Regression discontinuity designs in a latent variable framework.

Area of Science:

• Psychology
• Quantitative Psychology
• Econometrics

Background:

• Regression discontinuity (RD) designs are powerful quasi-experimental methods for causal inference.
• RD designs are underutilized in psychology despite suitability for interventions based on common psychological measures.
• Existing RD implementations may not adequately address measurement error common in psychological data.

Purpose of the Study:

• To present a straightforward method for implementing RD models within a structural equation modeling (SEM) framework.
• To demonstrate how SEM can account for measurement error and avoid misspecification in psychological RD analyses.
• To encourage broader adoption of RD designs and SEM-based approaches in psychological research.

Main Methods:

• The study proposes implementing RD designs as structural equation models (SEM).
• Monte Carlo simulation studies were conducted to evaluate the benefits of a latent variable RD model.
• An applied example with code and results is provided to illustrate the SEM-based RD approach.

Main Results:

• Latent variable RD models implemented via SEM can effectively account for measurement error.
• SEM provides a framework to avoid common issues of measurement model misspecification in psychological research.
• The proposed method integrates RD with SEM, enhancing its applicability in psychology.

Conclusions:

• Implementing regression discontinuity designs using structural equation modeling offers significant advantages for psychological research.
• This approach allows for robust causal inference while addressing critical issues like measurement error.
• The tutorial aims to make RD designs more accessible and beneficial for psychologists.