Skip to main content

Past ETX5900 exam coverage against the 2026 eight-week course

Current-scope update: This historical audit was counted against the earlier plan, which treated formal chi-square as legacy. The current Week 5 teaches and assesses the chi-square test of independence under LO4. Reclassify the chi-square items tagged “Legacy” below as Week 5 candidates when drafting the new blueprint, while independently checking their keys and expected-count conditions. The old counts and coverage table below have not been recomputed; full time-series forecasting remains outside the current core.

Scope and counting rule

Sources: ETX5900 Exam_2026.docx.md (called MD below) and Nov12 Exam_ETX5900.docx.pdf (called PDF below). Each has 10 MCQs, 10 numerical entry questions, and three multi-part short answers, for 23 numbered questions per paper. The MD file marks MCQ keys in bold and gives numerical keys after Answer:; the PDF marks MCQ keys in bold and written answers in red. The answer markings were used to review the items, but not to decide their curriculum tags.

Each numbered question receives one primary topic, week, and learning outcome so the counts add to 23 per paper. Mixed short-answer questions are listed by subpart in their notes; secondary outcomes are mentioned but do not increase the totals. A dash means the item assesses a legacy technique outside the planned 2026 core. Week 8 is integration and revision, so a question using previously taught techniques is assigned to its substantive teaching week rather than counted again in Week 8. These are question counts, not mark-weighted coverage.

MD paper: question-level map

Q Type Recalled topic or task 2026 week Primary LO Scope note
1 MCQ Correlation sign and simple regression slope 5 LO5 Also draws on LO4 association. Marked key is wrong; see audit.
2 MCQ Compare time-series forecast errors (MAD, MSFE) Legacy — Full forecasting is outside the planned core.
3 MCQ Direction of conditional probability 2 LO2
4 MCQ Interpret a row percentage in a contingency/pivot table 2 LO2 Also uses Week 1 pivot-table literacy.
5 MCQ Normality of the sampling distribution of a mean 3 LO3
6 MCQ Type I/II errors and significance level 4 LO3
7 MCQ Chi-square observed versus expected frequencies Legacy — Chi-square test procedure.
8 MCQ Chi-square independence test conclusion from p-value Legacy — Categorical independence itself is core; this formal test is not.
9 MCQ Effect of sample mean on confidence-interval width 3 LO3
10 MCQ Chi-square statistic when observed equals expected Legacy — Chi-square test procedure.
11 Numerical Probability for a sample mean 3 LO3 Uses Week 2 normal calculations; key needs correction.
12 Numerical Sample size needed to reduce standard error 3 LO3
13 Numerical Trend and seasonal forecast from regression output Legacy — Full time-series forecasting; no answer key shown.
14 Numerical Interpret a seasonal index relative to trend Legacy — Full time-series forecasting.
15 Numerical Joint probability from a conditional probability 2 LO2
16 Numerical Conditional probability from a contingency table 2 LO2
17 Numerical Point estimate of a population proportion 3 LO3
18 Numerical Expected cell count under chi-square independence Legacy — Closely related to Week 2 independence, but the expected-frequency calculation is not explicit core content.
19 Numerical Chi-square test statistic from a table Legacy — Key needs correction.
20 Numerical Trend and seasonal forecast Legacy — Full time-series forecasting.
21 Short answer (1) Interpret a mean CI; (2–4) proportion test and Type II error 4 LO3 Also assesses Week 3 estimation.
22 Short answer (1–2) Multiple-regression fit and prediction; (3) coefficient test; (4) adding a predictor 6 LO5 Also assesses LO4; answer text has inconsistent p-values.
23 Short answer (1–2) Sample-mean distribution and percentile; (3–4) proportion CI and claim 3 LO3 Also uses Week 2 normal percentile work.

PDF paper: question-level map

Q Type Recalled topic or task 2026 week Primary LO Scope note
1 MCQ Sample-size effect on mean/proportion CI width 3 LO3
2 MCQ Standard normal versus Student’s t distribution 2 LO2
3 MCQ Select t distribution for a mean CI 3 LO3 Also uses Week 2 distribution knowledge.
4 MCQ Frame a chi-square independence hypothesis Legacy — Formal chi-square test.
5 MCQ Type I/II error tradeoff at a small alpha 4 LO3
6 MCQ Repeated-sampling interpretation of 95% CIs 3 LO3 Stem has inconsistent sample counts.
7 MCQ Direction of conditional probability 2 LO2
8 MCQ Chi-square statistic when observed equals expected Legacy — Chi-square test procedure.
9 MCQ Interpret a row percentage in a pivot table 2 LO2 Also uses Week 1 pivot-table literacy.
10 MCQ Interpret chi-square statistic, p-value, and expected counts Legacy — Formal chi-square test.
11 Numerical Sample size needed to reduce standard error 3 LO3
12 Numerical Probability for a sample mean 3 LO3 Uses Week 2 normal calculations.
13 Numerical Normal upper-tail percentile 2 LO2 Key needs correction.
14 Numerical Missing seasonal index in multiplicative model Legacy — Full time-series forecasting.
15 Numerical Inverse standard-normal probability 2 LO2
16 Numerical Point estimate of a population proportion 3 LO3 Key needs correction.
17 Numerical Simple-regression R-squared from correlation 5 LO5 Key needs correction.
18 Numerical Expected value of a sample mean 3 LO3
19 Numerical Conditional percentage from a contingency table 2 LO2
20 Numerical Perfect positive correlation for an exact linear rule 5 LO4 LO5 foundation.
21 Short answer (a–c) One-sided mean test; (d–e) Type II error and consequences 4 LO3 Hypothesis label and calculation/key conflict.
22 Short answer (a) Sampling distribution; (b–c) mean CI and interpretation 3 LO3 Part (a) confuses sample estimate with population mean.
23 Short answer (a–b) Multiple-regression equation and fit; (c) coefficient test; (d) model comparison 6 LO5 Also assesses LO4.

Primary learning-outcome counts

2026 outcome MD PDF Combined
LO1 — data types and visual presentation 0 0 0
LO2 — probability and distributions 4 6 10
LO3 — sampling, estimation, and hypothesis testing 8 10 18
LO4 — relationships between business data 0 1 1
LO5 — simple and multiple regression 2 2 4
LO6 — Excel business modelling 0 0 0
Legacy topics outside the planned core 9 4 13
Total numbered questions 23 23 46

LO1 has secondary exposure through two pivot-table items (MD Q4, PDF Q9), and LO4 through regression/association items, but neither receives much direct assessment. LO6 has no exposure. The short-answer mark share is more concentrated than these question counts suggest: MD Q22 and PDF Q23 each devote 10 marks to multiple regression.

Primary topic and week counts

Topic / 2026 teaching week MD PDF Combined
Week 1 — data literacy, visualisation, descriptive statistics 0 0 0
Week 2 — probability, contingency tables, distributions 4 6 10
Week 3 — sampling and confidence intervals 6 8 14
Week 4 — hypothesis testing 2 2 4
Week 5 — association and simple regression 1 2 3
Week 6 — multiple regression 1 1 2
Week 7 — spreadsheet business modelling 0 0 0
Week 8 — integration as a separate primary topic 0 0 0
Legacy — chi-square test procedure 5 3 8
Legacy — full time-series forecasting 4 1 5
Total numbered questions 23 23 46

Answer-key and item-quality audit before reuse

  • MD Q1: Positive correlation in simple linear regression implies a positive fitted slope. The bolded C says correlation indicates strength but not direction; A is the defensible answer.
  • MD Q11: With standard error 12/√36 = 2, P(X̄ > 33) = P(Z > 1.5) ≈ 0.0668, which rounds to 0.07 to two decimal places, not the marked 0.06.
  • MD Q13: No Answer: entry follows the numerical forecast question. The image-based regression coefficients would need to be captured in a reproducible source before reuse.
  • MD Q19: The displayed 3 × 3 table gives a chi-square statistic of 27.31 (to two decimal places); the marked 36.54 does not match that table.
  • MD Q22(c): The worked coefficient-test answer uses two different p-values (9.42E-15/2 and 2.55E-17) for the same decision. Recheck against the displayed regression output. The wording also says education has a positive effect, which is a causal claim stronger than an observational regression establishes.
  • PDF Q6: The stem says 40 samples were drawn but asks how many of 60 intervals miss the mean. The marked 2 applies to 40; 60 would give 3 in expectation.
  • PDF Q13: The 90th percentile of N(4565, 1063²) is about 5927 dollars, not the red answer 5900.
  • PDF Q16: The point estimate is 140/200 = 0.70, not the red answer 0.43.
  • PDF Q17: For simple regression, R² = 0.85² = 0.7225, so the answer is 72.25%, not the red 56.25%.
  • PDF Q21: The answer repeats H0 for the alternative hypothesis. The reported test statistic −1.4798 and p-value 0.0730 conflict with the stated data: (6.45 − 7)/(1.36/√100) ≈ −4.044 (t with 99 degrees of freedom). Recalculate the p-value and conclusion before reuse.
  • PDF Q22(a): The answer labels 6.45, the observed sample mean, as the mean of the sampling distribution. That theoretical mean is the unknown population mean μ; 6.45 is an estimate of it.

Blueprint implication: These papers strongly favour Weeks 2–4 and preserve 13 legacy items. A 2026 exam needs deliberate Week 1/LO1 and Week 7/LO6 questions, direct simple and multiple regression coverage, and an explicit balance check by marks and subparts. The old keys should not be imported into Moodle without independent verification.