OCR MEI S1 2012 June — Question 6 18 marks

Exam BoardOCR MEI
ModuleS1 (Statistics 1)
Year2012
SessionJune
Marks18
PaperDownload PDF ↗
Mark schemeDownload PDF ↗
TopicData representation
TypeDraw histogram then estimate mean/standard deviation
DifficultyModerate -0.3 This is a standard grouped data question covering routine S1 techniques: histogram drawing with unequal class widths, midrange discussion, mean/SD estimation from grouped data, outlier detection using 1.5×IQR rule, and a straightforward proportion calculation. All parts follow textbook methods with no novel problem-solving required, making it slightly easier than average.
Spec2.02b Histogram: area represents frequency2.02f Measures of average and spread2.02g Calculate mean and standard deviation2.02h Recognize outliers

6 The engine sizes \(x \mathrm {~cm} ^ { 3 }\) of a sample of 80 cars are summarised in the table below.
Engine size \(x\)\(500 \leqslant x \leqslant 1000\)\(1000 < x \leqslant 1500\)\(1500 < x \leqslant 2000\)\(2000 < x \leqslant 3000\)\(3000 < x \leqslant 5000\)
Frequency72226187
  1. Draw a histogram to illustrate the distribution.
  2. A student claims that the midrange is \(2750 \mathrm {~cm} ^ { 3 }\). Discuss briefly whether he is likely to be correct.
  3. Calculate estimates of the mean and standard deviation of the engine sizes. Explain why your answers are only estimates.
  4. Hence investigate whether there are any outliers in the sample.
  5. A vehicle duty of \(\pounds 1000\) is proposed for all new cars with engine size greater than \(2000 \mathrm {~cm} ^ { 3 }\). Assuming that this sample of cars is representative of all new cars in Britain and that there are 2.5 million new cars registered in Britain each year, calculate an estimate of the total amount of money that this vehicle duty would raise in one year.
  6. Why in practice might your estimate in part (v) turn out to be too high?

Question 6:
Part (i):
AnswerMarks Guidance
AnswerMarks Guidance
Frequency density calculations: \(500\leq x\leq 1000\): fd = 0.014; \(1000 < x\leq 1500\): fd = 0.044; \(1500 < x\leq 2000\): fd = 0.052; \(2000 < x\leq 3000\): fd = 0.018; \(3000 < x\leq 5000\): fd = 0.0035M1 At least 4 fds correct. M1 can also be gained from freq per 1000: 14, 44, 52, 18, 3.5 (at least 4 correct)
All fds correctA1 Accept any suitable unit for fd, e.g. freq per 1000, BUT NOT FD per 1000. Allow fds correct to at least 3dp
Linear scales on both axes, label on vertical axisG1(L1) Label required on vert axis IN RELATION to first M1 mark. Accept f/w or f/cw. Ignore horizontal label
Width of bars drawn at 500, 1000 etc. NO GAPS ALLOWED. Linear scale.G1(W1) Must be drawn at 500, 1000 etc NOT 499.5 or 500.5 etc
Height of bars correct, dep on at least 3 heights correctG1(H1) Visual check only (within one square). If fds not given and one height wrong then max M1A0G1G1G0
[5]
Part (ii):
AnswerMarks Guidance
AnswerMarks Guidance
Do not know exact highest and lowest values so cannot tell what the midrange is. OR counterexample to show it may not be 2750. OR \((500+5000)/2 = 2750\) but very unlikely to be absolutely correct, probably close to true value — element of doubt needed.E1 Allow comment such as 'Highest value could be 5000 and lowest could be 500 therefore midrange could be 2750'. NO mark if incorrect calculation. Sight of 1750 AND 3000 (min and max of midrange) scores E1
[1]
Part (iii):
AnswerMarks Guidance
AnswerMarks Guidance
\(\text{Mean} = \frac{(750\times7)+(1250\times22)+(1750\times26)+(2500\times18)+(4000\times7)}{80}\)M1 For midpoints (at least 3 correct). No marks for mean or sd unless using midpoints
\(= \frac{151250}{80} = 1891\)A1 CAO. Accept correct answers for mean (1890 or 1891) and sd (850 or 846 or 845.5) from calculator
\(\Sigma x^2 f = (750^2\times7)+(1250^2\times22)+(1750^2\times26)+(2500^2\times18)+(4000^2\times7) = 342437500\)M1 For sum of at least 3 correct multiples \(fx^2\). Allow M1 for anything which rounds to 342400000
\(S_{xx} = 342437500 - \frac{151250^2}{80} = 56480469\)
\(s = \sqrt{\frac{56480469}{79}} = \sqrt{714943} = 846\)A1 Allow SC1 for RMSD 840.2 or 840 from calculator
Only an estimate since the data are grouped.E1 (indep) Or any mention of midpoints or 'don't have actual data' or 'data are not exact'
[5]
Part (iv):
AnswerMarks Guidance
AnswerMarks Guidance
\(\bar{x} - 2s = 1891 - (2\times846) = 199\); Allow 200M1 For either. FT any positive mean and their positive sd/rmsd. No marks unless using \(\bar{x}+2s\) or \(\bar{x}-2s\)
\(\bar{x} + 2s = 1891 + (2\times846) = 3583\); Allow 3580 or 3600A1 For both (FT). Do NOT penalise over specification here
So there are probably some outliersE1 Must include element of doubt. Dep on upper limit in range 3000–5000
[3]
Part (v):
AnswerMarks Guidance
AnswerMarks Guidance
Number of cars over \(2000\text{ cm}^3 = \frac{25}{80}\times 2.5\text{ million} = 781250\)M1 For \(\frac{25}{80}\times 2.5\) million or \(\frac{(18+7)}{80}\times 2.5\) million
Duty raised \(= 781250\times £1000 = £781\text{ million}\)M1 (indep) For something \(\times\) £1000 even if this is the first step
A1CAO. NB £781250000 is over specified so only 2/3
[3]
Part (vi):
AnswerMarks Guidance
AnswerMarks Guidance
Because the numbers of cars sold with engine size greater than \(2000\text{ cm}^3\) might be reduced due to the additional duty.E1 Allow any other reasonable suggestion. Condone 'sample may not be representative'. Allow 'sample is not of NEW cars'
[1]
# Question 6:

## Part (i):

| Answer | Marks | Guidance |
|--------|-------|----------|
| Frequency density calculations: $500\leq x\leq 1000$: fd = 0.014; $1000 < x\leq 1500$: fd = 0.044; $1500 < x\leq 2000$: fd = 0.052; $2000 < x\leq 3000$: fd = 0.018; $3000 < x\leq 5000$: fd = 0.0035 | M1 | At least 4 fds correct. M1 can also be gained from freq per 1000: 14, 44, 52, 18, 3.5 (at least 4 correct) |
| All fds correct | A1 | Accept any suitable unit for fd, e.g. freq per 1000, BUT NOT FD per 1000. Allow fds correct to at least 3dp |
| Linear scales on both axes, label on vertical axis | G1(L1) | Label required on vert axis IN RELATION to first M1 mark. Accept f/w or f/cw. Ignore horizontal label |
| Width of bars drawn at 500, 1000 etc. NO GAPS ALLOWED. Linear scale. | G1(W1) | Must be drawn at 500, 1000 etc NOT 499.5 or 500.5 etc |
| Height of bars correct, dep on at least 3 heights correct | G1(H1) | Visual check only (within one square). If fds not given and one height wrong then max M1A0G1G1G0 |

**[5]**

## Part (ii):

| Answer | Marks | Guidance |
|--------|-------|----------|
| Do not know exact highest and lowest values so cannot tell what the midrange is. OR counterexample to show it may not be 2750. OR $(500+5000)/2 = 2750$ but very unlikely to be absolutely correct, probably close to true value — element of doubt needed. | E1 | Allow comment such as 'Highest value could be 5000 and lowest could be 500 therefore midrange could be 2750'. NO mark if incorrect calculation. Sight of 1750 AND 3000 (min and max of midrange) scores E1 |

**[1]**

## Part (iii):

| Answer | Marks | Guidance |
|--------|-------|----------|
| $\text{Mean} = \frac{(750\times7)+(1250\times22)+(1750\times26)+(2500\times18)+(4000\times7)}{80}$ | M1 | For midpoints (at least 3 correct). No marks for mean or sd unless using midpoints |
| $= \frac{151250}{80} = 1891$ | A1 | CAO. Accept correct answers for mean (1890 or 1891) and sd (850 or 846 or 845.5) from calculator |
| $\Sigma x^2 f = (750^2\times7)+(1250^2\times22)+(1750^2\times26)+(2500^2\times18)+(4000^2\times7) = 342437500$ | M1 | For sum of at least 3 correct multiples $fx^2$. Allow M1 for anything which rounds to 342400000 |
| $S_{xx} = 342437500 - \frac{151250^2}{80} = 56480469$ | | |
| $s = \sqrt{\frac{56480469}{79}} = \sqrt{714943} = 846$ | A1 | Allow SC1 for RMSD 840.2 or 840 from calculator |
| Only an estimate since the data are grouped. | E1 (indep) | Or any mention of midpoints or 'don't have actual data' or 'data are not exact' |

**[5]**

## Part (iv):

| Answer | Marks | Guidance |
|--------|-------|----------|
| $\bar{x} - 2s = 1891 - (2\times846) = 199$; Allow 200 | M1 | For either. FT any positive mean and their positive sd/rmsd. No marks unless using $\bar{x}+2s$ or $\bar{x}-2s$ |
| $\bar{x} + 2s = 1891 + (2\times846) = 3583$; Allow 3580 or 3600 | A1 | For both (FT). Do NOT penalise over specification here |
| So there are probably some outliers | E1 | Must include element of doubt. Dep on upper limit in range 3000–5000 |

**[3]**

## Part (v):

| Answer | Marks | Guidance |
|--------|-------|----------|
| Number of cars over $2000\text{ cm}^3 = \frac{25}{80}\times 2.5\text{ million} = 781250$ | M1 | For $\frac{25}{80}\times 2.5$ million or $\frac{(18+7)}{80}\times 2.5$ million |
| Duty raised $= 781250\times £1000 = £781\text{ million}$ | M1 (indep) | For something $\times$ £1000 even if this is the first step |
| | A1 | CAO. NB £781250000 is over specified so only 2/3 |

**[3]**

## Part (vi):

| Answer | Marks | Guidance |
|--------|-------|----------|
| Because the numbers of cars sold with engine size greater than $2000\text{ cm}^3$ might be reduced due to the additional duty. | E1 | Allow any other reasonable suggestion. Condone 'sample may not be representative'. Allow 'sample is not of NEW cars' |

**[1]**

---
6 The engine sizes $x \mathrm {~cm} ^ { 3 }$ of a sample of 80 cars are summarised in the table below.

\begin{center}
\begin{tabular}{ | l | c | c | c | c | c | }
\hline
Engine size $x$ & $500 \leqslant x \leqslant 1000$ & $1000 < x \leqslant 1500$ & $1500 < x \leqslant 2000$ & $2000 < x \leqslant 3000$ & $3000 < x \leqslant 5000$ \\
\hline
Frequency & 7 & 22 & 26 & 18 & 7 \\
\hline
\end{tabular}
\end{center}

(i) Draw a histogram to illustrate the distribution.\\
(ii) A student claims that the midrange is $2750 \mathrm {~cm} ^ { 3 }$. Discuss briefly whether he is likely to be correct.\\
(iii) Calculate estimates of the mean and standard deviation of the engine sizes. Explain why your answers are only estimates.\\
(iv) Hence investigate whether there are any outliers in the sample.\\
(v) A vehicle duty of $\pounds 1000$ is proposed for all new cars with engine size greater than $2000 \mathrm {~cm} ^ { 3 }$. Assuming that this sample of cars is representative of all new cars in Britain and that there are 2.5 million new cars registered in Britain each year, calculate an estimate of the total amount of money that this vehicle duty would raise in one year.\\
(vi) Why in practice might your estimate in part (v) turn out to be too high?

\hfill \mbox{\textit{OCR MEI S1 2012 Q6 [18]}}