June 2023 Unit 2 Q5
5. A doctor knows that 34% of his patients currently take too little exercise. He believes he can reduce this proportion by putting patients through a behaviour change programme.
To test his belief, he takes a random sample of 156 of his patients and puts each of them through the programme.
It is found that 40 of these patients take too little exercise following the behaviour change programme.
(a) The doctor initially considers using a two-tailed test. Explain why this is not appropriate. [1]
(b) The doctor then decides to use a one-tailed test.
(i) Explain what is meant by a \(p\)-value in the context of this question. [1]
(ii) Test the doctor’s belief at the 1% level of significance, stating clearly the conclusion that the doctor should reach. [6]
| Scheme | Marks | Comments |
|---|---|---|
| Valid explanation e.g. The doctor is only interested in a reduction in the proportion of his patients who take too little exercise. e.g. The behaviour change program is designed to reduce the proportion of patients who take too little exercise. | E1 | Reference to upper tail E0 E0 for only referencing or implying increase/decrease in amount of exercise. |
| Scheme | Marks | Comments |
|---|---|---|
| (i) Valid explanation e.g. the \(p\)- value is the probability of 40 or fewer patients taking too little exercise after the programme, given that there has been no change. e.g. \(p\)-value is the probability of 40 or fewer patients taking too little exercise after the programme, given that \(p=0.34\). e.g. \(p\)-value is the probability of 40 or fewer patients taking too little exercise after the programme, given \(H_0\) is true. | E1 | oe must be in context. E0 for “given that there has been no change” or equivalent omitted. |
| (ii) \(H_0\colon p=0.34\) \(H_1\colon p\lt 0.34\) | B1 | Allow use of \(\theta\) or \(\pi\), or other letters if defined. Do not allow \(x=0.34\). Allow worded hypotheses or use of 34%. B0 for 0.34%. B0 for omission of \(p\) or a non-strict inequality in \(H_1\). |
| (Let the random variable \(X\) represent the number of patients who take too little exercise.) Under \(H_0\), \(X\sim\text{B}(156,0.34)\) | B1 | si (e.g. from correctly calculated \(P(X\le 40)\) or \(P(X=40)\)) |
| Method 1 (\(\boldsymbol{p}\)-value): \(\text{P}(X\le 40)=\ 0.015488\) | M1A1 | M0A0m0A0 for \(\text{P}(X=k)\), e.g. sight of \(P(X=40)=0.00572\) earns M0 M1 for attempt to find \(\text{P}(X\le 40)\) si. |
| Since 0.015488 > 0.01, there is insufficient evidence to reject \(H_0\). | m1 | FT their ‘\(\text{P}(X\le 40)\)’ – probability needed and correct comparison with 0.01. Do not need to see insufficient evidence for this mark. |
| There is insufficient evidence at the 1% level to suggest the behaviour change programme has had the desired effect. | A1 | CSO Must have fully correct solution. Do not allow categorical statements without reference to insufficient evidence or suggests within solution. Method 1: Withhold this mark if the p-value is not compared with the significance level. Method 2: Withhold this mark if \(P(X\le 39)\) given as CR or failing to consider both probs. Withhold this mark if the test statistic is not compared with the CV. |
| Total [8] |
Alternative (Method 2)
| Scheme | Marks | Comments |
|---|---|---|
| Method 2 (critical region): The CR is \(X\le k\) where \(P(X\le k)\le 0.01\). \(P(X\le 39)=0.009768\), \(P(X\le 40)=0.015488\) | (M1) | Either probability (evaluating \(P(X=39)\) or \(P(X=40)\) earns M0A0m0A0). |
| Hence, \(k=39\). The CR is \(X\le 39\). | (A1) | CV or CR identified (condone as probability statement, i.e. \(P(X\le 39)\)). Award M1A1 for correct CV/CR with no working. |
| Since 40 is not in the CR, there is insufficient evidence to reject \(H_0\). | (m1) | FT their calculated CV |