Psychological determinants of adolescent exercise adherence.
According to a recent study by the Centers for Disease Control (1985), the typical American school-age child displays fitness and activity profiles well below the levels believed to be necessary to significantly lower health risks. In fact, only one-third of all American school-age children meet minimum fitness standards (Castrone, 1991). According to Varni (1983), one out of every five American children will develop clinical symptoms of coronary heart disease before age 16.
A wide array of factors can affect behavior; intrapersonal and environmental factors can combine in myriad ways to produce a given behavior. Likewise, the decision of whether or not to exercise has best been explained by interactional models where influences from several areas are considered. Environmental influences such as peer or parental support (Martin & Dubbert, 1985) as well as intrapersonal factors such as self-motivation (Dishman & Gettman, 1980; Dishman & Ickes, 1981), and perceived self-competency (Powell, 1988) have been found to have a significant impact on exercise behavior.
Some variables are beyond the control of researchers working in different settings. For example, although participation in organized sports as a child has been shown to have a positive relationship to adult participation in similar activities (Greendorfer, 1983), there is little that a trainer in a corporate fitness facility can do to change those previously established patterns. Likewise, although parental activity level has been shown to have a significant impact on a child's behavior (Gould & Horn, 1984), there is little a practitioner or researcher in a physical education classroom can do about parental activity level. However, if a student has low self-motivation, a physical education teacher could provide extrinsic rewards for continued exercise participation. Additionally, it is relatively easy to provide opportunities for student success in physical education programs, thereby influencing exercise self-efficacy or perceived physical competence.
Unfortunately, most of the actual exercise time in a typical physical education class is devoted to the performance of team sports and the development of motor skills which do not appeal to the majority of students (Williams, 1988). This excessive emphasis on such activities is likely to lead to future alienation from physical activity. Rather than promoting a lifelong fitness pattern, it seems that many public school physical education programs are laying the foundation for inactivity.
It is therefore important to be cognizant of psychological factors which increase the likelihood of youth adherence to exercise and begin to incorporate that knowledge into physical education settings early in life. Unfortunately, only in the last few years has anyone bothered to ask our youth what kind of physical activity they like (Fox & Biddle, 1988). Traditionally it has been the physiological rather than the psychological variables that have been measured. Many fitness tests provide important information about the physiological state of our youngsters' bodies, but what about their minds? Intrapersonal psychological variables have received little, if any, attention in physical education settings.
The intent of this study was to examine selected psychological variables in order to determine their influence on the exercise behavior of secondary school students. Of special interest in this study were differences in variable effects according to gender and level of exercise competitiveness.
Previous research on the determinants of exercise adherence has included comparisons of many different subgroups, including distinctions based on gender, age, and socioeconomic status, but very limited attention has been paid to differences based on race and type of physical exercise (Iverson, Fielding, Crow, & Christenson, 1985; Martin & Dubben, 1982, 1985; Powell, 1988). One area which has received little, if any, systematic attention is the comparison of subjects' exercise adherence determinants based on differences in levels of competitiveness. There has been descriptive research on competitive adult and young athletes, adults in unsupervised as well as supervised settings, and youths in supervised settings. However, there is a noticeable gap in the literature in the area of noncompetitive, unstructured youth exercise adherence. Many exercising adolescents are not competitive athletes under the supervision of a coach or a teacher. Hence, it is important to understand the motivation of this noncompetitive, unsupervised group of adolescents.
The present study not only examined differences by gender and exercise setting, but attention was focused on whether subjects were competitive or noncompetitive exercisers. Perhaps future research could further subdivide this analysis to include an examination of gender differences within level of competitiveness.
A relatively unexplored psychological variable in youth is personality/sport congruence (Gavin, 1987, 1988a, 1988b). This is a measure of the degree of "match" between an individual's personality style and the type of exercise the person pursues. Since this variable can be easily manipulated in physical education classes as well as in unstructured exercise settings, it was included in this study.
The three additional psychological variables evaluated for their predictiveness of adolescent exercise adherence were self-motivation (Dishman & Gettman, 1980; Dishman & Ickes, 1981), perceived self-competency (Harter, 1988), and perceived control (Howard, Adelan, Smith, Nelson, Taylor, & Phanes, 1986). Self-motivation, as defined by Dishman & Gettman 1980; Dishman & Ickes, 1981, has been found to be a relatively consistent predictor of exercise adherence in adults, but it has not been well evaluated in youth. Perceived physical competency has been a moderately effective predictor of youth exercise activity in a few studies (Roberts, Kleiber, & Duda, 1981; Tappe, Duda, & Menges-Ehrnwald, 1990), but the results are inconclusive. The amount of control over the environment and exercise activity that one perceives has been shown to increase the frequency of exercise participation in adults (Langer, 1983; Thompson & Wankel, 1980). However, since no research was found which examined the effect this variable (perceived control) may have on adolescent exercise adherence, it was selected for the current study.
The Habitual Physical Activity Questionnaire (Baecke, Burema, & Frijters, 1982) was used here to define exercise adherence. This instrument is composed of three component scores: the Work Index, which measures job-related physical activity; the Sport Index, which is a measure of sport-related exercise activity; and the Leisure Index, which measures physical activity during leisure time (such as riding a bike to the store instead of driving a car). Since the focus of this study was on adherence to sport exercise activity, only the Sport Index score was used to determine the subjects' dependent variable scores.
The original research from which this data was excerpted (Douthitt, 1992) was more comprehensive; however, only the predictive value of the psychological independent variables for the dependent variable of exercise adherence was considered here. Specifying the determinants of physical activity/inactivity is fundamental to establishing necessary interventions aimed at reducing obesity and other hypokinetic disease in youth.
A correlational research design was used in which the independent variables of self-perception, self-motivation, perceived control, and personality/sport congruence were examined for their predictive value for the dependent variable of adolescent exercise adherence. The data collected were from written survey instruments representing the independent or predictor variables in this study: (1) The Adolescent Self-Perception Profile (Harter, 1988); (2) The Self-Motivation Inventory (Dishman & Ickes, 1981); (3) The Perceived Control at School Scale (Howard et al., 1986); and (4) The Psychosocial Activity Dimensions Profile (Gavin, 1987, 1988a, 1988b). The instrument which defined the dependent or criterion variable was The Habitual Physical Activity Questionnaire (Baecke, Burema, & Frijters, 1982). These instruments were selected because of their potential practical implications to physical education and their sound psychometric data.
Informed consent was obtained from both the physical education students and their parents. There were two data collection sessions. The first occurred in May at the end of the 1989-1990 school year and represented data from a supervised classroom setting. The second session was held in September of 1990, the beginning of the 1990-1991 school year. Both data collections were held at the high school during a regularly scheduled class time. Subjects were asked to complete the same five written surveys in each session. However, subjects were instructed to answer the five questionnaires as they pertained to their physical education classroom experiences in the first data collection, and to answer them as the questions related to their summer vacation experiences for the last collection of data.
Once the data were collected and analyzed, the results were presented to a panel of experts in physical education and sport psychology. They reviewed the results and presented their recommendations to this author. Stepwise multiple regression was used as the statistical variable selection technique-due to a lack of prior research using all these variables which would suggest the order of variable entry.
The sample for the first data collection was 132 9th-, 10th-, and 11th-grade male (n = 94) and female (n = 38) public school students who were enrolled in physical education classes at a large suburban high school in metropolitan Denver, Colorado. They were enrolled in classes taught by four different physical education teachers. A wide range of physical education activities was represented by these students in their classes including cycling, weight training, jogging, swimming, jazz and modern dance, golf, martial arts, racquet sports, and fitness walking.
For the second phase of data collection, 110 of the original subjects were available for follow-up procedures; 22 had either moved or declined to take part in the data collection. Of the 110 subjects, 77 were male and 33 were female. Again, a wide range of activities were represented, this time over their summer vacation: aerobic classes, cycling, dance, fitness walking, golf, jogging, martial arts, racquet sports, swimming, waterskiing, and weightlifting.
The results of the data analysis indicated predictive differences for males and females. For females, Perceived Athletic Competency was an inversely significant predictor of exercise adherence in the physical education class setting, accounting for 26% of the variance.
For males in the same setting, Perceived Romantic Appeal was inversely significant, yet accounted for a small amount of the variance (8%) in exercise adherence.
In the unstructured exercise setting, none of the variables had significant predictive value for exercise adherence in males. In females, Perceived Global Self-Worth and Perceived Physical Appearance collectively accounted for 32% of the total variance in exercise adherence.
Competitive and noncompetitive subgroups were also examined. Those subjects who responded "yes" to the question "Do you participate in junior varsity athletics, varsity athletics, or other competitive athletic programs?" were classified as competitive athletes, and those who responded "no" were classified as noncompetitive. The results of the prediction of exercise adherence for the two levels of competitiveness indicated differences in the behavior motivation of the two groups, as well as the behavior motivation from one setting to another. None of the variables was predictive of exercise adherence for the competitive subjects in either the structured or unstructured settings. However, for the noncompetitive subjects, Perceived Romantic Appeal was inversely predictive of exercise adherence (10% of the variance) in the physical education class setting. In the unstructured summer vacation setting, Personality/Sport Congruence was inversely predictive (29% of the variance) of the exercise adherence of noncompetitive subjects.
TABULAR DATA OMITTED
Table 2 Variables Predictive of Exercise Adherence In Females: Unstructured Setting Variable r [R.sup.2] F Sig. F 1. Perceived Global Self-worth -.43 .18 6.0 .02 2. Perceived Physical Appearance -.09 .32 6.1 .007
TABULAR DATA OMITTED
In order to interpret the findings, two general categories of interest were addressed: (1) Gender differences in determinants of adolescent exercise adherence; and (2) Differences in determinants of exercise adherence in subjects with different levels of competitiveness. A panel of experts in physical education and exercise science reviewed the data and made recommendations.
Gender Differences in Determinants of Youth Exercise Adherence
One variable, Perceived Romantic Appeal, was predictive of exercise adherence in the physical education setting for males. Although the amount of variance was modest (8%), it is interesting to note that the relationship was negative (-.28). This suggests that as perceived romantic appeal decreased, adolescent males increased their physical activity. It seems that male subjects who perceived themselves to be low in romantic appeal viewed more exercise as a way to improve their appeal. For females, Perceived Athletic Competency was a significant predictor of exercise adherence (26% of the variance), and the relationship was inverse (-.51). Perhaps they also sought to improve their competency perception through greater activity. If a female had a low perception of her athletic competency, she may have viewed an increase in her exercise level as a way to improve this competency. This finding is in conflict with that of Tappe, Duda, and Menges-Ehrnwald (1990) who found a moderate (.34) positive relationship between the two variables for adolescent females, but perceived athletic or physical competency was defined differently in the earlier work, which could account for the conflicting results.
For males in the unstructured exercise setting, none of the variables reached predictive significance. For females, two variables were predictive of exercise adherence: perceived global self-worth (18% of the variance) and perceived physical appearance (14% of the variance). The results of this analysis for females were particularly interesting. The initial simple correlations indicated an inverse relationship among exercise adherence and the two predictor variables, with a very low correlation between physical appearance and exercise adherence. However, once perceived global self-worth had entered the regression equation, the partial correlation between physical appearance and exercise adherence changed to a positive relationship (.41), and increased in magnitude. Harter (1990) has stated that the most consistently strong contributor to perceived global self-worth in adolescents is perceived physical appearance. The relationship between these two variables in this instance was the strongest among the research variables, but not one of multicollinearity (.76). Multicollinearity does not become problematic until correlations of .90 or higher are reached (Tabachnick & Fidell, 1989).
Thus, components of perceived global self-worth, which when partialed out or controlled, must alter the relationship of perceived physical appearance to exercise adherence in this group of females. For example, if an individual was secure or felt competent in her global self-esteem, she might not have perceived a need for increased physical activity. However, if the feelings of global self-worth are controlled for, the positive relationship between perceived physical appearance and exercise adherence becomes clear. One cannot know from the correlational relationship between exercise adherence and perceived physical appearance which variable came first. For example, does an individual with a more positive perceived physical appearance tend to seek out exercise experiences to maintain the positive perception, or does a higher level of physical activity foster an increase in perceived physical appearance?
The results of the present research are therefore consistent with previous work which has indicated that the determinants of exercise adherence vary between the sexes (Gould & Horn, 1984).
Differences in Determinants of Exercise Adherence in Subjects with Different Levels of Competitiveness
The findings of this research, in the supervised (physical education classes) as well as in the unstructured (summer vacation) exercise setting, indicated that none of the predictor variables was significant for the competitive group of subjects. Previous research has indicated that "to have fun," "to improve and learn skills," "to make and be with friends," and "to improve physical fitness" are consistent determinants of competitive young athletes' exercise adherence levels in supervised settings (Fox & Biddle, 1988; Gill, Gross, & Huddleston, 1983; Gould & Horn, 1984; Sapp & Haubenstricker, 1978). Perhaps this study did not tap the psychological variables which were most related to competitive adolescent exercise adherence.
For noncompetitive subjects in the supervised exercise setting, perceived romantic appeal was statistically significant, accounting for 10% of the variance. Although this leaves the majority of the variance in exercise adherence unaccounted for, this result and the results of the same analysis for the unstructured exercise setting do indicate that there probably are some differences in the exercise behavior motivation of competitive and noncompetitive adolescents. This particular finding regarding the relationship between exercise adherence and perceived romantic appeal has not been found in other research, hence future studies should investigate this relationship. As with the male subgroup discussed previously, the nature of the relationship was inverse, again suggesting an attempt by noncompetitive subjects with low perceived romantic appeal to improve that self-perception through increased exercise levels.
For the unstructured exercise setting, personality/sport congruence was predictive of the noncompetitive subjects' exercise adherence levels. The amount of variance in exercise adherence accounted for by this variable (29%) was greater than that of any other variable for any other analysis in this study. This correlational relationship was also inverse (-.54), yet it had a different meaning from an inverse relationship in other analyses with other variables. A lower personality/sport congruence score is associated with a greater degree of "match" between the individual's personality and the type of exercise activity in which he/she is engaged. In essence, this is a discrepancy score. Thus, the finding indicates that the lower the discrepancy between a subject's personality and exercise activity, the greater the individual's exercise adherence. This supports the theory proposed by Gavin (1987, 1988a, 1988b).
Given the accumulated research which indicates that "having fun" is the most important determinant of youth exercise adherence, the lack of a definitive answer as to exactly what is meant by "having fun," and given the strength of the association of personality/sport congruence to exercise adherence in this study, it could be that a match between an adolescent's personality and type of exercise activity results in more fun. Exercise advisors to people of all ages tend to suggest that they choose an activity they enjoy. This finding seems to support the contention that, when placed in a situation with the most options, (unsupervised setting; not having to think about being prepared for the next competitive season), adolescents chose activities which were the most like them. Other research with this instrument (Gavin, 1987, 1988a, 1988b) was performed using undergraduate college students; thus, future research should explore the relationship between personality/sport congruence and exercise adherence in adolescents in greater depth.
Even with the obvious limitations of self-report questionnaires, the results of this study indicate a significant link between some domains of perceived self-competency and exercise adherence, and between personality/sport congruence and exercise adherence in this group of adolescents.
This research was unique in that some of the psychological variables with the most predictive strength had inverse relationships with exercise adherence. Another unique finding was that whereas in previous studies of adolescent populations (Tappe, Duda, & Menges-Ehrnwald, 1990) the subjects were motivated to higher levels of exercise by feelings of competence in athletic endeavors, the adolescents in this study were motivated to increase exercise levels based on self-competence factors related to lower perceived physical appearance and lower perceived romantic appeal to others.
The implications for physical educators and those developing curricula are significant whether the relationship between domains of perceived self-competency and exercise adherence are positive or inverse, In either situation, perceived self-competency does seem to be important for improving levels of exercise activity. Hence, it is important to provide opportunities for success in physical education experiences for all children in order to lay a foundation for lifelong exercise activity.
Of the remaining psychological factors considered in this study, only personality/sport congruence had any predictive strength. This variable would be relatively easy to incorporate into a physical education setting by assessing the child's personality type and providing him/her with experiences in activities which "match" the personality profile.
Although self-motivation has shown a consistent positive relationship with adult exercise adherence (Dishman & Gettman, 1980; Dishman & Ickes, 1981), there was no similar finding in the present study. Perhaps self-motivation, as defined in Dishman's instrument is not valid for use with adolescents; or, perhaps self-motivation is more domain-dependent in adolescence than it was defined to be for adults in the original conception of the Self-Motivation Inventory.
The school where these data were collected was unusual in the amount of choice and control students had in their physical education programs. This may have resulted in the lack of predictive strength of the perceived control questionnaire.
The models which typically account for the most variance in exercise adherence are those which are multifaceted. It is generally necessary to take many different categories of factors into consideration, such as intrapersonal, interpersonal, and environmental. In this study, one component, intrapersonal factors, was examined. This could account for much of the unexplained variance in exercise adherence.
In future exercise adherence research, it would be interesting to use these variables with other adolescent populations to determine if the results would be replicated.
Baecke, J. A. H., Burema, J., & Frijters, J. E. R. (1982). A short questionnaire for the measurement of habitual physical activity in epidemiological studies. American Journal of Clinical Nutrition, 36, 936-942.
Castrone, L.(1991). Keeping kids fit. The Rocky Mountain News. pp. 63, 65.
Centers for Disease Control (1985). The status of the 1990 objectives for physical fitness and exercise. Morbidity and Mortality Weekly Report, 34, 521-524, 529-531.
Dishman, R. K., & Gettman, L. R. (1980). Psychobiological influences on exercise adherence. Journal of Sport Psychology, 2, 295-310.
Dishman, R. K., & Ickes, W. (1981). Self-motivation and adherence to therapeutic exercise. Journal of Behavioral Medicine, 4, 421-438.
Douthitt, V. (1992). Psychological determinants of exercise adherence in secondary school students. Doctoral Dissertation: University of Denver.
Fox, K., & Biddle, S. (1988). The child's perspective in physical education part 2: Children's participation motives. British Journal of Physical Education, 19, 79-83.
Gavin, J. (1987). Some observations on the relationship between activity preferences, participation, and personal profiles. (Technical manuscript.) Montreal: Concordia University.
Gavin, J. (1988a). Psychological issues and exercise prescription. Sports Medicine, 6, 1-10.
Gavin, J. (1988b). Body moves: The psychology of exercise. Harrisburg, PA: Stackpole Books.
Gill, D. L., Gross, J. D., & Huddleston, S. (1983). Participation motivation in youth sports. International Journal of Sport Psychology, 14, 1-14.
Gould, D, & Horn, T. (1984). Participation motivation in young athletes. In J. M. Silva, & R. S. Weinberg (Eds.), Psychological foundations of sport (pp. 359-370). Champaign, IL: Human Kinetics Books.
Greendorfer, S. L. (1983). Shaping the female athlete: The impact of the family. In M. Boutslier, & L. San Giovani (Eds.), The sporting woman (pp. 135-155). Champaign, IL: Human Kinetics Books.
Harris, L. (1979). The Perrier study: Fitness in America. New York: Perrier Great Waters of France, Inc.
Harter, S. (1988). Manual for the Self-Perception Profile for Adolescents. Denver, CO: University of Denver.
Harter, S. (1990). Self and identity development. In S. S. Feldman, & G. R. Elliott (Eds.), At the threshold: The developing adolescent. Cambridge, MA: Harvard University Press.
Howard, S., Adelman, H., Smith, D. C., Nelson, P., Taylor, L., & Phanes, V. (1986). An instrument to assess students' perceived control at school. Educational and Psychological Measurement, 46, 1005-1017.
Iverson, D. C., Fielding, J. E., Crow, R. S., & Christenson, G. M. (1985). The promotion of physical activity in the United States population: The status of programs in medical, worksite, community, and school settings. Public Health Reports, 100(2), 212-224.
Langer, E. (1983). The psychology of control. Beverly Hills, CA: Sage Publications.
Martin, J. E., & Dubbert, P. M. (1982). Exercise applications and promotion in behavioral medicine: Current status and future directions. Journal of Consulting and Clinical Psychology, 50(6), 1004-1017.
Martin, J. E., & Dubbert, P. M. (1985). Adherence to exercise. Exercise and Sport Science Reviews, 13 137-167.
Powell, K. E., (1988). Habitual exercise and public health: An epidemiological view. In R. K. Dishman (Ed.), Exercise adherence: Its impact on public health (pp. 15-39). Champaign, IL: Human Kinetics Books.
Roberts, G. C., Kleiber, D. A., & Duda, J. L. (1981). An analysis of motivation and children's sport: The role of perceived competence and participation. Journal of Sport Psychology, 3, 206-216.
Sapp, M., & Haubenstricker, J. (1978). Motivation for joining and reasons for not continuing in youth sports programs in Michigan. Paper presented at the American Association of Health, Physical Education, and Recreation National Convention, Kansas City, Missouri.
Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics (2nd ed.). San Francisco, CA: Harper Collins Publishers.
Tappe, M. K., Duda, J. L., & Menges-Ehrwald, P. M. (1990). Personal investment predictors of adolescent motivational orientation toward exercise. Canadian Journal of Sport Science, 15(3), 185-192.
Thompson, C. E., & Wankel, L. M. (1980). The effects of perceived activity choice upon frequency of exercise behavior. Journal of Applied Social Psychology, 10(5), 436-443.
Varni, J. W. (1983). Clinical behavioral pediatrics: An interdisciplinary approach. New York: Pergamon.
Williams, A. (1988). Physical activity patterns among adolescents--Some curriculum implications. Physical Education Review, 11, 1, 28-39.