Florian Hartig

Hi Sara, sorry for the late reply. With this kind of questions, it's hard to give precise answers without interacting with the data, but I will share a few ideas:...

OK, your first 4 models essentially look like this (this is M1).  My concern with this figure is that, as you see, there are no data that reach to...

Reply from the user: > I tried your suggestions to improve the fit : > > 1) By transforming the response (with powerTransform function) and fit a linear mixed model...

Just for reference: this was the original beta model:  So, just to summarise: if I got everything correctly your results regarding Shannon are that - simple LMM suggest no...

Mea culpa, I have been wanting to write this paper for so long. Maybe this summer. I had some other people asking about it. About other references: I think there...

See discussion here https://twitter.com/jebyrnes/status/1309564548207325185

This is similar to https://github.com/florianhartig/DHARMa/issues/34 ... adding categorical or ordinal responses is on my mind, but currently not top priority, also because it would require a complete new class of...

It seems tidybayes has something that goes in this direction https://cran.r-project.org/web/packages/tidybayes/vignettes/tidybayes-residuals.html#randomized-quantile-residuals-for-a-discrete-distribution

Suggestion from Louise McMillan > I wondered whether you've heard of the "sure" R package, related to the > 2018 paper by Liu & Zhang which proposes a "surrogate residuals"...

Here a piece of code that calculates DHARMa residuals for a fitted psem object ``` library(piecewiseSEM) mod = psem( lm(rich ~ distance + elev + abiotic + age + hetero...