OCR Further Statistics 2022 June — Question 9 10 marks

Exam BoardOCR
ModuleFurther Statistics (Further Statistics)
Year2022
SessionJune
Marks10
PaperDownload PDF ↗
Mark schemeDownload PDF ↗
TopicChi-squared goodness of fit
TypeChi-squared goodness of fit: Given ratios
DifficultyChallenging +1.2 Part (a) is a standard chi-squared goodness of fit test with given expected ratios—routine calculation requiring converting ratios to expected frequencies, computing test statistic, and comparing to critical value. Part (b) adds a layer requiring students to work backwards to find the minimum sample size for rejection, which involves algebraic manipulation and understanding of how the test statistic scales with sample size, elevating this above a purely procedural question.
Spec5.06b Fit prescribed distribution: chi-squared test

9 The head teacher of a school believes that, on average, pupil absences on the days Monday, Tuesday, Wednesday, Thursday and Friday are in the ratio \(3 : 2 : 2 : 2 : 3\). The head teacher takes a random sample of 120 pupil absences. The results are as follows.
Day of weekMondayTuesdayWednesdayThursdayFriday
Number of absences2816241636
  1. Test at the \(5 \%\) significance level whether these results are consistent with the head teacher's belief. A significance test at the \(5 \%\) level is also carried out on a second, independent, random sample of \(n\) pupil absences. All the numbers of absences are integers. The ratio of the numbers of absences for each day in this sample is identical to the ratio of the numbers of absences for each day in the original sample of size 120.
  2. Determine the smallest value of \(n\) for which the conclusion of this significance test is that the data are not consistent with the head teacher's belief.

Question 9:
AnswerMarks Guidance
9(a) H : data consistent with belief, H : not
0 1
Expected frequencies 30, 20, 20, 20, 30
X 2 values 0.1333, 0.8, 0.8, 0.8, 1.2

Total 3.73(333…) (= 51 65 )

< 9.488
Do not reject H . Insufficient evidence that
0
AnswerMarks
data inconsistent with head’s beliefB1
B1
M1
A1
A1
M1ft
A1ft
AnswerMarks
[7]1.1
1.1
3.4
1.1
1.1
1.1
AnswerMarks
2.2bOE, e.g. “follows the given ratio”
Stated or implied
At least two correct, or correct subs in formula
Fully correct
Compare with 9.488, or p = 0.443 > 0.05
Consistent first conclusion, FT on their X 2
Contextualised, not over-definite
AnswerMarks
Sufficient evidence that belief is correct: A0Can be implied by answer
SC: Some cells merged:
can get B2M1A0A0 M1A1
Needs CV 9.488 or 11.14 or 7.815
Withhold A1 if hypotheses reversed
Insuff evidence to reject H0, belief correct M1A1
AnswerMarks
(b)“Determine”, so needs proper working
3.73(333…) × n/120 > 9.488
( 7n/225 > 9.488)  n  305 (3 sf)
AnswerMarks
Must be multiple of 30, so least n is 330M1
A1
B1ft
AnswerMarks
[3]3.1b
1.1
AnswerMarks
3.2aInequality for n, using TS and CV
n > 305 or n  305 stated or implied, allow 306
AnswerMarks
Next multiple of 30 above their 3053.73 gives 305.2, 3.733… gives 304.97
e.g. answer of 305 or 306
Needs reason if rounded up by < 10 or to
multiple of 300
AnswerMarks
OR( 73 n0 − 31 n2 ) 2 ( 43 n0 − 21 n2 ) 2 ( 93 n0 − 31 n2 ) 2
+ 3  +
AnswerMarks Guidance
31 n2 21 n2 31 n2M1 Inequality for n, using 120, TS and CV, needs n
in both O and E terms
AnswerMarks
()Or T& I: n = 300 → 9.333 Any n in range
300 to 360, with x 2 in [9.33, 11.2]: M1
n = 330 (9.489) A2 (n = 305 → 9.489 A1)
= 1900 n + 3  1 n + 1 n  9 .4 8 8
150 100
AnswerMarks Guidance
( 7n/225 > 9.488)  n  305 (3 sf)( 7n/225 > 9.488)  n  305 (3 sf) A1
Must be multiple of 30, so least n is 330B1ft Next multiple of 30 above their 305, needs reason
Inequality for n, using 120, TS and CV, needs n
in both O and E terms
()
Or T& I: n = 300 → 9.333 Any n in range
300 to 360, with x 2 in [9.33, 11.2]: M1
n = 330 (9.489) A2 (n = 305 → 9.489 A1)
AnswerMarks
AH : no correlation between variables, H : positive correlation between variables [no context]
0 1B0
BH : no correlation between number of failures and temperature, H : correlation between number of failures and
0 1
temperature [not one-tailed: needs “positive” in H ]
AnswerMarks
1B0
CH : no evidence of correlation between runners’ positions, H : evidence of positive correlation between number of
0 1
AnswerMarks
failures and temperature [“evidence” does not belong in hypotheses; without “evidence of”2 this would be B1]B0
DH :  = 0, H :   0 [two errors: H wrong, no interpretation of ]
0 1 1B0
EH :  = 0, H :   0 [two errors: H wrong, no interpretation of ]
0 1 1B0
FH : r = 0, H : r > 0 [one error: no interpretation of . Allow use of r]
0 1B1
GH :  = 0, there is no correlation between number of failures and temperature; H :  > 0, there is positive correlation
0 1
AnswerMarks
[two different answers, either of which would score 1 but neither gains 2 and we don’t give 1+1 here]B1
HH :  = 0, H :  > 0, where  is the pmcc [neither “population” nor context]
0 1B1
IH :   0, H :  > 0, where  is the population pmcc
0 1
OR H :  = 0, H :  > 0, where  is the pmcc between temperature and number of failures
AnswerMarks
0 1B2
AReject H . There are more computer failures when temperatures are higher [too definite]
0M1A0
BReject H There is significant evidence of correlation [no context]
0.M1A0
CSufficient evidence to reject H . There are more computer failures when temperatures are higher
0
AnswerMarks
[“Evidence” used, albeit in wrong sentence, but BOD]M1A1
DWrong but validly obtained TS leading consistently to Do not reject H . There is insufficient evidence that there are
0
AnswerMarks
more computer failures when temperatures are higher [standard FT]M1A1
EWrong but validly obtained TS leading consistently to Do not reject H There is evidence that the number of computer
0.
failures is not correlated with temperature. [Non-rejection doesn’t give positive evidence that H is correct. When H is
0 0
AnswerMarks Guidance
not rejected, candidates should aim for a “double negative”]M1A0
F(from correct TS and CV) Do not reject H . (+ anything) [inconsistent]
0M0A0
GIf H , H the wrong way round, but calculations right: Reject H . (+ anything)
0 1 0M1A0
ResponseMark
A12 is very high so it does cast doubt [relevant comparison needs to be seen] B0B1
BThe probability is so low that 12 is worrying so it does cast doubt [unclear comparison] B0B1
C0.0653 is not equal to 0.12 so probably not valid [don’t allow “not equal”] B1
D0.0653 is not close to 0.12 B1
E12 is much higher than 6.53 so it is not valid [too definite] B1B0
FP(= 12) from B(100, 0.0653) = 0.0166 which is very small so it does cast doubt B0B1
GP(> 12) from B(100, 0.0653) = 0.00784 which is very small so it does cast doubt B0B1
HP( 12) from B(100, 0.0653) = 0.00784 which is very small so it does cast doubt B0B1
IP( 12) from B(100, 0.0653) = 0.0299 which is small so it casts doubt [0.035 from N(6.53, 6.1036)] B1B1
JExpected number is 1000.0653 = 6.53 which is very different from 12 so it does cast doubt B1B1
KWrong p, e.g.: 0.142 not far from 0.12 so it does not cast doubt (0.142 from Po( 6), 0.0267 from Po( 7) B1B1ft
LLikewise: Expected number is 14.2 which is not very close to 12 so it does cast doubt B1B1ft
MLikewise: P( 12) from B(100, 0.142) = 0.319 which is not too low so it does not cast doubt B1B1ft
PMT
Need to get in touch?
If you ever have any questions about OCR qualifications or services (including administration, logistics and teaching) please feel free to get in
touch with our customer support centre.
Call us on
01223 553998
Alternatively, you can email us on
support@ocr.org.uk
For more information visit
ocr.org.uk/qualifications/resource-finder
ocr.org.uk
Twitter/ocrexams
/ocrexams
/company/ocr
/ocrexams
OCR is part of Cambridge University Press & Assessment, a department of the University of Cambridge.
For staff training purposes and as part of our quality assurance programme your call may be recorded or monitored. © OCR
2022 Oxford Cambridge and RSA Examinations is a Company Limited by Guarantee. Registered in England. Registered office
The Triangle Building, Shaftesbury Road, Cambridge, CB2 8EA.
Registered company number 3484466. OCR is an exempt charity.
OCR operates academic and vocational qualifications regulated by Ofqual, Qualifications Wales and CCEA as listed in their
qualifications registers including A Levels, GCSEs, Cambridge Technicals and Cambridge Nationals.
OCR provides resources to help you deliver our qualifications. These resources do not represent any particular teaching method
we expect you to use. We update our resources regularly and aim to make sure content is accurate but please check the OCR
website so that you have the most up-to-date version. OCR cannot be held responsible for any errors or omissions in these
resources.
Though we make every effort to check our resources, there may be contradictions between published support and the
specification, so it is important that you always use information in the latest specification. We indicate any specification changes
within the document itself, change the version number and provide a summary of the changes. If you do notice a discrepancy
between the specification and a resource, please contact us.
Whether you already offer OCR qualifications, are new to OCR or are thinking about switching, you can request more
information using our Expression of Interest form.
Please get in touch if you want to discuss the accessibility of resources we offer to support you in delivering our qualifications.
Question 9:
9 | (a) | H : data consistent with belief, H : not
0 1
Expected frequencies 30, 20, 20, 20, 30
X 2 values 0.1333, 0.8, 0.8, 0.8, 1.2
Total 3.73(333…) (= 51 65 )
< 9.488
Do not reject H . Insufficient evidence that
0
data inconsistent with head’s belief | B1
B1
M1
A1
A1
M1ft
A1ft
[7] | 1.1
1.1
3.4
1.1
1.1
1.1
2.2b | OE, e.g. “follows the given ratio”
Stated or implied
At least two correct, or correct subs in formula
Fully correct
Compare with 9.488, or p = 0.443 > 0.05
Consistent first conclusion, FT on their X 2
Contextualised, not over-definite
Sufficient evidence that belief is correct: A0 | Can be implied by answer
SC: Some cells merged:
can get B2M1A0A0 M1A1
Needs CV 9.488 or 11.14 or 7.815
Withhold A1 if hypotheses reversed
Insuff evidence to reject H0, belief correct M1A1
(b) | “Determine”, so needs proper working
3.73(333…) × n/120 > 9.488
( 7n/225 > 9.488)  n  305 (3 sf)
Must be multiple of 30, so least n is 330 | M1
A1
B1ft
[3] | 3.1b
1.1
3.2a | Inequality for n, using TS and CV
n > 305 or n  305 stated or implied, allow 306
Next multiple of 30 above their 305 | 3.73 gives 305.2, 3.733… gives 304.97
e.g. answer of 305 or 306
Needs reason if rounded up by < 10 or to
multiple of 300
OR | ( 73 n0 − 31 n2 ) 2 ( 43 n0 − 21 n2 ) 2 ( 93 n0 − 31 n2 ) 2
+ 3  +
31 n2 21 n2 31 n2 | M1 | Inequality for n, using 120, TS and CV, needs n
in both O and E terms
() | Or T& I: n = 300 → 9.333 Any n in range
300 to 360, with x 2 in [9.33, 11.2]: M1
n = 330 (9.489) A2 (n = 305 → 9.489 A1)
= 1900 n + 3  1 n + 1 n  9 .4 8 8
150 100
( 7n/225 > 9.488)  n  305 (3 sf) | ( 7n/225 > 9.488)  n  305 (3 sf) | A1 | n > 305 or n  305 stated or implied, allow 306 | e.g. answer of 305 or 306
Must be multiple of 30, so least n is 330 | B1ft | Next multiple of 30 above their 305, needs reason | Or n/120 must be a multiple of 0.25
Inequality for n, using 120, TS and CV, needs n
in both O and E terms
()
Or T& I: n = 300 → 9.333 Any n in range
300 to 360, with x 2 in [9.33, 11.2]: M1
n = 330 (9.489) A2 (n = 305 → 9.489 A1)
A | H : no correlation between variables, H : positive correlation between variables [no context]
0 1 | B0
B | H : no correlation between number of failures and temperature, H : correlation between number of failures and
0 1
temperature [not one-tailed: needs “positive” in H ]
1 | B0
C | H : no evidence of correlation between runners’ positions, H : evidence of positive correlation between number of
0 1
failures and temperature [“evidence” does not belong in hypotheses; without “evidence of”2 this would be B1] | B0
D | H :  = 0, H :   0 [two errors: H wrong, no interpretation of ]
0 1 1 | B0
E | H :  = 0, H :   0 [two errors: H wrong, no interpretation of ]
0 1 1 | B0
F | H : r = 0, H : r > 0 [one error: no interpretation of . Allow use of r]
0 1 | B1
G | H :  = 0, there is no correlation between number of failures and temperature; H :  > 0, there is positive correlation
0 1
[two different answers, either of which would score 1 but neither gains 2 and we don’t give 1+1 here] | B1
H | H :  = 0, H :  > 0, where  is the pmcc [neither “population” nor context]
0 1 | B1
I | H :   0, H :  > 0, where  is the population pmcc
0 1
OR H :  = 0, H :  > 0, where  is the pmcc between temperature and number of failures
0 1 | B2
A | Reject H . There are more computer failures when temperatures are higher [too definite]
0 | M1A0
B | Reject H There is significant evidence of correlation [no context]
0. | M1A0
C | Sufficient evidence to reject H . There are more computer failures when temperatures are higher
0
[“Evidence” used, albeit in wrong sentence, but BOD] | M1A1
D | Wrong but validly obtained TS leading consistently to Do not reject H . There is insufficient evidence that there are
0
more computer failures when temperatures are higher [standard FT] | M1A1
E | Wrong but validly obtained TS leading consistently to Do not reject H There is evidence that the number of computer
0.
failures is not correlated with temperature. [Non-rejection doesn’t give positive evidence that H is correct. When H is
0 0
not rejected, candidates should aim for a “double negative”] | M1A0
F | (from correct TS and CV) Do not reject H . (+ anything) [inconsistent]
0 | M0A0
G | If H , H the wrong way round, but calculations right: Reject H . (+ anything)
0 1 0 | M1A0
Response | Mark
A | 12 is very high so it does cast doubt [relevant comparison needs to be seen] | B0B1
B | The probability is so low that 12 is worrying so it does cast doubt [unclear comparison] | B0B1
C | 0.0653 is not equal to 0.12 so probably not valid [don’t allow “not equal”] | B1
D | 0.0653 is not close to 0.12 | B1
E | 12 is much higher than 6.53 so it is not valid [too definite] | B1B0
F | P(= 12) from B(100, 0.0653) = 0.0166 which is very small so it does cast doubt | B0B1
G | P(> 12) from B(100, 0.0653) = 0.00784 which is very small so it does cast doubt | B0B1
H | P( 12) from B(100, 0.0653) = 0.00784 which is very small so it does cast doubt | B0B1
I | P( 12) from B(100, 0.0653) = 0.0299 which is small so it casts doubt [0.035 from N(6.53, 6.1036)] | B1B1
J | Expected number is 1000.0653 = 6.53 which is very different from 12 so it does cast doubt | B1B1
K | Wrong p, e.g.: 0.142 not far from 0.12 so it does not cast doubt (0.142 from Po( 6), 0.0267 from Po( 7) | B1B1ft
L | Likewise: Expected number is 14.2 which is not very close to 12 so it does cast doubt | B1B1ft
M | Likewise: P( 12) from B(100, 0.142) = 0.319 which is not too low so it does not cast doubt | B1B1ft
PMT
Need to get in touch?
If you ever have any questions about OCR qualifications or services (including administration, logistics and teaching) please feel free to get in
touch with our customer support centre.
Call us on
01223 553998
Alternatively, you can email us on
support@ocr.org.uk
For more information visit
ocr.org.uk/qualifications/resource-finder
ocr.org.uk
Twitter/ocrexams
/ocrexams
/company/ocr
/ocrexams
OCR is part of Cambridge University Press & Assessment, a department of the University of Cambridge.
For staff training purposes and as part of our quality assurance programme your call may be recorded or monitored. © OCR
2022 Oxford Cambridge and RSA Examinations is a Company Limited by Guarantee. Registered in England. Registered office
The Triangle Building, Shaftesbury Road, Cambridge, CB2 8EA.
Registered company number 3484466. OCR is an exempt charity.
OCR operates academic and vocational qualifications regulated by Ofqual, Qualifications Wales and CCEA as listed in their
qualifications registers including A Levels, GCSEs, Cambridge Technicals and Cambridge Nationals.
OCR provides resources to help you deliver our qualifications. These resources do not represent any particular teaching method
we expect you to use. We update our resources regularly and aim to make sure content is accurate but please check the OCR
website so that you have the most up-to-date version. OCR cannot be held responsible for any errors or omissions in these
resources.
Though we make every effort to check our resources, there may be contradictions between published support and the
specification, so it is important that you always use information in the latest specification. We indicate any specification changes
within the document itself, change the version number and provide a summary of the changes. If you do notice a discrepancy
between the specification and a resource, please contact us.
Whether you already offer OCR qualifications, are new to OCR or are thinking about switching, you can request more
information using our Expression of Interest form.
Please get in touch if you want to discuss the accessibility of resources we offer to support you in delivering our qualifications.
9 The head teacher of a school believes that, on average, pupil absences on the days Monday, Tuesday, Wednesday, Thursday and Friday are in the ratio $3 : 2 : 2 : 2 : 3$. The head teacher takes a random sample of 120 pupil absences. The results are as follows.

\begin{center}
\begin{tabular}{ | c | c | c | c | c | c | }
\hline
Day of week & Monday & Tuesday & Wednesday & Thursday & Friday \\
\hline
Number of absences & 28 & 16 & 24 & 16 & 36 \\
\hline
\end{tabular}
\end{center}
\begin{enumerate}[label=(\alph*)]
\item Test at the $5 \%$ significance level whether these results are consistent with the head teacher's belief.

A significance test at the $5 \%$ level is also carried out on a second, independent, random sample of $n$ pupil absences. All the numbers of absences are integers. The ratio of the numbers of absences for each day in this sample is identical to the ratio of the numbers of absences for each day in the original sample of size 120.
\item Determine the smallest value of $n$ for which the conclusion of this significance test is that the data are not consistent with the head teacher's belief.
\end{enumerate}

\hfill \mbox{\textit{OCR Further Statistics 2022 Q9 [10]}}