Arkansas Tech University The Culture Wars & Political Polarization in Perspective

categories is a decreasing trend in percent contribution for the poor (-0.225), accounting for about a 50% reduction in error over the mean model (R² = .495). This trend holds up even when we hold group size constant in the unweighted models (-0.208), though the model fit is not nearly as good (R² = .216). Presidential administration partisanship is also a significant predictor of the poor’s percent contribution (mean deviation), with that contribution increasing in Republican administrations (2.91). The contributions of the poor were much larger during the Reagan and George H.W. Bush administrations than they were during the Clinton administration. However, as we can see in the unweighted model, this result is spurious in the sense that it is an artifact of the group size for the poor in this administrations rather than, as it might appear, a polarizing reaction to the presidential partisanship. In the ‘poor’ model, presidential partisanship falls out as a significant predictor when we treat the groups as equal in size and assess merely categorical contribution to the group deviations on partisanship.

One consistent effect of presidential partisanship in both the weighted and unweighted models is its affect on the contribution to group polarization on partisanship for the lower middle class. In the weighted model the percent contribution of the lower middle class decreased by 9.416 percent for every survey year during Republican administrations relative to the Democratic administrations. And even ignoring the relative sizes of the class groups, there was still a significant negative coefficient for presidential partisanship (-5.792). The upper middle class has increasingly polarized on partisanship in the unweighted model (0.071), though the fit isn’t particularly good (R² = .191). Still, this could be a consequence of the southern white middle class shifting towards the Republican Party.

Also showing a significant coefficient for presidential partisanship is the working class model (6.971), even though the trend variable does not reach statistical significance. Furthermore, the positive coefficient indicates that the working class has contributed more to group polarization on partisanship

The story doesn’t get in any better for Franks when we turn our attention to the ideological dimension. Table 10.6 reports the group polarization for income categories on ideology. The group polarization for class on ideology exhibits little to no apparent linear trend in the raw, total score. The peak polarization for class on ideology is in 1980 (0.588). The lower polarization scores come mostly in the middle of the time series, with a low in 1992 (0.044). While the polarization in the 2000’s isn’t as large as that in the 1970’s, the classes were more polarized on ideology then than in the 1990’s. The polarization scores for the poor and working class groups are relatively stable with slight oscillations over the time period. However, there is an apparent uptick in the polarization for the ‘poor’ category during the George W. Bush administration. The peak contribution for the Poor was in 1980 (0.201), but the second largest contribution comes in 2004 (0.162).

Examining the regression models for ideology in Table 10.7, the trend among income groups indicates a slight decrease in ideological polarization over the past three decades. The only significant coefficient of polarization is for the “rich” category (-0.001), but for all of the categories the polarization coefficient is negative. These results give little support for the argument that the poor and working class citizens have become more conservative since the 1970’s. Rather than a lower income class of voters buying into the Republican message on the culture wars, the evidence on the ideological and partisan dimensions indicates that the political conflict between economic classes has been remarkably stable over the last thirty-some years.

Note that unlike with partisan polarization, many of the lower scores for ideology occur at points of time proximate to the identification of the culture wars and some of the more significant years for social issue polarization, as I have shown in previous analyses. For the “poor” category, once of its lowest polarization scores was in 1998 (0.02), a start-off point for much of the polarization we have seen on gay rights, abortion, and partisanship. The “working” class category peaked in terms of polarization in 1974 (0.184). The negative trend is particularly apparent in the “rich” category, with this category contributing almost nothing to the overall polarization score from 1984 on. These negative trends are apparent in the parameter estimates reported in Table 10.7. The goodness-of-fit statistic reflects the absence of a polarization trend for all of the category models and the income group polarization score.