Rasch-based instrument validations for an augmented reality vocabulary acquisition experiment

Abstract

The main purpose of this study was to provide validity evidence to support the creation and use of three instruments theorised to measure the existence of the form-meaning link between three sets of nonwords, and their assigned meanings within an experiment. The experimental study used a counterbalanced Latin square design to examine and compare three conditions of deliberate vocabulary study (word cards and two variations of tablet-based augmented reality) across three thematically constrained sets of controlled nonwords. The collected data gathered with the three delayed post tests conducted for each of the sets from the experiment were subjected to separate Rasch analyses, which are described in this study. The findings serve as evidence of construct validity and show: (1) the instruments performed as predicted by two a priori hypotheses, (2) the items were found to display good fit as to the expectations of the Rasch model, (3) each of the three instruments were found to be unidimensional, and (4) with no changes, these three instruments are expected to yield similar results if they were to be used to measure participants within a similar context from the same population.