Worked examples of Residuals to try before you open each solution, then practice
problems with answers.
Worked Examples
Example 1
easy
Given y^=2+3x, and observed point (4,15): calculate the residual and interpret whether the model over- or under-predicts.
Show solution
Answer
e=15−14=1 (positive). The model under-predicts by 1 unit.
First step
1.
Calculate predicted value: y^=2+3(4)=2+12=14
Full solution
2.
Calculate residual: e=y−y^=15−14=1
3.
Positive residual: actual value (15) is ABOVE the predicted value (14)
4.
Interpretation: the model under-predicts by 1 unit for this observation
A residual e=y−y^ measures the vertical distance between observed and predicted. Positive residual = point above the line (model under-predicts); negative residual = point below the line (model over-predicts). Residuals should average to zero for a good model.
Example 2
medium
Five observed and predicted values: (y,y^): (10,8),(15,14),(12,13),(20,19),(8,11). Calculate all residuals, verify they sum to 0, and compute ∑ei2.
Example 3
medium
Data: (1,3),(2,5),(3,4),(4,8). Fitted line: y^=1.5x+1.5. Compute all residuals.
Example 4
medium
Define standardized residuals and explain why they help identify outliers.
Example 5
hard
Given a residual of 5 at x=3 where y^=2+4x, find y. Then determine the residual if the line had been y^=2+5x instead.
Example 6
hard
Two students fit different lines to the same data. Student A reports residuals summing to 0; Student B reports residuals summing to 5. Whose line could be the LSRL?
Example 7
hard
A residual plot shows one residual much larger than the others. What is this point called and how should it be handled?
Practice Problems
Try each one, then open the solution to compare your method.
Example 1
easy
A regression model gives residuals: {3,−2,4,−5,0}. Are these residuals consistent with a valid LSRL? Calculate ∑ei and ∑ei2.
Example 2
hard
A residual plot shows residuals increasing in magnitude as y^ increases (fan shape). What assumption is violated, and what does this mean for the validity of hypothesis tests in regression?
Example 3
easy
A model predicts 50 but the observed value is 47. Compute the residual.
Example 4
easy
Observed =20, predicted =14. Find the residual.
Example 5
easy
A residual is 0. What does that mean about the data point?
Example 6
easy
Using y^=2+3x, find the residual for the point (4,15).
Example 7
easy
A positive residual means the observed value is above or below the prediction?
Example 8
easy
What is the sum of all residuals from a least-squares regression line?
Example 9
easy
A residual plot shows points scattered randomly around 0 with no pattern. Does this support the linear model?
Example 10
easy
A residual plot shows a clear U-shaped curve. What does this suggest about the linear model?
Example 11
medium
For y^=10−2x, find the residual at the point (3,2).
Example 12
medium
Three points have residuals 4, −1, and r3. If LSRL residuals must sum to 0, find r3.
Example 13
medium
A data point has the largest absolute residual in a dataset. What is such a point called, and does a large residual alone make it influential?
Example 14
medium
Observed values y=(10,14,22) have predictions y^=(12,14,20). Compute all three residuals.
Example 15
medium
A residual plot shows residuals increasing in spread as x grows (a fan shape). What model assumption is violated?
Example 16
medium
If a residual is −5 at a point where the predicted value is 30, what was the observed value?
Example 17
medium
Two models fit the same data: Model A has all residuals near zero with random scatter; Model B has high r2 but a curved residual plot. Which model is more trustworthy?
Example 18
medium
The standard deviation of the residuals (denoted s) is reported as 4. Roughly interpret this value.
Example 19
medium
Using y^=6+x, find the residual for the point (5,9).
Example 20
challenge
A line y^=1+2x fits points (1,4),(2,4),(3,8). Compute the residuals and verify they sum to zero. If not, what does that tell you?
Example 21
challenge
A point with a large positive residual lies at the far right edge (large x). Removing it changes the slope substantially. Classify the point.
Example 22
challenge
Given residuals with values 3,−2,−2,1,r5 from an LSRL, find r5, then state the residual sum of squares (RSS) using all five.
Example 23
easy
A regression line predicts y^=12 for x=5. The actual observation is y=9. Find the residual.
Example 24
easy
Given y^=4x−3 and observed point (2,7), compute the residual.
Example 25
easy
For a least-squares regression line, ∑(yi−y^i) always equals what value?
Example 26
medium
Residuals from a regression: {4,−3,2,x,−1} where the residual sum is zero. Find x.
Example 27
medium
For points (2,5),(4,7),(6,11) and line y^=1.5x+1, compute SSR (sum of squared residuals).
Example 28
medium
Predicted y^ values: 10,12,15. Observed y: 11,9,18. Compute the mean of the residuals.
Example 29
hard
An LSRL fit gives residuals with ∑ei2=80 and n=12. Estimate the residual standard error (use n−2 degrees of freedom).
Example 30
hard
A residual plot shows a curved (parabolic) pattern. What kind of model might better fit the data?
Example 31
hard
A regression has n=20, SST =200, SSR =50. Find R2.
Example 32
easy
Predicted y^=100, observed y=100. What is the residual?
Example 33
medium
For an LSRL with intercept b0=2 and slope b1=3 fit on xˉ=5, yˉ=?. Verify the line passes through (xˉ,yˉ).
Example 34
medium
A residual plot shows residuals fanning out as y^ grows. What is this called and what does it suggest?
Example 35
medium
A residual plot has all positive residuals on the left and all negative residuals on the right. What kind of pattern is missed?
Example 36
easy
Given y^=10−2x and observed (3,6), find the residual.
Example 37
medium
From a regression with se=2.5 and a residual of e=6, find the standardized residual.
Example 38
hard
Suppose a quadratic relationship y=x2 is fit with a line y^=ax+b. What kind of residual plot would result?
Example 39
challenge
For LSRL with no intercept (line forced through origin: y^=bx), is ∑ei=0 guaranteed?
What to focus on
Core idea:A residual is how far one data point falls above or below the regression line: e=y−y^.
Common stuck point:The procedure for residuals is the easy part; the trap is computing predicted minus observed. Asking "Am I taking one point's actual value minus the line's predicted value to measure its individual miss?" first is what keeps a correct-looking calculation from being attached to the wrong concept.
Ask yourself:Am I taking one point's actual value minus the line's predicted value to measure its individual miss?
Concept recap
The difference between an observed value and its predicted value from a regression model: residual=y−y^ (observed minus predicted).
A residual is like the gap between a weather forecast and the actual temperature. Forecast 20 °C, actual 23 °C: the forecast was 3 degrees too low, a residual of +3. One gap says little, but a forecast that is too low every afternoon reveals a flaw in the model, just as a pattern in a residual plot does.