Interpreting Models Without Interactions
What you'll learn: How to understand what regression coefficients mean when your model assumes predictors work independently of each other.
The Standard Regression Assumption
When you build a basic multiple regression model without interaction terms, you're making a crucial assumption: each predictor has an independent, additive effect on the response variable. This means that the effect of changing one predictor by one unit is always the same, regardless of the values of your other predictors.
What "Additive" Means
Think of predictors like ingredients in a recipe where each adds flavor independently. If adding salt increases tastiness by 2 points and adding sugar increases it by 3 points, then adding both gives you exactly 5 points total—no more, no less. The effect of salt doesn't depend on whether sugar is present.
In regression terms, if you have:
y = β₀ + β₁x₁ + β₂x₂
The coefficient β₁ tells you: "For every 1-unit increase in x₁, y changes by β₁ units—period." This holds whether x₂ is zero, 100, or a million.
Why This Matters
Without interactions, interpretation is straightforward. Each coefficient represents a constant, unchanging effect. You can say "holding all other variables constant, a one-unit increase in x₁ leads to a β₁ change in y." This ceteris paribus (all else equal) interpretation is clean and simple.
But reality is often messier—sometimes the effect of one variable does depend on another. That's where interactions come in (your next lesson).
Key Takeaway: Models without interactions assume each predictor affects the response independently and additively—the effect of one variable never changes based on the value of another variable.