October 2020 Paper 2 Q10
10 Pierre is a chef. He claims that 90% of his customers are satisfied with his cooking. Yvette suspects that Pierre is over-confident about the level of satisfaction amongst his customers. She talks to a random sample of 15 of Pierre’s customers, and finds that 11 customers say that they are satisfied. She then performs a hypothesis test.
Carry out the test at the 5% significance level. [7]
| Scheme | Marks | AO |
|---|---|---|
| \(\mathrm{H_0}\): \(p = 0.9\), where \(p = \mathrm{P}\)(a random customer is satisfied) | B1 | 1.1 |
| \(\mathrm{H_1}\): \(p \lt 0.9\) | B1 | 2.5 |
| \(X \sim \mathrm{Bin}(15, 0.9)\) and \(X \leqslant 10\) or 11 or 12 (condone \(\lt\) or \(=\) or \(\gt\) or \(\geqslant\)) | M1 | 3.3 |
| \(\mathrm{P}(X \leqslant 11)\) oe \(= 0.0556\) | \(\mathrm{P}(X \gt 11)\) oe \(= 0.944\) | A1 | 3.4 |
| Comp 0.05 | Comp 0.95 | A1 | 1.1 |
| Do not reject \(\mathrm{H_0}\) Condone Accept \(\mathrm{H_0}\) | M1 | 1.1 |
| There is insufficient evidence that Pierre is overconfident (or that < 90% are satisfied) oe, eg There is insufficient evidence that Yvette’s suspicion is correct | A1f | 2.2b |
| [7] |
Notes
Allow rounded or truncated to 2 sf throughout
B1 (\(\mathrm{H_0}\)): or \(p\) is proportion satisfied Allow other letters
OR \(p = 90\%\), where \(p\) is % of customers satisfied
B1 (\(\mathrm{H_1}\)): Subtract B1 for each error eg:
2-tail B1B0; undefined \(p\) B1B0; not in terms of parameter B1B0; Use of \(\leqslant\) with definition B1B0; Not include value 0.9 B0B0; \(\mathrm{H_0} = 0.9\) etc: B0B0
M1: Stated or implied eg by 0.0556 or 0.184 or 0.0127 or 0.944 or 0.816 or 0.987 or 0.0428 or 0.129
A1: BC cao
A1: Dep 0.0556 or 0.184 or 0.0127 or 0.944 or 0.816 or 0.987
Must be correct comparison, eg not 0.944 comp with 0.05
M1: Dep 0.0556 or 0.184 or 0.0127 (2 sf) or \(\mathrm{P}(X \leqslant 10\) or 11 or 12) seen or 0.944 or 0.816 or 0.987 \(\mathrm{P}(X \gt 10\) or 11 or 12) seen
And dep correct comparison, eg, not 0.944 comp with 0.05
A1f: In context. Not definite. Full statement
Not: There is evidence that Pierre is not overconfident oe
Alternative method for middle two A-marks
| Scheme | Marks |
|---|---|
| \(\mathrm{P}(X \leqslant 10) = 0.0127\) \(\mathrm{P}(X \leqslant 11) = 0.0556\) (Both needed) | A1 |
| Hence rejection region is \(X \leqslant 10\) (or \(X \lt 11\)) or critical value is \(X = 10\) (Dep on M1) | A1 |
Alternative (normal approximation)
| Scheme | Marks |
|---|---|
| N~Bin: \(\mu = 13.5\), \(\mu \lt 13.5\) (dep defined \(\mu\). If undefined: B0B0) | B1B0 |
| \(\mathrm{N}(13.5, 1.35)\) & \(X = 11.5\) or 11 (soi) | M1 |
| \(p = 0.0426\) | A1 |
| compare 0.05 (dep 0.0426 or 0.0157) | A1 |
| Conclusion | M0A0 |
2-tail
| Scheme | Marks |
|---|---|
| \(\mathrm{H_0}\): \(p = 0.9\) defined \(p\) \(\mathrm{H_1}\): \(p \ne 0.9\) | B1 B0 |
| \(X \sim \mathrm{Bin}(15, 0.9)\) and \(X \leqslant 11\) or 12 (condone \(\lt\) or \(=\) or \(\gt\) or \(\geqslant\)) Stated or implied eg by 0.0556 or 0.184 or 0.944 or 0.816 or 0.0428 or 0.129 | M1 |
| \(\mathrm{P}(X \leqslant 11)\) oe \(= 0.0556\) | A1 |
| Comp 0.025 (Dep 0.0556 or 0.184) | A1 |
| Conclusion | M0 A0 |