Construct Validity of the Revised High-Level Mobility Assessment Tool for Children Who Are Developing Typically.

Abstract

The aim of this project was to evaluate the construct validity of the high-level mobility assessment tool (HiMAT) for children who are developing typically with no underlying condition that may affect mobility. The HiMAT is a revised 8-item measure of high-level mobility. Rasch analysis was used to assess fit to the Rasch model indicating unidimensionality, person separation index local dependency, targeting of items, and differential item functioning. Children with typical development aged 5 to 12years (n = 1091; 554 girls and 537 boys) were recruited from 8 schools. The mean HiMAT score was 22.4/32 (range, 8-32). Data fit the Rasch model, indicating that the HiMAT was a unidimensional scale assessing a single construct (high-level mobility). The person separation index was .86, indicating good reliability. The skip item exhibited differential item functioning by sex. Although items were well targeted, the fit residual mean for persons was 2.40, indicating most found the tasks easy to complete. Unlike in adult populations, 2 items (walk and walk over obstacle) exhibited local dependency >0.2. The HiMAT is a unidimensional targeted performance measure of high-level mobility for children with typical development aged 5 to 12years. Further examination of the relationship between the test items "walk" and "walk over an obstacle" may be required to ensure that scores on these items are truly independent of one another.