Question 1
1 markA gym recorded the resting heart rates of 12 members who swim and of 12 different members who cycle.
Jack wants to calculate Spearman's rank correlation coefficient between the swimmers' heart rates and the cyclists' heart rates.
Explain why this is not appropriate.
Model answer
A correlation coefficient needs bivariate (paired) data, with two values for each person. These are two separate groups of different people, so a swimmer's heart rate is not paired with any particular cyclist's heart rate.
Explanation
Spearman's coefficient (like the PMCC) compares two variables measured on the same individuals. Two separate samples should be compared using averages and measures of spread instead. Saying the members do different sports is not enough: the key point is that the values are not paired.
Mark scheme
-
1 mark
B1 for the data are not paired (bivariate): each swimmer's heart rate is not linked to a particular cyclist's heart rate, as they are different people
oe, e.g. 'the two heart rates are not measured on the same people', 'there are no pairs of values'. Must show the idea that correlation needs paired data. Do not accept 'it will measure correlation', 'the samples are too small' or 'they are different sports' on its own.