Predicted Values for glm

[tegedik]

I saw this issue and encountered an error for glm using easystats::model_dashboard.

df <- carData::Mroz
glmfit <- glm(lfp ~ k618 + wc + hc + inc, df, family=binomial(link="logit"))
easystats::model_dashboard(glmfit)

Now, this model returns an error for plot under predicted values tab:

Error in if (ncol(vcovmat) < ncol(mm)) {: argument is of length zero
Error in lapply(text_modelbased, function(i) {: object 'text_modelbased' not found
Error in is.ggplot(x): object 'all_plots' not found

And for tabular summary:

Error in eval(expr, envir, enclos): object 'text_modelbased' not found

I realized that this has to do with the outcome (lfp) being a factor. When I convert it into numeric:

df$lfp <- as.numeric(df$lfp) -1

The problem disappears:

I don't think this is a serious issue but I want to let you know.

[strengejacke]

Minimal reprex:

library(insight)
library(modelbased)

df <- carData::Mroz
model <- glm(lfp ~ k618 + wc + hc + inc, df, family = binomial(link = "logit"))

grid <- insight::get_datagrid(model, at = "k618", range = "grid", preserve_range = FALSE, verbose = FALSE)
modelbased::estimate_expectation(model, data = grid, verbose = FALSE)
#> Warning in `[<-.factor`(`*tmp*`, ri, value = c(0, 0, 0, 0, 0, 0, 0, 0, 0, :
#> invalid factor level, NA generated
#> Error in if (ncol(vcovmat) < ncol(mm)) {: argument is of length zero

^{Created on 2022-08-31 with reprex v2.0.2}

[strengejacke]

Thanks, should be fixed once we merge easystats/insight#627 - you need to update easystats, too.