Parametric assumptions are tests that need to be run to decide whether the data can be tested using a parametric or a non-parametric test. The parametric tests are the stronger of the two. However, the data must fulfil all of the assumptions if it is to be used. In correlations, Pearson’s R is the parametric test and Spearman’s Rho is the non-parametric test.
The four parametric assumptions tests for correlations are:
- Data must be interval (scale/ordinal)
- Data must be normally distributed
- Data must have homoscedasticity
- Data must be linear
We will learn more about these in this topic.