June 2025 Unit 2 Q3
3. Martin is designing a new question for a multiple-choice quiz. The question has three options, only one of which is correct. Martin designs the question to be misleading and difficult to answer correctly. He claims that people would do better if they select their answer to the question at random rather than trying to work out the correct answer.
Martin decides to carry out a hypothesis test at the 5% level of significance to investigate his claim. He takes a random sample of 63 people and asks them to try to work out the correct answer to the new question, instead of just guessing randomly. He finds that 15 people answer the new question correctly and 48 people answer incorrectly.
| Scheme | Marks | Comments |
|---|---|---|
| (Let \(p\) be the probabiltiy of a correct answer.) \(H_0: p=\frac{1}{3}\) \(H_1: p<\frac{1}{3}\) | B1 | |
| [1] |
| Scheme | Marks | Comments |
|---|---|---|
| (i) (Let the random variable \(X\) represent the number of people who answer the new question correctly.) Under \(H_0\), \(X\sim\text{B}\left(63,\frac{1}{3}\right)\) | B1 | si FT their (a) if stated. |
| \(p-\text{value}=\text{P}(X\leq 15)=0.0679\) | M1A1 | M1 for \(\text{P}(X\leq 15)\) written or used. |
| (ii) Since \(0.0679>0.05\) There is insufficient evidence to reject \(H_0\). | M1 | |
| There is insufficient evidence at the 5% level to suggest that people do worse by trying to work out the correct answer rather than selecting the answer at random. e.g. Insufficient evidence to support Martin’s claim. | A1 | cso (all previous marks in (b) awarded) |
| [5] |
| Scheme | Marks | Comments |
|---|---|---|
| If \(X\sim\text{B}\left(63,\frac{1}{3}\right)\) Critical region \(\text{P}(X\leq 14)=0.0379\) | M1 | M0 for \(\text{P}(X=x)\) |
| Therefore CR is \(X\leq 14\) | A1 | |
| Prob (Type II error) = \(\text{P}\big(X\geq 15\,|\,X\sim\text{B}(63,0.2)\big)\) \(=1-0.7318\) | M1 | |
| \(=0.2682\) | A1 | cao 3sf or better |
| Valid interpretation. e.g. People are worse at answering the new question correctly than if they had selected the answer randomly, but Martin fails to conclude that is the case. e.g. There is about a 27% chance that Martin will make the incorrect conclusion if \(p=0.2\). e.g. The probability of failing to reject the hypothesis that the probability of people answering the new question correctly is worse than average when, in fact, it should have been rejected is about 27%. e.g. If \(p=0.2\), Martin should reject \(H_0\). He should conclude that people would indeed have been better off selecting their answer at random. There’s about a 25% chance that he doesn’t conclude that. | E1 | FT their Type II error |
| [5] | ||
| Total for Question | 11 |