I think it's important to try to convert the reason to a consequentialist reason every time actually; it's just that one isn't done at that point, you have to step back and decide if the reason is enough. Like the murder example one needs to avoid dismissing reasons for being in the wrong format.

"I don't want to tell my boyfriend because he should already know" translates to: in the universe in which I tell my boyfriend he learns to rely on me to tell him these things a little more and his chance of doing this sort of thing without my asking decreases in the future. You then have to ask if this supposed effect is really true and if the negative consequence is strong enough, which depends on things like the chances that he'll eventually figure it out. But converting the reason gets you answering the right questions.

Sunk cost fallacy could be a sign that you don't trust your present judgement compared to when you made the original decision to put the resources in. The right question is to ask why you changed your mind so strongly that the degree isn't worth it even at significantly less additional cost. It is because of new information, new values, new rationality skills or just being in a bad mood right now.

An advantage is that you feel just as clever for coming up with the right questions whatever you decide, which out to make this a bit easy to motivate yourself to implement.