Course contentsShow
← Data Science
Lesson 566 of 2,14515. Regression DiagnosticsFree lesson

Reading Q-Q Plots: Interpreting Points Along the Reference Line

Understand how to read Q-Q plots by examining whether residuals follow the 45-degree reference line indicating normality.

Reading Q-Q Plots: Interpreting Points Along the Reference Line

What you'll learn: How to interpret the pattern of points on a Q-Q plot to assess whether your regression residuals follow a normal distribution.

The Core Idea

A Q-Q (quantile-quantile) plot displays your residuals' quantiles against theoretical normal distribution quantiles. The key diagnostic feature is a 45-degree reference line that represents perfect normality. Your job is to examine how closely the plotted points hug this line.

Think of it like a "stay in your lane" test: if all the points march neatly along the diagonal reference line, your residuals are normally distributed. Deviations from this line reveal specific patterns of non-normality.

What Different Patterns Mean

Points along the line: Perfect or near-perfect normality. Your residuals behave as the regression model assumes—this is what you want to see.

S-shaped curve: Heavy tails. Points deviate upward at the top and downward at the bottom, indicating your residuals have more extreme values than a normal distribution predicts.

Inverted S-shape: Light tails. The distribution is more concentrated around the center than normal.

Points curving upward at the right end: Right skew. Large positive residuals are larger than expected under normality.

Points dropping downward at the left end: Left skew. Large negative residuals are more extreme than normal theory predicts.

The Practical Interpretation

Minor wiggles around the line are normal—real data is never perfectly normal. Focus on systematic, pronounced departures. Points near the middle of the line matter most; slight deviations at the extreme ends are less concerning with smaller sample sizes.

Key Takeaway: Points that closely follow the 45-degree reference line on a Q-Q plot indicate normally distributed residuals; systematic departures (S-curves, upward/downward trends) reveal specific violations of the normality assumption.