`createInputForDescriptives()` - identify items based on `group`, not by `label`

The argument `itemExpr` identifies the items of a scale by looking for the word `Indikator` in the label. BT labels usually don't have that. So scale items get the `type` `variable` instead of `item`.

My workaround was identifying every `scale` entry (without imputated variables/scales), saving the `group` names and then looking for entries that match the group name, but that have `variable` written in `type`; and then change that to `item`.

``` r
### Hotfix: change scale items from "variable" to "item"

# for data sets with pooled/imputated variables
group <- grep("_pooled", descriptives_sus[descriptives_sus$type == "scale",]$group, invert = TRUE, value = TRUE)
for(var in group){
descriptives_sus[descriptives_sus$group == var & descriptives_sus$type == "variable",]$type <- "item"
}

# for data sets without pooled/imputated variables
group <- descriptives_lfb_allg[descriptives_lfb_allg$type == "scale",]$group
for(var in group){
descriptives_lfb_allg[descriptives_lfb_allg$group == var & descriptives_lfb_allg$type == "variable",]$type <- "item"
}

```

Start by inspecting createInputForDescriptives() and how itemExpr uses labels containing "Indikator" to assign types. Reproduce the scale-item case described in the issue, then verify that entries sharing a scale's group are classified as type "item" rather than "variable", including pooled or imputed data.

Written by the indexing model from the issue text.