2

I like the tab_model from sjPlot very much but I can't seem to get past a hurdle.

I am using a very large mixed model, millions of data points, and somehow getting my tab_model output is taking forever. I've tried changing the df.method to "Wald", getting rid of confidence intervals, etc. but working with a smaller model I am showing no speed increase.

Can someone point me to either the switches I should be using to speed this up, or to an alternative which gives me the comparably nice output but more quickly? Thanks in advance.

Mitch M
  • 21
  • 1
  • Hi! Please add to your question tthe code of how are you trying to run tab_model and, if possible, a small sample of data. This way it would be easier to suggest improvements or alternatives. – Juan Bosco Apr 20 '22 at 18:08
  • Can't share the data, but what I've created is a linear mixed model, via lmer, based on about 1,000,000 observations for 400,000 persons. So it takes some time to run. I assume the time drain in tab_model is that there are some things it does based on re-running the model, and I would like to tell it not to do that! – Mitch M Apr 21 '22 at 20:51
  • I see, it's quite a lot of data. Does this issue occur if you assign the output of your model to an object, and then use tab_model, so you don't have to re run it? And, in case you are using knitr or something similar, have you tried to cache the results of lmer? – Juan Bosco Apr 21 '22 at 21:23
  • I save mdl <-lmer(formula, data) if that's what you mean. And I've tried running it outside of markdown so it's not a cache issue. – Mitch M Apr 25 '22 at 01:43

0 Answers0