I'm not sure usefulness has anything to do with the results of romantic priming. The simplest explanation seems to be that STEM careers are associated in our culture with unsexy attitudes and groups: aesthetically uninspiring, analytical as opposed to emotional, attractive to socially clumsy people, etc. If you've got romance on your mind, you're likely to find those associations at least mildly aversive without needing to go into cost/benefit analysis -- and indeed I predict that the results of a cost/benefit analysis, even limited to romantic opportunities, would be a lot less favorable to the humanities.
If you've got romance on your mind, you're likely to find those associations at least mildly aversive without needing to go into cost/benefit analysis --
This happens as a cost-benefit-like calculation anyway, somewhere in some part of the human brain not selection-pruned to be used for this at all.
After this, it becomes a struggle of definitions and lines drawn in the air to delimit one Thingspace cluster from another. Is it a cost-benefit analysis if it's run by the "Instinct" sub-processes in the brain? Is it a good cost-benefit analysis if...
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.