FS2.3.3 Testing the hypothesis that a correlation is zero using either…
Section
FS2.3 Correlation
Level
AS & A level
Statement
Testing the hypothesis that a correlation is zero using either Spearman’s rank correlation or the product moment correlation coefficient.
Guidance
Hypotheses should be in terms of ρ or ρ_s and test a null hypothesis that ρ or ρ_s = 0. Use of tables for critical values of Spearman’s and product moment correlation coefficients. Students will be expected to know that the critical values for the product moment correlation coefficient require that the data comes from a population having a bivariate normal distribution. Formal verification of this condition is not required.
June 2025 · AS Paper 2 (Further Statistics 2) · Q3(c)3 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
For the hypothesis test in part (c) some didn’t give their hypotheses in terms of ρ but very few attempted a tw-tail test. Most used the correct critical value from the tables with only a small minority giving the Spearman’s critical value. The most frequent cause for losing marks here was over the conclusion. Many gave a correct comparison such as – 0.667 < – 0.3783 and went on to state that there was evidence of a negative correlation. Such a comparison was not required but if attempted it should be correct so those who stated – 0.667 < 0.3783 for example lost the final mark.
Key points from the mark scheme and examiner's report
H0: ρ = 0, H1: ρ < 0 (in terms of ρ)
one-tailed 5% pmcc cv for n = 20 is −0.3783 (not Spearman's)
significant: evidence of negative correlation between BMI and speed
any comparison written must be correct
June 2025 · Paper 4B (Further Statistics 2) · Q2(b)3 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
This question was on Spearman’s rank and testing for correlation and most candidates found this a very accessible question and were able to score most, if not all, of the marks. In part (b), candidates were usually successful in setting up correct null and alternative hypotheses. The most common errors were using r instead of ρ with some incorrectly deducing that agreement meant a two-tailed instead of a one-tailed test. Candidates usually found the correct critical value and made a correct comparison and those who did set up a two- tailed test were able to score this mark for a correct critical value under that test. Others lost the final mark for not mentioning the context in the conclusion.
Key points from the mark scheme and examiner's report
H₀: ρs = 0, H₁: ρs > 0 (one-tailed for agreement)
CV (n = 9, 5%) = 0.6
0.8 > 0.6, evidence the lecturers agree about the rank order
using r instead of ρ, two-tailed tests, no context
June 2024 · AS Paper 2 (Further Statistics 2) · Q27 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
Usually, the calculation of Spearman’s rank correlation coefficient is a good source of marks for all students. It was unfortunate here that some could not see how to get started and there were attempts to write down coordinates for each of the 8 points and this made the question unnecessarily difficult. Those who didn’t even try this could often obtain a mark for writing the hypotheses in terms of ρ or ρs and usually the correct critical value was written down from the tables. Those who did realise that they could simply order the ranks often did so correctly and were able to find the correct value for rs and usually make a correct conclusion for the test. A few students used their coordinates to find the product moment correlation coefficient which used up precious time and usually scored no marks.
Key points from the mark scheme and examiner's report
read the ranks from the scatter diagram (no need for coordinates): Σd² = 162, rs = −0.929
H0: ρs = 0, H1: ρs < 0
cv −0.8333
reject H0: evidence of negative rank correlation
comparisons like −0.929 > −0.833 lose the final mark
June 2024 · Paper 4B (Further Statistics 2) · Q2(b)3 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
In part (b), candidates were usually successful in setting up correct null and alternative hypotheses. The most common errors were using r instead of with some incorrectly deducing that agreement meant a two-tailed instead of a one-tailed test. Candidates usually found the correct critical value and made a correct comparison, although some candidates had reached a negative value in (a) which should have resulted in going back and realising that the
Key points from the mark scheme and examiner's report
using r not ρ, or a two-tailed test, were common errors
June 2023 · AS Paper 2 (Further Statistics 2) · Q1(b)4 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
than the ranks. In part (b) most gave correct hypotheses in terms of and the majority used the correct critical value of 0.7833. A correct statement about the significance of the test was almost always seen and many also interpreted this clearly mentioning “scores” and the “tasks”.
Key points from the mark scheme and examiner's report
H0: ρs = 0, H1: ρs > 0
Spearman cv 0.7833 at 1% for n = 9 (pmcc 0.7498 is B0)
reject H0: positive correlation between ranks of scores for tasks P and Q, mentioning scores and tasks
June 2022 · AS Paper 2 (Further Statistics 2) · Q1(b)2 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
Part (b)(i) was less successful with the majority of candidates stating a critical value rather than a critical region. A few candidates incorrectly used the PMCC table to find their critical value. In (b)(ii) candidates on the whole gave sufficient context in their conclusion stating there was not enough evidence to reject the null hypothesis, and hence not enough evidence of agreement in the film ranks.
Key points from the mark scheme and examiner's report
(i) critical region rs > 0.5636 (a region, not just a critical value
do not use the pmcc table)
(ii) not in the CR: insufficient evidence of agreement between their film ranks, in context
June 2022 · AS Paper 2 (Further Statistics 2) · Q1(c)4 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
Part (c) was again well attempted with many demonstrating understanding of how the additional piece of data would affect the Spearman’s rank correlation coefficient. On a few occasions, ∑ 2 d candidates incorrectly believed that since the value of would be unchanged, so would the value of the correlation coefficient, not realising the fact that n was not equal to 11 instead of 10. A correct critical value found in most cases as was a correct contextualised conclusion.
Key points from the mark scheme and examiner's report
the new film adds d = 0 so Σd² stays 84 but n = 11: rs = 0.618
new cv 0.5364
now sufficient evidence of agreement
some wrongly kept rs unchanged
Nov 2021 · Paper 4B (Further Statistics 2) · Q1(c)4 marks
Mark scheme (whole question; this part is marked with the orange bar)
Key points from the mark scheme and examiner's report
H₀: ρs = 0, H₁: ρs > 0
CV (n = 7, 5% one-tailed) = 0.7143
0.7106 is not in the critical region, insufficient evidence
conclusion in context of ranks
Oct 2020 · AS Paper 2 (Further Statistics 2) · Q13 marks
Mark scheme (whole question; this part is marked with the orange bar)
Key points from the mark scheme and examiner's report
H0: ρ = 0, H1: ρ > 0 (in terms of ρ)
one-tailed 5% cv for n = 22 is 0.3598, critical region r > 0.3598
0.3892 is significant so evidence supports the agent's belief, in context
June 2019 · AS Paper 2 (Further Statistics 2) · Q1(b)4 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
Part (b) was also well attempted with the majority of candidates being able to find the critical value and draw a correct conclusion. The main error that occurred was not writing the hypotheses in terms of Rho. would decrease although we allowed d or d2 decrease for D and E. The second part was to indicate that the ranks of D and E would be the same. The most common explanation given that did not have the required detail was “the judges are in more agreement therefore Spearman’s rank correlation coefficient would decrease”
Key points from the mark scheme and examiner's report
H0: ρs = 0, H1: ρs > 0 (hypotheses in terms of ρ, the commonest loss)
cv 0.8333 at 1%
reject H0: evidence the judges agree
June 2019 · Paper 4B (Further Statistics 2) · Q8(a)4 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
Part (a) was well answered with only a minority of candidates using the critical value for the pmcc rather than for Spearman’s. In part (b) it was pleasing to see that the majority of
Key points from the mark scheme and examiner's report
H₀: ρs = 0, H₁: ρs > 0
CV (n = 9, 5% one-tailed) = 0.6
0.85 is in the critical region, evidence of positive correlation
some used the pmcc critical value
June 2018 · AS Paper 2 (Further Statistics 2) · Q3(b)4 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
This question was designed to test understanding of how to deal with tied ranks when calculating the Spearman’s rank correlation coefficient. It was clear that this is one aspect of Part (b) was a standard hypothesis test for positive correlation in the ranks. Some students lost the mark by writing out their hypotheses in words, others by using r. Most went on to find the correct critical value from the tables and made a correct comparison with their calculated value from part (a). The conclusion of the hypothesis test needed to be in context
Key points from the mark scheme and examiner's report
H0: ρs = 0, H1: ρs > 0 (not in words, not r)
cv 0.8286
reject H0: positive rank correlation between high jump and long jump results, in context
June 2018 · AS Paper 2 (Further Statistics 2) · Q3(c)2 marks
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
This question was designed to test understanding of how to deal with tied ranks when calculating the Spearman’s rank correlation coefficient. It was clear that this is one aspect of Part (c) was generally well attempted with a correct critical value and a correct conclusion often seen. The condition required for this test to be valid was not well known and there were not many correct responses seen in part (d). Likewise in part (e), students struggled to show they understood that the answers to parts (c) and (d) indicate that the results in high jump and long jump do not fit a positive linear pattern.
Key points from the mark scheme and examiner's report
pmcc cv 0.7293 for n = 6 one-tailed
0.678 < 0.7293 so no evidence of positive correlation
June 2018 · AS Paper 2 (Further Statistics 2) · Q3(d)1 mark
Mark scheme (whole question; this part is marked with the orange bar)
Examiner's comment
This question was designed to test understanding of how to deal with tied ranks when calculating the Spearman’s rank correlation coefficient. It was clear that this is one aspect of not many correct responses seen in part (d). Likewise in part (e), students struggled to show they understood that the answers to parts (c) and (d) indicate that the results in high jump and long jump do not fit a positive linear pattern.
Key points from the mark scheme and examiner's report
data must come from a bivariate normal distribution
not well known
Find an Edexcel Further Maths tutor
Stuck on this? Book a trial lesson with our Further Maths tutor.