Teacher's Corner: An R Shiny App for Sensitivity Analysis for Latent Growth Curve Mediation.

Area of Science:

• Behavioral Science
• Psychology
• Biostatistics

Background:

• Understanding mechanisms of behavior change is crucial for effective interventions.
• Latent growth curve mediation models (LGCMM) are a recommended statistical approach.
• Temporal precedence of change from mediator to outcome is key in LGCMM.

Purpose of the Study:

• To introduce and evaluate the Correlated Augmented Mediation Sensitivity Analyses (CAMSA) App.
• To assess the robustness of mediating paths in LGCMM against potential confounding variables.
• To provide a tool for rigorous analysis of behavior change mechanisms.

Main Methods:

• Implementation of sensitivity analysis for LGCMM using the CAMSA App.
• Application of the CAMSA approach to simulated data.
• Validation of the CAMSA approach using real-world data from a substance use disorder treatment study.

Main Results:

• The CAMSA App provides a method to evaluate the robustness of mediating pathways in LGCMM.
• Sensitivity analysis helps determine if a mechanism of change is reliable despite potential confounders.
• The approach was successfully applied to both simulated and empirical data.

Conclusions:

• The CAMSA approach enhances the validity of LGCMM for studying behavior change mechanisms.
• This method is vital for ensuring that identified mechanisms are not artifacts of unmeasured confounding.
• The CAMSA App is a valuable tool for researchers in behavioral science and clinical psychology.