Development and testing of a metacognition scale using computer multimedia for nursing students: An application of multidimensional item response theory

Abstract

The aims of this research were to develop and test a metacognition scale using computer multimedia for nursing students (MSCM-NS) by applying the Multidimensional Item Response Theory (MIRT). The research was carried out in 2 phases: Phase 1 Instrument Development and Phase 2 Quality Assessment. The participants consisted of 600 fourth-year nursing students in Year 2018 who were selected using multi-stage sampling. Multidimensional analysis and confirmatory factor analysis were performed to analyze the data. The results showed that (1) the MSCM-NS consisted of a 21-item objective test with 3 choices and a 3-item subjective test, and (2) the IOC was between 0.6–1; the OUTFIT MNSQ (unweighted mean square) value was between 0.52–1.36 and the INFIT MNSQ (weighted mean square) value was between 0.51–1.38; the marginal reliability for response pattern scores was .65, the multidimensional approach was a more fitting model than the composite approach (multidimensional approach: G2 = 24,772.99, AIC = 24,820.99, and composite approach: G2= 24,791.28, AIC = 24,835.28, x2 = 18.29, df = 2, p < .05), and the consecutive approach (G2 = 24,792.16, AIC = 24,838.16, x2 = 19.17, df = 1, p < .05), which indicates that the metacognition scale for nursing students had the highest multidimensional construct validity. In addition, the results of confirmatory factor analysis show that the metacognition model was consistent of empirical data (x2 = 7.724, df = 8, p = .461, GFI = .996, AGFI = .989, RMSEA = .000)