| Exam Board | OCR MEI |
|---|---|
| Module | S1 (Statistics 1) |
| Marks | 17 |
| Paper | Download PDF ↗ |
| Mark scheme | Download PDF ↗ |
| Topic | Hypothesis test of binomial distributions |
| Type | Multiple binomial probability calculations |
| Difficulty | Standard +0.3 This is a straightforward S1 binomial hypothesis test question with standard calculations. Part (i) involves routine binomial probability computations and expectation. Parts (ii)-(iii) require a one-tailed test comparing P-values to given significance levels—all standard textbook procedures with no novel problem-solving or conceptual challenges. Slightly above average difficulty only due to the multi-part nature and need for careful interpretation of significance levels. |
| Spec | 2.04b Binomial distribution: as model B(n,p)2.04c Calculate binomial probabilities2.05a Hypothesis testing language: null, alternative, p-value, significance2.05b Hypothesis test for binomial proportion2.05c Significance levels: one-tail and two-tail |
| Answer | Marks | Guidance |
|---|---|---|
| Answer | Marks | Guidance |
| \(P(\text{Exactly 20 cured}) = \binom{20}{20} \times 0.78^{20} \times 0.22^{0} = 0.0069\) | M1 | For \(0.78^{20}\) oe |
| \(P(\text{At most 18 cured}) = 1 - (0.0069 + 0.0392)\) | M1 | For \(P(19) + P(20)\); not necessarily correct, but both correct attempts at binomial, including coefficient in (i) and no extra terms (such as \(P(X=18)\)); condone use of \(p = 0.8\) |
| \(= 0.954\) (0.95385) | A1 [3] | CAO; allow 0.95 with working |
| Answer | Marks | Guidance |
|---|---|---|
| Answer | Marks | Guidance |
| \(E(X) = np = 20 \times 0.78 = 15.6\) | B1 [1] | CAO; do not allow final answer of 15 or 16 even if correct 15.6 given earlier |
| Answer | Marks | Guidance |
|---|---|---|
| Answer | Marks | Guidance |
| Let \(X \sim B(20, 0.78)\) | ||
| Let \(p\) = probability of a patient being cured (for population) | B1 | For definition of \(p\) in context |
| \(H_0: p = 0.78\) | B1 | For \(H_0\) |
| \(H_1: p > 0.78\) | B1 | For \(H_1\); no further marks if point probabilities used |
| Answer | Marks | Guidance |
|---|---|---|
| Answer | Marks | Guidance |
| \(P(X \geq 19) = 0.0392 + 0.0069\) | B1 | For notation \(P(X \geq 19)\) or \(P(X > 18)\) or \(1 - P(X \leq 18)\) or \(1 - P(X < 19)\); notation \(P(X = 19)\) scores B0 |
| \(= 0.0461\) | B1* | CAO; allow 0.0462 |
| \(0.0461 > 1\%\) | M1* dep | For comparison with 1% |
| So not significant. | A1 | Allow 'accept \(H_0\)' or 'reject \(H_1\)' |
| Conclude that there is not enough evidence to suggest that the new drug is more effective than the old one. | E1 [8] | Must include 'insufficient evidence' or similar element of doubt; in context; do not allow 'sufficient evidence to suggest proportion cured is 0.78' |
| Answer | Marks | Guidance |
|---|---|---|
| Answer | Marks | Guidance |
| \(P(X \geq 20) = 0.0069 < 1\%\) | M1 | For at least one comparison with 1% |
| So critical region is \(\{20\}\) | B1* | CAO dep on the two correct probabilities |
| (19 not in CR so) not significant | A1* dep | Dep on correct CR |
| Conclude that there is not enough evidence to suggest that the new drug is more effective than the old one. | E1* dep | Ignore any work on lower critical region |
| Answer | Marks | Guidance |
|---|---|---|
| Answer | Marks | Guidance |
| With a 5% significance level rather than a 1% level, the null hypothesis would have been rejected. OR: 'there would be enough evidence to suggest that the new drug is more effective than the old one.' | B1* | oe; FT their probability from (ii) but NO marks if point probabilities used; must have sensible attempt to use \(P(X=19) + P(X=20)\) or must have correct CR |
| This is because \(0.0461 < 5\%\) | B1* dep [2] | oe; dep on correct answer of 0.0461 compared with 5% or 0.9539 compared with 95% or correct CR |
## Question 2:
### Part (i)(B)
| Answer | Marks | Guidance |
|--------|-------|----------|
| $P(\text{Exactly 20 cured}) = \binom{20}{20} \times 0.78^{20} \times 0.22^{0} = 0.0069$ | M1 | For $0.78^{20}$ oe |
| $P(\text{At most 18 cured}) = 1 - (0.0069 + 0.0392)$ | M1 | For $P(19) + P(20)$; not necessarily correct, but both correct attempts at binomial, including coefficient in (i) and no extra terms (such as $P(X=18)$); condone use of $p = 0.8$ |
| $= 0.954$ (0.95385) | A1 [3] | CAO; allow 0.95 with working |
### Part (i)(C)
| Answer | Marks | Guidance |
|--------|-------|----------|
| $E(X) = np = 20 \times 0.78 = 15.6$ | B1 [1] | CAO; do not allow final answer of 15 or 16 even if correct 15.6 given earlier |
### Part (ii)
| Answer | Marks | Guidance |
|--------|-------|----------|
| Let $X \sim B(20, 0.78)$ | | |
| Let $p$ = probability of a patient being cured (for population) | B1 | For definition of $p$ in context |
| $H_0: p = 0.78$ | B1 | For $H_0$ |
| $H_1: p > 0.78$ | B1 | For $H_1$; no further marks if point probabilities used |
### Part (ii) continued
| Answer | Marks | Guidance |
|--------|-------|----------|
| $P(X \geq 19) = 0.0392 + 0.0069$ | B1 | For notation $P(X \geq 19)$ or $P(X > 18)$ or $1 - P(X \leq 18)$ or $1 - P(X < 19)$; notation $P(X = 19)$ scores B0 |
| $= 0.0461$ | B1* | CAO; allow 0.0462 |
| $0.0461 > 1\%$ | M1* dep | For comparison with 1% |
| So not significant. | A1 | Allow 'accept $H_0$' or 'reject $H_1$' |
| Conclude that there is not enough evidence to suggest that the new drug is more effective than the old one. | E1 [8] | Must include 'insufficient evidence' or similar element of doubt; in context; do not allow 'sufficient evidence to suggest proportion cured is 0.78' |
**99% method:** $P(X \leq 18) = 0.9539$; B1B1* CAO; $0.9539 < 99\%$ M1* then as per scheme
**Alternative Method:**
| Answer | Marks | Guidance |
|--------|-------|----------|
| $P(X \geq 20) = 0.0069 < 1\%$ | M1 | For at least one comparison with 1% |
| So critical region is $\{20\}$ | B1* | CAO dep on the two correct probabilities |
| (19 not in CR so) not significant | A1* dep | Dep on correct CR |
| Conclude that there is not enough evidence to suggest that the new drug is more effective than the old one. | E1* dep | Ignore any work on lower critical region |
### Part (iii)
| Answer | Marks | Guidance |
|--------|-------|----------|
| With a 5% significance level rather than a 1% level, the null hypothesis would have been rejected. OR: 'there would be enough evidence to suggest that the new drug is more effective than the old one.' | B1* | oe; FT their probability from (ii) but NO marks if point probabilities used; must have sensible attempt to use $P(X=19) + P(X=20)$ or must have correct CR |
| This is because $0.0461 < 5\%$ | B1* dep [2] | oe; dep on correct answer of 0.0461 compared with 5% or 0.9539 compared with 95% or correct CR |
---
2 A drug for treating a particular minor illness cures, on average, 78\% of patients. Twenty people with this minor illness are selected at random and treated with the drug.
\begin{enumerate}[label=(\roman*)]
\item (A) Find the probability that exactly 19 patients are cured.\\
(B) Find the probability that at most 18 patients are cured.\\
(C) Find the expected number of patients who are cured.
\item A pharmaceutical company is trialling a new drug to treat this illness. Researchers at the company hope that a higher percentage of patients will be cured when given this new drug. Twenty patients are selected at random, and given the new drug. Of these, 19 are cured. Carry out a hypothesis test at the $1 \%$ significance level to investigate whether there is any evidence to suggest that the new drug is more effective than the old one.
\item If the researchers had chosen to carry out the hypothesis test at the $5 \%$ significance level, what would the result have been? Justify your answer.
\end{enumerate}
\hfill \mbox{\textit{OCR MEI S1 Q2 [17]}}