OCR MEI S3 2013 June — Question 2 17 marks

Exam BoardOCR MEI
ModuleS3 (Statistics 3)
Year2013
SessionJune
Marks17
PaperDownload PDF ↗
Mark schemeDownload PDF ↗
TopicT-tests (unknown variance)
TypePaired sample t-test
DifficultyStandard +0.3 This is a standard paired t-test application with clear structure: students calculate differences, find mean and standard deviation, then perform a one-tailed test against a specific value (10 litres). The calculations are straightforward with small sample size (n=10), and the confidence interval in part (iv) follows directly from the same statistics. While it requires understanding of paired tests and proper hypothesis formulation, it's a textbook example with no novel insights required—slightly easier than average due to its routine nature and clear guidance through parts (i)-(iv).
Spec5.05c Hypothesis test: normal distribution for population mean5.05d Confidence intervals: using normal distribution

2 A company supplying cattle feed to dairy farmers claims that its new brand of feed will increase average milk yields by 10 litres per cow per week. A farmer thinks the increase will be less than this and decides to carry out a statistical investigation using a paired \(t\) test. A random sample of 10 dairy cows are given the new feed and then their milk yields are compared with their yields when on the old feed. The yields, in litres per week, for the 10 cows are as follows.
CowABCDEFGHIJ
Old feed144130132146137140140149138133
New feed148139138159138148146156147145
  1. Why is it sensible to use a paired test?
  2. State the condition necessary for a paired \(t\) test.
  3. Assuming the condition stated in part (ii) is met, carry out the test, using a significance level of \(5 \%\), to see whether it appears that the company's claim is justified.
  4. Find a 95\% confidence interval for the mean increase in the milk yield using the new feed.

Question 2:
Part (i)
AnswerMarks Guidance
AnswerMarks Guidance
A paired test would eliminate any differences between individual cattleE1
[1]
Part (ii)
AnswerMarks Guidance
AnswerMarks Guidance
Must assume: Normality of population …B1
… of differencesB1
[2]
Part (iii)
AnswerMarks Guidance
AnswerMarks Guidance
\(H_0: \mu_D = 10\), \(H_1: \mu_D < 10\)B1 Both. Accept alternatives e.g. \(\mu_D > -10\) for \(H_1\), or \(\mu_A - \mu_B\) etc provided adequately defined. Hypotheses in words only must include "population". Do NOT allow "\(\bar{X} = \ldots\)" or similar unless clearly stated to be a population mean
Where \(\mu_D\) is the (population) mean increase/difference in milk yield. MUST be PAIRED COMPARISON \(t\) testB1 For adequate verbal definition. Allow absence of "population" if correct notation \(\mu\) is used
Differences (increases) (after \(-\) before) are: 4 9 6 13 1 8 6 7 9 12M1 Allow "before \(-\) after" if consistent with alternatives for hypotheses above
\(\bar{x} = 7.5\), \(s_{n-1} = 3.566(8)\) \((s_{n-1}^2 = 12.722(2))\)A1 Do not allow \(s_n = 3.3837\) \((s_n^2 = 11.45)\)
Test statistic is \(\dfrac{7.5 - 10}{\frac{3.5668}{\sqrt{10}}}\)M1 Allow candidate's \(\bar{x}\) and/or \(s_{n-1}\). Allow reversed numerator compared with 2.2164
\(= -2.2164\)A1 c.a.o. but ft from here in any case if wrong. Use of \(10 - \bar{x}\) scores M1A0, but ft
Refer to \(t_9\)M1 No ft from here if wrong
Single-tailed 5% point is \(-1.833\)A1 Must be minus 1.833 unless absolute values are being compared. No ft from here if wrong. \(P(t < -2.2164) = 0.0269\)
SignificantA1 ft only candidate's test statistic
Sufficient evidence to suggest that the mean milk yield has not increased by 10 litres (per cow per week)A1 ft only candidate's test statistic. Conclusion in context to include "on average" o.e. Accept "Sufficient evidence to suggest that the company's claim is not justified." o.e.
[10]
Part (iv)
AnswerMarks Guidance
AnswerMarks Guidance
CI is given by \(7.5 \pm\)M1 ZERO/4 if not same distribution as test. Same wrong distribution scores maximum M1B0M1A0. Recovery to \(t_9\) is OK. Allow candidate's \(\bar{x}\)
\(2.262\)B1 2.262 seen
\(\times \dfrac{3.5668}{\sqrt{10}}\)M1 Allow candidate's \(s_{n-1}\)
\(= 7.5 \pm 2.5514 = (4.948, 10.052)\)A1 c.a.o. Must be expressed as an interval
[4]
# Question 2:

## Part (i)

| Answer | Marks | Guidance |
|--------|-------|----------|
| A paired test would eliminate any differences between individual cattle | E1 | |

**[1]**

## Part (ii)

| Answer | Marks | Guidance |
|--------|-------|----------|
| Must assume: Normality of population … | B1 | |
| … of differences | B1 | |

**[2]**

## Part (iii)

| Answer | Marks | Guidance |
|--------|-------|----------|
| $H_0: \mu_D = 10$, $H_1: \mu_D < 10$ | B1 | Both. Accept alternatives e.g. $\mu_D > -10$ for $H_1$, or $\mu_A - \mu_B$ etc provided adequately defined. Hypotheses in words only must include "population". Do NOT allow "$\bar{X} = \ldots$" or similar unless clearly stated to be a population mean |
| Where $\mu_D$ is the (population) mean increase/difference in milk yield. MUST be PAIRED COMPARISON $t$ test | B1 | For adequate verbal definition. Allow absence of "population" if correct notation $\mu$ is used |
| Differences (increases) (after $-$ before) are: 4 9 6 13 1 8 6 7 9 12 | M1 | Allow "before $-$ after" if consistent with alternatives for hypotheses above |
| $\bar{x} = 7.5$, $s_{n-1} = 3.566(8)$ $(s_{n-1}^2 = 12.722(2))$ | A1 | Do not allow $s_n = 3.3837$ $(s_n^2 = 11.45)$ |
| Test statistic is $\dfrac{7.5 - 10}{\frac{3.5668}{\sqrt{10}}}$ | M1 | Allow candidate's $\bar{x}$ and/or $s_{n-1}$. Allow reversed numerator compared with 2.2164 |
| $= -2.2164$ | A1 | c.a.o. but ft from here in any case if wrong. Use of $10 - \bar{x}$ scores M1A0, but ft |
| Refer to $t_9$ | M1 | No ft from here if wrong |
| Single-tailed 5% point is $-1.833$ | A1 | Must be minus 1.833 unless absolute values are being compared. No ft from here if wrong. $P(t < -2.2164) = 0.0269$ |
| Significant | A1 | ft only candidate's test statistic |
| Sufficient evidence to suggest that the mean milk yield has not increased by 10 litres (per cow per week) | A1 | ft only candidate's test statistic. Conclusion in context to include "on average" o.e. Accept "Sufficient evidence to suggest that the **company's** claim is not justified." o.e. |

**[10]**

## Part (iv)

| Answer | Marks | Guidance |
|--------|-------|----------|
| CI is given by $7.5 \pm$ | M1 | ZERO/4 if not same distribution as test. Same wrong distribution scores maximum M1B0M1A0. Recovery to $t_9$ is OK. Allow candidate's $\bar{x}$ |
| $2.262$ | B1 | 2.262 seen |
| $\times \dfrac{3.5668}{\sqrt{10}}$ | M1 | Allow candidate's $s_{n-1}$ |
| $= 7.5 \pm 2.5514 = (4.948, 10.052)$ | A1 | c.a.o. Must be expressed as an interval |

**[4]**

---
2 A company supplying cattle feed to dairy farmers claims that its new brand of feed will increase average milk yields by 10 litres per cow per week. A farmer thinks the increase will be less than this and decides to carry out a statistical investigation using a paired $t$ test. A random sample of 10 dairy cows are given the new feed and then their milk yields are compared with their yields when on the old feed. The yields, in litres per week, for the 10 cows are as follows.

\begin{center}
\begin{tabular}{ | l | c | c | c | c | c | c | c | c | c | c | }
\hline
Cow & A & B & C & D & E & F & G & H & I & J \\
\hline
Old feed & 144 & 130 & 132 & 146 & 137 & 140 & 140 & 149 & 138 & 133 \\
\hline
New feed & 148 & 139 & 138 & 159 & 138 & 148 & 146 & 156 & 147 & 145 \\
\hline
\end{tabular}
\end{center}

(i) Why is it sensible to use a paired test?\\
(ii) State the condition necessary for a paired $t$ test.\\
(iii) Assuming the condition stated in part (ii) is met, carry out the test, using a significance level of $5 \%$, to see whether it appears that the company's claim is justified.\\
(iv) Find a 95\% confidence interval for the mean increase in the milk yield using the new feed.

\hfill \mbox{\textit{OCR MEI S3 2013 Q2 [17]}}