Thanks! What you are pointing out is really the point of departure where I tried and failed. There are basically two modes of thought about such things in political science. The economics-inspired mode of thinking takes the revealed preference very seriously, and combined with the assumption that preferences obey Euclidean geometry, have built a big industry that seeks to “measure” what people think with precision. The other school, inspired by psychology (or, as some psychologists have described, pop-psychology) goes overboard in the opposite direction: by claiming that people lack stable preferences and can be easily manipulated.
This is perhaps a bit unfair, since a lot of classic works in political science and voting behavior, took a more nuanced view — that (at least some) people are not that unreasonable in how they think, but are hardly mathematical automatons. In a sense, that’s where a lot of my inspiration came from, but, this is also the school of thought that gets rather limited attention since many of their propositions are not neatly testable, whereas the assumption of people as mathematical automata yields neat measurements that say X is 0.67 in the liberal direction (people actually do this!) and most results that show seemingly “irrational” behavior can be used to prop up “people are irrational, period” faction. To be fair, more nuanced results are difficult, and people do not want more complexities than is “necessary.”