I read the "methods" section of the study, and saw no reference to controlling for occupation nor socioeconomic status (as opposed to marital status, self-reported health, and other things that are listed).

However, I only read it briefly, so I may have missed something.