In a model with only contrast-based aggregate data, study names are incorrectly converted to integers when setting the resdev names in fnames_oi. This happens because the .study column is a factor, which is converted into an integer vector when concatenating with NULL. This PR proposes to convert to a character vector, which can be safely concatenated with NULL.
This causes the labels in DIC plots to appear as seemingly random integers, instead of the study names.
dmphillippo
commented
2 years ago
In a model with only contrast-based aggregate data, study names are incorrectly converted to integers when setting the
resdevnames infnames_oi. This happens because the.studycolumn is a factor, which is converted into an integer vector when concatenating withNULL. This PR proposes to convert to a character vector, which can be safely concatenated withNULL.This causes the labels in DIC plots to appear as seemingly random integers, instead of the study names.