I'd say the opposite: with a parties system you should count the popular vote because what people are doing is voting for a part; if people were voting for elec

I'd say the opposite: with a parties system you should count the popular vote because what people are doing is voting for a part; if people were voting for electors (which is a more adherent to fact possibility in other systems) then you should focus on data reflecting that - but it does not seem the case in this context. The «relevant conclusion that can be derived» from the popular vote is "more people voted for this part rather than the other, and how much". It seems indicative. That it may not reflect the "electoral vote" is actually a "problem" of Political Science.