I understand there are no p-values in ggeffect output. But I wonder if you could be more specific about how kable or pander tables could be used to run the two-tailed t-tests that I need?
It seems tsum.test might do the trick (using summary statistics as input rather than the full "population level data") but I'm unsure how to make this work in a way that allows me to replicate and repeat the test several times across different model specifications. Any input would be much appreciated.
In my reprex case above (2 binary variables) I would need to 3 statistical tests (see updated table in original post, where numbers now match the (fictions) numbers in my reprex).
thanks, Scott