Course contentsShow
← Data Science
Lesson 551 of 2,14514. Simple Linear RegressionFree lesson

No Perfect Multicollinearity in Simple Regression

Understanding why this assumption is trivial in simple regression but important conceptually.

No Perfect Multicollinearity in Simple Regression

What you'll learn: Why the "no perfect multicollinearity" assumption is automatically satisfied in simple regression, even though it becomes critically important in multiple regression later.

Why This Assumption Exists

Multicollinearity refers to a situation where predictor variables are highly correlated with each other, making it impossible to isolate each variable's individual effect on the outcome. Perfect multicollinearity means two or more predictors are perfectly linearly related—one can be expressed as an exact linear combination of the others.

Why It's Trivial in Simple Regression

Here's the key insight: in simple linear regression, you only have ONE predictor variable (X).

A variable cannot be perfectly correlated with itself in a problematic way—it simply is itself. You need at least two predictor variables for multicollinearity to even be possible.

An Analogy

Imagine you're baking a cake with only one ingredient: flour. Someone warns you, "Don't use two ingredients that are the same thing!" That warning doesn't apply—you literally can't have duplicate ingredients when you only have one ingredient to begin with.

Why We Still Mention It

Even though this assumption is automatically satisfied in simple regression, we include it in the list because:

  1. It completes the conceptual framework you'll need for multiple regression
  2. It helps you understand that regression assumptions exist for mathematical and statistical reasons
  3. It prepares you for the more complex scenarios ahead where multicollinearity becomes a real diagnostic concern

Key Takeaway: In simple linear regression with one predictor, perfect multicollinearity is mathematically impossible—you need at least two predictors for this assumption to matter. This assumption is mentioned for completeness and future extension.