A Two-Sample Analysis of the Construct Validity and Temporal Stability of a Process-Oriented Measure of Exercise...

The Exercise Habit Strength Questionnaire (EHSQ) has shown excellent psychometric properties. However, children in different weight categories exhibit distinct patterns of exercise engagement. Therefore, examining the EHSQ's psychometric properties across different weight groups across time may provide additional evidence of its discriminatory power and temporal stability.

This study was conducted in two samples. Sample 1 involved 450 secondary school students (mean age: 14.72 ± 1.19 years), consisting of 52% males and 48% females across different body mass index (BMI) categories-normal (17.70%), overweight (23.44%), and obese (58.85%). Data collection involved the EHSQ, Body Appreciation Scale and the Implicit Association Test. Sample 2 consisted of 60 female students aged 16 years from all-girls schools, with 48.3% in the normal category and the remaining underweight, overweight, and obese categories. The EHSQ was administered twice, with a three-month interval between administrations.

Confirmatory factor analysis provided evidence of factorial validity (χ ² = 210.06, df = 129, P < 0.05; χ ²/df = 1.63; CFI = 0.96; GFI = 0.95; RMSEA = 0.04; AIC = 314.69). Item loadings were above 0.5, except for one item, with latent factor intercorrelations ranging from 0.2 to 0.5. The results of the Intraclass correlation coefficient analysis demonstrated excellent test-retest reliability. Additionally, internal consistency scores in Sample 2 mirrored those of Sample 1, with Cronbach's alpha values of 0.756 for patterned action, 0.757 for automaticity, 0.89 for negative consequences, and 0.84 for stimulus cue.

Overall, the results provided adequate support for the hypothesised psychometric properties of the ESHQ.