…ges are tools; by extension, programming paradigms are sets of tools. Figure out the problem first. Picking your tool is the result of figuring out the problem, not part of it. Get the theory done first. Work through the problem in the abstract. Think about data flow and what…
Well, actually it is. Nobody I know has the luxury of picking the tool that’s right for the job based on the technical considerations of a chosen solution. Or worse, the chosen solution itself has other constraints.
That is (among a lot of non-technical stuff) availability of people with knowledge of the platform/domain.
But even that’s not the end of it because FP vs OOP is not constrained by the language. So you choose Java (or PHP or whatnot). The method you choose is something that should give you a boost is efficiency, efficacy and so on. But plenty of people prefer the devil they know despite the potential benefits of even trying the other side even when arguments may call for it.
Of course it’s largely a psychological thing when you can’t really make the most of your knowledge.