NCR SDAL Event

Garrett Howardson, Consortium Research Fellow, U.S. Army Research Institute, Foundational Sciences Research Unit (FSRU)

Abstract: Ecological momentary assessment research studies psychological/behavioral phenomena in situ where over-burdening participants introduces significant measurement error. Undue burden is avoided with one of two measurement strategies: (1) more precisely measure a small portion of the construct using repeated measures of similar items, or (2) measure the full construct less precisely using fewer measures. The motivating logic for such strategies is classical test theory, where precisely measuring the full construct requires a burdensome number of items. This problem is particularly acute for emotions research in that the possible range of emotional experiences is quite broad forcing researchers to make less than ideal measurement decisions. In this talk, I introduce a two-dimensional, confirmatory item response theory (IRT) model and show that the full breadth of emotional experiences can be captured without overburdening participants. I administered a comprehensive but parsimonious 17-item emotions checklist before and after a short training. Results indicated: (1) item fit was acceptable; (2) marginal reliabilities across both dimensions were acceptable; (3) item and test information functions followed the theoretically expected pattern, and (4) the checklist was sensitive to pre-/post-training changes in emotions.