|Jago, Russell - UNIVERSITY OF BRISTOL, UK|
|Klesges, Lisa - ST. JUDE'S CHILD HOSPITAL|
Submitted to: Health Education Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: April 1, 2006
Publication Date: December 1, 2006
Citation: Watson, K., Baranowski, T., Thompson, D., Jago, R., Baranowski, J., Klesges, L.M. 2006. Innovative application of a multidimensional item response model in assessing the influence of social desirability on the pseudo-relationship between self-efficacy and behavior. Health Education Research. 21(Suppl 1):i85-i97. Interpretive Summary: There has been some concern that current psychosocial measures may lack validity and reliability. As a result, we submitted our measure of fruit and vegetable (FV) self-efficacy, which met usual classical test theory criteria for a valid and reliable scale to the latest psychometric procedure: Item Response Modeling (IRM). IRM revealed that the scale did not cover the expected range for distribution of similar scale scores, and that the items did not cover the more difficult end of the distribution. The measure of FV self-efficacy needs to be modified to meet the enhanced IRM requirements for a valid and reliable scale.
Technical Abstract: This study examined multidimensional item response theory (MIRT) modeling to assess social desirability (SocD) influences on self-reported physical activity self-efficacy (PASE) and fruit and vegetable self-efficacy (FVSE). The observed sample included 473 Houston-area adolescent males (10–14 years). SocD (nine items), PASE (19 items) and FVSE (21 items) were measured with previously validated self-report instruments containing Likert-type responses. Physical activity was objectively measured using the Computer Science Application Incorporated/Manufacturing Technology Incorporated (CSA/MTI) accelerometer. Total fruit, juice and vegetable consumption was measured with a food frequency questionnaire. Correlations between self-efficacy and behaviors were minimal, regardless of controlling for SocD. However, in a simulated sample derived to demonstrate the utility of MIRT when relationships exist, the pseudo-relationships between self-efficacy and behaviors were substantially weaker after controlling for SocD. MIRT provided disattenuated correlations between SocD and self-efficacy, thereby providing more precise estimates of the real influence of SocD on the relationship between self-efficacy and behavior. However, as shown in the observed sample, more research is needed to understand the influence of SocD on the relationship between self-efficacy and behaviors for different populations and for different degrees of SocD response bias.