dividing white folks equally into three income groups indicates that new deal support came largely from lower third/middle third, hence, success was to, a degree, based on some greater portion of middle third understanding its interests in relation to third below it...
longitudinal data indicates that upper third whites were more likely to identify/ vote republican both back then and now, same goes for lower third vis-a-vis democrats...
hypothetical white voter choice based solely on income/interest relationship with electorate is comprised of equally of upper/middle/lower third voters: all upper third votes republican, all lower third votes democrat, lower half of middle votes democrat, upper half of middle votes republican, takes only small change among those literally in middle of middle third to swing outcome one way of another...
in 'real life', white upper third appears to have always had higher turnout than lower third, but discrepancy is greater now than when first national elections studies were conducted at mid-20th century (btw: significant turnout drop-off among low income population has been unique to u.s. with its absence of labor-oriented party, weak welfare state)...
today, upper third whites comprise about 45% of white turnout, middle third about 33% and lower third about 22%, income/interest relationship is still best indicator of party identification and vote choice, therefore, republicans can come out ahead while receiving *minority* of middle third white votes...
democrats do *not* attempt to mobilize lower third, like republicans, their efforts are directed at higher income folks when it comes to mobilization, democratic 'reach out' to more affluent has quadrupled in last 25 years (i would mark change at point in early 80s when then-california congressman tony coehlo convinced party that it needed to go to corporate sector/wealthy donors with hat-in-hand asking what can we do in order to get your support/contributions), in process, democratic party abandoned it strongest by-income partisans because lower third wasn't going to be of much help in 'money race'...
declining union membership has been important factor, and smaller percentage of whites are union members than is case for minorities (current union organizing success that is occurring is largely among both male & female minorities and white women), while union vote was never monolithic for democrats as some people seem to think, about 60% of union households identified themselves as democrats fifty years ago and similar percentage does so today, while union households increased as percentage of electorate in 2004 to 20%, that remains well below almost 30% in early 60s, even the 25% that it was in early 80s...
moreover, while union household party identification has remained constant, union household vote choice for democratic party presidential candidates over republican candidates has declined from almost 20% fifty years to about 10% today, of importance, perhaps, is that republican 'reach out' to union houeholds is about three times greater than in the past... mh