Granted that Meg McKean or someone else can (and probably will) clarify your puzzle much better than yours truly, allow me to give a response while I'm still possessed with this (hopefully temporary but
definitely) irrational urge to stick my head out answering posts.

The way I see it, the difference between rationalist views (such as realists or neo-institutionalists) and constructivist views is in the analytical
framework: the former is an economics based positivist approach while the latter is a sociology based interpretive (or post-modern, if you wish). The difference is (I'm not sure but going out on a limb) whether the utility function is quantifiable, not whether materialistic or not, or whether one deals with interests and the other preferences. After all, the bases of the utility function is pleasure and pain, which are not materialistic. Also spatial voting theory (which derives from social choice theory), for instance,conceive preferences as a (one-, two-, or
multiple) dimensional space, which means that rationalist theorizing includes preferences.

As such, the rivalry between rationalist and constructivist views is a choice of epistemology and not that rationalists can only explain material interest and constructivists non-material interests. I have not found a case that cannot be explained by both rationalists and constructivists: its just that they entail totally different explanations based on different notions of what can be regarded as data, material,or proof. Hence, never the twain....?

Nobuhiro Hiwatari