error from tab_model function

[JM-RG]

Hi,
sjPlot package is fascinating so first of all thank you for this amazing contribution to R users community.

I would like to create an lmer output table using the tab_model function. However I get this (btw not so informative) error: Error: Column 6 must be named

The model is in the format
mod <-lmer( y ~ (-1) + Sex + (1|ID) + (1|Year), data = df, REML = TRUE, verbose=TRUE, na.action="na.fail", control=lmerControl(optimizer="Nelder_Mead"))

I get all the common outputs from the model without problems. Also, using tidy_lmer function I can get a more simple output table of the model without problems.

I have inspecting some time the tab_model function inside, trying to find where the fix is comming but with no luck.

Do you have some idea where the problem could be?

I have no reproduccible example but the output of the model can be found here:

https://www.dropbox.com/s/29922nqycgjbst7/mod.Rdata?dl=1

and I can provide more information if needed.

Thank you so much in advance.

[strengejacke]

I think the (-1) on your formula is causing troubles. Maybe with a function from sjstats, like sjstats:: model_frame() or sjstats::pred_vars(). I'm currently traveling, but will look into this after I'm back.
Maybe you can check if the same issue occurs due the example model in lme4::lmer(), with (-1) in the formula?

[JM-RG]

Hi Daniel,

thanks for your quick reply. Surprisingly to me I´ve tried tab_model with the example models provided in lme4::lmer() and got the same error with both intercept structure sintax (with and without (-1)) .
I also checked out with my own model without (-1) and same problem pops up: Error: Column 6 must be named

I was initially thinking the problem may came from my data but looks like the problem comes from other aspect since it also happens with different data.

I hope you can provide help with this issue in the next days.

thank you again

[JM-RG]

Hi again,

I have tried the same approach in another computer with different R sessions and got the same error.
In case it is informative you can find my session info here:
https://www.dropbox.com/s/bwzyv4q8jjyfu8k/sessionInfo.txt?dl=0

In another computer with a different session (see sessionInfo document linked), the function worked out well with the example data from lme4, restarting the session and charging only regression-modeling related libraries lme4 and sjPlot:
https://www.dropbox.com/s/bjhwee4c3idai8g/sessionInfo_zz.txt?dl=0

However, runing my model especifying lme4::lmer give me this error:

Computing p-values via Wald-statistics approximation (treating t as Wald z).
Error in UseMethod("family") : 
  no applicable method for 'family' applied to an object of class "logical"

Whereas the class of the object model with mi initial session was lmerModLmerTest (from lmerTest), the class of the model launching this new error is lmerMod (from lme4)
So it may be related to interaction among packages or something similar? Just to trying to give you as much information as possible....

[strengejacke]

When I re-fit the model user lme4::lmer(), it works for me. The error lies in the broom-package:

broom::tidy(mod)
#> Error: Column 6 must be named

[strengejacke]

I opened an issue: tidymodels/broom#446

[strengejacke]

Can you confirm that tab_model() works when you use lme4 to fit the model?

[JM-RG]

Yes, it does work when I run the model using lme4 ("lmerMod" class) and only have loaded in the session the packages lme4 and sjPlot. However if I load other modelling packages such as lmerTest (i.e. I need to run models with it and thus the class of the model is "lmerModLmerTest") or broom, tab_model function doesn't work.

[strengejacke]

broom should be no problem if loaded, only lmerTest, I think. I'll close this issue now, we need to wait for the broom.mixed package to be released.

[strengejacke]

I added a workaround (i.e. an own tidier) for lmerModLmerTest-Objects, so this issue should be fixed in sjPlot now. Will commit tonight.