Structural Equation Modeling in HCI Research using SEMinR

Abstract

Structural equation models (SEMs) are statistical techniques that help to identify models of latent variables in survey data. This allows researchers to test both the quality of the measurement instrument - the survey - as well as the hypothesized relationships using a single model. Partial least squares structural equation modeling (PLS-SEM) is a subset of SEM that works well with small sample sizes and non-parametric data, which frequently occur in HCI research. In this course, we will provide a short introduction into SEMinR, an open-source library for the R language. SEMinR is an easy-to-use domain-specific language for defining, estimating, visualizing, and validating SEMs using the PLS method. SEMinR provides means for scientific reporting and can be used by academics and practitioners alike.