ABSTRACT
The statistical tests described in this chapter are used to analyse ordinal data for one or more groups of cases, or data which do not meet the assumptions for using parametric tests. These tests determine whether the distribution of cases differs significantly from chance across the groups or from some theoretical distribution. For example, instead of treating coursework marks as a ratio or interval scale, we could treat these scores as ordinal in the sense that a mark of 6, for example, is simply higher than the marks below it and that the difference between, say, marks of 2 and 3 is not necessarily the same as the difference between 3 and 4. We could then use some of the tests discussed in this chapter to find out whether having individual tutorials results in higher coursework marks than having group tutorials or no tutorials. We could use other tests contained in this chapter to see whether there is a significant increase in coursework marks after attending tutorials (i.e. at post-test) compared to before having them (i.e. at pre-test) and whether any increase which occurs is maintained later on (i.e. at follow-up). These tests are less powerful than parametric tests in the sense that differences are less likely to be statistically significant. We can check this by comparing the significance levels of the tests in this chapter with the significance level of parametric tests carried out on the same data in previous chapters.
