Skip to content
AP Statistics · Regression Analysis

Residual

Residual appears in 6 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 Residual is tested

Example 1Medium
A regression analysis of x = hours of training and y = marathon finish time (minutes) produces the LSRL: Time-hat = 280 - 0.9*(Hours). The residual for a runner who trained 50 hours and finished in 240 minutes is:
  1. A-5 minutes
  2. B85 minutes
  3. C-85 minutes
  4. D5 minutesCorrect

Why D is right

Predicted finish time = 280 - 0.9*(50) = 280 - 45 = 235 minutes. Residual = observed - predicted = 240 - 235 = +5 minutes. A positive residual means the runner finished slower than predicted.
Why the other 3 choices miss
A
Incorrect because -5 would be the negative of the correct residual. This error occurs when computing predicted - observed instead of observed - predicted.
B
Incorrect because 85 minutes corresponds to the difference between the intercept (280) and the observed time (240), ignoring the effect of training hours on the prediction.
C
Incorrect because -85 would result from predicted - observed (reversed) using an incorrect predicted value, possibly computing 280 - 0.9*50 = 235 and then taking 235 - 240 with the wrong sign convention, or using a wrong formula entirely.
Example 2Medium
A regression line is fit to data and found to have the equation y-hat = 50 + 3x. A point (x = 8, y = 80) lies on the regression line. Which of the following statements is true?
  1. AThe point is an outlier because y = 80 is far from the mean of y.
  2. BThe point is influential because x = 8 is an extreme value.
  3. CThe residual for this point is 0, since the point lies on the regression line.
  4. DThe residual for this point is 80 - (50 + 3*8) = 80 - 74 = 6.Correct

Why D is right

The predicted value at x = 8 is y-hat = 50 + 3*(8) = 50 + 24 = 74. The point's observed y-value is 80, so it does not actually lie on the regression line (despite the wording in the stem). The residual = observed - predicted = 80 - 74 = 6, which makes choice D correct. Choice C is wrong because the point is not on the line, so the residual is not 0. Choices A and B cannot be concluded: outlier and influential status depend on the residual and the point's position relative to the full pattern of data, not the absolute value of y or x alone.
Why the other 3 choices miss
A
Incorrect because whether a point is an outlier depends on its residual and position relative to the pattern of all data, not just the absolute value of y. Without knowing the distribution of all y-values, this cannot be concluded.
B
Incorrect because x = 8 may not be extreme relative to the range of x-values in the dataset. Influential points require both an extreme x-value (high leverage) and a large residual.
C
Incorrect because the point (8, 80) does not lie on the regression line. The line predicts y-hat = 74 when x = 8, and 80 ≠ 74, so the residual is not zero.
Example 3Easier
A researcher has a data set with a strong positive association between x and y. The residual for observation i is -12.4. This means that
  1. Aboth B and C are correct descriptionsCorrect
  2. Bobservation i has a y-value that is 12.4 units below the predicted value from the regression line
  3. Cthe regression line overestimates the response by 12.4 units for observation i
  4. Dobservation i has an x-value that is 12.4 units below the mean of x

Why A is right

A residual is defined as residual = observed - predicted. Here that equals -12.4, so the observed y-value lies 12.4 units below the value predicted by the regression line (choice B). Equivalently, the regression line overpredicts (overestimates) the response by 12.4 units for this observation (choice C). Choice D is false: the residual measures the vertical gap between observed and predicted y, not the horizontal distance of x from its mean. Since both B and C correctly describe the same residual from different perspectives, the complete and unique answer is A.
Why the other 3 choices miss
B
Incorrect as a standalone answer because choice C is also correct. Both A and C are equivalent statements about the same residual.
C
Incorrect as a standalone answer because choice A is also correct. Both A and C are equivalent statements about the same residual.
D
Incorrect. The residual measures the vertical distance between the observed and predicted y-values, not the horizontal distance of x from its mean. The x-distance relates to leverage, not the residual.

Studied alongside

Other concepts from Regression Analysis.