July 26, 2024Open Access

To Interact or not to Interact: Pros and Cons of Including Interactions in Linear Regression Models

View Full Paper

Ask AI

Bookmark

Authors

ARAljoscha RimplerLeiden University HKHenk A. L. KiersHiroshima University DRDon van RavenzwaaijEötvös Loránd University

Discussion

Member takes

Implication

Simulation demonstrates that omitting interaction effects severely biases linear regression estimates in psychological datasets, highlighting the necessity of theory-guided model selection.

Key Points

Misspecification in linear regression models drastically biases parameter estimates, frequently resulting in null findings or completely reversed simple effects.
Across 55,296,000 analyses per model, correctly specified interaction effects explained population data better, whereas misspecification reduced generalizability.
Simulation evaluating 9,216 parameter conditions confirms that sound theory must guide whether interaction effects are retained or omitted in regression models.

Cite This Study

Rimpler et al. (2024) studied this question.

synapsesocial.com/papers/68e5ef86b6db64358758452d https://doi.org/10.31234/osf.io/6utws

View Full Paper

Ask AI

Bookmark

Implication

Simulation demonstrates that omitting interaction effects severely biases linear regression estimates in psychological datasets, highlighting the necessity of theory-guided model selection.

To Interact or not to Interact: Pros and Cons of Including Interactions in Linear Regression Models

Discussion

Member takes

Implication

Key Points

Cite This Study

To Interact or not to Interact: Pros and Cons of Including Interactions in Linear Regression Models

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider