Arkansas Tech University The Culture Wars & Political Polarization in Perspective

The party comparisons indicate that the placement of the Democratic and Republican parties has become more Pro-Choice since the 1980’s. This is a strange result for Republicans, considering the partisan polarization on abortion that was apparent in the Chapter 9 analysis. George H.W. Bush was viewed as more pro-life than Ronald Reagan, either due to the shifting abortion views of the public or perhaps given Reagan’s inconsistent record on abortion from his days as governor of California. Bob Dole was seen as more Pro-Choice than Reagan, though the difference is small (-0.083). The difference between George W. Bush and Reagan’s placement on the abortion scale is largest, likely owing to Bush’s self-identification as a religious conservative, a Christian-convert, and the prominence he gave faith and faith-based politics in his administration.

There was only one statistically significant difference in the relative partisan differences. The comparison of 1980 to 1992 suggests that respondents saw Clinton as further distant from George H.W. Bush on abortion when compared to Reagan and Jimmy Carter. In the total absolute distances, there was little evidence that average voters perceived the parties or the candidates much differently relative to 1980. Only in the 1980 to 1996 comparison was there a decrease in the total distance between the average respondent and the presidential candidates (0.152) with a t-test that met the standard for statistical significance (.0051). Overall, while it was apparent that views on abortion and the placements

TABLE 11.3: Trend Models for Perceived Distance Measures on Jobs

MODEL: Distance = Year	N	Intercept	Parameter Estimate		Standard Error	R2
JOBS – Respondent Placements
R Jobs Self Placement	12	-32.918	0.018	***	0.005	.673
Democratic Party Position	12	-14.121	0.009	*	0.005	.248
Democratic Prez Cand Position	9	-40.870	0.022	**	0.009	.460
Republican Party Position	12	-36.363	0.021	***	0.004	.736
Republican Prez Cand Position	9	-40.759	0.023	***	0.006	.676
JOBS - Respondent Relative Distance from Parties & Candidates
R – DP (Dem Party)	12	16.656	-0.008		0.006	.165
R – DPC (Dem Prez Cand)	9	34.029	-0.017		0.009	.320
R – DHC (Dem House Cand)	6	15.686	-0.007		0.010	.133
R – DSC (Dem Sen Cand)	-	-	-		-	-
R – RP (Rep Party)	12	43.206	-0.022	***	0.007	.488
R – RPC (Rep Prez Cand)	9	36.703	-0.019	*	0.009	.341
R- RHC (Rep House Cand)	6	44.081	-0.022	*	0.011	.497
R – RSC (Rep Sen Cand)	-	-	-		-	-
JOBS - Respondent Relative Partisan Proximity (Rep Distance – Dem Distance)
|R – RP| – |R – DP|	12	3.583	-0.002		0.005	.013
|R – RPC| – |R – DPC|	9	-17.929	0.009		0.007	.172
|R – RHC| – |R – DHC|	6	30.870	-0.016		0.011	.378
|R – RSC| – |R – DSC|	-	-	-		-	-
JOBS – Total Relative Partisan Distance (Rep Distance + Dem Distance)
|R – RP| + |R – DP|	12	41.590	-0.021		0.027	.159
|R – RPC| + |R – DPC|	9	31.635	-0.014		0.016	.103
|R – RHC| + |R – DHC|	6	7.525	-0.001		0.010	.004
|R – RSC| + |R – DSC|	-	-	-		-	-

* significant at .10 level

** significant at .05 level

***significant at .01 level

had changed, there was no compelling evidence that citizens were placing the parties or the candidates at positions on the abortion dimension polarized from their own positions on abortion.

The government role in jobs provision linear trend models are presented in Table 11.3. Immediately apparent should be the statistically significant shift in the average position of the public, both parties, and both candidates. However, all these trends are in the same direction. The public has shifted towards a position of less government responsibility in the provision of jobs, but their collective perception of the Democratic Party (0.009), the Democratic presidential candidates (0.022), the Republican Party (0.021), and the Republican presidential candidates (0.023) is that they’ve all shifted in that direction as well. However, while of the positional models evidence positive linear coefficients, the Democratic Party lags behind relative to the other parties and candidates and the public itself, and this model has the lowest goodness-of-fit for the linear trend (R² = .248). The model trend that tracks closest to that of the American public is the Republican Party trend and, this is also the best performing of the positional linear trends. The Republican party model explains almost 75% of the variation in the average placement of the party on the jobs scale (R² = .746). The models of the public positional trend (R² = .673) and the placement of the Republican presidential candidates (R² = .676) perform equally well, accounting for approximately 67% of the model variation relative to the mean position on the scale. Finally, while the Democratic presidential candidate model has a beta coefficient with a near-equal magnitude of that of the Republicans, the model doesn’t perform nearly as well, accounting for less than 50% of model variation (.460).

The relative partisan difference models are perfectly comprehensible given the positional trends identified above. While the Democratic models are all insignificant, their coefficients slope in the same negative direction indicating declining distance between the respondents and the Democrats. However, the Republican models are also negatively sloped and they are statistically significant. The relative perceived distances between the average public position and their placement of the Republican Party (-0.022), the Republican presidential candidates (-0.019), and the Republican House candidates (-0.022) have all declined. The Republican models explain between 40% and 50% of the variation in relative partisan distance.

Given that both parties and the party candidates are perceived to have become less supportive of government intervention in the jobs market along with the American public, the results from the relative and total absolute distance measures should be expected. The public does not appear to have grown closer to either party, given the absence of statistically significant linear trends in the relative absolute partisan distance measures. And, naturally, the American public does not believe the Republican and Democrats have polarized on jobs relative to their own average position on the scale, as is apparent from the statistically insignificant total absolute partisan difference models. Overall, while there was a sizable shift in public opinion on jobs and their placement of both parties and candidates on the issue, since that movement was all in the same direction there is no apparent perceived polarization on the government role in jobs provision. The fault line on jobs has shifted, but all the partisan political actors are perceived to have shifted right along with it.

At first blush, the total distance measures would appear to be evidence in favor of Fiorina’s conjecture that elites have polarized away from the masses. However, jumping to that conclusion is unwarranted. Recall that political polarization implies greater dispersion. The total distance measures are relative to the average (mean) respondent. Hence these measures indicate that the parties have polarized relative to the average member of the mass public. The culture wars thesis (political polarization) predicts that the Republican and Democratic Parties will move towards the poles of the distribution relative to the center point. So the question isn’t whether the parties have moved away from the center on any particular issues. The question is whether the Parties have diverged independent of the voters.

For that, we need to take a closer look at the mass public. One possible explanation for the greater perceived distance between the average member of the mass public and the political parties is that they have diverged relative to the average voter but in response to a concomitant shift towards the poles of the distribution by partisan identifiers. In other words, rather than elites shifting to the poles despite the mass public, the party elites may be shifting in response to the segment of the mass public that they most count on for votes and funds. If this is the case, then not only would this divergence be

TABLE 11.4: Total Partisan Distance Measures for Party Identifiers

MODEL: DV (dist) = B0 + B1(year) + e		N	Intercept	Parameter Estimate		Standard Error	R2
IDEOLOGY
R	|R – RP| + |R – DP|	16	-51.846	0.027	***	0.006	.637
R	|R – RPC| + |R – DPC|	9	-49.457	0.026	**	0.011	.456
R	|R – RHC| + |R – DHC|	10	-39.174	0.021		0.012	.265
R	|R – RSC| + |R – DSC|	-	-	-		-	-
I	|R – RP| + |R – DP|	16	-23.670	0.013		0.009	.120
I	|R – RPC| + |R – DPC|	9	-3.248	0.002		0.013	.005
I	|R – RHC| + |R – DHC|	10	-73.347	0.038	*	0.017	.385
I	|R – RSC| + |R – DSC|	-	-	-		-	-
D	|R – RP| + |R – DP|	16	-32.192	0.017	**	0.006	.339
D	|R – RPC| + |R – DPC|	9	-13.326	0.008		0.010	.078
D	|R – RHC| + |R – DHC|	10	-54.203	0.028	***	0.009	.539
D	|R – RSC| + |R – DSC|	-	-	-		-	-
JOBS
R	|R – RP| + |R – DP|	12	39.905	-0.019		0.012	.201
R	|R – RPC| + |R – DPC|	9	-16.756	0.011		0.008	.229
R	|R – RHC| + |R – DHC|	6	26.306	-0.012		0.019	.094
R	|R – RSC| + |R – DSC|	-	-	-		-	-
I	|R – RP| + |R – DP|	12	28.827	-0.014		0.011	.137
I	|R – RPC| + |R – DPC|	9	64.571	-0.032	***	0.010	.584
I	|R – RHC| + |R – DHC|	6	101.035	-0.050		0.030	.412
I	|R – RSC| + |R – DSC|	-	-	-		-	-
D	|R – RP| + |R – DP|	12	-23.602	0.013	*	0.007	.228
D	|R – RPC| + |R – DPC|	9	2.245	-0.001		0.012	.001
D	|R – RHC| + |R – DHC|	6	-12.445	0.007		0.007	.178
D	|R – RSC| + |R – DSC|	-	-	-		-	-

* significant at .10 level

** significant at .05 level

***significant at .01 level

intelligible as an electoral strategy, but also it could directly contribute to the increase in distance reflected in the total distance measure. Because as the Democratic and Republican identifiers shifted towards the poles, they would perceive a greater distance between themselves and the other party irrespective of whether the party had changed its position on the issue dimension at all. Of course, if they did shift towards their own constituents, that would also contribute to a greater perceived distance between party identifiers for one party and the perceived distance between them and the other party.

** significant at .05 level

***significant at .01 level

One indicator that this might be the case is found in Table 11.4. Note that it is the Republican and Democrat identifiers that are driving the significant trend results on total perceived distance. For all the issue dimensions, independents perceive an increase in total distance on only one measure (Republican House candidates on the ideological dimension). And even there, the model barely achieves significance and has a relatively moderate goodness-of-fit (.385). The significant increases in total perceived partisan distance from elite actors are exclusively in the party identifier models: mass Republicans perceiving increasing distance between themselves and Democratic elite actors, and likewise for mass Democrats and Republican elites. Democrats perceive increasing total distance on jobs, and both Republican and Democratic identifiers see increasing total distance between themselves and the parties and candidates on the ideological dimension.

In Table 11.5, we see the same trend. Independent identifiers perceive a significant decline in the distance between themselves and the parties (-0.037) and the candidates (-0.035) on defense spending. Whereas the linear trend of increasing distance is in an oppositional model: Democratic identifiers perceiving increasing distance between themselves and House candidates on defense spending (0.041). All three of the models account for between 50% and 60% of the variance in total absolute distance. While the total distance measures suggest partisan polarization rather than alienation given the perceptions of Independent identifiers relative to the partisan identifiers, it isn’t direct evidence. It is possible, though unlikely, that partisan identifiers attribute the increasing total distance on these political dimensions equally to both political parties or even exclusively to their own party. We could have disaffection rather than party polarization.

In order to directly test partisan polarization, I examine the absolute distances between party identifiers and the parties and candidates individually. In Table 11.6 I report on the absolute distances between party identifiers and the Republican and Democrat parties. Note, if the partisan constituent responsiveness hypothesis is correct, we would expect Republican identifiers to perceive little to no increase in the distance between themselves and the Republican Party while, at the same time, we would expect them to perceive a substantively large increase in the distance between themselves and the Democratic Party (likewise for Democratic party identifiers vis-à-vis the two parties). The results in Table 11.8 are consistent with this prediction across all of the issue dimensions and for most of the individual measures. The results for the ideological dimension are particularly noteworthy. Republican and Democratic identifiers perceive either no trend or a declining trend in the distance between

** significant at .05 level

***significant at .01 level

themselves and their own parties. Whereas both Republican and Democratic identifiers reported significant and substantively large increases in the distance between themselves and the opposite party over the time series. This table in particular demonstrates that the polarization on ideology, accurately perceived by the mass public and partisan identifiers, is a rational response to the issue positions of their respective constituents. The Republicans perceive a declining distance between themselves and Republican elites (-0.006), and mass Democrats identify a similar negative trend in the distance between themselves and Democratic elites, though it is statistically insignificant (-0.006). Furthermore, by a magnitude between two and three times, the explanatory power of the oppositional trend models outperforms the constituent models. The linear trend in the distance between mass Democrats and the Republican party explains almost 68% of the model variation (R² = .675). While not as strong, the Republican oppositional model relative to the Democratic Party (R² = .432) outperforms the Republican constituent model (R² = .320) by a significant margin. For spending and defense spending, the only significant models are the constituent models. Both the model on government spending (-0.012) and the model on defense spending (-0.017) show declining distance between the Democratic party identifiers and their elites, explaining between 25% and 50% of the model variation. The jobs model has significant decreasing distance between both the Republican constituents and the Republican Party (-0.017)and the Democratic constituents and the Democratic Party (-0.015). Furthermore, Democratic identifiers perceived increasing distance between themselves and the Republican Party (0.028). All three of these models explained between 40% and 50% of the model variation.

As noted in Chapter 2, culture war skeptics argue that elites have polarized independent of and despite of mass moderation over the past few decades. And more recently, Fiorina expresses doubt that even partisan sorting can explain why elites have ideologically diverged from the mass electorate. The puzzle for the skeptics is thus why elites have polarized away from the mass electorate they are supposed to be responsive to. There is an obvious and easy answer to this apparent puzzle: the elites haven’t grown more distant from the mass electorate, at least not in terms of alienation from their constituents. Indeed, the opposite has occurred. Partisans at the mass level believe their own party elites have either remained constant or even grown closer to them since the 1970’s. As measured by the perceived Euclidian distance from the candidates, neither the parties nor the candidates have diverged from the mass public on several issue dimensions (e.g. jobs). And where they have diverged, the divergence of the parties is explained almost entirely by the perceptions of the oppositional party identifiers. Mass Republicans believe that the Democratic elites have polarized relative to their own positions on the issues and ideology, and the same goes for mass Democrats vis-à-vis Republican elites.

Conclusion

I examined the perceived distance on the political issues between the mass public and the parties and candidates. On jobs and ideology, however, there is a significant trend in the average respondent self-placement. Both trends are in a conservative direction, with the American public becoming more conservative and less in favor of government responsibility for job creation over the past three decades. On the ideological dimension, respondents report greater distance between themselves and the Democratic Party and its candidates, while the regression coefficients on the absolute distances between respondents and Republicans are negative, suggesting that the mass public has moved closer to the Republican Party on the ideological dimension. This is coupled by a perception among the mass public that the Republican Party has become more conservative over the last three decades, according to the significant coefficients reported in the first section of Table 11.4. Just as with abortion, the total distance measures indicate that the parties are more distant from the average respondent today than in the past. On jobs, government spending, and defense spending, there are no significant findings of increasing polarization between respondents and the parties or candidates. Indeed, there are negative coefficients for most of the total distance models reported on these issues, indicating a declining perceived distance between the respondents and the partisan elites. This is particularly so on the ‘government responsibility for job creation’ dimension, with respondents reporting that both the Republicans and the Democrats have moved closer to them on this dimension.

There is strong evidence that, to the degree that the parties have diverged from one another and from the center of the mass of the electorate it is in response to divergence on the same issue dimensions by their own party identifiers. Party identifiers perceive little to no distance between themselves and their parties on the issue dimensions on average, while they perceive a substantial increase in the distance between themselves and the opposite party over the course of the time series. Thus what perceived polarization between the mass electorate and the party elites I find is accounted for, not by an alienation of the elites from the voters, but rather by partisan polarization. It is a function of the partisan polarization that has occurred since the 1970’s. Republican identifiers have polarized on the issues and ideological dimension relative to Democratic identifiers. Interestingly, given this fact, their perceptions of the parties and candidates could have remained constant, and there still would have been an increase in the distance between partisans of one party and the elites of the opposite party, as that constant position would be further distant from the new and polarized position of the partisan identifiers. What I find is that Republicans at the mass level view the Democrats as increasingly distant from their own positions, and likewise for mass Democrats and Republican elites. Polarization, not alienation, has occurred. Both the mass and elites, the party in the electorate and the party in government, have polarized on the issue and ideological dimension. In the next chapter, I continue to assess the relationship between mass and elite polarization. I will use an objective measure of elite ideology in Chapter 12, rather than perceived ideological placement of the elite partisans employed in this chapter. Using the D-W Nominate scores for Congress, I attempt to assess the causal direction between mass and elite polarization as well as continue to assess the degree to which both have polarized on the ideological dimension.

CHAPTER 12: Who is Wagging Whom? Responsive Publics, Receptive Elites


“