Are signifcane asterisks reliable when using robust linear models (rlm) in combination with estimate plots (plot_model)? about sjplot HOT 3 OPEN

See discussion here:
https://stat.ethz.ch/pipermail/r-help/2006-July/108659.html

I'd say, as long as you're transparent about the approximation method (and therefore, you could try parameters::model_parameters() and look at the output, which should indicate the approximation method), it's ok to compute them. Compare to mixed models, where you have many different methods (wald, satterthwaite, kenward-rogers, ...), none of which one would say it's perfectly accurate - they're all approximations.

from sjplot.

Thanks a lot for your swift reply.
I was not aware of the possibility to use parameters::model_parameters(model)
It works perfectly and gives me indeed p-values.
Also the approximation method is mentioned: "p-values (two-tailed) computed using a Wald t-distribution
approximation."

I will mention this assuming that this is also what the plot_model() output displays.

Cheers!

from sjplot.

Yes, plot_model() uses the parameters package to compute the summary tables

from sjplot.