Estimated marginal means and standard errors of the primary health behavior measures are shown by group and time in . Omnibus repeated measures MANOVAs were performed for eight groupings of health behavior measures. These analyses were significant for drinking driving behaviors, F(2,136)=4.43, p=.01, exercise behaviors, F(5,140)=6.12, p=.00), nutrition habits, F(3,143)=5.37, p=.00, sleep habits, F(2,144)=5.03, p=.01, and health quality of life, F(5,140)=3.09, p=.01, with improvements on each of these behaviors across time. No differences were seen over time on alcohol, cigarette, and marijuana consumption measures.
Estimated marginal means of health behavior measures by group and time
Univariate analyses showed decreases in the frequency of riding with a drunk driver, F(1,145)=9.63, p=.01, and a near significant decrease in driving while drunk, F(1,137)= 3.64, p=.06. In addition, a repeated measures ANOVA showed the frequency of substance use self-control behaviors increased over time for all groups, F(1,142)=12.71, p=.00. Univariate tests also showed increases in length of time engaged in exercise, F(1,144)=5.60, p=.02, 30-day frequency of moderate physical activity, F(1,144)=14.96, p=.00, and 7-day moderate exercise F(1,144)=13.67, p=.00, over time for all three groups. An increase in the consumption of foods containing healthy fats in the past 7-days, e.g., vegetable oil, seeds, nuts, olive oil, or fish, was found over time, F(1,145)=4.67, p=.03. In addition, an increase in both the number of hours of sleep each night, F(1,145)=7.68, p=.01, and frequency of getting enough sleep, F(1,145)=7.15, p=.01, was seen for all groups over time. Univariate tests also indicated a significant reduction over time for all groups in days that their mental health, F(1,144)=13.36, p=.00, and spiritual health were not good, F(1,144)=5.09, p=.03, as well as a reduction in days that poor health kept them from doing usual activities, F(1,144)=7.49, p=.01, indicating improvements in health quality of life. Lastly, group by time interaction effects showed an increase in the use of techniques to manage stress, F(2,144)=5.48, p=.01, and the number of health behavior goals set in the last 30 days, F(2,143)=5.35, p=.01, but only among those adolescents receiving the consult, or consult plus contract.
Effect sizes calculated for significant univariate tests ranged from small to medium, but were primarily medium in size for adolescents receiving the consult plus contract. Larger medium effects were seen for 30-day moderate physical activity and 7-day average moderate exercise for adolescents receiving the consult or consult plus contract, with effects approaching a large size for 30-day moderate physical activity for adolescents receiving the consult. Medium size effects for students receiving the consult plus contract were also seen on measures of eating good fats, quality of sleep, self-control behaviors, stress management, setting specific health goals, and activity limitation. Meanwhile, medium effects were seen for students in all groups on improved mental health, and for those receiving the consult and consult plus contract on spiritual health. Lastly, a medium effect was found for adolescents receiving the contract on riding with a drinking driver.
To better understand the differential change in health behavior goal setting by students, we examined the frequency and percentages of participants setting or trying to achieve specific health behavior goals during the past 30 days by group, shown in . Among participants receiving the contract alone, an increasing proportion set goals related to two health behaviors, while those receiving the consult, and those receiving the consult plus contract, showed increases in goal setting for eight behaviors.
Frequency and percentage of participants setting health behavior goals by group and time
Repeated measures ANOVAs were used to examine a number of belief and image-related factors associated with the underlying Behavior-Image Model. Significant treatment by time interaction effects were found on the self-image of being physically active and athletic, with a decreasing image of being physically active and athletic among students receiving the contract alone, but an increasing image of being physically active and athletic for students receiving the consult, and the consult plus contract, F(2,144)=4.69, p=.01. A related measure of body image satisfaction was found to differ over time, with those participants in the contract only group showing no change from pre to post test, but those participants in the consult, and consult plus contract groups showing increased satisfaction with how their body looked, F(1,144)=4,19, p=.04. No differences were found on self-image associated with being a partier (e.g., alcohol drinker) or a drug user (p’s>.05). There was a significant difference over time, however, on the alcohol health behavior coupling belief, with participants in all groups more likely to agree that drinking too much alcohol interferes with other health behaviors, F(1,144)=26.09, p=.00. Meanwhile, there was a near significant difference over time on the exercise health behavior coupling belief, with students receiving the consult, and those receiving the consult plus contract, more likely to agree that regular physical activity or exercise improves other health behaviors, F(1,144)=3.70, p=.06.