In 2012 a CNN/ORC poll asked "Which of these statements comes closest to your view?...Taxes on wealthy people should be kept low because they invest their money in the private sector and that helps the economy and creates jobs. Taxes on wealthy people should be kept high so the government can use their money for programs to help lower-income people." 36% favored "kept low," 56% "kept high," and 8% had no opinion. In 2011 the figures were 34% and 62%. The question was also asked in 1990, 1991, and 1992, when 18-20% said "kept low" and 69% said "kept high."
That is, there was a substantial move towards saying that taxes on wealthy people should be "kept low" between the early 1990s and the 2010s. My guess is that it resulted from Republican voters following their leaders. In 1990, a bill raising top tax rates was passed with support from George H. W. Bush and a substantial number of Republicans in Congress. After Bush lost his bid for re-election in 1992, a bill raising top tax rates further was passed over the unanimous opposition of Republicans in Congress. In 1994, the Republicans gained majorities in both houses of Congress for the first time in 40 years. Those events seem to have convinced Republican politicians that it's never a good idea to support increasing taxes, even on people with high incomes. It seems reasonable that when ordinary voters hear all of the people they trust advocating the same position, some of them will switch over to that position.
This has a connection to the larger issue of whether a moderate candidate might get support in the Republican primaries. David Brooks says yes, while Paul Krugman says no: "Trump, Cruz, and Carson with the support of roughly two-thirds of likely Republican primary voters . . . do we really imagine that any significant fraction . . . consists of moderate voters who just don’t realize what they would be getting?" I think Brooks is right, or at least could be right--to some extent, more moderate candidates would create more moderate voters.