Skip to content
AP Statistics · Regression Analysis

Residuals

Residuals appears in 9 reviewed AP Statistics questions and sits in Regression Analysis. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

How Residuals is tested

Example 1Medium
A normal probability plot (normal quantile plot) of residuals from a regression model shows points following a roughly straight diagonal line. What does this indicate?
  1. AThe residuals are normally distributed.Correct
  2. BThe residuals have non-constant variance.
  3. CThere is a strong linear relationship between the variables.
  4. DThe regression model is overfit.

Why A is right

A normal probability plot is used to assess whether data follows a normal distribution. When the points fall approximately along a straight diagonal line, this indicates that the data (in this case, residuals) are approximately normally distributed. Deviations from the straight line indicate departures from normality.
Why the other 3 choices miss
B
Non-constant variance (heteroscedasticity) would be detected in a plot of residuals versus fitted values, not in a normal probability plot.
C
A normal probability plot assesses normality, not the strength of a linear relationship. A scatterplot or correlation coefficient would assess linearity.
D
Overfitting refers to a model fitting noise in the data and is not assessed by a normal probability plot of residuals.
Example 2Easier
In a regression analysis, the residual for a data point is defined as:
  1. Aobserved y − predicted yCorrect
  2. Bobserved y − mean of y
  3. Cobserved x − predicted x
  4. Dpredicted y − observed y

Why A is right

A residual is the difference between the observed response value (actual data point) and the predicted value from the regression line: residual = y − ŷ. Positive residuals indicate the point is above the line; negative residuals indicate it is below.
Why the other 3 choices miss
B
Incorrect because observed y − ȳ is a deviation from the mean (used in computing total sum of squares), not a residual from the regression line.
C
Incorrect because residuals are measured in the y-direction (vertical distance from the line), not in the x-direction. Simple linear regression minimizes vertical, not horizontal, distances.
D
Incorrect because predicted y − observed y gives the negative of the residual. By convention, residuals are computed as observed minus predicted.

Studied alongside

Other concepts from Regression Analysis.