OCR MEI S1 2011 June — Question 7 18 marks

Exam BoardOCR MEI
ModuleS1 (Statistics 1)
Year2011
SessionJune
Marks18
PaperDownload PDF ↗
Mark schemeDownload PDF ↗
TopicHypothesis test of binomial distributions
TypeState hypotheses with additional parts
DifficultyStandard +0.3 This is a standard S1 binomial distribution and hypothesis testing question with straightforward calculations. Part (i) requires direct application of binomial probability formulas, parts (ii)-(iii) involve routine one-tailed hypothesis tests with given parameters, and part (v) tests understanding of significance levels. All techniques are textbook exercises requiring no novel insight, though the multi-part structure and hypothesis testing context place it slightly above the most basic recall questions.
Spec2.04b 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

Any patient who fails to turn up for an outpatient appointment at a hospital is described as a 'no-show'. At a particular hospital, on average 15% of patients are no-shows. A random sample of 20 patients who have outpatient appointments is selected.
  1. Find the probability that
    1. there is exactly 1 no-show in the sample, [3]
    2. there are at least 2 no-shows in the sample. [2]
The hospital management introduces a policy of telephoning patients before appointments. It is hoped that this will reduce the proportion of no-shows. In order to check this, a random sample of \(n\) patients is selected. The number of no-shows in the sample is recorded and a hypothesis test is carried out at the 5% level.
  1. Write down suitable null and alternative hypotheses for the test. Give a reason for your choice of alternative hypothesis. [4]
  2. In the case that \(n = 20\) and the number of no-shows in the sample is 1, carry out the test. [4]
  3. In another case, where \(n\) is large, the number of no-shows in the sample is 6 and the critical value for the test is 8. Complete the test. [3]
  4. In the case that \(n \leqslant 18\), explain why there is no point in carrying out the test at the 5% level. [2]

(i)
Answer/Working: \(X \sim B(20, 0.15)\)
(A) Either \(P(X = 1) = \binom{20}{1} \times 0.15^1 \times 0.85^{19} = 0.1368\) or \(P(X = 1) = P(X \leq 1) - P(X \leq 0) = 0.1756 - 0.0388 = 0.1368\)
AnswerMarks
Marks: M1 \(0.15^1 \times 0.85^{19}\), M1 \(\binom{20}{1} \times p^1 q^{19}\), A1 CAO3
Guidance: With \(p + q = 1\). Allow answer 0.137 with or without working or 0.14 if correct working shown. See tables at the website http://www.mei.org.uk/files/pdf/formula_book_mf2.pdf. For misread of tables 0.3917 – 0.1216 = 0.2701 allow M1M1A0 also for 0.1304 – 0.0261 = 0.1043
Answer/Working: (B) \(P(X \geq 2) = 1 - P(X \leq 1) = 1 - 0.1756 = 0.8244\)
AnswerMarks
Marks: M1 for 1 – their 0.1756, A1 CAO2
Guidance: Provided 0.1756 comes from \(P(X=0) +P(X=1)\). Allow answer 0.824 with or without working or 0.82 if correct working shown. Point probability method: \(P(1) = 0.1368\), \(P(0) = 0.0388\). So \(1 - P(X \leq 1) = 1 - 0.1756\) gets M1 then mark as per scheme. M0A0 for \(1 - P(X \leq 1) = 1 - 0.4049 = 0.5951\). For misread of tables \(1 - 0.3917 = 0.6083\) allow M1A1 also for \(1 - 0.1304 = 0.8696\) provided consistent with part (A) OR M1A0 if formula used in part (A)
Question 7(ii)
Answer/Working: Let \(X \sim B(n, p)\). Let \(p =\) probability of a 'no-show' (for population). \(H_0: p = 0.15\), \(H_1: p < 0.15\)
\(H_1\) has this form because the hospital management hopes to reduce the proportion of no-shows.
AnswerMarks
Marks: B1 for definition of \(p\), B1 for \(H_0\), B1 for \(H_1\), E14
Guidance: Allow \(p = P(\text{no-show})\) for B1. Definition of \(p\) must include word probability (or chance or proportion or percentage or likelihood but NOT possibility). Preferably as a separate comment. However can be at end of \(H_0\) as long as it is a clear definition ' \(p =\) the probability of no-show, NOT just a sentence 'probability is 0.15'. \(H_0 : p(\text{no-show}) = 0.15\), \(H_1 : p(\text{no-show}) < 0.15\) gets B0B1B1. Allow \(p=15\%\), allow \(0\) or \(\pi\) and \(p\) but not \(x\). However allow any single symbol if defined. Allow \(H_0 = p=0.15\), Do not allow \(H_0: P(X=x) = 0.15\), \(H_1 : P(X=x) < 0.15\) (unless \(x\) correctly defined as a probability). Do not allow \(H_0\) and \(H_1\) reversed for B marks but can still get E1. Allow NH and AH in place of \(H_0\) and \(H_1\). For hypotheses given in words allow Maximum B0B1B1E1 Hypotheses in words must include probability (or chance or proportion or percentage) and the figure 0.15 oe.
Question 7(iii)
Answer/Working: \(P(X \leq 1) = 0.1756 > 5\%\)
So not enough evidence to reject \(H_0\). Not significant. Conclude that there is not enough evidence to indicate that the proportion of no-shows has decreased.
AnswerMarks
Marks: M1 for probability seen, but not in calculation for point probability; M1 dep for comparison; A14
Guidance: Zero for use of point prob - \(P(X = 1) = 0.1368\). Do NOT FT wrong \(H_1\). Allow accept \(H_0\), or reject \(H_1\). Full marks only available if 'not enough evidence to...' oe mentioned somewhere. Do not allow 'evidence to reject \(H_1\)' for final mark but can still get 3/4. Upper end comparison: \(1 - 0.1756 = 0.8244 < 95\%\) gets
Question 7 continued
Answer/Working: Note: use of critical region method scores M1 for region [0], M1 for 1 does not lie in critical region, then A1 E1 as per scheme
AnswerMarks
Marks:
Guidance: E1 dep for conclusion in context. M2 then A1E1 as per scheme. Line diagram method: M1 for squiggly line between 0 and 1 with arrow pointing to left, M1 0.0388 seen on diagram from squiggly line or from 0, A1E1 for correct conclusion. Bar chart method: M1 for line clearly on boundary between 0 and 1 and arrow pointing to left, M1 0.0388 seen on diagram from boundary line or from 0, A1E1 for correct conclusion
(iv)
Answer/Working: \(6 < 8\). So there is sufficient evidence to reject \(H_0\). Conclude that there is enough evidence to indicate that the proportion of no-shows appears to have decreased.
AnswerMarks
Marks: M1 for comparison seen, A1, E1 for conclusion in context3
Guidance: Allow '6 lies in the CR'. Do NOT insist on 'not enough evidence' here. Do not FT wrong \(H_1\); p>0.15 but may get M1. In part (iv) ignore any interchanged \(H_0\) and \(H_1\) seen in part (ii)
(v)
Answer/Working: For \(n \leq 18\), \(P(X \leq 0) > 0.05\) so the critical region is empty.
AnswerMarks
Marks: E1 for \(P(X \leq 0) > 0.05\), E1 indep for critical region is empty2
Guidance: E1 also for sight of 0.0536. Condone \(P(X = 0) > 0.05\) or all probabilities or values, (but not outcomes) in table (for \(n \leq 18\)) > 0.05. Or 'There is no critical region'. For second E1 accept '\(H_0\) would always be accepted'. Do NOT FT wrong \(H_1\). Use professional judgement - allow other convincing answers
## (i)
**Answer/Working:** $X \sim B(20, 0.15)$

(A) Either $P(X = 1) = \binom{20}{1} \times 0.15^1 \times 0.85^{19} = 0.1368$ or $P(X = 1) = P(X \leq 1) - P(X \leq 0) = 0.1756 - 0.0388 = 0.1368$

**Marks:** M1 $0.15^1 \times 0.85^{19}$, M1 $\binom{20}{1} \times p^1 q^{19}$, A1 CAO | 3

**Guidance:** With $p + q = 1$. Allow answer 0.137 with or without working or 0.14 if correct working shown. See tables at the website http://www.mei.org.uk/files/pdf/formula_book_mf2.pdf. For misread of tables 0.3917 – 0.1216 = 0.2701 allow M1M1A0 also for 0.1304 – 0.0261 = 0.1043

**Answer/Working:** (B) $P(X \geq 2) = 1 - P(X \leq 1) = 1 - 0.1756 = 0.8244$

**Marks:** M1 for 1 – their 0.1756, A1 CAO | 2

**Guidance:** Provided 0.1756 comes from $P(X=0) +P(X=1)$. Allow answer 0.824 with or without working or 0.82 if correct working shown. Point probability method: $P(1) = 0.1368$, $P(0) = 0.0388$. So $1 - P(X \leq 1) = 1 - 0.1756$ gets M1 then mark as per scheme. M0A0 for $1 - P(X \leq 1) = 1 - 0.4049 = 0.5951$. For misread of tables $1 - 0.3917 = 0.6083$ allow M1A1 also for $1 - 0.1304 = 0.8696$ provided consistent with part (A) OR M1A0 if formula used in part (A)

---

# Question 7(ii)

**Answer/Working:** Let $X \sim B(n, p)$. Let $p =$ probability of a 'no-show' (for population). $H_0: p = 0.15$, $H_1: p < 0.15$

$H_1$ has this form because the hospital management hopes to reduce the proportion of no-shows.

**Marks:** B1 for definition of $p$, B1 for $H_0$, B1 for $H_1$, E1 | 4

**Guidance:** Allow $p = P(\text{no-show})$ for B1. Definition of $p$ must include word probability (or chance or proportion or percentage or likelihood but NOT possibility). Preferably as a separate comment. However can be at end of $H_0$ as long as it is a clear definition ' $p =$ the probability of no-show, NOT just a sentence 'probability is 0.15'. $H_0 : p(\text{no-show}) = 0.15$, $H_1 : p(\text{no-show}) < 0.15$ gets B0B1B1. Allow $p=15\%$, allow $0$ or $\pi$ and $p$ but not $x$. However allow any single symbol if defined. Allow $H_0 = p=0.15$, Do not allow $H_0: P(X=x) = 0.15$, $H_1 : P(X=x) < 0.15$ (unless $x$ correctly defined as a probability). Do not allow $H_0$ and $H_1$ reversed for B marks but can still get E1. Allow NH and AH in place of $H_0$ and $H_1$. For hypotheses given in words allow Maximum B0B1B1E1 Hypotheses in words must include probability (or chance or proportion or percentage) and the figure 0.15 oe.

---

# Question 7(iii)

**Answer/Working:** $P(X \leq 1) = 0.1756 > 5\%$

So not enough evidence to reject $H_0$. Not significant. Conclude that there is not enough evidence to indicate that the proportion of no-shows has decreased.

**Marks:** M1 for probability seen, but not in calculation for point probability; M1 dep for comparison; A1 | 4

**Guidance:** Zero for use of point prob - $P(X = 1) = 0.1368$. Do NOT FT wrong $H_1$. Allow accept $H_0$, or reject $H_1$. Full marks only available if 'not enough evidence to...' oe mentioned somewhere. Do not allow 'evidence to reject $H_1$' for final mark but can still get 3/4. Upper end comparison: $1 - 0.1756 = 0.8244 < 95\%$ gets

---

# Question 7 continued

**Answer/Working:** Note: use of critical region method scores M1 for region [0], M1 for 1 does not lie in critical region, then A1 E1 as per scheme

**Marks:** | 

**Guidance:** E1 dep for conclusion in context. M2 then A1E1 as per scheme. Line diagram method: M1 for squiggly line between 0 and 1 with arrow pointing to left, M1 0.0388 seen on diagram from squiggly line or from 0, A1E1 for correct conclusion. Bar chart method: M1 for line clearly on boundary between 0 and 1 and arrow pointing to left, M1 0.0388 seen on diagram from boundary line or from 0, A1E1 for correct conclusion

## (iv)
**Answer/Working:** $6 < 8$. So there is sufficient evidence to reject $H_0$. Conclude that there is enough evidence to indicate that the proportion of no-shows appears to have decreased.

**Marks:** M1 for comparison seen, A1, E1 for conclusion in context | 3

**Guidance:** Allow '6 lies in the CR'. Do NOT insist on 'not enough evidence' here. Do not FT wrong $H_1$; p>0.15 but may get M1. In part (iv) ignore any interchanged $H_0$ and $H_1$ seen in part (ii)

## (v)
**Answer/Working:** For $n \leq 18$, $P(X \leq 0) > 0.05$ so the critical region is empty.

**Marks:** E1 for $P(X \leq 0) > 0.05$, E1 indep for critical region is empty | 2

**Guidance:** E1 also for sight of 0.0536. Condone $P(X = 0) > 0.05$ or all probabilities or values, (but not outcomes) in table (for $n \leq 18$) > 0.05. Or 'There is no critical region'. For second E1 accept '$H_0$ would always be accepted'. Do NOT FT wrong $H_1$. Use professional judgement - allow other convincing answers

---
Any patient who fails to turn up for an outpatient appointment at a hospital is described as a 'no-show'. At a particular hospital, on average 15% of patients are no-shows. A random sample of 20 patients who have outpatient appointments is selected.

\begin{enumerate}[label=(\roman*)]
\item Find the probability that
\begin{enumerate}[label=(\Alph*)]
\item there is exactly 1 no-show in the sample, [3]
\item there are at least 2 no-shows in the sample. [2]
\end{enumerate}
\end{enumerate}

The hospital management introduces a policy of telephoning patients before appointments. It is hoped that this will reduce the proportion of no-shows. In order to check this, a random sample of $n$ patients is selected. The number of no-shows in the sample is recorded and a hypothesis test is carried out at the 5% level.

\begin{enumerate}[label=(\roman*)]
\setcounter{enumi}{1}
\item Write down suitable null and alternative hypotheses for the test. Give a reason for your choice of alternative hypothesis. [4]
\item In the case that $n = 20$ and the number of no-shows in the sample is 1, carry out the test. [4]
\item In another case, where $n$ is large, the number of no-shows in the sample is 6 and the critical value for the test is 8. Complete the test. [3]
\item In the case that $n \leqslant 18$, explain why there is no point in carrying out the test at the 5% level. [2]
\end{enumerate}

\hfill \mbox{\textit{OCR MEI S1 2011 Q7 [18]}}