Moral Conservatism and Voting

Abstract

This chapter studies the effects of moral and practical considerations on political sentiments. Both have important effects but here, the moral trumps the practical. The voters’ practical considerations are indicated by their feelings about the economy and the problems of their state of residence, and their moral considerations by their moral conservatism. This chapter asks: How do moral factors affect the linkages in the graphical models of practical voting presented in Chap. 4? Moral conservatism reduces the effects of a state’s blue-purple-red political classification and the voters’ warm economic feelings. It also directly influences the endogenous political variables (operational political ideology and party affiliation) that influence practical voting. The final graphical models depict the joint effects of moral conservatism and a state’s human developmental context as these directly and indirectly influence voting decisions. The total effects of state-level typological variables based on human development and income equality are much weaker than the total effects of such individual-level variables as moral conservatism, the endogenous political variables, and the exogenous social attributes. These findings suggest a disconnection between the needs of the states for improved health, education, and economic wellbeing and the national electoral process; moral politics worsens this disconnection.

Essentially, all models are wrong, but some are useful.

—George E. P. Box and Norman R. Draper (1987, p. 424)

In the aftermath [of the 1994 mid-term elections], President Clinton blamed angry white voters upset with the Democrats on “guns, God and gays.”

—Jonathan Alter (2013)

Notes

• 1. Bartels (2008, 65); Frank (2004, 8, 136, 196, 245).

• 2. Bartels (2008, 78).

• 3. Bartels (2008, 79).

• 4. Bartels (2008, 86–89).

• 5. Gries (2014, 5).

• 6. Gries (2014, 15–16).

• 7. Gries (2014, 49–76). Biblical literalism is strongly associated with Evangelicalism. From 2007 to 2014, the percentage of Evangelicals among Protestants increased from 51% to 55%. Evangelicals now compose 19% of the adult population in the U.S. (Cooperman et al. (2015).

• 8. Lakoff ([1996] 2002); Gries (2014, 39).

• 9. Many disadvantaged people do not vote or cannot vote. Consequently, there is a disconnection between political and economic elites, many of whom are well-paid, and the needs of people in the United States, many of whom are poor (McCarty et al. 2006; Fiorina and Abrams 2009). Many citizens are susceptible to influence from political leaders and adjust their ideological views to conform to those of the elites (Levendusky 2009). The targeting by political campaigns toward sentiments indicative of authoritarianism can persuade citizens to vote for candidates who promise simple solutions to nuanced problems (Hillygus and Shields 2008; Hetherington and Weiler 2009). Moral issues regarding same-sex marriage rights, women’s reproductive rights, and gun-use rights often trump such fundamental practical issues as inequality, healthcare, education, climate change, economic growth, and political voice (Schlozman et al. 2012). All of these factors contribute to the disconnection between the policies of the leadership class and the needs of the poor and middle-class.

• 10. Ordinal variables can be rescaled from zero to unity; see Burns et al. (2001). Smith (2008, 239–242) reviews this use of rescaling and also provides an example (2011, 82).

• 11. The odds of variable x can be interpreted in terms of a distance between two points. The numerator is the distance already travelled and the denominator is the distance not yet travelled. Let the total distance be eight miles. If the distance travelled is six miles then the odds of x is the distance travelled = 6 divided by the distance not yet travelled = 2, the odds therefore are 3. If the distanced travelled is half way to the destination, then the odds are 4/4 = 1. If the trip has just begun and the distance travelled is only 2 miles, then the odds are 2/6 = 0.33. Ratios of proportions also produce the odds: 6/8 = 0.75, 2/8 = 0.25, therefore the odds of x are 0.75/0.25 = 3. Similarly, 0.5/0.5 = 1, and (2/8)/(6/8) = 0.25/0.75 = 1/3 = 0.33.

• 12. This book views the SEMs as formalizations of the more exploratory recursive modeling that precede the SEM modeling. These formalizations readily provide goodness-of-fit measures and estimates of the direct, indirect, and total effects. This sequential procedure is roughly analogous to Simon ’s (1957, 99–144) mathematical formalizations of inventories of empirical findings. Ideally, the exploratory and SEM analyses should use separate data sets. Limitations of data and resources ruled out the application of this analytic procedure here. The skeptical reader is invited to test these models on other data sets, perhaps for the 2012 and 2016 elections.

• 13. Kline (2005, 133–149) explicates these and other goodness-of-fit indexes.

• 14. Weber ([1904–5] 1958). This interpretation is consistent with Etzioni’s (1988, 63–64, 93–94) theoretical paradigm.

• 15. The Supreme Court decision of June 26, 2013 invalidating the federal Defense of Marriage act (DOMA) that disparaged same-sex marriages recognized by the states was made in the context of the public’s rising support of such marriages. Desilver (June 26, 2013) of the Pew Research Center reports findings from a survey taken in May, 2013 that found for the first time more than half (51%) of Americans favored allowing gay men and lesbians to marry and two-thirds (67%) of Americans favored allowing legal agreements that would give the same rights as [heterosexual] marriages.

• 16. The Supreme Court’s split decision limiting abortion rights is consistent with the ambivalence of U.S. citizens. The Gallup poll finds that 26% of Americans say abortion should be legal under any circumstance and 20% say it should be illegal in all circumstance. The majority (52%) chooses something in between (Saad 2013; Vavreck 2015). Fifty-two percent of Texas voters think that abortion should be legal in all or most cases and 32% think that abortion should be illegal in all or most cases. Eight in ten voters want the Texas legislature to focus on education, jobs, and the economy instead of social issues like abortion. A majority of those surveyed hold that decisions about abortion are personal, and should be made by a woman in consultation with her family and physician (Greenberg Quinlan Rosner Research 2013).

• 17. The public is now less likely to want to control gun ownership (46%) than to protect the right of Americans to own guns (52%). However, there are strong partisan differences with Republican’s against gun control and Democrats for gun control. A majority of Democrats (60%) say that guns do more to put people’s safety at risk while only a third (35%) say guns do more to protect people from being victims of crime. Republicans show the opposite percentages, about 20% to about 80%. (Pew Research Center, December 10, 2014). Gun ownership is more prevalent in the South and Midwest and less prevalent in the Northeast and West. Among gun owners, an increasing percentage from August 1999 to February 2013 say they own a gun for protection, +22%, and a decreasing percentage say they own a gun for hunting, −17% (Pew Research Center, June 10, 2013).

• 18. If the proportions were calculated directly from the logistic estimates then there would be no cross-check, which the regression estimates provide.

• 19. Lazarsfeld (1955, xi).

• 20. When Republican affiliation is regressed on the trichotomous state classification the latter’s effect is b = +0.085 (p = 0.0103) but when moral conservatism is controlled, this effect becomes b = −0.052 (p = 0.0572).

• 21. Models assuming reciprocal effects suggest that the priorities of the variables in Fig. 12.6 are appropriate: moral conservatism → operational ideology; warm economic feelings → operational ideology; moral conservatism → warm economic feelings, and not the opposite effects.

• 22. Smith (1999), Figs. 4.9 and 4.10 show that the reciprocal effect of ideology on party affiliation is larger than the feedback effect of party affiliation on ideology.

• 23. When the model of Fig. 12.6 is replicated by substituting for economic feelings a four-item index of 425 voters’ views about economic problems, the total effects on voting of moral conservatism (b = 0.654) again trumps the total effects of economic sentiments (b = 0.195). Moral conservatism has larger direct effects than economic problems on political ideology (b = 0.444 to b = 0.229), party affiliation (b = 0.277 to b = 0.230), and voting for McCain (b = 0.312 to b = 0). All of these nonzero effects are statistically significant (p < 0.0001). The warm economic feelings question was asked of 2000 voters; the economic problems questions were asked of only a quarter of the total sample. These missing data limit the use of this index.