| Exam Board | Edexcel |
|---|---|
| Module | S3 (Statistics 3) |
| Year | 2012 |
| Session | June |
| Marks | 12 |
| Paper | Download PDF ↗ |
| Mark scheme | Download PDF ↗ |
| Topic | Hypothesis test of Spearman’s rank correlation coefficien |
| Type | Hypothesis test for association |
| Difficulty | Standard +0.3 This is a straightforward application of Spearman's rank correlation with standard hypothesis testing. Part (a) requires ranking data and applying a formula (routine calculation), part (b) is a standard hypothesis test with critical value lookup, and part (c) tests conceptual understanding of tied ranks. The question is slightly easier than average because it's a direct application of a standard procedure with no novel problem-solving required, though the tied ranks consideration in part (c) adds minor complexity. |
| Spec | 5.08e Spearman rank correlation5.08f Hypothesis test: Spearman rank |
| Candidate | \(A\) | \(B\) | \(C\) | \(D\) | \(E\) | \(F\) | \(G\) | \(H\) |
| Manager \(X\) | 62 | 56 | 87 | 54 | 65 | 15 | 12 | 10 |
| Manager \(Y\) | 54 | 47 | 71 | 50 | 49 | 25 | 30 | 44 |
| Answer | Marks | Guidance |
|---|---|---|
| Answer/Working | Marks | Guidance |
| Attempt to rank score \(X\) and score \(Y\) | M1 | For an attempt to rank both variables |
| Attempting \(d^2\) for their ranks, must be using ranks | M1 | |
| \(\sum d^2 = 18\) | A1 | For sum of 18 |
| \(r_s = 1 - \frac{6 \times 18}{8(64-1)} = 0.7857...\) | M1A1 | awrt 0.786 |
| Answer | Marks | Guidance |
|---|---|---|
| Answer/Working | Marks | Guidance |
| \(H_0 : \rho = 0\) | B1 | Null hypothesis in terms of \(\rho\) or \(\rho_s\) |
| \(H_1 : \rho > 0\) | B1 | Alt hypothesis as given |
| Critical region \(r_s > 0.6429\) | B1 | cv of \(+0.6429\) (or 0.7381 if two-tailed from hyp) |
| \((0.7857 > 0.6429)\) sufficient evidence to reject \(H_0\) | M1 | Correct statement relating \(r_s\) with cv, cv must be such that \(\ |
| There is evidence of agreement between the scores awarded by each manager | A1ft | Correct contextualised comment. Must mention "scores/rankings" and "manager". Use of "association" is A0 |
| Answer | Marks | Guidance |
|---|---|---|
| Answer/Working | Marks | Guidance |
| (\(A\) and \(D\) are now) tied ranks (for Manager \(Y\)) | B1 | Tied ranks implied by 2.5, 6.5 or both 2 or 6 or description |
| Average rank awarded to \(A\) and \(D\), and use \(r_s = \frac{S_{xy}}{\sqrt{S_{xx}S_{yy}}}\) | B1 | Average rank implied by 2.5 or 6.5 or description and 'use of pmcc' |
# Question 1:
## Part (a)
| Answer/Working | Marks | Guidance |
|---|---|---|
| Attempt to rank score $X$ and score $Y$ | M1 | For an attempt to rank both variables |
| Attempting $d^2$ for their ranks, must be using ranks | M1 | |
| $\sum d^2 = 18$ | A1 | For sum of 18 |
| $r_s = 1 - \frac{6 \times 18}{8(64-1)} = 0.7857...$ | M1A1 | awrt 0.786 |
**(5 marks)**
## Part (b)
| Answer/Working | Marks | Guidance |
|---|---|---|
| $H_0 : \rho = 0$ | B1 | Null hypothesis in terms of $\rho$ or $\rho_s$ |
| $H_1 : \rho > 0$ | B1 | Alt hypothesis as given |
| Critical region $r_s > 0.6429$ | B1 | cv of $+0.6429$ (or 0.7381 if two-tailed from hyp) |
| $(0.7857 > 0.6429)$ sufficient evidence to reject $H_0$ | M1 | Correct statement relating $r_s$ with cv, cv must be such that $\|cv\| < 1$ |
| There is evidence of agreement between the scores awarded by each manager | A1ft | Correct contextualised comment. Must mention "scores/rankings" and "manager". Use of "association" is A0 |
**(5 marks)**
## Part (c)
| Answer/Working | Marks | Guidance |
|---|---|---|
| ($A$ and $D$ are now) tied ranks (for Manager $Y$) | B1 | Tied ranks implied by 2.5, 6.5 or **both** 2 or 6 or description |
| Average rank awarded to $A$ and $D$, **and** use $r_s = \frac{S_{xy}}{\sqrt{S_{xx}S_{yy}}}$ | B1 | Average rank implied by 2.5 or 6.5 or description and 'use of pmcc' |
**(2 marks)**
---
\begin{enumerate}
\item Interviews for a job are carried out by two managers. Candidates are given a score by each manager and the results for a random sample of 8 candidates are shown in the table below.
\end{enumerate}
\begin{center}
\begin{tabular}{ | l | c | c | c | c | c | c | c | c | }
\hline
Candidate & $A$ & $B$ & $C$ & $D$ & $E$ & $F$ & $G$ & $H$ \\
\hline
Manager $X$ & 62 & 56 & 87 & 54 & 65 & 15 & 12 & 10 \\
\hline
Manager $Y$ & 54 & 47 & 71 & 50 & 49 & 25 & 30 & 44 \\
\hline
\end{tabular}
\end{center}
(a) Calculate Spearman's rank correlation coefficient for these data.\\
(b) Test, at the $5 \%$ level of significance, whether there is agreement between the rankings awarded by each manager. State your hypotheses clearly.
Manager $Y$ later discovered he had miscopied his score for candidate $D$ and it should be 54 .\\
(c) Without carrying out any further calculations, explain how you would calculate Spearman's rank correlation in this case.\\
\hfill \mbox{\textit{Edexcel S3 2012 Q1 [12]}}