Statistics & Probability — Intermediate
Level 3 Paper — Production (Derivations, Reasoning, Explanations)
Time limit: 45 minutes
Total marks: 50
Instructions: Show all working. Where a derivation is requested, build it from first principles — stating axioms/definitions used. Marks are indicated per part.
Question 1 — Derive Bayes' theorem and apply (9 marks)
(a) Starting from the definition of conditional probability, derive Bayes' theorem for two events and with . State every step and the reason for it. (3)
(b) A diagnostic test detects a disease that affects of a population. The test is sensitive (correctly positive if diseased) and specific (correctly negative if healthy). A randomly chosen person tests positive. Compute the probability they actually have the disease, giving the answer to 4 decimal places. (4)
(c) In one sentence, explain out loud why this posterior probability is so much lower than the test's sensitivity. (2)
Question 2 — Central tendency and dispersion from grouped data (11 marks)
The times (minutes) taken by 40 students to complete a task:
| Time (min) | 0–10 | 10–20 | 20–30 | 30–40 | 40–50 |
|---|---|---|---|---|---|
| Frequency | 4 | 10 | 14 | 8 | 4 |
(a) Estimate the mean. (2)
(b) Derive the grouped median using the interpolation formula . State what each symbol represents. (4)
(c) Estimate the modal value using the modal-class formula. (3)
(d) State whether the distribution is skewed and justify using your three averages. (2)
Question 3 — Variance from first principles (8 marks)
(a) Show algebraically that the population variance identity follows from the definition . Justify each manipulation. (4)
(b) For the data set , compute the mean, variance, and standard deviation using your identity. (4)
Question 4 — Binomial distribution: mean & variance derivation (9 marks)
(a) Write the PMF of a binomial random variable and state its two parameters. (2)
(b) Derive from first principles (you may use the identity ). Explain the key step. (4)
(c) A fair six-sided die is rolled 12 times. Let be the number of sixes. State , , and compute to 4 decimal places. (3)
Question 5 — Counting & binomial theorem (8 marks)
(a) Explain the difference between a permutation and a combination, and derive . (3)
(b) In how many ways can the letters of the word BALLOON be arranged so that the two L's are never adjacent? Show your reasoning. (3)
(c) Find the coefficient of in the expansion of using the general term. (2)
Question 6 — Probability rules & explain-out-loud (5 marks)
(a) State the three Kolmogorov axioms of probability. (2)
(b) Two events , satisfy , , . Determine whether and are independent, mutually exclusive, or neither. Show the test for each. (3)
Answer keyMark scheme & solutions
Question 1
(a) (3 marks)
By definition of conditional probability: and . (1)
From the second, . (1)
Substitute into the first: . (1) This is Bayes' theorem.
(b) (4 marks)
Let = disease, = positive test.
, , . (1)
Total: . (1)
(1) . (1)
(c) (2 marks)
Because the disease is rare (2%), the large healthy population produces many false positives that swamp the smaller number of true positives, so a positive result is still probably a false alarm. (2)
Question 2
(a) (2 marks)
Midpoints: 5,15,25,35,45. . (1)
Mean min. (1)
(b) (4 marks)
. Cumulative freq: 4,14,28,36,40 → median class 20–30. (1)
Symbols: (lower boundary), (cf before class), , . (1)
Median . (1)
min. (1)
(c) (3 marks)
Modal class 20–30: . (1)
Mode . (1)
min. (1)
(d) (2 marks)
Mean > Median > Mode : meanmedianmode indicates a slight positive (right) skew. (2)
Question 3
(a) (4 marks)
. (1)
. (1)
Since , middle term . (1)
(1)
(b) (4 marks)
. (1)
. (1)
. (1)
. (1)
Question 4
(a) (2 marks)
, . (1) Parameters: (trials), (success prob). (1)
(b) (4 marks)
; term vanishes. (1)
Use : . (1)
Factor and let : . (1)
The sum , so . (1)
(c) (3 marks)
: . . (1)
. (1)
. (1)
Question 5
(a) (3 marks)
Permutation = ordered selection; combination = unordered. (1)
Each combination of objects can be ordered in ways, and these together give all permutations: . (1)
Hence . (1)
(b) (3 marks)
BALLOON: 7 letters with L×2, O×2. Total arrangements . (1)
LL together (glue): 6 units, O×2: . (1)
Never adjacent . (1)
(c) (2 marks)
General term . (0.5)
. (0.5)
Coeff . (1)
Question 6
(a) (2 marks)
(1) for all events; (2) ; (3) for mutually exclusive , . (2)
(b) (3 marks)
. (1)
Independent test: ✓ → independent. (1)
Mutually exclusive would need , false. So independent, not mutually exclusive. (1)
[
{"claim":"Bayes posterior P(D|+)=0.1624","code":"num=Rational(95,100)*Rational(2,100); den=num+Rational(10,100)*Rational(98,100); result=round(float(num/den),4)==0.1624"},
{"claim":"Grouped mean=24.5, median=24.2857","code":"mean=Rational(980,40); med=20+Rational(20-14,14)*10; result=(mean==Rational(49,2)) and (round(float(med),4)==24.2857)"},
{"claim":"Variance of {2,4,4,6,9} is 5.6","code":"xs=[2,4,4,6,9]; m=Rational(sum(xs),5); v=Rational(sum(x*x for x in xs),5)-m**2; result=v==Rational(28,5)"},
{"claim":"Binomial P(X=2), n=12,p=1/6 = 0.2961","code":"p=binomial(12,2)*Rational(1,6)**2*Rational(5,6)**10; result=round(float(p),4)==0.2961"},
{"claim":"BALLOON non-adjacent L arrangements=900","code":"total=factorial(7)/(factorial(2)*factorial(2)); glue=factorial(6)/factorial(2); result=(total-glue)==900"},
{"claim":"Coefficient of x^5 in (2x-3/x)^9 is 41472","code":"c=binomial(9,2)*2**7*(-3)**2; result=c==41472"}
]