The method worked better than simple aggregation for validation words such as "fifty-fifty chance", "never" or "always" (indicated in the plot by the black intervals compared to the gray interval areas).

That's interesting - thanks for comparing the methods and for letting me know!

Analytical solutions are of course more elegant than numerical integration.

Still, it is worthwile to have a paper elaborating on this in detail!

Just a remark regarding "The same incorrect computation appears in implementations of latent-trait multinomial processing tree models":

The issue has been mentioned in the literature (shorturl.at/Ajx0W), and the R package TreeBUGS has the function probitInverse as a solution (shorturl.at/mRp1S).

Looks and reads great, I especially like the concise figure!

This is an important issue which is often overlooked, so the contribution will be useful for many modelers. 👍

It would be good to learn already in school that uncertainty is a fundamental aspect of science.

I guess people generally dislike being in a cognitive state of uncertainty. But I vaguely remember some studies showing that scientists are better at coping with it.