Answer

None—These are all measures of correlation. Kendall’s is a non-parametric measure of correlation that involves measuring the degree of correspondence in rank order between two variables. Spearman’s is also a non-parametric measure of correlation based on rank order. Both methods have the same statistical power and are based on the same underlying assumptions. Their magnitude is often different, however, because they are computed in different ways and measure different concepts. Kendall’s is based on the probability that the rank orders are different, whereas Spearman’s is based on the degree of variability in rank order of one variable accounted for by the other.

Spearman’s is equivalent to Pearson’s (product moment) correlation coefficient computed on rank rather than value. Pearson’s correlation coefficient, r, is the most familiar. It measures the degree of linear association between two variables and is a parametric statistic; that is, it requires some degree of assumption of a normal distribution when interpreting 95% confidence intervals and P values.

Cronbach’s is commonly used to measure the internal consistency of a measurement scale made up of several components, and is based on internal correlation. For example, overall patient satisfaction might be measured by adding together the response to four questions on satisfaction: satisfaction with the consultation, the information given, the treatment offered, and the politeness of the clinician. Cronbach’s would measure how well answers to these four questions correlate in order to measure a single underlying construct or idea. High values of (say over 0.8) imply high internal consistency, suggesting the questions are measuring the same idea and it makes sense to add them together in one summary score. Low degrees of (say below 0.5) imply the questions are measuring different ideas and should not be added together.

Cite this as: BMJ 2009;339:b2578