Item response theory for before-after designs in interprofessional education research

Matthew J Kerry,Jan J Reinders,Wim P Krijnen,Marion Huber
DOI: https://doi.org/10.1080/13561820.2024.2395979
2024-09-10
Abstract:Although Item Response Theory (IRT) has been recommended for helping advance interprofessional education (IPE) research, its use remains limited. This may be partly explained by potential misconceptions regarding IRT`s "limitation" to cross-sectional data. The aim of this study is to demonstrate how Item Response Theory (IRT) can be applied effectively in before-and-after designs in IPE research. Specifically, a two-week before-after design with survey methodology using the Extended Professional Identity Scale (EPIS), an interprofessional identity measure, was conducted among n = 146 mixed health-science students. Results indicated that EPIS increased significantly before-after intervention by .74 standardised mean differences, t146 = 7.73, p < .05. The before-after IRT model also gave a test-retest reliability estimate of .60 which was considered acceptable. Comparison of the IRT model with a conventional paired-t-test indicated similar effect size estimates of Cohen's d = .56 and .54, respectively. We demonstrate IRT`s flexibility to before-after studies in IPE. Application of this model can yield accurate changes in target IPE constructs, and it is advantageous to classical test theory vis-à-vis baseline differences.
What problem does this paper attempt to address?