|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
THE INFLUENCE OF PARTICIPANT PREFERENCE AND PERCEIVED
DIFFICULTY ON EXERCISE ADHERENCE
KATIE ALLISON RICKEL
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
UNIVERSITY OF FLORIDA
Katie Allison Rickel
I would like to thank my mentor, Dr. Michael Perri, as well as Dr. Patricia
During, for their continued support and guidance in the compilation and completion of
This study was supported by National Heart, Lung, and Blood Institute Grant HL
58873 and by the National Institutes of Health for the University of Florida's General
Clinical Research Center Grant RR00082.
TABLE OF CONTENTS
ACKNOWLEDGMENT S .........__... ......._. .............._ iii..
LIST OF TABLES ........._.__........_. ...............v....
AB STRAC T ................ .............. vi
1 INTRODUCTION ................. ...............1.......... ......
2 MATERIALS AND METHODS .............. ...............5.....
Participants .............. ...............5.....
Proc edure s ................ ...............6............ ....
Recruitm ent .............. ...............6.....
Baseline assessment............... ...............
Experimental Design .............. ...............7.....
Exercise Intervention ................. ...............8.................
Exercise Logs................ ...............9..
Predictor M measures .............. ...............9.....
Criterion Measures ................. ...............10.................
Statistical Analyses ................. ...............11.................
3 RE SULT S ................. ...............12.......... .....
Participant Characteristics .............. ...............12....
Treatment Preference ................. ...............12.................
Perceived Difficulty ........._.__............. ... ... ........ ........ ...........1
Relation between Treatment Preference and Perceived Difficulty.............................14
Secondary Analy ses ................. ...............14.................
4 DI SCUS SSION ................. ................. 19......... ....
LIST OF REFERENCES ................. ...............25........... ....
BIOGRAPHICAL SKETCH .............. ...............28....
LIST OF TABLES
1 Baseline Demographic Characteristics of Participants Assigned to Preferred and
Non-Preferred Groups ........... ........... ...............16....
2 Adherence of Participants Assigned to Preferred or Non-Preferred Exercise .........16
3 Ratings of Perceived Difficulty with Participants Grouped by Intensity
Assignment .........._. ...... ..._ ...............17....
4 Ratings of Perceived Difficulty with Participants Grouped by Exercise
Preference .........._. ...... ..._ ...............17....
5 Standardized Beta Weights for Factors Contributing to Variance in Exercise
Adherence .........._. ...... ..._ ...............18....
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
THE INFLUENCE OF PARTICIPANT PREFERENCE AND PERCEIVED
DIFFICULTY ON EXERCISE ADHERENCE
Katie Allison Rickel
Chair: Michael G. Perri
Major Department: Clinical and Health Psychology
This study examined whether (a) assignment to preferred versus non-preferred
exercise treatments or (b) perceived difficulty of prescribed exercise predicted adherence
to walking programs that varied in intensity and frequency. This study also examined the
relationship between pre-exercise treatment preference and perceived difficulty.
Participants were 68 female and 84 male sedentary adults (65.1% Caucasian, 20.4%
African American, 6.6% Asian, 5.9% Hispanic, 2.0% unspecified; mean age = 49.2
years). Participants were randomly assigned in a 2X2 design to walk 30 min/day at an
intensity of either 45-55% or 65-75% of maximum heart rate reserve, at a frequency of
either 3-4 days/week or 5-7 days/week. Prior to notification of treatment assignment,
each participant indicated a preferred level (moderate versus high) of intensity and
frequency for exercise training. After notification of actual treatment assignment but
before beginning the exercise program, each participant rated the perceived difficulty
associated with completing the assigned treatment. Dependent measures were derived
from daily exercise logs completed over six months. Analyses were completed for the
proportion of total prescribed minutes completed and for the proportion of total
prescribed minutes completed while in a target heart rate training zone. Results showed
that participants assigned to their preferred treatment did not exhibit different rates of
frequency or intensity adherence than those assigned to non-preferred treatments.
Perceived difficulty showed a significant negative correlation with the proportion of
prescribed minutes of exercise completed (r = -.20, p = .04); however, this relationship
was mediated by the level of intensity assigned. Participants who preferred programs of
moderate frequency, moderate intensity walking made significantly higher ratings of
perceived difficulty of their assigned prescription than did participants who preferred
regimens of higher frequency and higher intensity (p = .03). Post-hoc analyses revealed
that baseline aerobic fitness (VOzmax adjusted for age and sex) was largely responsible for
this relationship, with less fit individuals preferring regimens of moderate intensity.
However, the association between baseline fitness and exercise adherence was not
significant (frequency adherence, p = .09; intensity adherence, p = .20). Collectively,
these findings suggest that program-related variables, such as the intensity level of
prescribed exercise, may be more influential in predicting exercise adherence than
person-related variables such as participant preference, perceived difficulty, or baseline
Sedentary behavior has been linked to the development of a variety of chronic
illnesses, including coronary heart disease, type 2 diabetes, hypertension, colon cancer,
and obesity (United States Department of Health and Human Services, 2002). Thus,
recognizing that an inactive lifestyle is a modifiable risk factor, public health officials
have made initiatives promoting increased physical activity a priority in recent years.
While the percentage of Americans reporting no leisure-time physical activity has
decreased from 31% in 1989 to 26% in 2004 (Centers for Disease Control and
Prevention, 2004), a great deal of research is still being dedicated to investigating ways to
bring this statistic down even more. Further, even when individuals do adopt exercise
programs, long-term adherence to increased physical activity is problematic (Marcus et
al., 2000). Thus, much of the current literature has focused on investigating the variables
that might be responsible for promoting or compromising adherence.
A potential exerciser's perceptions and preferences regarding a prescription of
structured exercise may influence subsequent adherence to the program. Social cognitive
theory (Bandura, 1997) and behavioral choice theory (Epstein, 1998) posit that
individuals engaging in a preferred modality of exercise will show better adherence than
those engaging in a non-preferred type of exercise. A variety of studies have investigated
the factors that participants may consider when making these appraisals of a
recommended exercise program. Preferences based on type of exercise have been
examined in both the U. S. (King et al., 2000) and in Australia (Booth et al., 1997);
studies concur that walking is the preferred type of exercise among both middle-aged and
older adults. Further, a population-based mail survey of 1,332 adults found that 53% of
respondents preferred unstructured physical activity, while only 31% preferred structured
activity (Salmon et al., 2003). This survey research has focused largely on inquiring
about the qualitative aspects of exercise. However, there is little data on preference for
the quantitative aspects of exercise, namely the preferred frequency and intensity levels
After determining what kinds of exercise individuals prefer, a number of
researchers have examined whether recommending preferred programs of exercise might
be an avenue through which exercise adherence may be bolstered. Some studies have
shown that assigning participants to their preferred modality of weight loss (Adams,
1998) and exercise (Parfitt & Gledhill, 2004) have led to more successful behavior
change, as well as to other benefits such as heightened self-efficacy and improved
psychological health. Yet other studies have failed to show improved adherence based on
tailoring treatment to individual preferences. For instance, when sedentary adults were
given either standardized or tailored physical activity prescriptions, participants did not
differ significantly in the amount of activity completed (Renjilian et al., 2001). In
addition, participants who were given a choice of activities in an exercise class did not
show higher rates of attendance than individuals assigned to a particular activity (Patten
et al., 2000). Hence, the influence of individuals' a priori preferences on subsequent
exercise adherence remains unclear.
Self-efficacy may be another cognitive factor that influences adherence. A number
of studies have found that higher self-efficacy during an exercise program (Brassington et
al., 2002) is associated with superior adherence. Further, self-efficacy has emerged as a
more powerful predictor of exercise maintenance than other psychological constructs
such as exercise intention, attitude, and affect (Courneya & McAuley, 1994). One
component of self-efficacy particularly important in exercise adherence is perceived
difficulty of the program (Garcia & Mann, 2004; Hagger et al., 2004). A number of
studies have investigated which factors a participant might consider when assessing the
difficulty of a recommended program of exercise. It has become clear that individuals
use a variety of criteria in making these appraisals. While the general population cites
lack of time (Godin et al., 1994), lack of facilities (Salmon et al., 2003), and laziness
(Godin et al., 1991) as significant barriers to adhering to an exercise regimen, older adults
endorse health status as an important determinant of difficulty (Booth et al., 1997).
Perceived difficulty during exercise and concurrent adherence have been related in
several studies. In population-based survey research, correlations between perceived
difficulty and exercise participation have been detected (De Bourdeaudhuij & Sallis,
2002). This pattern has been detected in intervention studies as well; for example, ratings
of perceived difficulty made after five weeks of participation in an exercise program
accounted for 3 8% of the variance in adherence to the 11 week supervised regimen.
While these findings may be important when working with individuals who are already
engaged in an exercise program, there may be clinical utility in assessing perceived
difficulty before exercise actually begins.
Some studies have found that baseline ratings of the broader construct of self-
efficacy can also predict subsequent exercise adherence. For instance, in a study of
overweight adults receiving counseling for changes in physical activity in a primary care
setting, baseline self-efficacy predicted positive changes in activity level after 4 months
and after 12 months of the intervention (Steptoe et al., 2000). However, other studies
have found that the dose of exercise recommended was more predictive of adherence
than was exercise self-efficacy assessed at baseline (Petrella et al., 2003). Thus, it
remains unclear whether baseline ratings of perceived difficulty, a component of self-
efficacy, might be predictive of subsequent exercise adherence.
This study examined the influence of participant preference and perceived
difficulty on adherence to walking regimens. We hypothesized that these cognitive
constructs, when assessed at baseline, would predict subsequent exercise adherence.
Specifically, we predicted that those participants assigned to preferred exercise programs
would show different rates of adherence than those participants assigned to non-preferred
programs of exercise. In addition, we hypothesized that participants' pre-exercise ratings
of perceived difficulty of their assigned program would be negatively related to their
adherence to that program over the course of the exercise trial. Lastly, we hypothesized
that there would be a relationship between these two pre-exercise cognitions; namely,
those participants preferring prescriptions of moderate intensity and frequency would
make higher ratings of perceived difficulty to their assigned programs than would those
participants preferring programs of high intensity and frequency. We attempted to
advance the existing literature by assessing these characteristics before the initiation of
exercise and by employing an obj ective and cumulative measure of adherence over a six-
MATERIALS AND METHODS
Participants were 152 healthy but sedentary women (n = 68) and men (n = 84)
participating in a 2-year study examining the effects of walking regimens that differed in
frequency and intensity. (The participant data considered in the present study comprise a
subset of the participants from the parent study (Perri et al., 2002). Because only the
fourth and fifth cohorts of this 5 cohort study completed the questionnaires from which
the data were extrapolated, only these participants' data could be considered. The
sub sample chosen did not differ from the total sample in terms of relevant baseline
characteristics.) Participants were adults, aged between 30 and 69 years, who endorsed a
sedentary lifestyle (defined as less than one hour per week of leisure-time physical
activity over the prior 12 months), had a body mass index (BMI) between 19 and 45 and
a resting blood pressure below 140 mm Hg systolic and 90 mm Hg diastolic. Participants
also consented to maintaining their typical physical activity patterns and to not beginning
any new exercise programs other than the one assigned in the study. Individuals who
failed to give informed consent, who refused to accept random assignment, or who were
already involved in another research study were excluded. Potential participants were
also excluded if, at screening, their medical history, clinical examination, or laboratory
results revealed any of the following conditions: coronary heart disease; chronic or
recurrent respiratory or gastrointestinal conditions; hypercholesterolemia (for women,
LDL-cholesterol > 190 mg/dl; for men, > 160 mg/dl) or use of antihyperlipidemia
medication; diabetes mellitus; fasting blood glucose > 110 mg/dl; cancer within 10 years;
or bone, muscle, or j oint conditions that would prevent walking on a regular basis.
Women in the study agreed to use appropriate birth control to avoid pregnancy during the
course of the study.
Participants were recruited through newspaper advertisements, direct mailings,
community-based presentations, and television and radio announcements. After
completing an initial screening by telephone, potentially eligible individuals attended an
informational meeting during which the investigators explained the study in detail and
obtained informed consent from the participants. After consenting, participants
completed a medical history questionnaire that inquired about illness and surgery history,
as well as about current medication use.
During a baseline testing session, each participant completed a symptom-limited
maximal exercise test to determine maximum heart rate. Prior to this treadmill test, blood
pressure and a standard 12-lead electrocardiogram were obtained in the sitting and
standing positions. Three ECG leads were monitored continuously during the exercise
test and a 12-lead ECG was repeated every 60 seconds of exercise and at 1, 3, and 5
minutes into recovery. Participants between the ages of 30 and 49 years performed the
standard Bruce protocol (Bruce et al., 1973). Participants between the ages of 50 and 69
years performed a modified Bruce protocol (Pollock & Willmore, 1990), with workloads
increasing by approximately three metabolic equivalents every 3 minutes. The tests
continued until participants reached voluntary maximal exertion or until they became
symptomatic or developed significant hemodynamic or electrocardiographic endpoints.
The Karvonen formula (American College of Sports Medicine, 1986) was used to
calculate heart rate reserve, HRees (HRees = HRmax resting HR).
Following stratification by age (30-49 years and 50-69 years), sex, and peak
oxygen consumption (VOzmax), participants were randomly assigned to one of the
following four exercise training conditions: (1) higher intensity, higher frequency; (2)
higher intensity, moderate frequency; (3) moderate intensity, higher frequency; (4)
moderate intensity, moderate frequency, or to a fifth group, (5) a physician's advice
group. The computer randomization was carried out using SAS's PROC PLAN (Spector
et al., 1985).
Intensity prescriptions. After determining each participant' s maximum and
resting heart rate during the baseline treadmill testing, each individual was given an
individualized training heart rate (THR). Participants in the moderate intensity groups
were assigned a THR corresponding to 45-55% HRees; participants in the higher intensity
were given a THR corresponding to 65- 75% HRees. These designations of 'moderate' and
'higher' correspond to the intensity ranges characterized as 'moderate' and 'hard' in the
1996 Surgeon General's Report (USDHHS, 1996). Participants were taught how to
monitor their heart rate during the orientation to the program, and they were instructed
how to adjust their walking pace to adhere to their intensity prescription, so that their
heart rate would fall in their assigned THR zone. In order to confirm self-reported heart
rate values during unsupervised walking sessions, each participant completed a 30 minute
walk during the weekly meetings on three separate occasions during weeks 5, 12, and
24 while wearing a monitor that recorded heart rate at one minute intervals.
Frequency prescriptions. With regard to the frequency prescriptions, participants
were instructed to walk in their THR zone for 30 minutes a day (either in a continuous
session or in up to three sessions, each at least 10 minutes in duration) for either 3-4 days
(in the moderate frequency conditions) or 5-7 days (in the higher frequency conditions).
Walking could be completed at home or at a work site or both. Further, each participant
was given a tailored program to gradually attain his or her full intensity and frequency
prescription over the first month of the study.
Following randomization, the participants in exercise training groups (1-4)
participated in an exercise intervention using walking as the mode of training for a total
of 24 months. The first six months consisted of the active treatment, and the last eighteen
months focused on maintenance. Participants were instructed to attend each of the 1 1
group intervention sessions offered over the first six months. The sessions were led by
counselors (counterbalanced by condition) with graduate training in exercise science
and/or behavioral science who followed a structured manualized protocol. The sessions
focused on various aspects of social cognitive theory (Bandura, 1986), and included
lessons on such topics as goal setting, cognitive restructuring, self-monitoring, and
problem solving. Sessions were conducted weekly during Month 1, biweekly during
Months 2 and 3, and monthly during Months 4-6. Participants could accumulate walking
minutes during the day in either structured 10 minute walking increments or more
continuously (for a total duration of 30 structured minutes/day). Further, training
occurred wherever was most convenient for the participant (i.e. at home, work, or some
Each day, participants were instructed to use daily training logs for self-monitoring
of their exercise. In these logs, participants recorded the duration (i.e. number of
minutes) and the intensity (i.e. average heart rate) of all leisure-time bouts of walking of
at least 10 minutes duration that occurred during that day. Each participant was given a
Polar Beat Heart Rate monitor (Polar Electro, Inc., Port Washington, N.Y.), which
displayed the participant' s heart rate continuously. Participants were told to wear the
monitor during each bout of walking and to record in their exercise logs the heart rate
they observed most frequently. The exercise logs were collected by staff at the
intervention sessions, and when participants were absent from these sessions, staff
collected the information by telephoning the participants.
Exercise preference. After attending an informational session during which an
investigator explained how the four walking programs in the study would differ in the
frequency and intensity of exercise that would be required, participants answered the
following question: As you know, in this study it will be necessary to randomize you to a
training group. However, if you could choose, which training program would you choose
to be in? Participants answered in a forced-choice format, endorsing one of the
following: (a) walking at a moderate intensity, 3-4 days/week; (b) walking at a moderate
intensity, 5-7 days/week; (c) walking at a high intensity, 3-4 days/week; (d) walking at a
high intensity, 5-7 days/week; (e) walking at any intensity that you want on each day you
walk; or (f) other. Participants answered this question before they were notified of their
actual randomization assignment, and thus before they began exercise.
Perceived difficulty. After participants were informed of which of the four
walking programs they had been randomized to, they were asked the following question:
"Now that you know which training program you will be asked to participate in, how
diffcult do you think it will be for you to accomplish the expected changes in your
physical activity? Participants made a global rating of difficulty, on a scale ranging
from 0 (not difficult at all) to 100 (maximally difficult). Ratings were made before
participants began their exercise programs.
Two indicators of exercise adherence were calculated: (a) frequency adherence, the
mean proportion of prescribed walking minutes completed per week, and (b) intensity
adherence, the mean proportion of prescribed walking minutes completed in target heart
rate zone per week. Previous studies have suggested that heart rate is a valid and reliable
indicator of the intensity level of physical activity (King et al., 1991). Each of these
proportions was calculated using data recorded on the daily exercise logs kept by the
participants over the 6-month intervention. For participants in the moderate frequency
groups, the total number of walking minutes completed per week (and total number of
minutes completed in their target heart rate zone) was divided by 90 minutes (30 min on
3 days); for participants in the higher frequency groups, the total number of walking
minutes per week (and total number for minutes completed in their target heart rate zone)
was divided by 150 minutes (30 min on 5 days). Missing data were treated in a
conservative fashion, by assuming that when participants did not record exercise in their
logs, the exercise did not occur. In addition, if a participant dropped out of treatment, it
was assumed that he or she had stopped exercising.
Treatment preference. Those participants who had been assigned to the walking
program for which they had indicated a preference were dummy coded in the analysis
with a value of 0 and labeled "matched" participants. Those participants who had been
assigned to one of the three walking programs for which they had not indicated a
preference were dummy coded as 1 in the analysis and labeled "mismatched"
participants. Subsequently, the "matched" and "mismatched" participants were
compared in independent sample t-tests with regard to measures of frequency adherence
and measures of intensity adherence. An alpha level of .05 was used for significance
Perceived difficulty. Two Pearson bivariate correlations were conducted; one
examining the relation between ratings of perceived difficulty and frequency adherence
and the other examining the relation between ratings of perceived difficulty and intensity
adherence. An alpha level of .05 was used for significance testing.
Relation between treatment preference and perceived difficulty. A one-way
analysis of variance (ANOVA) was conducted to compare differences in perceived
difficulty ratings when participants were grouped according to which of the four walking
programs they had endorsed as their preferred group. The between-subj ects variable was
which walking program a participant indicated as preferred; there were four levels of this
variable, one for each of the four programs offered. The dependent variable was ratings
of perceived difficulty. An alpha level of .05 was used for significance testing and
Bonferroni corrections were used in post hoc testing.
The sample was composed of 152 middle aged, overweight, but well educated men
(n = 84) and women (n = 68). The baseline demographic characteristics of participants
when grouped according to whether they were assigned to their preferred prescription are
displayed in Table 1. The ethnic-racial makeup of the sample was 65.1% Caucasian,
20.4% African American, 6.6%Asian, 5.9% Hispanic, and 2.0% unknown. ANOVAs
showed that the treatment groups did not differ significantly in baseline measures of age,
educational level, or BMI. After six months of the intervention, 9 women and 6 men had
dropped out, Hyve of whom were African-American and ten of whom were Caucasian, and
these participants had a mean age of 46.93 (SD = 9.4) and a mean body mass index of
24.80 (SD = 5.01). The racial composition, sex composition, BMI, and age of the group
before individuals dropped out and the group after the individuals dropped out were not
After comparing participant preferences and actual randomization assignments, 23
individuals were labeled "matched" participants while 82 were labeled "mismatched"
participants. Those participants who did not complete the questionnaire were excluded
from this analysis. The two groups that emerged did not differ significantly in mean age,
education level, gender composition, racial composition, fitness level, or BMI (ps > .05).
Independent samples t-tests revealed that those assigned to preferred exercise
prescriptions did not differ from those not assigned to preferred prescriptions in terms of
frequency adherence or intensity adherence. The results of the t-tests are shown in Table
2. The effect sizes of the frequency and intensity adherence analyses were d = 0.26 and
0. 11, respectively; both are small effect sizes (Cohen, 1988).
A significant negative relationship was detected between perceived difficulty and
frequency adherence (r = -.20, R2= .04, p = .04). The relationship between perceived
difficulty and intensity adherence was marginally significant (r = -0. 17, p = .09).
However, after conducting a 2 X 2 ANOVA, with assigned intensity level(moderate
versus high) and assigned frequency level (moderate versus high) as between-subj ect
factors and perceived difficulty as the dependent variable, a significant main effect
emerged for intensity level assigned [F(1, 100) = 9. 12, p = 0.003]. Individuals assigned
to higher intensity prescriptions made significantly higher perceived difficulty ratings
than those assigned to moderate intensity prescriptions, as described in Table 3. There
was no main effect of assigned frequency nor an interaction between frequency and
To determine whether the association between perceived difficulty and adherence
was mediated by the intensity prescribed, a partial correlation was conducted controlling
for the assigned intensity level. The findings showed a weakened association between
perceived difficulty and frequency adherence (r = -. 16, p = 0. 11) and intensity adherence
(r = -.07, p = .50). Thus, intensity assignment appeared to partially mediate the
relationship between perceived difficulty and adherence.
Relation between Treatment Preference and Perceived Difficulty
A one-way ANOVA, conducted to determine whether a relationship existed
between prescription preference and perceived difficulty of assigned prescription,
revealed significant group differences [F(3, 96) = 3.01, p = .03]. Means and standard
deviations of ratings of perceived difficulty (made by participants when grouped by
exercise preference) are shown in Table 4. Post-hoc analyses using Bonferroni
corrections showed that those who preferred a prescription of moderate frequency and
moderate intensity reported significantly higher ratings of perceived difficulty than did
participants who preferred a prescription of higher frequency and higher intensity (p <
A series of one-way ANOVAs were conducted to establish whether relevant
baseline characteristics might be related to either treatment preference or to ratings of
perceived difficulty. To examine the relation between treatment preference and baseline
characteristics, a series of 2X2 ANOVAs were conducted. The between-subjects
variables for all ANOVAs were preferred walking frequency and preferred walking
intensity. The dependent variables examined were age, sex, baseline fitness (VOzmax ,
and exercise history (minutes of exercise/year). Bonferroni corrections were used in post
hoc testing. The ANOVA examining baseline fitness revealed a significant main effect
for intensity preference, but no main effect was uncovered for either frequency
preference or the interaction between frequency and intensity preferences. Those
individuals preferring moderate intensity exercise were significantly less fit at baseline
than those preferring high intensity programs [F(1,116) = 4.23, p=.04)].
To examine the relation between perceived difficulty ratings and baseline
characteristics, a series of Pearson bivariate correlations were conducted to look for
associations between perceived difficulty and age, sex, baseline fitness, or exercise
history. The only significant negative relation that emerged was between baseline fitness
and the perceived difficulty of assigned exercise prescription (r = -.21, p = .03). Thus, it
appeared that baseline fitness partially mediated the relation between intensity preference
and perceived difficulty. However, baseline fitness and adherence (the original construct
of interest) were not related (frequency adherence, r= .16, p= 0.09; intensity adherence,
r=. 12, p=.20).
Lastly, to determine the amount of variance that each examined variable (match to
preferred group, perceived difficulty, frequency assignment, intensity assignment)
contributed to each measure of adherence, two linear regressions were conducted. In
both regressions, the level of intensity of exercise prescribed accounted for the greatest
proportion of the variance when the variance from the other factors were accounted for
(frequency adherence, P = -. 16; intensity adherence, P = -.40). The only factors that
contributed significant unique variance were the level of intensity assigned and the level
of frequency assigned in the intensity adherence analysis. The beta weights associated
with each factor in each regression are displayed in Table 5. In addition, it was found
that match to preferred group, perceived difficulty, frequency assignment, and intensity
assignment (when considered simultaneously in a regression model) collectively
accounted for 53% of the variance in frequency adherence and for 62% of the variance in
Table 1. Baseline Demographic Characteristics of Participants
Assigned to Preferred and
M SD p
Characteri sti c
VO2 max (ml/kg/min)
Table 2. Adherence of Participants Assigned to Preferred or Non-Preferred Exercise
-1.47 .14 .26
Note. Adherence refers to the proportion of the minimum prescription fulfilled by the
participant over six months.
Table 3. Ratings of Perceived Difficulty with Participants Grouped by Intensity
Intensity Assignment M SD
Moderate 24.17 25.70
High 38.21 25.89
Note. Ratings of perceived difficulty were made on a global scale from 0 (no difficulty)
to 100 (maximally difficult).
Table 4. Ratings of Perceived Difficulty with Participants Grouped by Exercise
Exercise Preference M SD
Moderate Frequency/Moderate Intensity 43.96a 27.78
High Frequency/Moderate Intensity 28.84 ab 27.51
Moderate Frequency/High Intensity 28.79ab 22.94
High Frequency/High Intensity 19.58 b 19.94
Note. Ratings of perceived difficulty were made on a global scale from 0 (no difficulty)
to 100 (maximally difficult); means with different subscripts are significantly different
from each other at p < .05.
Table 5. Standardized Beta Weights for Factors Contributing to Variance in Exercise
Adherence Type Factor P p
Match to Preferred Exercise .14 .17
Perceived Difficulty -.13 .23
Frequency Assigned -.14 .17
Intensity Assigned -. 16 .13
Match to Preferred Exercise .10 .27
Perceived Difficulty -.02 .85
Frequency Assigned -.20 .04
Intensity Assigned -.40 .001
Note. Adherence refers to the proportion of the minimum prescription fulfilled by the
participant over six months.
The present study examined how a potential exerciser' s initial impressions of a
physical activity program might be associated with subsequent adherence to the program.
The associations between various baseline participant characteristics and cognitions were
also investigated. Although we hypothesized that participants assigned to preferred
levels of walking frequency and intensity would show different rates of adherence than
participants who were not, our data did not reveal such a difference. However, the effect
size of the frequency adherence analysis (Cohen' s d = .26) suggests that there may have
been insufficient power to detect this relatively small effect size. Nonetheless, these
Endings imply that tailoring treatments to participant preference may not effectively
improve adherence in clinical settings.
Additionally, although we found support for our hypothesis that frequency
adherence and perceived difficulty were negatively related, the relationship was only
marginally significant for intensity adherence. Also, these relationships appeared to be
partially explained by prescription assignment. When level of assigned intensity was
controlled for, the relationships between perceived difficulty and adherence were
weakened. These results concur with prior studies (Dishman & Buckworth, 1996; Perri
et al., 2002) that found participants in higher intensity groups showed poorer adherence
than participants assigned to moderate intensity exercise. Therefore, intensity assignment
may be a more important factor than perceived difficulty when considering adherence to
Lastly, our data supported the hypothesis that prescription preference and perceived
difficulty are related. Participants who preferred prescriptions of moderate frequency and
intensity reported significantly higher ratings of perceived difficulty than did those
participants who preferred higher frequency and intensity. Participants' fitness was
related to both prescription preference and perceived difficulty; thus, baseline fitness
level may be responsible for this relationship. However, there emerged no relationship
between fitness and adherence.
Collectively, these Eindings suggest that characteristics ofan exercise program,
such as the intensity level of exercise prescribed, may be more influential in predicting
exercise adherence than characteristics of the individual, such as participant preference,
perceived difficulty, or baseline fitness. Further, the characteristics of the program
appeared to be more predictive of intensity adherence than frequency adherence, as the
level of frequency and level of intensity assigned each contributed significant, unique
variance to intensity adherence. Because the intensity adherence measure captured
information regarding both frequency and intensity of exercise, it is likely that this
measure was more sensitive and therefore was able to detect existing relationships.
Thus, prescribing moderate intensity exercise may be more effective in improving
adherence than tailoring programs to participant-related variables. This finding may be
particularly important when prescribing exercise for highly fit individuals, who may
perceive less difficulty and thus prefer programs of higher intensity. Since our results
suggests that individuals assigned to higher intensity programs may show compromised
adherence, these potential exercisers might benefit from encouragement to begin with
moderate intensity programs.
This study represents a methodological improvement over previous research
conducted in this area. The maj ority of studies assessing the factors governing changes
in physical activity have gathered their data at a single time point. For example, one
study assessing osteoporosis health beliefs and the practice of preventative weight-
bearing exercises collected information about both factors in a single testing session
(Wallace, 2002), as did another study investigating the association between older
women' s beliefs about exercise barriers and the amount of physical activity practiced
(O'Brien Cousins, 2000). By contrast, participants in the present study indicated their
perceived level of difficulty before commencing exercise; therefore we are able to make
some conj ectures about the direction of causality between perceived difficulty and
In addition, previous studies have generally not utilized objective measures of
adherence, relying solely upon general, retrospective self-reports that are often
descriptive in nature (i.e. Corwyn & Brenda, 1999). To address this limitation, the
present study used daily logs to track exercise, with participants utilizing heart rate
monitors to gather an obj ective measure of intensity and using pedometers to conduct
validity testing of reported frequency.
There were, however, a number of potential limitations associated with the present
study. The first limitation involves the measure used to gather data concerning treatment
preference. Because participants were simply asked to indicate which of the four walking
programs they preferred, they had no method to indicate the strength of their preference.
Therefore, especially for those participants whose strength of preference was weak, the
test-retest reliability of the measure would likely be low, bringing its validity into
question. In addition, relying on a single item to encompass potentially complicated
constructs may have rendered our analyses too crude and insensitive to detect existing
relationships. Another limitation concerning the treatment preference investigation was
that participants were not randomly assigned to either their preferred walking program or
their non-preferred walking program. These analyses were conducted after the trial had
been completed; thus, the "matched"/ "mismatched" distinction was made post-hoc. This
necessarily led to a larger group of participants who were "mismatched" with their
preferred group. However, Levene's test for homogeneity of variance was not
significant, supporting the assumption that the variance of the "matched" group did not
differ significantly from the variance of the "mismatched" group.
There were also a variety of potential limitations associated with the design of the
parent study. First, the measures of adherence were derived completely from self-
reported data, subject to intentional and accidental inaccuracies. However, previous
research has shown self-reported heart rate to be a reliable and valid indicator of intensity
during physical activity (King et al., 1991; Koltyn & Morgan, 1992; Pollock & Wilmore,
1990). Also, past studies have demonstrated high convergent validity for exercise logs
shown through agreement with ambulatory HR monitoring (King et al., 1991). Further,
since participants were given heart rate monitors to monitor intensity during exercise, we
were able to obtain a sample of digitally recorded heart rate data during selected exercise
bouts. An examination of these data revealed that most participants were indeed walking
within their prescribed heart rate zones. Second, before generalizing our findings to a
larger population of sedentary adults, one must consider the fact that our participants had
a considerably higher level of educational achievement (M~= 15.9 years of education, SD
= 3.7) than would be expected overall for adults in the United States, and that the
majority of our sample was Caucasian (65.1%). Further, the climate in the community
where the study was conducted (Gainesville, FL) may have been particularly favorable
for outdoor exercise, without the ice and snow that may represent barriers to exercise in
other locations. On the other hand, however, the heat and humidity present in the
summer months may have represented a barrier that those in other communities may not
have to face.
In light of this Einal limitation, extensions of this work should investigate whether
these Eindings generalize to other specific populations for which exercise
recommendations are being made. For instance, it may be useful to assess whether the
initial impressions of an exercise program influence adherence in a more obese
population, for which the perceived barriers and probable preferences likely differ. In
addition, because pediatric obesity is becoming an increasingly serious problem in the
United States, it may behoove public health officials to determine how initial impressions
of an exercise program influence children's adherence to activity programs. Another
possible extension of this research could investigate the effect of these initial impressions
on long-term maintenance of exercise, since the Eindings of the present study are limited
to the "adoption" phase of an exercise program.
In summary, the present study suggests that, during the adoption phase of
exercise, the characteristics of the recommended program are more predictive of
adherence than are characteristics of potential exercisers. Specifically, the data suggest
that participants exhibited higher adherence rates to programs of moderate, rather than
high, intensity; this relation appeared to exist regardless of participant exercise
preference, perceived difficulty, or fitness at baseline. Thus, as public health officials
continue to revise recommendations for physical activity, it is suggested that
recommending programs of moderate intensity may lead to better adherence among the
maj ority of previously sedentary, middle-aged, unfit American adults.
LIST OF REFERENCES
Adams, P.M. (1998). Taking charge of one's life: A theoretical model for weight
management success. Dissertation Abstracts hIternational, 58, 6483.
Bandura A. (1997). Self-efficacy: the exercise of control. New York: W.H. Freeman and
Booth, M.L., Bauman, A., Owen, N., & Gore, C.J. (1997). Physical activity preferences,
preferred sources of assistance, and perceived barriers to increased activity among
physically inactive Australians. Preventive M~edicine, 26, 131-137.
Brassington, G.S., Atien, A., Perczek, R.E., DiLorenzo, T.M., & King, A.C. (2002).
Intervention related cognitive versus social mediators of exercise adherence in the
elderly. American Journal of Preventive M~edicine, 23, 80-86.
Centers for Disease Control and Prevention. (2004). Physical activity and good nutrition:
Essential elements to prevent chronic diseases and obesity: 2004. Retrieved
November 1, 2004, from http://www. cdc.gov/ncedphp/aag/pdf/aag_dnpa2004.pdf
Cohen, J. (1988). Statistical power analyses for the social sciences. Hillsdale, NJ:
Corwyn, R.F., & Brenda, B.B. (1999). Examination of an integrated theoretical model of
exercise behavior. American Journal of Health Behavior, 23, 38 1-392.
Courneya, K.S., & McAuley, E. (1994). Are there different determinants of the
frequency, intensity, and duration of physical activity? Behavioral1\~edicine, 20,
De Bourdeaudhuij, I., & Sallis, J. (2002). Relative contribution of psychosocial variables
to the explanation of physical activity in three population-based adult samples.
Preventive M~edicine, 34, 279-288.
Dishman, R., & Buckworth, J. (1996). Increasing physical activity: a quantitative
synthesis. Medicine & Science in Sport & Exercise, 28, 706-719.
Duncan, G.E., Perri, M.G., Anton, S.D., Limacher, M.C., Martin, A.D., Lowenthal, D.T.,
Airing, E., Bottiglieri, T., & Stacpoole, P.W. (2004). Effects of exercise on
emerging and traditional cardiovascular risk factors. Preventive M~edicine, 39, 894-
Epstein, L.H. (1998). Integrating theoretical approaches to promote physical therapy.
American Journal of Preventive M~edicine, 15, 257-265.
Garcia, K., & Mann, T. (2004). From "I wish" to "I will": Social cognitive predictors of
behavioral intentions. Journal of Health Psychology, 8, 347-360.
Godin, G., Desharnais, R., & Valois, P. (1994). Differences in perceived barriers to
exercise between high and low intenders: Observations among different
populations. American Journal of Health Promotion, 8, 279-285.
Godin, G., Valois, P., & Jobin, J. (1991). Prediction of intention to exercise of individuals
who have suffered coronary heart disease. Journal of Clinical Psychology, 4, 762-
Hagger, M.S., Chatzisarantis, N.L., & Biddle, S.J. (2004). A meta-analytic review of the
theories of reasoned action and planned behavior in physical activity: Predictive
validity and the contribution of additional variables. Journal of Sport and Exercise
Psychology, 24, 3-32.
King, A.C., Castro, C., Wilcox, S., Eyler, A.A., Sallis, J.F., & Bronson, R.C. (2000).
Personal and environmental factors associated with physical inactivity among
different racial-ethnic groups of middle-aged and older-aged women. Health
Psychology, 19, 354-364.
King, A.C., Haskell, W.L. Taylor, C.B., Kraemer, H.C., & DeBusk, R.F. (1991). Group
versus home-based exercise training in healthy older men and women: A
community-based clinical trial. Journal of the American M~edical Association, 266,
Koltyn, K.F., & Morgan, W.P. (1992). Efficacy of perceptual vs. heart rate monitoring in
the development of endurance. British Journal of Sports M~edicine, 26, 132-134.
Marcus, B.H., Dubbert, P.M., Forsyth, L.H., McKenzie, T.L., Stone, E.J., Dunn, A.L, &
Blair, S.N. Physical activity behavior change: Issues in adoption and maintenance.
Health Psychology, 19 (suppl.), 32-41.
O'Brien Cousins, S. (2000). "My heart couldn't take it": Older women's beliefs about
exercise benefits and risks. Journal ofGerontological & Psychological Sciences,
Parfitt, G., & Gledhill, C. (2004). The effect of choice of exercise mode on psychological
responses. Psychology of Sport and Exercise, 5, 111-117.
Patten, C.A., Armstrong, C.A., Martin, J.E., Salis, J.F., & Booth, J. (2000). Behavioural
control of exercise in adults: Studies 7 and 8. Psychology & Health. 15, 571-581.
Perri, M.G., Anton, S., Durning, P.E., Ketterson, T.U., Sydeman, S.J., Berlant, N.E.,
Kanasky, W.F., Newton, R.L, Limacher, M.C., & Martin, A.D. (2002). Adherence
to exercise prescriptions: Effects of prescribing moderate versus higher levels of
intensity and frequency. Health Psychology, 21, 452-458.
Petrella, R.J., Koval, J.J., Cunningham, D.A., & Paterson, D.H. (2003). Can primary care
doctors prescribe exercise to improve fitness? American Journal of Preventive
Medicine, 24, 316-322.
Pollock, M.L., & Will more, J.H. (1990). Exercise in health and' disease: Evaluation and'
prescription for prevention and' rehabilitation (2nd ed.). Philadelphia: Saunders.
Renjilian, D.A., Perri. M., Nezu, A.M., McKelvey, W.F., Shermer, R., & Anton, S.
(2001). Individual versus group therapy for obesity: Effects of matching
participants to their treatment preferences. Journal of Consulting and' Clinical
Psychology, 69, 717-721.
Salmon, J., Owen, N., Crawford, D., Bauman, A., & Sallis, J.F. (2003) Physical activity
and sedentary behavior: A population-based study of barriers, enjoyment, and
preference. Health Psychology, 22, 178-188.
Steptoe, A., Rink, E., & Kerry, S. (2000). Psychosocial predictors of changes in physical
activity in overweight sedentary adults following counseling in primary care.
Preventive M~edicine, 31, 183-194.
United States Department of Health and Human Services. (2002). Physical activity
fundamental to preventing disease. Retrieved November 1, 2004, from
Wallace, L. S. (2002) Osteoporosis prevention in college women: Application of the
expanded health belief model. American Journal of Health Behavior, 26, 163-172.
Katie Allison Rickel graduated summa cum laude from Duke University in May
2003 with a Bachelor of Science degree in psychology. She plans to pursue a doctoral
degree in clinical and health psychology at the University of Florida. Her academic
interests lie in medical and health psychology, with a focus on obesity and weight