I can't access the underlying article, but generally researchers use some form of confirmatory factor analysis to evaluate whether the questions form a valid me

I can't access the underlying article, but generally researchers use some form of confirmatory factor analysis to evaluate whether the questions form a valid measure of the hypothetical construct (including whether there are any confounding variables). For instance, other questions could be included to distinguish unpopular supervisors from unpopular ones driven by the hypothesized bottom line mentality. Without looking at the actual study and complete questionnaire, it's hard to say whether the researcher did so correctly, but certainly the toolset exists to do so within the social sciences, making it a little unfair to blanket condemn the whole field.