There was research on this after the election - voting by income level. Clinton won in the 0-50K income range while Trump won biggest in the 50K-100K range, and also, by a small margin, the 100K on up range.
Trump won among whites in all income classes - including the poor ones - with the exception of college-educated poor whites a majority of which did not vote for Trump.
Also, correlation studies show that a much greater predictor of whether someone would vote for Trump was racism and hostility toward other cultures, rather than economic insecurity.
Related to this, is the belief on the left that a substantial, election-determining number of Trump voters would have instead voted for Bernie Sanders had he been in the general election. There are insurmountable difficulties of such a what-if scenario - namely, Sanders wasn’t in the general election and it would be very difficult to predict how the voters would have reacted to the withering attacks on Sanders (recall McGovern or Dukakis). I live in work in the central Appalachian region, and I know of no Trump voter - including most of my well-paid federal government employee co-workers, who would have instead voted for Sanders.
Mainly, elections are won by turnout, not people “changing their minds”. Sanders would have improved turnout among the discouraged young left, but then again, the turnout of the right may have been higher too as they would be very motivated to “stop a socialist”.