← Dennis Pezan
MA103 · WPR 1 Practice · 164 pts

WPR 1 Practice Examination

Descriptive and predictive modeling — the Company B ruck march problem

This is unofficial review material prepared by the course CIC for the study session. It is not the WPR and is not drawn from any secured assessment. Coverage follows the published lesson objectives for Lessons 1–21 and the conventions of Homeworks 1 and 2. Per the course director, the emphasis is conceptual understanding and precise mathematical interpretation, not Excel mechanics.

Course conventions used here: the outlier screen is |z| > 2.5 (HW1 Q16); residual plots are drawn against fitted values (HW2 Q11); KNN is a classification method decided by majority vote (HW2 §5); and the starting k is √ntrain rounded to the nearest odd integer (HW2 Q22).

Total points
164
Authorized
Calculator only
Coverage
Lessons 1–21
  1. Show your work. When in doubt, explain too much rather than too little. When asked to verify something, show the arithmetic.
  2. Round statistics to two decimal places unless told otherwise, carrying unrounded values through intermediate steps.
  3. Every interpretation must be written in context — with the variable name and its units. An answer with no units is incomplete.
  4. Several questions ask you to carry a value forward from an earlier question. If you could not get the earlier value, state your assumption and continue; follow-on credit applies.
  5. No formula sheet is provided on the WPR, and none is provided here. There is very little computation on this material — but what there is, you are expected to know cold.

Situation. You are the operations officer for Company B, 2nd Regiment. The company runs a 12-mile ruck march at the end of every term, and the commander wants to use last term's data to (1) describe how the company actually performed and (2) predict how individual cadets will do this term, so at-risk cadets can be put on a remedial training plan. All ruck times are in minutes; all cadets carried a 35-pound rucksack over the same route. The Army standard for the 12-mile foot march is 180 minutes — finish at or under it and you earn a GO.

Part I · 20 points

Modeling Framework & Data

Lessons 1–3, 8. Framing a question, population versus sample, classifying variables, partitioning data, and telling an unusual observation from a bad one.

1.
Descriptive or predictive?
4 pts

Label each question below. Then, for (b) only, state in one sentence what makes it that type.

  1. What was the average 12-mile ruck time in Company B last term?
  2. What ruck time should we expect from a cadet who trains 25 miles per week?
  3. Which of the four platoons had the largest spread in ruck times?
2.
Population, sample, and data dictionary.
7 pts

(a) The commander's question is about this term's cadets, but your data is last term's 64 records. Identify the population of interest and state whether your data is the full population or a sample. How do you know?

(b) Complete the data dictionary. Give the variable type (categorical, or numerical and then discrete or continuous) and the units, if any.

Company B roster — partial data dictionary
ColumnExampleTypeUnits
Platoon3rd
Ruck time184.5
Prior ruck marches completed7
Met the 12-mile standardGO

(c) In one sentence, why should the commander require a data dictionary before anyone analyzes this roster?

3.
Training and test sets.
5 pts

A cadet builds a model on all 64 company records, reports that it fits those 64 records very well, and recommends it to the commander. State what is wrong with that procedure and what the cadet should have done instead. Be specific about what the reported error is actually measuring.

4.
The modeling framework.
4 pts

For the commander's remedial-training problem, state the decision, the objective, and one assumption the model requires.

Part II · 50 points

Descriptive Modeling

Lessons 2–7. Center, spread, quartiles, standardizing, choosing the right summary, correlation, and captioning a figure.

Questions 5–9 use the eight recorded 12-mile ruck times, in minutes, from 1st Platoon, in ascending order:

1st Platoon — 12-mile ruck times (minutes)
x1x2x3x4x5x6x7x8
168172175175180184190236
5.
Mean and median.
6 pts

Calculate the mean and the median. Show your work.

6.
Sample standard deviation.
8 pts

Calculate the sample standard deviation s. Show the squared deviations.

s = √Σ(xi − x̄)2n − 1

s = (two decimal places)

7.
z-scores and the outlier screen.
8 pts

Use your own x̄ from Question 5 and s from Question 6. Follow-on credit applies, so carry your values forward and continue.

  1. Calculate the z-score of x8 = 236 and of x1 = 168. Show your work.
  2. Apply the course's |z| > 2.5 screening rule. Which of these two, if either, is flagged as a potential outlier? Then explain what the flag does and does not establish.
8.
Choosing a measure of center.
6 pts

The histogram shows all 64 recorded ruck times in Company B.

0 5 10 15 20 160 170 180 190 200 210 220 230 240 12-mile ruck time (minutes) Number of cadets
Figure 1 — Company B 12-mile ruck times, n = 64

(a) Circle the measure of center you would report as the "typical" ruck time, and justify it in one to two sentences using the language of the course.

MeanMedian

(b) Without computing anything, state whether the mean is larger or smaller than the median here, and why.

9.
Boxplot versus histogram.
4 pts

Below is a boxplot of the eight 1st Platoon ruck times from Questions 5–9.

160 170 180 190 200 210 220 230 240 12-mile ruck time (minutes)
Figure 2 — 1st Platoon, n = 8. The point beyond the right whisker is flagged by the 1.5 × IQR fence.

(a) Read the five-number summary off the figure: the box edges, the line inside the box, and the whisker ends. State the IQR.

(b) What does the boxplot show that the histogram does not, and what does the histogram show that the boxplot does not?

10.
Reading a correlation table.
8 pts

Excel returns the correlations below, computed on all 64 records. Every cadet carried a 35-pound ruck; the load column varies only because of small differences in how canteens were filled.

Pearson correlation with 12-mile ruck time (minutes), n = 64
Variabler
Weekly training mileage (mi/wk)−0.92
2-mile run time (min)0.78
Body weight (lb)0.15
Ruck load (lb)0.04
  1. Which variable has the strongest linear association with ruck time? Interpret the sign in context, and state whether building a linear model on it is both reasonable and necessary.
  2. A cadet concludes: "This proves that training more miles per week causes faster ruck times." Is that supported? Name a plausible third variable.
  3. Does r = 0.04 prove that ruck load has no effect on ruck time? Explain.
11.
Write a caption. Most-missed on Gate 2
10 pts

You are preparing a one-page read-ahead for the battalion commander, who has no statistics background and must decide whether to fund a remedial ruck program. Write a caption for Figure 1. A caption has a title and then, in three to five sentences, (1) describes the figure and (2) explains what the reader should take away and why it matters for the decision.

Part III · 54 points

Linear Regression & Residuals

Lessons 9–16. Reading a fitted line, interpreting slope, intercept and R2, residual diagnostics against fitted values, and out-of-sample error.

Six cadets form the training set. x is average weekly training mileage, y is the 12-mile ruck time.

Training set (n = 6)
Cadet123456
x — mileage (mi/wk)51015202530
y — ruck time (min)210195189183177180
165 170 180 190 200 210 0 5 10 15 20 25 30 35 Weekly training mileage (mi/wk) Ruck time (min)
Figure 3 — Training set with least-squares trendline. Minor gridlines are 2 minutes.
12.
Read the line of best fit off the graph.
6 pts

Do not use the data table — read Figure 3.

  1. Read the trendline (not the data points) at x = 10 and x = 30.
    at x = 10: ŷ ≈    at x = 30: ŷ ≈
  2. Use those two points to compute the slope, then write the equation.
    ŷ =

Excel reports the trendline as ŷ = 210 − 1.2x with R2 = 0.854. Use these for the rest of Part III.

13.
Interpretation.
6 pts
  1. Interpret the slope in context, with units.
  2. Interpret the intercept in context.
14.
Extrapolation. Most-missed on Gate 2
6 pts

The battalion S3 likes your model and proposes using it to identify cadets for a high-mileage training pilot. He asks what ruck time the model predicts for a cadet who trains 60 miles per week.

  1. Compute the model's prediction at x = 60. Show your work.
    ŷ = minutes
  2. Is this an appropriate use of the model? Justify your answer.
15.
R2.
4 pts

Complete the sentence, then state one thing R2 does not tell you.

Approximately % of the variation in is explained by its linear relationship with .

Strong, moderate, or weak?

16.
Residuals and training error.
8 pts

(a) Complete the four empty rows using ŷ = 210 − 1.2x.

Training residuals, e = y − ŷ
xy actualŷ predictede residual
5210
10195198−3
15189
20183186−3
25177180−3
30180

(b) Compute the training MAE from all six residuals.

(c) Why does MAE use absolute values? What would happen if you averaged the six signed residuals?

17.
Residual plot and the regression assumptions.
10 pts

(a) Plot the six residuals from Question 16 against the fitted values ŷ.

+8 +4 0 −4 −8 170 180 190 200 210 Fitted value ŷ (min) Residual (min)
Figure 4 — Plot your residuals here. Note the horizontal axis is ŷ, not x.

(b) Assess each assumption, citing evidence from your plot: linearity, constant variance, outliers.

18.
Distribution of residuals.
4 pts

Compute the mean, standard deviation, minimum, and maximum of the six training residuals.

The mean of the residuals should be very close to zero for a correctly fitted least-squares model. Is yours? What would it mean if the mean were substantially different from zero?

19.
Test error versus training error.
10 pts

Three cadets were held out of the fit entirely. Complete the table using ŷ = 210 − 1.2x.

Test set (n = 3) — held out of the fit
x (mi/wk)y actualŷ predicted|y − ŷ|
8208
18183
28185
Test MAE

In two to three sentences, compare your test MAE against the training MAE you computed in Question 16(b) and state what the two numbers mean for using this model on this term's cadets.

Part IV · 32 points

K-Nearest Neighbors

Lessons 19–21. Building a class label, standardizing with training statistics, distances, the majority vote, and the effect of k.

The commander now wants a simpler prediction: will a cadet earn a GO on the 12-mile march? KNN requires a categorical dependent variable, so the continuous ruck time is converted to a class at the Army standard of 180 minutes.

KNN training set (ntrain = 6)
CadetWeekly mileage (mi/wk)2-mile run (min)Ruck time (min)Class
A815205NO-GO
B1614190NO-GO
C2017198NO-GO
D2013182NO-GO
E2411170GO
F3214176GO

Cadet Q, whose class is unknown, trains 30 mi/wk and runs 2 miles in 12.0 minutes.

20.
Set up the model.
6 pts
  1. State the threshold used to create the class label, and why 180 minutes is a defensible choice here rather than, say, the median ruck time.
  2. Explain why the test records must be standardized using the training mean and standard deviation, rather than statistics computed from the test set itself.
21.
Standardize.
8 pts
  1. Compute the mean and sample standard deviation of weekly mileage across the six training cadets.
    x̄ =    s =
  2. For 2-mile run time you are given the training statistics x̄ = 14.0 min and s = 2.0 min. Standardize cadet Q on both features.
    z1* =    z2* =
22.
Standardized distances.
8 pts

Complete the table, then show one full distance calculation.

Distance from Q = (1.25, −1.00) in standardized space
Cadetz1z2d from QRankClass
A−1.500.50NO-GO
BNO-GO
CNO-GO
DNO-GO
EGO
FGO
23.
Choose k and classify.
6 pts

(a) Apply the rule of thumb k ≈ √ntrain, rounded to the nearest odd integer.

ntrain =   √ntrain ≈   initial k =

Why do we prefer an odd value of k for a two-class problem?

(b) Classify cadet Q by majority vote at each k, naming the neighbors used.

kNeighbors used (in order)VotePredicted class
1
3
5
24.
Sensitivity to k.
4 pts

Four held-out cadets were classified at each k. Carry your three predictions for cadet Q down from Question 23, mark them correct or incorrect, then fill in the three accuracies.

Test cadetmi/wk2-mi runTrue classk=1k=3k=5
Q3012.0GOQ23Q23Q23
T21217.0NO-GONO-GO ✓NO-GO ✓NO-GO ✓
T32813.0GOGO ✓GO ✓NO-GO ✗
T43216.0NO-GOGO ✗NO-GO ✓NO-GO ✓
Test accuracy

Which k would you use, and why? Explain both what goes wrong at k = 1 and what goes wrong at k = 5.

Part V · 8 points

Synthesis

Lessons 8, 16, 19–21. Comparing models on the right metric and knowing what each family actually assumes.

25.
Compare all models.
5 pts

Complete the summary table. Write N/A where a metric does not apply, and be prepared to say why.

ModelR2Train MAETest MAETest accuracy
Baseline — predict the training mean (189 min)N/A9.009.67
Linear regression, ŷ = 210 − 1.2x0.854Q16Q19
Degree-5 polynomial through all six training points1.0000.0014.50
KNN classification at your chosen kQ24

Which model do you recommend, and why? Address the degree-5 polynomial specifically.

26.
Contrast the two model families.
3 pts

In two to three sentences, describe one fundamental difference between the way linear regression makes a prediction and the way KNN makes a prediction. What does each assume, or not assume, about the relationship between the inputs and the output?

What actually costs points

The first two are not predictions. They are the two most-missed items on Gate 2, where the caption alone carried 10 of the 25 points.

1 — The captionCadets describe the axes and stop, or write for a statistician instead of the decision-maker. A caption needs a title, a description, and a takeaway that answers the question the reader actually has. On a residual plot, name the equal-variance problem precisely — a fan means the model is both over- and under-predicting, not just one of them.
2 — ExtrapolationCadets compute the prediction correctly, then defend using it. Being able to evaluate the equation at x = 60 says nothing about whether the answer means anything. Full credit requires the words: outside the training range, no evidence the relationship holds there.
3 — Dividing by nSample standard deviation divides by n − 1. On Q6 that is 21.75 versus 20.34, and every z-score downstream is then wrong.
2 — Flagging 236 as an outlierQ7 is a trap. z = 2.34 does not clear the course's |z| > 2.5 screen (HW1 Q16). Cadets who pattern-match "biggest number = outlier" lose it.
3 — Interpretation with no context or units"The slope is −1.2" is not an interpretation. Every interpretation item requires the number, both units, the direction, and association rather than causation language.
4 — Reading residual size instead of residual patternSmall residuals do not make a model correct. The U-shape in Q17 is the finding — and the plot is against fitted values, per HW2 Q11.
5 — Treating training error as the model's accuracyQ19 and Q25 both turn on this. A training MAE of 0.00 is a warning sign, not a result.
8 — Averaging in KNNKNN here is classification: the prediction is the majority class of the k neighbors, not an average of their values. A cadet who averages is answering a question the course does not ask.
7 — Standardizing the test set with test statisticsQ20(b). The training mean and SD define the scale; recomputing on the test set moves the yardstick and leaks test information into the model.

If you are short on time

Cut to a core set: Questions 5, 6, 7, 11, 12, 14, 16, 19, 21, 22, 23 — 84 points. That keeps every computation the class flagged as likely (mean, standard deviation, z-scores, line of best fit, residuals, train-versus-test MAE, and the full KNN chain) and both items Gate 2 proved were the hardest: the caption (11) and extrapolation (14). Assign the remaining interpretation items as overnight work — they need no calculator.

Overnight reps — from HW1 and HW2, not on this test

Work these straight out of the homeworks using Penguins, Ames Housing, or the Happiness Report:

Answer key at a glance

ItemAnswerItemAnswer
5. Mean / median185 / 177.5 min16b. Training MAE4.00 min
6. s21.75 min17. Residual patternU-shaped, linearity fails
7a. z(236) / z(168)2.34 / −0.7818. Residual mean / sd0.00 / 4.65 min
7b. |z| > 2.5 screenNeither flagged19. Test MAE7.20 min
8. Measure of centerMedian (right-skewed)21. mileage mean / s20 / 8 mi/wk
9. Boxplot summary173.5 / 177.5 / 187, IQR 13.521b. Q standardized(1.25, −1.00)
12. Line of best fitŷ = 210 − 1.2x22. Neighbor orderE, F, D, B, C, A
13a. Slope−1.2 min per mi/wk23. k = 1 / 3 / 5GO / GO / NO-GO
14. Extrapolation138 min — inappropriate24. Accuracy 1 / 3 / 575% / 100% / 50%
15. R285.4% of variation25. RecommendLinear regression