Basically, here we are testing if all the p-axis have the same variability or eigenvalues. There are two tests to choose in a menu form such as the Bartlett and Mauchly's sphericity tests.
Only needs to input the multivariate data matrix and the significance level (default = 0.05). It outputs the sample-size, the p-variables, the observed statistic used to test any deviation from an expected sphericity and the probability that null Ho: is true.
I noticed a clear difference in results by this function and a similar test in SPSS (significant vs. non-significant findings). My data contains scores for 17 participants (rows) across three conditions (columns) as input into a one-way ANOVA for repeated measures. What is weird is that the degrees of freedom are completely different, being 2 in SPSS and 5 with this function, as is also the case with the other computed statistics. Which one is right?