Statistical assumptions

Always check for statistical assumptions before running any inferential statistical tests

Whenever conducting inferential statistics, researchers make sure that the statistical assumptions associated with the statistical tests are met. If the statistical assumptions are not met, then the inferences yielded from the statistical tests cannot be interpreted and generalized to the population. The validity of any reported p-value is based upon the meeting of statistical assumptions.  

Click on a button below to learn more about each statistical assumption. These statistical assumptions are embedded into the decision process of Research Engineer, so you will also encounter them when using the Statistics Engine.

The steps for testing statistical assumptions in SPSS

The statistical tests used to test statistical assumptions in SPSS

There are several different statistical assumptions: Independence of observations, normality, homogeneity of variance, normality of difference scores, sphericity, and the chi-square assumption. There are statistical tests used to assess statistical assumptions: Skewness and kurtosis, logarithmic transformations, Levene's Test of Equality of Variances, and Mauchly's Test.