORY
People exercise for a variety of reasons. When people are asked why they exercise, their
answers are sometimes based on the reasons they believe they should have for exercising.
What we want to know are the reasons people actually have for exercising. Please
respond to the items below as honestly as possible. To what extent is each of the
following an important reason that you have for exercising? Use the scale below, ranging
from 1 to 7, in giving your answers.
1. To be slim
2. To lose weight
3. To maintain my current weight
4. To improve my muscle tone
5. To improve my strength
6. To improve my endurance, stamina
7. To improve my flexibility, coordination
8. To cope with sadness, depression
9. To cope with stress, anxiety
10. To increase my energy level
11. To improve my mood
12. To improve my cardiovascular fitness
13. To improve my overall health
14. To increase my resistance to illness and disease
15. To maintain my physical well-being
16. To improve my appearance
17. To be attractive to members of the opposite sex
18. To be sexually desirable
19. To meet new people
20. To social with friends
21. To have fun
22. To redistribute my weight
23. To improve my overall body shape
24. To alter a specific area of my body
LEISURE-TIME ACTIVITY QUESTIONNAIRE
Instructions. This is a scale which measures your leisure-time exercise (i.e. exercise that
was done during your free time such as intramural sports-NOT your sport/fitness class).
Considering a typical week, please indicate how often (on average) you have engaged in
strenuous, moderate, and mild exercise more than 20 minutes during your free time?
1. Strenuous exercise: heart beats rapidly (e.g. running, basketball, jogging, hockey,
squash, judo, roller blading, vigorous swimming, vigorous long distance bicycling,
vigorous aerobic dance classes, heavy weight training)
How may times per typical week do you perform strenuous exercise for 20 minutes or
2. Moderate exercise: not exhausting, light sweating (e.g. fast walking, baseball, tennis,
easy bicycling, volleyball, badminton, easy swimming, popular and folk dancing)
How may times per typical week do you perform moderate exercise for 20 minutes or
3. Mild exercise: minimal effort, no sweating (e.g. easy walking, yoga, archery, fishing,
bowling, horseshoes, golf)
How may times per typical week do you perform mild exercise for 20 minutes or
EXERCISE-INDUCED FEELING INVENTORY (EFI)
Instructions: Use the following scale to indicate the extent to which each word below
described how you feel at this moment in time. Record your answer in the blank space
0 = Do Not Feel
1 = Feel Slightly
2 = Feel Moderately
3 = Feel Strongly
4 = Feel Very Strongly
PROFILE OF MOOD STATES
Instructions: Below is a list of words that describe feelings people have. Please read
each one carefully. Circle ONE number that corresponds to the degree which best
describes HOW YOU HAVE BEEN FEELING DURING THE PAST WEEK
0 = Not at all
1 = A little
2 = Moderately
3 = Quite a bit
4 = Extremely
1. Friendly ............ 0 1 2 3 4
2. Tense.............. 0 1 2 3 4
3. Angry .............. 0 1 2 3 4
4. W orn out............ 0 1 2 3 4
5. Unhappy............ 0 1 2 3 4
6. Clear-headed ......... 0 1 2 3 4
7. Lively .............. 0 1 2 3 4
8. Confused ............ 0 1 2 3 4
9. Sorry forthingsdone. .0 1 2 3 4
10. Shaky ............. 0 1 2 3 4
11. Listless ............ 0 1 2 3 4
12. Peeved ............ 0 1 2 3 4
13. Considerate........ 0 1 2 3 4
14. Sad ............... 0 1 2 3 4
15. Active............. 0 1 2 3 4
16. Onedge........... 0 1 2 3 4
17. Grouchy........... 0 1 2 3 4
18. Blue.............. 0 1 2 3 4
19. Energetic .......... 0 1 2 3 4
20. Panicky............ 0 1 2 3 4
21. Hopeless........... 0 1 2 3 4
22. Relaxed............ 0 1 2 3 4
23. Unworthy .......... 0 1 2 3 4
24. Spiteful ............ 0 1 2 3 4
25. Sympathetic........ 0 1 2 3 4
26. Uneasy............ 0 1 2 3 4
27. Restless ............ 0 1 2 3 4
28. Unable to concentrate. 0 1 2 3 4
29. Fatigued........... 0 1 2 3 4
30. Helpful............ 0 1 2 3 4
31. Annoyed .......... 0 1 2 3 4
32. Discouraged ........ 0 1 2 3 4
33. Resentful ........... 0 1 2 3 4
34. Nervous ........... 0 1 2 3 4
35. Lonely ........... 0 1 2 3 4
36. Miserable .......... 0 1 2 3 4
37. Muddled .......... .0 1 2 3 4
38. Cheerful ............0 1 2 3 4
39. Bitter ............. 0 1 2 3 4
40. Exhausted .......... 0 1 2 3 4
41. Anxious........... 0 1 2 3 4
42. Ready to fight ....... 0 1 2 3 4
43. Good-natured ...... .0 1 2 3 4
44. Gloomy............ 0 1 2 3 4
45. Desperate .......... 0 1 2 3 4
46. Sluggish ........... 0 1 2 3 4
47. Rebellious .......... 0 1 2 3 4
48. Helpless ............0 1 2 3 4
49. W eary............. 0 1 2 3 4
50. Bewildered ......... 0 1 2 3 4
51. A lert.............. 0 1 2 3 4
52. Deceived .......... 0 1 2 3 4
53. Furious............ 0 1 2 3 4
54. Efficient........... 0 1 2 3 4
55. Trusting............ 0 1 2 3 4
56. Fullofpep......... 0 1 2 3 4
57. Bad-tempered ....... 0 1 2 3 4
58. Worthless.......... 0 1 2 3 4
59. Forgetful ......... 0 1 2 3 4
60. Carefree ........... 0 1 2 3 4
61. Terrified ........... 0 1 2 3 4
62. Guilty ............. 0 1 2 3 4
63. Vigorous........... 0 1 2 3 4
64. Uncertain about things 0 1 2 3 4
65. Blushed............ 0 1 2 3 4
Using the scale below, please indicate how you feel right now, at this moment
-5 -4 -3 -2 -1 0 1 2 3 4 5
very bad bad faily bad neutral fairly good good very good
STATE-TRAIT ANXIETY INVENTORY
Directions: A number of statements which people have used to
describe themselves are given below. Read each statement and
then circle the appropriate number to the right of the statement
to indicate how you feel right now, that is, at this moment.
There are no right or wrong answers. Do not spend too much o 0
time on any one statement but give the answer which seems to > m
describe your present feelings best.
r H 00
1. I feel calm ................................ .. ......................... 1 2 3 4
2. I feel secure ........................................................... 1 2 3 4
3. I am tense ......................................................... ........ 1 2 3 4
4. I feel strained ...................................... ................. ..... 1 2 3 4
5. I feel at ease ............... .................... ............. ... ...... 1 2 3 4
6. I feel upset ......................................................... ....... 1 2 3 4
7. I am presently worrying over possible misfortunes ..................... 1 2 3 4
8. I feel satisfied ................................. ............... ..... ..... 1 2 3 4
9. I feel frightened ................................ ............ ............ 1 2 3 4
10. I feel comfortable .................. ...................................... 1 2 3 4
11. I feel self-confident .................. ............. ........ ........... .. 1 2 3 4
12. I feel nervous ...................................... ................. ...... 1 2 3 4
13. I am jittery ........... ........................................ ......... 1 2 3 4
14. I feel indecisive ................. ................ ............ ............. 1 2 3 4
15. I am relaxed .................. ........... ................ ... ........ 1 2 3 4
16. I feel content ...................................... ................. ...... 1 2 3 4
17. I am w worried ...................................... ............... ........... 1 2 3 4
18. I feel confused ..................... ............. ............. ............. 1 2 3 4
19. I feel steady .................................................... ........ 1 2 3 4
20. I feel pleasant ................. .............. ............... 1 2 3 4
POSITIVE AND NEGATIVE AFFECT SCHEDULE
This scale consists of a number of words that describe different feelings and emotions.
Read each item and then mark the appropriate answer in the space next to that word.
Indicate to what extent each word applies to you right now. Use the following scale to
record your answers.
1 = very slightly or not at all
2 = a little
3 = moderately
4 = quite a bit
5 = extremely
RATING OF PERCEIVED EXERTION
7 Very, very light
9 Very light
11 Fairly light
13 Somewhat hard
17 Very hard
19 Very, very hard
BRUCE PROTOCOL TEST REPORT
Time Speed Grade Heart Rate Blood Pressure RPE Comment
0-1 1.7 10 XXXXXXXXX
3-4 2.5 12 XXXXXXXXX
6-7 3.4 14 XXXXXXXXX
9-10 4.2 16 XXXXXXXXX
12-13 5.0 18 XXXXXXXXX
15-16 5.5 20 XXXXXXXXX
18-19 6.0 22 XXXXXXXXX
XXXX XXX XXXX XXXXXXX XXXXXXXXX XXX XXXXXXXXXXXXXXXXXXXX
IP 0 XXX