Transcript
Page 1: Perceived Barriers to Physical Activity among High School Students

Preventive Medicine 28, 608–615 (1999)Article ID pmed.1999.0489, available online at http://www.idealibrary.com on

Perceived Barriers to Physical Activity among High School Students1

Kenneth R. Allison, Ph.D.,*,2 John J. M. Dwyer, Ph.D.,† and Susan Makin, M.Ed.†

*Department of Public Health Sciences, Faculty of Medicine, University of Toronto, Toronto, Ontario M5S 1A8 Canada; and†Toronto Public Health Division, City

Background. Perceived barriers to physical activity,the factor structure of perceived barriers, and the rela-tionship between perceived barriers and participationin vigorous physical activity were examined.

Methods. A two-stage cluster sample of high schoolstudents (N 5 1,041) in a large Metropolitan Torontoschool district was used. Students completed a ques-tionnaire (response rate 81.4%) dealing with participa-tion in physical activity in three settings. Factor analy-sis was used to examine the dimensionality ofperceived barriers. Multiple regression analysis wasthen used to examine the relationship between per-ceived barriers and participation.

Results. Time constraints due to school work, otherinterests, and family activities were three of the fourbarriers considered most important. Females citedconsistently higher levels of perceived barriers thanmales. Two empirically distinct and theoreticallymeaningful factors emerged from the analysis—perceived internal barriers and perceived externalbarriers. Perceived internal barriers were predictiveof physical activity in overall activity and outside ofschool activity. Perceived external barriers were pre-dictive of overall physical activity and other schoolactivity, but in the direction opposite to that hypothe-sized.

Conclusions. It was concluded that perceived barri-ers may be predictive of physical activity participationamong high school students only under specific condi-tions. q 1999 American Health Foundation and Academic Press

Key Words: barriers; physical activity; self-efficacystudents.

1 At the time of the study, the authors were project investigatorsof the North York Community Health Promotion Research Unit, sup-ported by a Health System-Linked Research Unit grant from theOntario Ministry of Health. This research was supported by Grant

942R017 from the Canadian Fitness and Lifestyle Research Institute.

2 To whom reprint requests should be addressed. Fax: (416) 978-2087. E-mail: [email protected].

60

of Toronto, Toronto, Ontario, Canada

INTRODUCTION

Given considerable empirical support for the role ofphysical activity and fitness in disease prevention andhealth promotion [1–4], several investigators haveidentified the need for additional studies that examinethe determinants of physical activity participation[5–8]. The information derived from such studies canbe used to further elucidate the factors influencing be-havior and to assist in developing appropriate interven-tion strategies to promote higher levels of participation.Several studies and reviews have documented lowerlevels of activity among older adolescents, especiallyfemales [9–11]. However, few studies have examinedthe factors influencing physical activity for this group[8,12].

A factor of considerable potential importance as adeterminant of physical activity is the notion of per-ceived barriers. According to the Health Belief Model,the presence of perceived barriers decreases the likeli-hood of engaging in preventive health practices, espe-cially if perceived barriers outweigh the perceived bene-fits of doing so [13,14]. Thus, perceived barriersrepresent obstacles to engaging in behavior whichmight otherwise help to prevent disease and enhancehealth. Perceived barriers may reflect environmentalfactors (external barriers), such as a lack of supportfrom friends and family, low resources, or a lack of timedue to other responsibilities. Additionally, perceivedbarriers may represent more individual, psychologi-cally based factors (internal barriers), such as a lack ofmotivation, other interests, or concerns about engagingin physical activity in public.

In an analysis of data from the Campbell’s Surveyon Well-Being [15] four perceived-barriers factors werespecified: lack of support and services, lack of confidenceand ability, lack of time, and fear of injury [16]. Ina subsequent multiple regression analysis predictingintention to engage in physical activity, lack of confi-dence and ability and fear of injury were the most pre-

dictive, with different sets of predictors depending onage and gender. The study of high school students by

8 0091-7435/99 $30.00Copyright q 1999 by American Health Foundation and Academic Press

All rights of reproduction in any form reserved.

Page 2: Perceived Barriers to Physical Activity among High School Students

T

ditional variables included in the analysis (self-efficacy,

PERCEIVED BARRIERS

Tappe et al. [17] reported the major barriers to exerciseto be time constraints, unsuitable weather, school andschoolwork, and lack of interest or desire. Males weremore likely to report use of alcohol and drugs and hav-ing a girlfriend as barriers to exercise, whereas femaleshad significantly higher reported levels of time con-straints as barriers.

Perceived barriers have been shown to be one of themost consistent predictors of a wide range of health-related behaviors [14,18,19]. (Perceived benefits, an-other Health Belief Model construct, have been shownto be more predictive of sick-role behavior than ofhealth-related behaviors [14].) This is also the case forphysical activity participation. In studies of physicalactivity, perceived barriers have been shown to be asso-ciated with exercise self-efficacy [20], exercise inten-tions [16,21], and exercise behavior among both adults[22–25] and, in some cases, adolescents [17,26,27].

Hofstetter et al. [20], examining the social learningcorrelates of exercise self-efficacy, found perceived bar-riers to be one of the most important predictors of self-efficacy when controlling for the effects of additionalvariables. Perceived barriers were shown to be thestrongest predictor of jogging versus nonexercising ina study by Slenker et al. [23] The study by Sallis et al.[22] of determinants of vigorous exercise in a sampleof 2,053 adults found perceived barriers to be thestrongest predictor of vigorous exercise. A study byTappe et al. [26] reported overall barriers to exercise tobe one of the strongest predictors of adolescent physicalactivity levels for males, but not females. Stucky-Roppand DiLorenzo found mother’s perceived barriers to bea salient predictor of physical activity levels for bothmale and female adolescents [27]. In summary, thereis considerable indication in the research literature thatperceived barriers is an important factor related tohealth behavior in general and participation in physicalactivity in particular. However, few studies have exam-ined, for adolescents, the nature and dimensionality ofperceived barriers and their relationship to physicalactivity in various settings.

Another major factor influencing participation inphysical activity is self-efficacy, a central construct ofSocial Cognitive Theory [28,29]. Self-efficacy representsthe confidence one has in being able to enact a certainbehavior in particular circumstances. There is substan-tial indication in the research literature that self-efficacy is an important correlate of physical activityparticipation for both adults [5,7,20,22,30] and childrenand adolescents [8,31,32].

In the study described in this paper, perceived barri-

ers to physical activity among high school students areexamined. The paper describes the importance of spe-cific perceived barriers for the whole sample, as wellas gender and grade level differences in perceived barri-ers. The factor structure of perceived barriers is also

O PHYSICAL ACTIVITY 609

analyzed. Finally, the relationship between perceivedbarriers and physical activity in three separate settingsis examined, controlling for the simultaneous effects ofperceived self-efficacy, age, and gender. In the begin-ning of the study, we hypothesized that lack of time dueto various reasons would be the most salient perceivedbarrier for subjects. We had no specific hypothesis re-garding the factor structure of perceived barriers, al-though we believed that two dimensions of perceivedbarriers are likely, based on their location within (inter-nal barriers) or outside (external barriers) the individ-ual. We also hypothesized that perceived barriers wouldbe inversely related to physical activity participation(overall and in the three settings) and that this relation-ship would remain when controlling for the effects ofadditional variables in the model. We believed that ad-

age, and gender) would influence physical activity.However, we believed that perceived barriers, repre-senting obstacles that may be difficult to overcome,would continue to influence physical activity behavior.

DESIGN AND METHODS

Sample Description

The study sample consisted of 1,041 grade 9 (N 5 688)and grade 11 (N 5 353) students in a large, ethnicallydiverse, board of education in Metropolitan Toronto.Males comprised 51% of the sample and the mean agewas 14.9 years. A two-stage cluster sample was usedfor the design [33,34]. In the primary stage of sampling12 of 18 secondary schools containing both grade 9 and11 students were randomly selected. From the selectedschools, classrooms of grades 9 and 11 were randomlyselected for inclusion in the survey. [Grades 9 and 11were selected for comparative purposes. For logisticalreasons, it was not possible to sample other secondarylevel classes (grades 10 and 12).] The school level re-sponse rate was 75% (8 of 12 schools) and the studentlevel response rate was 81.4% (1,041 eligible comple-tions/1,041 1 21 refusals 1 217 absentees).

Measurement Instrument

Perceived barriers. The measures of perceived bar-riers used in this study are based partly on those usedin previous studies of adults [15,30]. In an earlier pilotstudy and the current study, the items asked the impor-tance of 16 factors in “preventing you from participatingin vigorous physical activity.” Items are scored on a

5-point scale ranging from “not at all” (1) to “a greatdeal” (5). These items form the basis for the descriptiveanalysis reported. A subsequent analysis of the factorstructure of perceived barriers produced factor scoresthat were used in the multiple regression analysis.
Page 3: Perceived Barriers to Physical Activity among High School Students

610 ALLISON, DWYE

Self-efficacy. A 20-item scale of self-efficacy in rela-tion to overcoming perceived barriers to vigorous physi-cal activity was created for this study. The items areadapted versions of previously validated measures[8,20,35]. The adapted measures consist of items as-sessing the confidence respondents have that they canparticipate in vigorous physical activity in a number ofsituations (e.g., if there is a lack of support from familyor if you lacking energy). The items are measured ona 5-point scale ranging from “not at all confident” (1)to “very confident” (5). Factor analysis, used earlier toidentify the factor structure of physical activity self-efficacy, produced two factors—self-efficacy regardingparticipation in vigorous physical activity despite exter-nal barriers and self-efficacy despite internal barriers[36]. Factor scores derived from this factor analysiswere used in the multiple regression analysis re-ported here.

Participation in vigorous physical activity. Physicalactivity and exercise, measured in alternative ways inprevious studies [37–40], have been found to be reliablymeasured using self-reports on standardized question-naires [41]. The measures used in the current studywere modified versions of those used in the Campbell’ssurvey and the U.S. National Adolescent StudentHealth Survey [15,42]. This measure of physical activ-ity retains the standard description of vigorous physicalactivity endorsed by the American College of SportsMedicine [43]. An initial statement used to introducethe question reads, “A physical activity is vigorous if itlasts for 20 minutes or more, and makes your heart beatfaster, and makes you breathe a lot faster.” Respondentswere then asked to indicate the number of days perweek that they currently engage in vigorous physicalactivity for each of three settings—physical educationclasses (0–5 days), other activities in school (intramu-rals, interscholastic sports) (0–7 days), and outside ofschool (recreation, community sponsored sports) (0–7days). An additive index was computed reflecting thenumber of days of vigorous physical activity in the threesettings (coded 0–7).

The study received ethical approval from review com-mittees at the University of Toronto, the municipal pub-lic health department, and the board of education. Priorto completing a self-administered questionnaire, stu-dents indicated their willingness to participate by sign-ing a consent form. Data were collected in classroomsduring regular class time during the fall and early win-ter of 1994–1995.

Data Analysis

The first stage of the analysis examined the degreeto which respondents perceived the 16 perceived barri-ers to be important, using the mean score for each item.

R, AND MAKIN

Next, gender differences in perceived barriers were ex-amined, using analysis of variance. Exploratory factoranalysis was then conducted in order to examine theunderlying dimensions of perceived barriers. Principalcomponent analysis was utilized, with oblique rotation.Selection of factors with an eigenvalue of greater than1 and Catell’s scree test were used to determine thenumber of factors [44]. The derived factor scores wereused in the multiple regression analysis.

Finally, hierarchical multiple regression analysis wasutilized in order to examine the relationship betweenperceived barriers and physical activity. Physical activ-ity participation (dependent variable) was regressed onthe following independent variables, according to thestages of entry: (1) perceived barriers, (2) self-efficacy,(3) age and gender. This analysis was used in order to

determine the effects of perceived barriers when takinginto account the effects of additional variables knownto be important predictors of physical activity. The mul-tiple regression analysis was also run separately bygrade level.

RESULTS

Levels of Physical Activity

The mean number of days per week participating invigorous physical activity was highest for outside ofschool activity (M 5 2.9, SD 5 2.2), compared withphysical education class (M 5 2.2, SD 5 1.7) and otherschool-related activities (M 5 l.5, SD 5 1.9).

Mean Importance of Perceived Barriers

Descriptive statistics for the overall sample (N 51,041) indicate that the mean level of perceived barrierswas relatively low (only 1 of 16 had a mean over 3.0).Perceived barriers related to a lack of time were seenas having the highest importance (Table 1). A lack oftime due to schoolwork, to other interests, and to familyactivities represents three of the four highest rankedperceived barriers, supporting our hypothesis. Per-ceived barriers related to mood, lacking energy, lackingself-discipline, and discomfort also received meanscores of 2.0 or greater, but all other perceived barrierswere ranked below 2.0. Gender differences in meanlevels of perceived barriers are consistent, with femaleshaving higher levels than males for 8 of 9 instances inwhich significant differences exist (Table 1). Despitethese differences, the hierarchical order in ranking per-ceived barriers (using mean importance) for males andfemales is largely similar. Fewer significant differencesin perceived barriers were found by grade level. Being

self-conscious, having an injury, or lacking support fromfamily and friends were significantly higher for grade9 students, whereas cost and lacking time due to part-time work were significantly higher among grade 11students (table not shown).
Page 4: Perceived Barriers to Physical Activity among High School Students

Time—part-time work 1.62 1.07 1.69 1.14 1.54 1.00 0.026 Cost 0.49Illness 1.60 1.04 1.63 1.07 1.57 1.02 0.306 Time—school work 0.48

Injury 1.58 0.98 1.61 1.02 1.55 0.94 0.303

Note. N’s ranged from 1,012 to 1,038. P values are based on analysisof variance of gender differences in perceived barriers.

Factor Structure of Perceived Barriers

Principal components analysis (PCA) with obliquerotation was used to explore the factor structure ofperceived barriers. Oblique rotation, which allows thefactors to be correlated [44], was used because the fac-tors underlying perceived barriers were expected to becorrelated, according to theory. An initial analysis withPCA identified four factors with eigenvalues over 1.0.

However, since this extraction technique frequentlygenerates too many factors, the scree test was also usedto determine the number of factors to retain for factoranalysis. Using the scree test, and based on theoreticalgrounds and interpretability, a two-factor solution wasselected as most suitable.

The pattern matrix for the two-factor solution isshown in Table 2. Factor 1 includes items related pri-marily to perceived barriers related to personal/individ-ual issues: being self-conscious, having discomfort,lacking self-discipline, or lacking energy, for example.Factor 2 was composed of items dealing with perceivedbarriers related to the social environment surroundingthe individual, such as cost, lack of time due to familyactivities, lack of friends’ support, or a lack of time dueto schoolwork. Thus, the two factors represent concep-tually interpretable distinctions between perceived in-ternal barriers and perceived external barriers. Resultsfrom this analysis were saved as factor scores and used

in the subsequent analysis of the relationship betweenperceived barriers and participation in vigorous physi-cal activity.

The a reliability coefficient for the 10 items loadingmost heavily on the internal barriers factor was 0.79

Eigenvalue 4.4 1.4Percentage of variance 27.3 9.0

Note. Total sample 5 1,041. Items loading less than 0.30 not shown.

and, for the 6 items loading most heavily on externalbarriers, was 0.60.

Relationship between Perceived Barriers andPhysical Activity Participation

In examining the bivariate relationship between per-ceived barriers and participation in vigorous physicalactivity, the perceived internal barriers factor was in-versely related to overall physical activity (index), phys-ical activity outside of school, and activity in physicaleducation class (Table 3). However, the perceived exter-nal barriers factor was only related to other schoolactivity.

Self-efficacy regarding participation despite externalbarriers was positively correlated with overall physicalactivity as well as physical activity in the three settings,but self-efficacy for participating despite internal barri-ers was inversely related to physical activity (Table 3).Age was inversely related to physical activity participa-tion (overall, physical education class, outside ofschool), as was female gender (overall and in thethree settings).

The results of hierarchical multiple regression indi-cate that the effects of perceived barriers on participa-tion are somewhat diminished when taking into ac-count the effects of self-efficacy factors, age, and gender

PERCEIVED BARRIERS T

TABLE 1

Mean Importance of Perceived Barriers by Gender

All Males Females

Mean SD Mean SD Mean SD P,

Time—school work 3.01 1.23 2.90 1.25 3.12 1.19 0.004Time—other interests 2.87 1.25 2.79 1.29 2.95 1.20 0.040Mood 2.51 1.21 2.34 1.20 2.68 1.21 0.001Time—family activities 2.44 1.20 2.32 1.16 2.57 1.23 0.001Lack energy 2.20 1.27 2.06 1.25 2.35 1.28 0.001Lack self-discipline 2.05 1.18 1.93 1.17 2.18 1.17 0.001Discomfort 2.00 1.06 1.93 1.02 2.05 1.10 0.069Cost 1.98 1.18 1.96 1.19 2.00 1.16 0.674Not fun 1.96 1.23 1.91 1.22 2.01 1.24 0.202Self-conscious 1.91 1.14 1.74 1.02 2.09 1.22 0.001Stressed 1.89 1.10 1.73 0.98 2.07 1.18 0.001Lack family support 1.79 1.22 1.81 1.25 1.75 1.18 0.411Lack friends’ support 1.69 1.03 1.68 1.03 1.70 1.02 0.801

O PHYSICAL ACTIVITY 611

TABLE 2

Factor Loadings for Principal Components Analysis with ObliqueRotation of Perceived Barriers Items

Factor 1 Factor 2

Mood 0.78 20.31Discomfort 0.69Stressed 0.66Not fun 0.64Lack energy 0.60Lack self-discipline 0.54Injury 0.53Self-conscious 0.52Illness 0.36Time—other interests 0.35Lack family support 0.65Time—family activities 0.62Lack friends’ support 0.51Time—part-time work 0.51

(Table 4). Perceived internal barriers remained a sig-nificant predictor of overall physical activity and physi-cal activity outside of school but not in other settings.Perceived external barriers were a significant predictorof physical activity, both overall and in other school

Page 5: Perceived Barriers to Physical Activity among High School Students

SE—external barriers 0.10 0.004 0.22 0.001 0.28 0.001 0.23 0.001

s

barriers on participation in vigorous physical activity

SE—internal barriers 20.11 0.001 20.17Age 20.20 0.001 20.05Gender (female) 20.09 0.005 20.17

Note. The analysis was based on pairwise deletion of missing case

activities but not in the expected direction—higher lev-els of perceived barriers were associated with participa-tion. Thus, the findings provide only some support forthe hypothesized relationship between perceived barri-ers and participation in physical activity.

Self-efficacy despite external barriers was a signifi-cant predictor of overall physical activity, other schoolactivity, and outside of school activity. Self-efficacy de-spite internal barriers was a significant predictor ofoutside of school activity, but in the direction oppositeto that expected. Age was a significant predictor of phys-ical activity (overall and in the three settings) as wasgender. As expected, older and female students wereless likely to participate.

The combined effects of perceived barriers, self-effi-cacy, age, and gender explained 6% of the adjusted vari-ance in vigorous physical activity in physical educationclass, 8% in other school activities, 15% in outside ofschool activities, and 13% of overall activity.

The multiple regression analysis was also run sepa-

Age 20.21 0.001 20.08Gender (female) 20.07 0.044 20.15

Adj R square P, Adj R squa0.06 0.001 0.08

Note. The analysis was based on listwise deletion of missing cases.

0.001 20.24 0.001 20.20 0.0010.111 20.08 0.014 20.21 0.0010.001 20.22 0.001 20.15 0.001

. N’s ranged from 891 to 1,032.

612 ALLISON, DWYER, AND MAKIN

TABLE 3

Correlations of Participation in Vigorous Physical Activity with Perceived Barriers, Self-Efficacy, Age, and Gender

Physical educ class Other school Outside school Overall

Pearson Pearson Pearson PearsonCorr P, Corr P, Corr P, Corr P,

Perceived bars—internal 20.08 0.020 20.04 0.200 20.22 0.001 20.14 0.001Perceived bars—external 0.01 0.903 0.09 0.008 20.04 0.175 0.03 0.378

r

in physical education class were significant with respectto grade 9 students, whereas, in the whole sample(grades 9 and 11 combined), these effects were notsignificant.

DISCUSSION

The findings indicate that, overall, students did notperceive barriers to be a large problem (based on themean levels). Participants in the study attached thehighest importance to perceived barriers dealing witha lack of time—due to schoolwork, other interests, and(to a lesser extent) family activities. This concern abouttime is consistent with previous studies of perceivedbarriers to physical activity [15–17]. Lack of time dueto schoolwork corresponds to the principal perceivedbarrier for adults—lack of time due to work (employ-ment). However, schoolwork includes time spent in

classes as well as time outside of school devoted to

rately by grade level, with interesting results. Perceived studying and preparing assignments. Thus, the timepressures of student life are not restricted to schoolbarriers were predictive of vigorous physical activity

only for grade 9 students, not for grade 11 students hours.Not being in the mood and lack of energy were also(Tables 5 and 6). Also, the effects of internal perceived

TABLE 4

Multiple Regression of Vigorous Physical Activity on Perceived Barriers, Self-Efficacy, Age, and Gender

Physical educ class Other school Outside school Overall(N 5 856) (N 5 845) (N 5 846) (N 5 832)

Stand Regr Stand Regr Stand Regr Stand RegrCoeff P, Coeff P, Coeff P, Coeff P,

Perceived bars—internal 20.06 0.102 20.02 0.631 20.14 0.001 20.11 0.003Perceived bars—external 0.03 0.345 0.09 0.014 0.02 0.566 0.07 0.040SE—external barriers 0.05 0.287 0.18 0.001 0.19 0.001 0.16 0.001SE—internal barriers 20.07 0.123 20.05 0.208 20.09 0.031 20.08 0.059

0.015 20.08 0.009 20.22 0.0010.001 20.20 0.001 20.13 0.001

e P, Adj R square P, Adj R square P,0.001 0.15 0.001 0.13 0.001

Page 6: Perceived Barriers to Physical Activity among High School Students

Age 20.01 0.942 20.08 0.039 20.11 0.004 20.13 0.002Gender (female) 20.02 0.576 20.12 0.003 20.19 0.001 20.09 0.025

a

Adj R square P, Adj R squ0.01 0.062 0.09

Note. The analysis was based on listwise deletion of cases.

considered important relative to other perceived barri-ers to engaging in vigorous physical activity. In contrastto perceived barriers dealing with a lack of time, whichare largely situated “outside” of the individual, moodand energy exemplify more affective, personal states orqualities. In summary, in the hierarchy of perceivedbarriers seen as important, three can be consideredexternal barriers and two internal barriers.

While there were similarities between males and fe-males in the rank order of perceived barriers, therewere consistent differences by sex in the level of impor-tance attached to these barriers. Eight of nine instancesof significant gender differences in perceived barriersreflected higher levels among females. This pattern isconsistent with at least some of the previous researchon perceived barriers [15]. However, the reasons for

gender differences in perceived barriers are not entirely

Gender (female) 20.14 0.021 20.17

Adj R square P, Adj R squa0.05 0.019 0.05

Note. The analysis was based on listwise deletion of cases.

re P, Adj R square P, Adj R square P,0.001 0.16 0.001 0.12 0.001

group. Yet the factors that give rise to higher levels ofperceived barriers among females need to be exam-ined empirically.

Perhaps the objective life circumstances of femalestudents influence higher levels of perceived barriersto physical activity. For example, less discretionarytime or additional home/family responsibilities by fe-males may be the case. Alternatively, female studentsas a whole may simply not be as interested in participat-ing in vigorous physical activity as males. In order tounderstand gender differences in perceived barriers,further research on the possible factors influencing per-ceived barriers is needed, as well as an examination ofthe perceived benefits of engaging in physical activity,which also could differ by gender.

Exploratory analysis of the factor structure of per-

PERCEIVED BARRIERS TO PHYSICAL ACTIVITY 613

TABLE 5

Multiple Regression of Vigorous Physical Activity on Perceived Barriers, Self-Efficacy, Age, and Gender—Grade 9 Students

Physical educ class Other school Outside school Overall(N 5 565) (N 5 556) (N 5 557) (N 5 546)

Stand Regr Stand Regr Stand Regr Stand RegrCoeff P, Coeff P, Coeff P, Coeff P,

Perceived bars—internal 20.12 0.014 20.07 0.156 20.19 0.001 20.16 0.001Perceived bars—external 0.03 0.572 0.15 0.001 0.06 0.153 0.09 0.055SE—external barriers 0.05 0.373 0.13 0.014 0.10 0.050 0.12 0.023SE—internal barriers 20.02 0.680 20.11 0.038 20.15 0.003 20.12 0.024

ceived barriers revealed two theoretically meaningful

and interpretable factors—perceived internal barriersclear. Consistent examples of lower levels of participa-

tion in physical activity among females have been and perceived external barriers. Thus, the findings sug-gest the likelihood that perceived barriers to physicalreported in the scientific literature, indicating that

higher levels of barriers would be expected among this activity is a multidimensional concept. While previous

TABLE 6

Multiple Regression of Vigorous Physical Activity on Perceived Barriers, Self-Efficacy, Age, and Gender—Grade 11 Students

Physical educ class Other school Outside school Overall(N 5 291) (N 5 289) (N 5 289) (N 5 286)

Stand Regr Stand Regr Stand Regr Stand RegrCoeff P, Coeff P, Coeff P, Coeff P,

Perceived bars—internal 20.02 0.754 0.03 0.598 20.08 0.173 20.05 0.387Perceived bars—external 0.06 0.329 0.03 0.574 20.02 0.973 0.09 0.146SE—external barriers 0.09 0.186 0.22 0.001 0.31 0.001 0.23 0.001SE—internal barriers 20.09 0.208 0.02 0.779 20.01 0.988 20.02 0.781Age 20.02 0.736 20.05 0.436 20.02 0.700 20.03 0.633

0.003 20.20 0.001 20.20 0.001

re P, Adj R square P, Adj R square P,0.001 0.15 0.001 0.12 0.001

Page 7: Perceived Barriers to Physical Activity among High School Students

614 ALLISON, DWYE

research on the factor structure of perceived barriersalso indicated its multidimensionality, the specific fac-tor structure found here (internal and external barri-ers) has not been previously identified. Research onadditional samples is needed to confirm the factor struc-ture explored here.

The hypothesis that perceived barriers would predictparticipation in vigorous physical activity was onlypartly supported. The perceived internal barriers factorwas shown to be inversely related to participation, asexpected, but only for overall activity and outside ofschool activity (not for physical education class or otherschool activity). Thus, it appears that perceptions ofinternal barriers are predictive of physical activity par-ticipation in settings or situations that are less struc-tured than school-related activities. This makes sense,since students who are not in the mood for exercise,who feel stressed or lack energy, are less likely to partic-ipate in physical activity when the environment sur-rounding their decision is not constrained, i.e., whenthey have discretionary “free time.” In school settings,there may be institutional expectations (largely uncon-trollable) affecting their participation in physical activ-ity. This finding lends support to the construct validityof the barriers measures.

Perceived external barriers were a weak but signifi-cant predictor of physical activity participation (overalland in other school activities). However, the directionof this relationship was opposite to that hypothesized—higher levels of perceived barriers were positively re-lated to participation. This does not necessarily meanthat those with high perceived external barriers aremore likely to initiate physical activity. It may be thatthose engaged in higher levels of physical activity aremore aware of obstacles to further activity. Because ofthe cross-sectional data, it is not possible to determinethe causal direction of the relationships observed.

Interestingly, perceived barriers were only predictiveof physical activity among grade 9 (not grade 11) stu-dents. This was an unexpected result, since we expecteda consistent relationship across grades. Additional re-search is needed to examine grade level differences inperceived barriers and other factors influencing physi-cal activity during the high school years. Also, it may beuseful in future studies to examine the role of perceivedbenefits in the relationship between perceived barriersand physical activity participation.

The findings from this study indicate that self-efficacy despite perceived barriers to physical activityis a more consistent predictor of physical activity partic-ipation than perceived barriers. Specifically, self-effi-

cacy despite external barriers was a positive predictorof vigorous physical activity, both overall and in otherschool and outside of school settings. This is not surpris-ing, given previously reported empirical support for therelationship between self-efficacy and physical activity

R, AND MAKIN

[7,8,22,32]. Perceived self-efficacy to engage in vigorousphysical activity despite internal barriers was not pre-dictive of participation in the hypothesized direc-tion—it was a weak predictor of outside of school activ-ity, in a negative direction. Possible reasons for thisfinding are discussed in a previous report [32]. Thefindings reported here raise the question of the extentto which perceived barriers affect physical activitythrough self-efficacy. Structural equation modeling canbe used to examine these relationships further.

Age and gender were consistent predictors of physicalactivity participation, overall and in the three set-tings—physical education class, other school activity,and outside of school activity. These findings are consis-tent with both the relationships hypothesized for thisstudy and the previous research on age and genderdifferences in physical activity [9–11]. Further researchis needed in order to understand the social meaning ofage and gender differences in physical activity partici-pation.

Some of the limitations of this study are as follows:the cross-sectional design of the survey makes it diffi-cult to sort out the causal relationships among variablesstudied; subjects included in the study represented stu-dents in an urban, ethnically diverse school district—thus the findings cannot be generalized beyond thosein similar groups; and the measurement of physicalactivity was restricted to vigorous activity and thusdid not include activity of moderate intensity. Also, themodel tested accounted for a relatively small amount ofvariance in vigorous activity, indicating that additionalfactors need to be considered when the purpose is tobuild and test comprehensive frameworks to explainand predict physical activity participation. Finally, thestudy should be replicated with additional samples ofstudents.

Despite these limitations, findings from this studyhave tentative implications for public health policiesand programs. Since patterns of physical activity indi-cate lower levels among older and female students,these groups should be considered important priorities.Furthermore, attempts to reduce barriers to activity

(both internal and external) should be encouraged. Fi-nally, programs designed to increase self-efficacy shouldbe supported in order that students will develop in-creased confidence to participate despite perceived andactual barriers to physical activity.

ACKNOWLEDGMENTS

Dr. Kenneth R. Allison is a Career Scientist of the Ontario Ministryof Health, Health Research and Development Program. Appreciationis extended to Patricia McNair for assistance in data collection and

processing and to two anonymous reviewers for comments.

REFERENCES

1. U.S. Department of Health and Human Services: Physical activ-ity and health: a report of the Surgeon General. Atlanta (GA):

Page 8: Perceived Barriers to Physical Activity among High School Students

T

PERCEIVED BARRIERS

U.S. Department of Health and Human Services, Centers forDisease Control and Prevention, National Center for ChronicDisease Prevention and Health Promotion, 1996.

2. Blair S, Kohl H, Paffenbarger R, Clark D, Cooper K, Gibbons L:Physical fitness and all-cause mortality: a prospective study ofhealthy men and women. JAMA 1989;262:2395–401.

3. Powell K, Blair S: The public health burdens of sedentary livinghabits: theoretical but realistic estimates. Med Sci Sports Ex-erc 1994;26:851–6.

4. Lee I, Hsieh C, Paffenbarger R: Exercise intensity and longevityin men: the Harvard Alumni Health Study. JAMA 1995;273:1179–84.

5. Dishman R, Sallis J, Orenstein D: The determinants of physicalactivity and exercise. Public Health Rep 1985;100:158–71.

6. Godin G, Shephard R: Use of attitude–behaviour models in exer-cise promotion. Sports Med 1990;10:103–21.

7. Sallis J, Hovell M: Determinants of exercise behavior. Exerc SportSci Rev 1990;18:307–30.

8. Reynolds K, Killen J, Bryson S, Maron D, Taylor C, MaccobyN, Farquhar J: Psychosocial predictors of physical activity inadolescents. Prev Med 1990;19:541–51.

9. Sallis J: Epidemiology of physical activity and fitness in childrenand adolescents. Crit Rev Food Sci Nutr 1993;33:403–8.

10. Heath G, Pratt M, Warren C, Kann L: Physical activity patternsin American high school students. Arch Pediatr Adolesc Med1994;148:1131–6.

11. Allison K, Adlaf E: Age and gender differences in physical inactiv-ity among Ontario teenagers. Can J Public Health 1997;88:177–80.

12. Dishman R, Dunn A: Exercise adherence in children and youth:implications for adulthood. In: Dishman R, Dunn A, editors. Exer-cise adherence: its impact on public health. Champaign (IL): Hu-man Kinetics, 1988.

13. Becker M, Haefner D, Kasl S, Kirsht J, Maiman L, Rosenstock I:Selected psychosocial models and correlates of individual health-related behaviors. Med Care 1977;15:27–48.

14. Strecher V, Rosenstock I: The health belief model. In: Glanz K,Lewis F, Rimer B, editors. Health behavior and health education:theory, research, and practice. 2nd ed. San Francisco: Jossey-Bass, 1997.

15. Stephens T, Craig C: The well-being of Canadians: highlightsof the 1988 Campbell’s Survey. Ottawa: Canadian Fitness andLifestyle Research Institute, 1990.

16. Wankel L, Mummery W: Factors related to intention to be physi-cally active: information from the Campbell’s Survey of Well-Being. Paper presented at the 1992 International Conference onPhysical Activity, Fitness, and Health, Ottawa, 1992.

17. Tappe M, Duda J, Ehrnwald P: Perceived barriers to exerciseamong adolescents. J Sch Health 1989;59:153–5.

18. Janz N, Becker M: The health belief model: a decade later. HealthEduc Q 1984;11:1–47.

19. Norman R: The nature and correlates of health behavior. Ottawa:Health and Welfare Canada, 1986. [Health Promotion StudiesSeries No. 2]

20. Hofstetter C, Sallis J, Hovell M: Some health dimensions of self-efficacy: analysis of theoretical specificity. Soc Sci Med 1990;31:1051–6.

21. Godin G, Valois P, Jobin J: Prediction of intention to exercise ofindividuals who have suffered from coronary heart disease. JClin Psychol 1991;47:762–72.

22. Sallis J, Hovell M, Hofstetter C, Faucher P, Elder J, Blanchard

O PHYSICAL ACTIVITY 615

J, Caspersen C, Powell K, Christenson G: A multivariate studyof determinants of vigorous exercise in a community sample.Prev Med 1989;18:20–34.

23. Slenker S, Price J, Roberts S, Jurs S: Joggers versus nonexercis-ers: an analysis of knowledge, attitudes and beliefs about jogging.Res Q Exerc Sport 1984;55:371–8.

24. Yoshida K, Allison K, Osborn R: Social factors influencing per-ceived barriers to physical exercise among women. Can J PublicHealth 1988;79:104–8.

25. Ali N, Twibell R: Health promotion and osteoporosis preventionamong postmenopausal women. Prev Med 1995;24:528–34.

26. Tappe M, Duda J, Menges-Ehrnwald P: Personal investmentpredictors of adolescent motivational orientation toward exercise.Can J Sports Sci 1990;15:185–92.

27. Stucky-Ropp R, DiLorenzo T: Determinants of exercise in chil-dren. Prev Med 1993;22:880–9.

28. Bandura A: Social learning theory. Englewood Cliffs (NJ): Pren-tice Hall, 1977.

29. Bandura A: Social foundations of thought and action: a socialcognitive theory. Englewood Cliffs (NJ): Prentice Hall, 1986.

30. Hovell M, Sallis J, Hofstetter C, Spry V, Faucher P, CaspersenC: Identifying correlates of walking for exercise: an epidemiologicprerequisite for physical activity promotion. Prev Med 1989;18:856–66.

31. Trost S, Pate R, Dowda M, Saunders R, Ward D, Felton G: Genderdifferences in physical activity and determinants of physical ac-tivity in rural fifth grade children. J Sch Health 1996;66:145–50.

32. Allison K, Dwyer J, Makin S: Self-efficacy and participation invigorous physical activity by high school students. Health EducBehav 1999;26:10–22.

33. Warwick D, Lininger C: The sample survey: theory and practice.New York: McGraw–Hill, 1975.

34. Green L, Lewis F: Measurement and evaluation in health educa-tion and health promotion. Palo Alto (CA): Mayfield, 1986.

35. Sallis J, Pinski R, Grossman R, Patterson T, Nader P: The devel-opment of self-efficacy scales for health-related diet and exercisebehaviors. Health Educ Res 1988;3:283–92.

36. Dwyer JM, Allison KR, Makin S: Internal structure of a measureof self-efficacy in physical activity among high school students.Soc Sci Med 1998;46:1175–82.

37. Stephens T, Jacobs D, White C: A descriptive epidemiology ofleisure-time physical activity. Public Health Rep 1985;100:147–58.

38. Heath G, Pate R, Pratt M: Measuring physical activity amongadolescents. Public Health Rep 1993;108:42–6.

39. Taylor H, Jacobs D, Schucker B, Knudsen J, Leon A, DebackerG: A questionnaire for the assessment of leisure time physicalactivities. J Chron Dis 1978;31:741–55.

40. Jacobs D, Ainsworth B, Hartman T, Leon A: A simultaneousevaluation of 10 commonly used physical activity questionnaires.Med Sci Sports Exerc 1993;25:81–91.

41. Washburn R, Montoye H: The assessment of physical activity byquestionnaire. Am J Epidemiol 1986;123:563–76.

42. American School Health Association, Association for the Ad-vancement of Health Education, Society for Public Health Educa-

tion: The National Adolescent Student Health Survey: a reporton the health of America’s youth. Oakland (CA): Third PartyPubl., 1989.

43. American College of Sports Medicine: Guidelines for exercise test-ing and prescription. 4th ed. Philadelphia: Lea & Febiger, 1991.

44. Tabachnick B, Fidell L: Using multivariate statistics. 3rd ed.New York: Harper Collins, 1995.


Recommended