File(s) not publicly available
Construct validity of the Test of Visual-Motor Integration: An evaluation using Rasch analysis
journal contributionposted on 02.04.2019, 00:00 by T Brown, Carolyn UnsworthCarolyn Unsworth
Background: Rasch analysis can be used as a method of evaluating construct validity during test construction, validation, or evaluation. Objective: The aim of this study was to examine the construct validity of the Test of Visual-Motor Integration (TVMI) by applying Rasch analysis to evaluate its scalability, dimensionality, differential item functioning (DIF), hierarchical ordering, and scoring category structure. Methods: The participants included 400 children aged 5–12 years, recruited from six schools in the Melbourne metropolitan area, Victoria, Australia. The children completed the TVMI under the supervision of an occupational therapist. Results: Overall, 3 of 30 of the TVMI scale items were problematic: Items 1 and 2 exhibited a Rasch model misfit and Item 5 exhibited DIF based on gender. In addition, the TVMI items were not found to be hierarchically ordered as reported in its manual. The TVMI scoring category structure was satisfactory. Conclusion: Because only three TVMI items were found to be problematic, the construct validity of the TVMI appears satisfactory. However, given that the items are not hierarchically ordered, it is recommended that clinicians use the TVMI with caution when calculating participants’ performance scores. This is particularly relevant if testing is ceased if a participant reaches the TVMI ceiling point by drawing three items incorrectly in a row.