Electronic Thesis and Dissertation Repository


Doctor of Philosophy




Dr. Harry Prapavessis


The objective of this dissertation was to address salient conceptual and measurement issues related to physical activity and self-efficacy in adolescents. Building upon previous physical activity device validation studies, Study 1 investigated the measurement of agreement between the Actiheart (a combined heart rate monitor and accelerometer device) and doubly labeled water (DLW) for estimating free-living physical activity energy expenditure (PAEE) in a sample of adolescents. The Actiheart was found to overestimate PAEE compared to DLW by a mean difference of 9.80 kcal∙kg-1∙d-1 (95% limits of agreement: -21.22 to 1.72 kcal∙kg-1∙d-1). The Actiheart was found, however, to improve device wear time compliance in adolescents (i.e., the number of days valid measures of PAEE were obtained), an issue that has been raised previously with respect to using objective physical activity tools in this population.

In an attempt to improve the level of specificity of current physical activity efficacy measures, Study 2, focused on developing a conceptually based and psychometrically sound domain-specific physical activity efficacy questionnaire (The Self-Efficacy for Daily Physical Activity Questionnaire; SEPAQ). Results supported the tenability of two independent models. Model 1 consisted of five independent domain-specific factors of physical activity efficacy (leisure-time, household, transportation, school, and ambulatory transportation and school) while Model 2 included two additional domain-specific physical activity efficacy factors (occupation and active gaming).

Finally, the application of the SEPAQ was tested in a physical activity prediction study (Study 3). Model 1 explained 20% and 9% of the subjective (Physical Activity Questionnaire-Adolescents) and objective (Actiheart derived PAEE) physical activity variance, respectively. The addition of the domain-specific physical activity efficacy measures, occupation and active gaming in Model 2, increased the amount of variance explained in subjective and objective physical activity to 27% and 24%, respectively. Implications and directions for future research are discussed.

Included in

Kinesiology Commons