Exercise and Psychosocial Health
Biddle, Stuart, Research Quarterly for Exercise and Sport
The rapidly expanding literature on exercise and health has always included psychological and psychosocial health, albeit initially at a rather anecdotal level of describing the "feel good effect" participants often report after exercise (Biddle & Mutrie, 1991). In reviewing the literature, exercise will not be strictly delimited. Habitual physical activity, exercise, aerobic fitness training, and sport will all be included where appropriate. Psychosocial health is also broadly defined to include psychological and social-psychological outcomes. However, there is no accepted definition in the field, although mental health usually includes positive characteristics, such as high self-esteem and positive mood (Mutrie & Biddle, 1995; Stephens, 1988), as well as reduced levels of negative affect, such as anxiety and depression (Martinsen & Stephens, 1994). More recently, as far as exercise is concerned, other aspects have also been studied (see Mutrie & Biddle, 1995), such as cognitive functioning and stress reactivity, as well as the potential negative psychological effects of exercise (e.g., dependence).
In the meta-analyses referred to below, trends are quantified across studies through the calculation of an effect size (ES) which represents the meaningfulness of exercise effects. It is calculated as the difference between experimental (treatment) and comparison (control) group means divided by the control group or pooled standard deviation (see Salazar, Petruzzello, Landers, Etnier, & Kubitz, 1993; Thomas & French, 1986). For the sake of interpretation, ESs below 0.39 will be classified as small, 0.40-0.69 as moderate, and above 0.70 as large (Thomas, Salazar, & Landers, 1991).
Anxiety and Reactions to Stress
An area that has received a great deal of attention in this area is the potential of exercise to reduce anxiety and influence stress reactivity. Petruzzello, Landers, Hatfield, Kubitz, and Salazar (1991) conducted a meta-analysis investigating state, trait, and psychophysiological measures of anxiety in relation to exercise. An overall ES for state anxiety of 0.24 (n of ESs = 207) was found. Few possible moderating variables showed strong effects. For example, there was no difference between acute and chronic exercise, between state anxiety measures taken immediately and 30 min after exercise, or between participants differing in health status or age. However, the positive effect was noted only for aerobic exercise, although the number of anaerobic studies was low (n = 13).
For trait anxiety, Petruzzello et al. (1991) found an overall ES of 0.34 (n = 62), suggesting a small effect for chronic exercise. Again, few moderating variables had a strong effect, although training programs greater than 10 weeks seemed to provide the stronger effects. Finally, using psychophysiological indicators of anxiety, Petruzzello et al. found an ES of 0.56 (n = 138), depicting a moderate effect for exercise. All psychophysiological measures had significant ESs, although blood pressure and heart rate showed smaller effects than skin, electromyographic, and central nervous system measures. All periods of exercise duration showed significant effects, although larger effects were noted for periods up to 30 min.
Overall, these results show that exercise is associated with small-to-moderate reductions in anxiety when anxiety was assessed by either physiological or self-report measures. This effect was largely independent of key moderator variables and thus leaves open the issue of underlying mechanisms of the anxiety-reducing effects of exercise.
McDonald and Hodgdon (1991) also performed a meta-analytic review, looking only at studies investigating aerobic fitness training, in contrast to Petruzzello et al. (1991) who included all types of exercise. Nevertheless, McDonald and Hodgdon's results are comparable to those of Petruzzello et al., with ESs of 0.28 for state anxiety (n = 13) and 0. …