Testing Measurement Invariance using Multi-group Latent Class Analysis: A Cross-National Study of ICT Self-Efficacy

The purpose of this study is to provide guidelines for the process of testing measurement invariance using multi-group latent class analysis (MGLCA). To achieve this, data from the ICILS 2018 was utilized to select three countries with different levels of ICT development (South Korea, the United States, and Portugal). For these countries, the latent class classification patterns of middle school students’ ICT self-efficacy were identified, and the similarities and differences in latent classes among countries were examined at the scale and indicator levels. The analysis revealed several key findings. First, in all three countries, the number of latent classes for ICT self-efficacy was consistently determined as three, indicating a similar model structure of latent class across countries. Second, examination of measurement invariance at the scale level classified by country resulted in a partial invariant model. This suggests shared response patterns among latent classes of ICT self-efficacy across countries but also the existence of country-specific characteristics influencing responses, independent of latent classes. Lastly, exploration of the indicator variables revealed differences in response probabilities for three indicator variables. This study is significant in proposing an approach using latent class analysis for comparative group studies on ICT self-efficacy.