Residual plot worksheet with answer key
8 problems on residual plot, ready to print for class or homework. The answer key prints on its own page, so hand out everything before it.
Name: ______________________________ Date: ______________ Period: ______
Problem 1
A ferry company models crossing time in minutes, , from the wind speed on the strait in km/h, , using the least-squares line . (a) A crossing made in a 15 km/h wind took 26 minutes. Find its residual. (b) A crossing made in a 30 km/h wind took 37 minutes. Find its residual. (c) On a residual plot of these crossings, which one plots above the horizontal line at zero?
Problem 2
A city compost program records, for each collection week, the number of households enrolled, , and the compost collected in kilograms, . The least-squares line is . (a) A week with 25 households enrolled produced 130 kg. Find the residual and say what its sign means. (b) An earlier week with 12 households enrolled produced 70 kg. Find the residual and say what its sign means. (c) A program officer writes that the 25-household week shows the model is working, because its residual came out positive. What is wrong with that reasoning?
Problem 3
A hatchery models a young trout's length in centimeters, , from its age in months, , with the least-squares line . (a) One trout is 8 months old and has a residual of cm. How long is it, and what does the residual say about that fish? (b) A second trout measures 24 cm and has a residual of cm. How old is it? (c) A third trout has a residual of exactly . What does that tell you about that fish, and what does it not tell you about the model?
Problem 4
A caterer records the number of guests, , and the kilograms of food waste, , at five events: , , , , . The least-squares line is . (a) Compute all five residuals. (b) Show that they add to zero and explain why that had to happen. (c) Which events did the line overpredict, and by how much? (d) Describe the residual plot and say whether a linear model is appropriate here.
Problem 5
Three studies each fit a least-squares line and produce a residual plot of the residuals against the explanatory variable.
- Plot A, a print shop, finishing time against the number of copies in a job: the residuals fall on both sides of zero across the whole range of , and the vertical spread stays about the same, roughly to minutes, at every . Nothing bends.
- Plot B, a reservoir, water released against the rainfall two days earlier: every residual at the low end of is negative, every residual in the middle of the range is positive, and every residual at the high end is negative again, tracing a smooth arch.
- Plot C, a warehouse, order picking time against the number of items in the order: for the smallest orders the residuals sit within minutes of zero, and for the largest orders they run from about to minutes. The residuals stay centered on zero all the way across and nothing bends.
(a) For each plot, say whether a linear model is appropriate. (b) What does plot B tell you about where the line misses? (c) A student says plot C proves the line is the wrong shape. Correct that. (d) In which of the three plots do the residuals add to zero?
Problem 6
A parks department fits maintenance hours per season, , to trail length in kilometers, , for six sections and gets . Its residual plot shows one point per section.
| Section | Length (km) | Residual (hours) |
|---|---|---|
| Cedar | 2 | |
| Birch | 4 | |
| Fern | 6 | |
| Granite | 8 | |
| Heron | 10 | |
| Juniper | 12 |
(a) Which section does the line fit worst, and by how much? (b) What does the sign of that section's residual say about it? (c) Which section did the line underpredict by the most? (d) How many maintenance hours did the worst-fit section actually take? (e) Does the plot give you a reason to abandon the linear model?
Problem 7
A driving school measures the braking distance in meters, , of one car from five speeds in km/h, , on the same dry surface: , , , , . Technology reports the least-squares line with . (a) Compute the five residuals. (b) Describe the residual plot. (c) The instructor argues that is proof the linear model fits well. Decide whether that argument holds, and justify your answer from the residual plot. (d) Find what the line predicts at 90 km/h, then use the pattern in the data to say whether that prediction is too high or too low.
Problem 8
A park runs six clean-up events and records the number of volunteers, , and the litter collected in kilograms, : , , , , , . The least-squares line is . (a) Compute the six residuals and their sum. (b) A coordinator says the sum shows the line fits the data well. Respond. (c) Describe the residual plot and say whether a linear model is appropriate, justifying your answer. (d) A seventh event with 16 volunteers collected an amount that fell 9 kg below the line's prediction. Find that amount and say how it fits your answer to part (c).
From statslearn.org/worksheets/residual-plots-worksheet. Free to print and copy for classroom use.
Residual plot: answer key
(a) minutes. (b) minutes. (c) The 30 km/h crossing, because its residual is positive.
(a) kg, positive, so the model underpredicted by 15 kg. (b) kg, negative, so the model overpredicted by 6 kg. (c) The sign gives the direction of the miss, not the quality of the fit, and the kg week is the one the line fits worse.
(a) cm, and the line overpredicted this fish by 2.5 cm. (b) The prediction was cm, so and months. (c) That fish sits exactly on the line, so the prediction matched its length; it says nothing about how well the model fits the other fish.
(a) Residuals kg. (b) They sum to , which least squares forces because the line passes through . (c) The 10-guest event was overpredicted by 1 kg and the 40-guest event by 2 kg. (d) The residuals scatter on both sides of zero with no curve and no widening, so a line is appropriate, though five points is thin evidence.
(a) A: appropriate, random scatter with constant spread. B: not appropriate, the arch is curvature. C: the form is not contradicted, but the spread grows, so predictions for large orders are much less precise. (b) The line overpredicts at both ends of the rainfall range and underpredicts in the middle. (c) A fan reports changing spread, not a wrong form; curvature is the signal for a wrong form and plot C has none. (d) All three, because least squares forces the residuals to sum to zero.
(a) Granite, missed by 6 hours. (b) The residual is negative, so the line overpredicted Granite: it took 6 fewer hours than predicted. (c) Fern, underpredicted by 5 hours. (d) Predicted hours, so observed hours. (e) No. The residuals show no curve and no widening, so the form looks fine; Granite is one point worth investigating, not a reason to drop the model.
(a) Residuals m, summing to . (b) A clear U: positive at both ends, negative through the middle. (c) The argument fails. measures scatter about a line, not whether a line is the right form, and the U shows the misses are systematic, so a curved model fits better. (d) The line gives 45 m, and the data's second differences of 2 m point to about 59 m, so the line's prediction is far too low as well as an extrapolation.
(a) Residuals kg, summing to . (b) The sum is zero for every least-squares fit, since the line passes through , so it is an arithmetic check and not evidence of a good fit. (c) The residuals trace an arch, negative at both ends and positive in the middle, and that curvature means a linear model is not appropriate; the collected amounts rise by 10, 7, 6, 5, then 2 kg, showing shrinking returns. (d) Predicted kg, so the event collected kg, matching the flattening the arch already showed.
Every step is worked at statslearn.org/practice/residual-plots-practice. Free to print and copy for classroom use.
The key gives final answers only. The full worked solution for each problem is on the matching practice set, and every other sheet is on the worksheets page.