I read the coverage of this academic paper with some amusement. Something not mentioned by the reviewers is that the data set came from just one college. Should that then be applicable everywhere? I don’t know.
Also, the idea of correcting for brightness of students strikes me as misguided. Smart students know to apply themselves to majors that will pay off. Also, the concept that high-paying careers require more hours is also misguided. Most of the graduates in lower earning professions can’t work more hours, even if they want to; there is no demand for that.
The paper was written to deal with how one statistically deals with non-responses in surveying. That it deals with education is a happy accident. I would be careful generalizing from this paper. To me, it is another example of advanced research that is highly intelligent, but may lack common sense.