Support for Trump was strongly correlated with negative views of blacks/immigrants/Muslims and moderate/liberal views on economics. Support for Cruz was strongly correlated with conservative economic views. Trump generally did better with those who reported economic anxiety, Kasich and Rubio did worse with said group, and Cruz did about the same regardless of self-reported economic anxiety. Income itself wasn't especially important, though Trump did best with those making under $30k and Rubio/Kasich did best with those making >$50k (support for Cruz didn't change too much based on income).
My source here is Identity Crisis by John Sides et. al. One of their main points is that Trump won both the primary and general election in large part because of "racialized economic anxiety", which largely boils down to the attitude of: "non-whites are taking all the jobs and government benefits". Many also hold the view that "discrimination against whites is a bigger problem than discrimination against non-whites".
That's a poor way of putting it. If asked in a poll, Trump supporters would be more likely to say that Blacks are more criminal than Whites than supporters of other candidates, but that's just objectively true based on facts about arrest rates and things like that. Blacks as a group commit far more crimes per capita. It's like if you say that school shooters tend to be White, that doesn't mean you see Whites negatively, it just means that you recognize the objective fact that school shooters tend to be White. The way you word it is the vilification of Trump supporters that the MSM was dishonestly putting out there, and it's a big reason Trump won. I see you haven't learned your lesson.
The authors were looking at responses to opinion-based questions such as:
-"Do you attribute racial inequality to discrimination or to blacks' lack of effort"
-"Do you oppose interracial dating?"
-"Do you rate Muslims favorably or unfavorably?"
