Unbalanced two- and three-way repeated-measures ANOVA results have not been verified.
Example: more subjects in Group 1 than in Group 2.
Please interpret results for these designs with caution, and recognize that they might not be valid.
There are a few key differences between the MATLAB and Python versions of spm1d.
The main version of spm1d is the Python version. The MATLAB version incorporates much but not all of the functionaly of the Python version.
There will be discrepancies between the Python and MATLAB versions for two main reasons:
Practically the differences are likely to be small, but users should be aware that non-sphericity corrections, by definition, drive the critical threshold higher.
MATLAB results which just reach the critical threshold must therefore be interpreted with caution.
The second issue will result in slightly underestimated cluster-level p values.
Inside the ./spm1d/examples folder you will find Matlab scripts which implement a variety of statistical tests.
These examples correspond to the Python scripts of the same name, which are described in: the main documentation.
Statistical theory underlying the examples is documented here: Theory
We do not provide detailed documentation for the MATLAB version of spm1d.
If you require support consider the following options: