Skip to content

Residuals and Residual Plots

Analyze residual plots to assess the fit of a regression model.

Written and reviewed by the Study Mondo Education TeamLast updated
🎯⭐ INTERACTIVE LESSON

Try the Interactive Version!

Learn step-by-step with practice exercises built right in.

Start Interactive Lesson →

📉 Residuals and Residual Plots

What Is a Residual?

Residual: The difference between observed and predicted value

residual=observed−predicted=y−y^\text{residual} = \text{observed} - \text{predicted} = y - \hat{y}

Interpretation:

  • Positive residual: actual value above regression line
  • Negative residual: actual value below regression line
  • Small residuals: good fit; large residuals: poor fit

Reading Residual Plots

A residual plot graphs residuals on y-axis vs. predicted values (or x-values) on x-axis.

✅ Good Residual Plot (Random Scatter)

  • Residuals scattered randomly around 0
  • No clear pattern
  • Equal vertical spread (homoscedasticity)
  • Conclusion: Linear model is appropriate

❌ Pattern in Residual Plot (Non-linearity)

  • Residuals form a U-shape or inverted U
  • Or systematic curve
  • Conclusion: Relationship is NOT linear; try transformation

❌ Increasing Spread (Heteroscedasticity)

  • Residuals spread wider as x increases
  • Conclusion: Variance is not constant; may need transformation or weighted regression

Worked Example

Data: Height vs. Weight (n = 30 students)

Regression line: Weight^=50+3.2⋅Height\hat{\text{Weight}} = 50 + 3.2 \cdot \text{Height}

One student: Height = 70 in, actual weight = 240 lbs

Weight^=50+3.2(70)=274\hat{\text{Weight}} = 50 + 3.2(70) = 274 lbs

Residual =240−274=−34= 240 - 274 = -34 lbs (actual weight below prediction)

If residual plot shows random scatter, linear model fits well.

Conditions for Linear Regression (LINER)

  1. Linear: Scatterplot shows linear trend
  2. Independent: Observations independent
  3. Normal: Residuals approximately normal
  4. Equal SD: Constant vertical spread (homoscedasticity)
  5. Random: Random sample

Residual plots help check conditions 1, 3, 4.

Common Mistakes

❌ Confusing residuals with errors (same thing, different context) ❌ Ignoring patterns; thinking a small curve is "close enough" ❌ Not checking residuals before making predictions ❌ Using residuals to predict; they should center on 0

AP Exam Tip

Say "The residual plot shows random scatter with no pattern, so a linear model is appropriate." Or "The residual plot shows a curved pattern, indicating the relationship is non-linear."

📚 Practice Problems

1Problem 1easy

❓ Question:

For regression ŷ = 10 + 2x, calculate the residual for the point (5, 25).

💡 Show Solution

Step 1: Identify actual value Point (5, 25): x = 5, y = 25 (actual)

Step 2: Calculate predicted value ŷ = 10 + 2(5) = 10 + 10 = 20

Step 3: Calculate residual Residual = y - ŷ Residual = 25 - 20 = 5

Step 4: Interpret The residual is POSITIVE (+5), meaning:

  • Actual value is ABOVE predicted value
  • Point is 5 units above the regression line
  • Model UNDERESTIMATES by 5 units

Answer: Residual = 5 (point is above the line)

2Problem 2medium

❓ Question:

A residual plot shows points scattered randomly around zero with no pattern. What does this indicate?

💡 Show Solution

Step 1: Understand what random scatter means Good residual plot characteristics: ✓ Points scattered RANDOMLY ✓ No curved, U-shaped, or other patterns ✓ Roughly equal spread at all x values ✓ Centered around residual = 0

Step 2: What this indicates The linear model is APPROPRIATE:

  1. Linear relationship is valid (no curved pattern)
  2. Constant variance (homoscedasticity)
  3. No systematic errors
  4. Independence assumption met

Step 3: What to do ✓ Can proceed with predictions ✓ Can trust confidence intervals ✓ Linear regression is validated

Answer: Random scatter indicates the linear model is APPROPRIATE. The relationship is truly linear, variance is constant, and there are no systematic errors.

3Problem 3medium

❓ Question:

A residual plot shows a curved (U-shaped) pattern. What does this suggest and what should you do?

💡 Show Solution

Step 1: Identify the problem U-shaped or curved residual plot means: Linear model is INAPPROPRIATE

The relationship is actually nonlinear (curved).

Step 2: Why this is a problem

  • Linear model makes systematic errors
  • Underestimates in middle, overestimates at extremes (or vice versa)
  • Predictions will be biased
  • Violates linearity assumption

Step 3: Solutions Option 1: Transform the data

  • Try log(y) vs x, or √y vs x
  • Replot residuals - should become random

Option 2: Use nonlinear regression

  • Quadratic: y = a + bx + cx²
  • Exponential: y = ae^(bx)

Step 4: Check new model After transformation, residual plot should show random scatter.

Answer: Curved residuals indicate NONLINEAR relationship. Transform variables (log, square root) or use nonlinear regression. Recheck residuals after adjustment.

4Problem 4medium

❓ Question:

For points (1,3), (2,5), (3,6) with regression ŷ = 2 + 1.5x, verify residuals sum to zero.

💡 Show Solution

Step 1: Calculate predicted values Point 1: ŷ₁ = 2 + 1.5(1) = 3.5 Point 2: ŷ₂ = 2 + 1.5(2) = 5 Point 3: ŷ₃ = 2 + 1.5(3) = 6.5

Step 2: Calculate residuals Residual = y - ŷ

Point 1: e₁ = 3 - 3.5 = -0.5 Point 2: e₂ = 5 - 5 = 0 Point 3: e₃ = 6 - 6.5 = -0.5

Step 3: Sum residuals Σ(residuals) = -0.5 + 0 + (-0.5) = -1.0

This is close to zero (small rounding error).

Step 4: Why residuals sum to zero Mathematical property: For least-squares regression, Σ(y - ŷ) = 0 ALWAYS

  • Guaranteed by the formulas
  • Positive and negative errors balance
  • Line goes through "middle" of data

Answer: Residuals sum to approximately 0. For true least-squares line, they ALWAYS sum exactly to zero.

5Problem 5hard

❓ Question:

A residual plot shows increasing spread (fan shape) as x increases. What does this violate and what are the implications?

💡 Show Solution

Step 1: Identify the violation Fan-shaped residuals violate: CONSTANT VARIANCE (homoscedasticity)

The spread increases with x (heteroscedasticity).

Step 2: Implications for predictions

  • Predictions less reliable at high x (wide spread)
  • Predictions more reliable at low x (tight spread)
  • Standard errors are WRONG
  • Confidence intervals misleading

Step 3: Implications for inference

  • t-tests may be invalid
  • p-values unreliable
  • Hypothesis tests have wrong error rates
  • Can't trust significance levels

Note: Estimates (slope, intercept) are still unbiased, but uncertainty measures are wrong.

Step 4: Solutions

  • Transform y (try log(y) or √y)
  • Use weighted least squares
  • Use robust standard errors
  • Report with caution

Answer: Violates CONSTANT VARIANCE assumption. Standard errors and confidence intervals unreliable. Solutions: transform y, use weighted least squares, or robust standard errors.

Explain using:

⚠️ Common Mistakes: Residuals and Residual Plots

Avoid these 3 frequent errors

← Previous Topic
Least-Squares Regression
Next Topic →
Coefficient of Determination
🎴

Practice with Flashcards

Rate this topic's cards with spaced repetition. Cards join your deck when you finish a topic's lesson and take its exit quiz.

📖

Browse All Topics

Explore more AP Statistics topics

📌 Related Topics in Unit 2: Exploring Two-Variable Data

Scatter Plots and CorrelationLeast-Squares RegressionCoefficient of DeterminationTransformations for Linearity

❓ Frequently Asked Questions

What is Residuals and Residual Plots?▾
Analyze residual plots to assess the fit of a regression model.
How can I study Residuals and Residual Plots effectively?▾
Start by reading the study notes and working through the examples on this page. Then use the flashcards to test your recall. Practice with the 5 problems provided, checking solutions as you go. Regular review and active practice are key to retention.
Is this Residuals and Residual Plots study guide free?▾
Yes — all study notes, flashcards, and practice problems for Residuals and Residual Plots on Study Mondo are free to access. No account is needed.
What course covers Residuals and Residual Plots?▾
Residuals and Residual Plots is part of the AP Statistics course on Study Mondo, specifically in the Unit 2: Exploring Two-Variable Data section. You can explore the full course for more related topics and practice resources.
Are there practice problems for Residuals and Residual Plots?▾
Yes, this page includes 5 practice problems with detailed solutions. Each problem includes a step-by-step explanation to help you understand the approach.