Loo-adjusted R2 warning: In ypredloo - y : longer object length is not a multiple of shorter object length

[DominiqueMakowski]

I have the following warning when extracting loo adjusted r2. Any idea why? Thanks a lot 😃

Warning message:
In ypredloo - y :
  longer object length is not a multiple of shorter object length

The data is attached and the code is the following:

library(rstanarm)

df <- read.table("df.txt")
fit <- stan_lmer(y ~ group / poly(x, 2) +
                   (1|r1) + (1|r2),
                 data=df, seed=6, QR=TRUE)
sjstats::r2(fit, loo=TRUE)

[strengejacke]

Should be fixed, though I'm not sure why there's a difference between the lengh of y (values from the response) and the length of ypredloo (the predicted values for the response). Maybe it's not possible to predict all values for the response properly in your model?

[strengejacke]

ncol(posterior_linpred(fit))
#> [1] 3963