Applying mixed-effects modeling to single-subject designs: An introduction.

Area of Science:

• Behavior Analysis
• Statistical Inference
• Psychology

Background:

• Traditional statistical tests (e.g., t-tests, ANOVA) in behavior analysis often aggregate data, obscuring individual variability crucial for analysis.
• A long-standing conflict exists between behavior analysis and traditional statistical inference methods due to data aggregation issues.

Purpose of the Study:

• To introduce and evaluate the application of generalized linear mixed-effects modeling for analyzing single-subject data in behavior analysis.
• To demonstrate how mixed-effects modeling can overcome the limitations of traditional statistical tests by incorporating random effects.

Main Methods:

• A generalized linear mixed-effects model was applied to single-subject data from a reinforcement task study (Ackerlund Brandt et al., 2015).
• The model incorporated random effects to quantify and retain individual subject variability within the analysis.

Main Results:

• The results from the mixed-effects modeling were consistent with traditional visual analyses of the single-subject data.
• The study demonstrated that mixed-effects modeling provides a robust statistical framework for predicting individual behavior without requiring data aggregation.

Conclusions:

• Mixed-effects modeling offers a statistically sound method for analyzing single-subject designs in behavior analysis, preserving individual variability.
• The findings support the integration of mixed-effects models into the standard analytical practices for behavior analysis research.