So I was always told that (as beambot points out) the number of credits was based on lecture time, and moreover that you should expect to spend 2-3 times as much time outside of lectures as you did inside. So your generic 3-credit course (nominally three hours of lecture, actually more like 2 1/2) would require something like 9-12 hours total: 3 in lecture, 6-9 at home.

Now in my experience, this was much more accurate for humanities than for math/science classes, where the amount of time the class took depended strongly on one's ability and preparation. I remember in particular one math professor whose problems were tricky: if you had the right insight, the proof was usually very simple, but if you didn't have that insight, you'd struggle and struggle. (And then there were the classes for which I was way under-prepared, and on which I spent way more than 25 hours.)

Add on to that the fact the time spent in lecture (on which the number of credits depends) doesn't necessarily track the difficulty of the material, and you have a recipe for wide variation in the amount of time one spends for a certain amount of credits. (Labs are notorious for taking way more time than the number of credits.)

And then you get into how some classes are more efficient than others in terms of how much learning you get for the amount of time you spend, and this doesn't match up with the number of credits either...