Error in cond.var + Z %*% D %*% t(Z) : non-conformable arrays

[Sannemalee]

Hi @aloy,

I am trying to use lmeresampler on my multilevel model (package nlme). I downloaded the latest version of the package via remotes::install_github("aloy/lmeresampler"). However, I get error messages. When I use na.action=omit, I get:
Error in [<-.data.frame(*tmp*, , as.character(mod.fixd[[2]]), value = c(2.77420893174699, :
replacement has 4079 rows, data has 4615
When I use na.action=exclude, I get Error in cond.var + Z %% D %% t(Z) : non-conformable arrays.

This is my model:
NA_SimpleMorningNAprediction <- lme(MorningNegA ~ EveningNegA_within + day_cent, random=list(PP =pdNatural(~1 + day_cent)), data=Fulldata, na.action=na.omit, control=list(opt="optim"), method = "ML", correlation = corCAR1(form = ~ day_cent|PP))
summary(NA_SimpleMorningNAprediction)

And this is the code I use for bootstrap:
boot_SimpleMorningNAprediction<-bootstrap(NA_SimpleMorningNAprediction, .f=fixef, type="parametric", B=1000)

When I use a simpler model like these below, I get the same error messages.
NA_SimpleMorningNAprediction <- lme(MorningNegA ~ EveningNegA_within, random=~1|PP, data=Fulldata, na.action=na.exclude, method = "ML")
Error in cond.var + Z %% D %% t(Z) : non-conformable arrays

NA_SimpleMorningNAprediction <- lme(MorningNegA ~ EveningNegA_within, random=~1|PP, data=Fulldata, na.action=na.omit, method = "ML")
Error in [<-.data.frame(*tmp*, , as.character(mod.fixd[[2]]), value = c(2.77420893174699, :
replacement has 4079 rows, data has 4615

Do you know what the problem is? Is this a bug?
Unfortunately, I cannot give you the real data (due to regulations).

Best wishes,
Sanne

[aloy]

Hi @Sannemalee ,

I definitely want to get to the bottom of this. Could you use a data set similar to the one you have to share a reproducible example? I know that you can't share your data, but I am wondering if this happens with other data sets you have seen as well.

[Sannemalee]

Hi @aloy ,

I've found an ESM dataset online (here: https://osf.io/t9gzu/) that I used to test. I receive the same issue. I converted it to R.
Testdata.csv
This is the script I used now to test it.

Test <- lme(AffPos ~ Connected, random=~1|ID, data=Testdata, na.action=na.exclude, method = "ML")
summary(Test)
Test <- lme(AffPos ~ Connected, random=~1|ID, data=Testdata, na.action=na.omit, method = "ML")
summary(Test)

Hope this helps!

[Sannemalee]

Dear @aloy ,

I was wondering whether you have found the time to dig into this issue.

Best wishes,
Sanne

[aloy]

Hi @ Sannemalee,

This fell through the cracks a bit. I will get a fix for this up soon and ping you when it's there!