Latent Interaction Modeling with Planned Missing Data Designs

ABSTRACT Planned missing data (PMD) designs allow researchers to collect additional data under time constraints and to reduce participant burden, both of which can occur in social, behavioral, and educational research settings. The imposed missing data patterns, however, can hamper the efficiency of statistical models implemented to test hypotheses that are of interest to substantive researchers, including whether a treatment works the same for all students. Typically, PMD designs result in a modest power deficiency; however, this tenet has not been extended to latent interaction models. Such models are of increasing importance as researchers investigate moderated relationships involving continuous latent variables. Monte Carlo simulations were used to assess the efficacy of various latent interaction estimation methods under PMD designs.