Interaction Effects

Definition

An Interaction Effect occurs in an experiment when the influence of one independent variable (factor) on the dependent variable (response) depends on the level of another independent variable.

Why It Matters

In the real world, factors don’t act in isolation. If you ignore interaction effects, you will misattribute success or failure, missing the “synergy” or “interference” that actually drives the outcomes in your experiments or business.

Core Concepts

• Synergy and Interference: Interactions can be positive (synergy) or negative (interference).
• Non-Additivity: In the presence of an interaction, the total effect is not simply the sum of the individual main effects.
• Visual Detection: Interactions appear as non-parallel lines on an interaction plot.
• Statistical Significance: Interaction terms are included in ANOVA tables and regression models (e.g., $Y = \beta_1 X_1 + \beta_2 X_2 + \beta_{12} X_1 X_2$).
  • How to read: “The value Y is equal to beta one times X one, plus beta two times X two, plus beta one two times the product of X one and X two.”
  • Meaning / when to use: The product term $X_1 X_2$ captures whether the effect of one factor changes depending on the level of the other; a significant $\beta_{12}$ means factors are not independent.

Connected Concepts

• 2^k Factorial Design
• Experimental Design in Simulation
• Interdependence