[Feature request] Support output of test_predictions from ggeffects

[roaldarbol]

Describe the solution you'd like
I recently discovered the ggeffects package (by @strengejacke, so maybe you're the right person to address this), and still can't get my arms down - that's how useful it is. Especially test_predictions is a brilliant tool to make targeted tests of your predictions by the model. And essentially, I'd just love to get the output from test_predictions into a report sentence. It's great to get a long sentence about the model as a whole, but getting help to convey the comparisons that are targeting your specific question would be great!

How could we do it?
I'm not really aware of the inner workings of report, so I'll just leave it here.

[roaldarbol]

I guess the wider issue is how to best report marginal effects in writing - that's where this would be extremely useful.

[strengejacke]

Thanks for suggesting this feature. Maybe (at least) one thing we need to think about is how to qualify the "strength" of an effect/difference/contrast in terms of some kind of standardized effect size.

[roaldarbol]

That makes sense - and feel free to rename the issue if it would be better as a starting point on reporting marginal effects in general.

Are there any current standard ways to report marginal effects in research papers? Would you have a reference at hand I could look through, or a guide of sorts?