I've written at least one other answer on coherentism and correspondence which I'll incorporate as reference, but this question is somewhat different.
The relative attractiveness of coherence and correspondence accounts of truth is mostly going to hinge on the answer to the following question:
Do we have unmediated access to the "out there"?
If your answer is essentially, yes, then correspondence theories of truth are going to be very attractive. On such an account, we are sufficiently equipped to directly access the world that is out there and we find truth when the beliefs in our head correspond to what is out there.
Moreover, if your answer is a strong yes, then coherence is going to seem pretty incoherent as an idea of truth. In essence, this is because we can skip the games we play in our heads and just get "to the things themselves."
If your answer is essentially no or a dim yes, then coherence becomes more important as a way of defining truth. In other words, am I striking bedrock, when I look out into the world or am I just hitting my own concepts. If I think that everyone has a frame through which they look at the world and this frame colors their ability to perceive reality, then it makes a lot more sense to view "truth" as that which matches the frame rather than that which matches the out there.
To make things a bit more concrete, when I go to the zoo and look at a giraffe, what is happening?
- I experience photons striking my eyes in various wavelengths?
- I see a giraffe-1 (I possess an idea of the giraffe and match the shadow I see to it)
- I see a giraffe-2 (I see the giraffe and intuit the form of giraffeness from it)
- I see a giraffe-3 (I bring an idea of giraffe and am told this fits the bill by the label)
View 1 seems very compatible a correspondence account, because all we have are scientific measurements involved. You also don't a giraffe, which seems like a bit of a problem vis-a-vis our every day experience.
View 2 is Plato. Sure, we get giraffes. We also get great correspondence theory. We also have a large number of entities (Ideas) and questions about how we possess these.
View 3 is Aristotle. We're empirically getting the same forms. But how do we intuit these forms? We do get correspondence -- does the thing I see match the form from my previous experience.
View 4 gets us closer to coherence. This is a giraffe because we've agreed to giraffe as the name for this sort of thing and truth is to correctly call things giraffes that everyone else does (coincidentally, there's an interesting feature of the Japanese word for giraffe キリン -- it refers to both the common animal and a mythical beast).
In a weird way, view 1 about science often jumps back to view 4 in many other domains.
Maybe to just try and get at it another way, if you think we bring categories to things, then it's a lot harder to maintain this optimism. Thus, people often present Hegel as a coherentist (though I would argue he's a complicated sort of realist).
tl;dr - correspondence theories require an optimism about our access to the world as it is; coherence theory does not. Usually if people go for coherence, it's because they have an epistemology or theory of mind where we don't get access to the thing themselves.
