So, just in the premise the study, by needing to throw away almost all the survey responses where everyone agrees and get the policies what they want, actually goes a far way to debunking the populist premise that "economic elites" and "the People" are opposing factions with inherently different political opinions.
Further, the study considered the group as "disagreeing" if the margin in survey answers was at least 10%. So if 90% of one group wanted something and 80% of the other, that was considered as them disagreeing.
When a follow-up study further filtered responses to questions where more that margin spans the 5% mark, i.e. most people in each group disagree, then the effect completely disappears and it's more or less a coin toss which way policy goes.