ABSTRACT

A correlation coefficient describes the relationship between two variables. In addition, it can be used to find the statistical significance of a given relationship. It is necessary to produce a scattergram of the data for the two variables and to think about the nature of the sample being tested, otherwise there is a danger of missing a relationship because it is non-linear or suggesting a relationship which is actually an artefact of the sample.

The degree to which a test will produce the same score from one occasion to another – its reliability – and the degree to which judges agree in the way they use a scoring system – interrater reliability – can be ascertained using tests which are based on correlation. In addition, certain forms of validity of a measure can be checked by correlation.