No, it shouldn't tell us everything, but if someone makes a very data-oriented claim ("millenials will be the first generation poorer than their parents") and you return with data that shows the opposite, you can make a claim that the data is poorly gathered, etc.

But pivoting to the "but it's not my lived experience, bro" is weak.

If you made the claim that "millenials have it harder than their parent" then we're talking something where experience can be more useful.

Yes, but I guess I'd say that we should not attempt to capture an exact sentiment in making policy decisions. The bigger the decision, the more people are involved and affected, and the more people, the greater the variation in their sentiments. It simply doesn't make sense to try to somehow please each individual to address something like housing affordability in the US (or California, or Los Angeles, or even Monterey). The only way to do that is to aggregate sentiments into data. In doing so you lose precision about those sentiments, but that's good, because some of that precision is measuring idiosyncratic stuff that shouldn't play a role in solving the problem in question.