The Physical Activity Enjoyment Scale (PACES) is commonly used to assess enjoyment during physical activities, and its measurement properties have been evaluated using Rasch models in healthy adults. However, the use of PACES to assess physical activity (PA) enjoyment has not been cross validated among high school students. This study used Rasch analysis to evaluate construct validity and reliability of the PACES in a high school student sample.
Method:
A total of 349 9th-grade students (180 male; 169 female), aged 13 to 15 years (M= 14.13 ± .793), completed the 18-item PACES scale at the conclusion of physical education classes. Responses to PACES items were analyzed using Rasch partial credit model with Winsteps software. The construct validity of PACES was determined by examining dimensionality, item Infit and Outfit statistics, and person-item map. Reliability was evaluated by checking the Rasch separation indexes. Finally, the appropriateness of the 6-point Likert response (1= “Totally true” to 6 = “Totally false”) format was determined by examining the Rasch step difficulty.
Analysis/Results:
Of the 18 PACES items, five (i.e., “I feel bored”, “It makes me depressed”, “It’s very invigorating”, “I am very frustrated by it”, and “It’s not at all stimulating”) were identified as misfit items (Infit and Outfit statistics < 0.75 or > 1.33 logits) and were removed. The remaining 13 good-fit items (difficulty level ranged from .81 for Item “I feel good physically while doing it” to .66 logits for Item “I felt as though I would rather be doing something else”) demonstrated one dimension which accounted for 62.4% of the variance in the measure. PA enjoyment levels of 11.7% of the students were not covered by item difficulty levels at the higher end of the Rasch person-item map. The separation index for person measures was 3.01 logits, with a reliability of 0.90 (equivalent to traditional Cronbach’s Alpha), and the separation index for item measures was 6.12 logits with a reliability of 0.97. Lastly, the Rasch threshold estimates (i.e., step difficulty) were disordered for six items.
Conclusions:
Rasch analysis supported the construct validity and reliability of PACES to assess PA enjoyment among high school students after removing five misfit items. However, the 6-point Likert response format was not appropriate for some remaining items. Future studies are needed to develop more easier-to-endorse items in the scale so that the broad range of PA enjoyment levels among high school students can be adequately assessed.