By David Gopoian and Ralph Whitehead, Jr.

One of the puzzles that faces Democratic electoral strategists is what to make of white working-class voters. Many Democrats believe that it is in the objective economic self-interest of white working-class voters to give a majority of their presidential votes to the Democratic ticket. In the last two elections, however, it does not appear that white working-class voters have done this--over even come close to doing so.

Now, for those who follow the discussion of the voting behavior of the white working class, a recent paper by Larry Bartels of Princeton University, "What's the Matter with What's the Matter with Kansas", offers good news and bad news. The good news: "Has the white working class abandoned the Democratic Party," asks Bartels. "No. White voters in the bottom third of the income distribution have actually become more reliably Democratic in presidential elections..."

The bad news: Bartels's definition of the white working class--white voters whose incomes put them in the lower-third of the household income distribution--is quite different from the prevailing definition of the white working class.

In a broad version of the prevailing definition, the white working class consists of white voters whose education has stopped short of a four-year college degree. (A narrower definition might focus on the voters within this segment who are ages 30 to 60, work in blue-collar and pink-collar and lower-white-collar jobs in the service sector and what's left of the manufacturing sector, and have household incomes that surround the median household income for the nation, $44,000.) In 2004, voters who fit the broader definition--noncollege whites-- favored George W. Bush by a margin of 23 points in the NEP national exit poll.

The Bartels paper doesn't include a demographic profile of the voters who fit his definition of the white working class. But if you turn to the NES, the same survey he uses for the data in his paper, you can determine some of the demographic features of the Bartels group. And they don't bear a strong resemblance to the demography of the prevailing definition. For one thing, the median household income for his group is $21,000, less than half as high as the median household income in the prevailing definition. And the income figure is lower largely because only one-third (35%) of the Bartels voters were actively doing paid work. Also, of those who were working nearly half were under the age of 30. This leaves only 19 percent of the voters in Bartels' group who were at least 30 years old and also actively on the job. The plurality of voters in Bartels' definition were retired (35%) or permanently disabled (8%). An additional 4% were unemployed or laid-off.

Bartels' paper does underscore a couple of important conditions: For many Americans, retirement means economic hardship. And a lot of workers who are under the age of 30 are likely to be slogging through the lower reaches of the earnings scale. More to the point, he is surely correct in implying a notable truth about these voters, the close-to-the-edge elders and the still-struggling 20-somethings: namely, it is in the objective economic self-interest of these voters to support the Democratic ticket -- and they actually do support it. But his findings don't necessarily solve the puzzle of what Democrats ought to make of "white working class" voters as the term is generally understood.

Indeed, in the same 2004 NES data set that Bartels used for his analysis, the white working class, using the white noncollege definition, comprised nearly half of the voting electorate, had median household incomes of $47,500, and provided Kerry with only 40% of their votes, compared to about 59% for all other voters in the NES sample.

The Democrats' troubles with white working-class voters are further substantiated by state-level exit polls. Kerry failed to gain a plurality of noncollege white voters in any of the 22 states where data for respondents' education levels were obtained. By contrast, Kerry did win a majority of votes from white college-educated voters in 8 of those states and topped 48% among these voters in 4 other states. For all of the comfort that Bartels' analysis provides regarding surprisingly strong Democratic performance among the least affluent white voters, there is still plenty of justified anxiety about how to reach the forgotten majority of white working-class voters.