trend line for the poor to unweighted group polarization. It decline sharply from 1970 to 1980, and then stabilizes until a slight spike upward during the George W. Bush administration (Figure 10.5). But note this uptick doesn’t even reach the 1970’s level of percent contribution to group polarization for the poor. The two biggest categorical contributors in the unweighted scores are the working class on the one hand and the rich on the other. The peak for the rich was in 1988 (40.27%) while the peak for the poor was in 1970 (25.83%). It does appear that the rich contributed more in the 1990’s (local max = 40.27%) than they have in the 2000’s (local max = 19.72%). However, there is little by way of an overall trend in contribution to group polarization on partisanship since 1970 by the rich.
In Table 10.5 I report a multiple regression of the weighted and unweighted group polarization measure using survey year (time trend) and the party of the presidential administration as independent variables. As in previous analyses, the logic of including a presidential variable is to test whether the public polarizes in reaction to the party of the president, given their generally distinct policy positions and the consequence of having a president who shares or does not share your own partisanship in office.
Figure 10.5: Unweighted Mean Deviation Trends in Average Party Identification for Rich and Poor Citizens
Table 10.5: Weighted & Unweighted Mean Dev. Trend Regressions of Class GP on Party ID, 1970-2004
