NO OTHER GODS BEFORE ME: THE SUPREME COURT, PUBLIC OPINION, AND THE 10 COMMANDMENTS

NO OTHER GODS BEFORE ME: THE SUPREME COURT, PUBLIC OPINION, AND THE 10 COMMANDMENTS Ryan Cannon Abstract: Over the past three decades, scholarship regarding the effect of Supreme Court decisions on public opinion has produced several working hypotheses. One in particular, the structural response hypothesis, predicts that when the Court rules on issues in opposition to public opinion, support for or against the issue will polarize. To test this hypothesis, changes in opinion on 6 individual characteristic levels were measured prior to and after the release of two Supreme Court decisions regarding the public display of religious symbols. Contrary to the results predicted using the "structural response hypothesis," analysis of changes within groups did not show increases in polarization. Many groups showed a decrease in strong support or strong opposition to the public display of religious symbols, with increases in moderated positions. This raises some doubts as to the Court s ability to affect public opinion, as well as the validity of the structural response hypothesis. Keywords: Supreme Court, Public Opinion, Structural Response Hypothesis, Polarization, Religious Symbols, Public Display, Individual Characteristic Levels, Support, Opposition 1. Introduction Americans combine the notions of Christianity and of liberty so intimately in their minds, that it is impossible to make them conceive of one without the other -Alexis de Tocqueville, Democracy in America, 1935 On June 27, 2005, the Supreme Court announced decisions in the cases of Van Orden v. Perry and McCreary County v. ACLU. These two legal battles concerned the public display of religious symbols in Texas and Kentucky. For over four decades prior to the decision, a monument of the 10 Commandments had stood outside of the Texas state capitol building. Two Kentucky courthouses also started displaying the 10 Commandments beginning around the year 2000. In both instances, these monuments sparked controversy and litigation, eventually making it to the nation s highest court. While the Texas display in Perry served a clear enough secular purpose to be upheld, the display in McCreary constituted a violation of the Establishment Clause. In the majority opinion, Justice Souter stressed the importance of a clear demarcation between church and state. He noted that, this is no time to deny the prudence of understanding the Establishment Clause to require the Government to stay neutral on religious belief (McCreary County v. ACLU: 33). While these rulings did not vary from Ryan Cannon 2016. Originally published in Explorations: The UC Davis Undergraduate Research Journal, Vol. 18 (2016). http://explorations.ucdavis.edu The Regents of the University of California.

previous Supreme Court decisions regarding the issue of religious establishment, American attitudes towards religious displays seem to deny the prudence which Justice Souter prescribed in the decision. A Pew survey conducted just six months earlier found that 75% of respondents supported public displays of the 10 Commandments (Allen 2005). This apparent divide between support for religious displays and the outcomes in Perry and McCreary raises questions about the ability of the Supreme Court to affect public attitudes. Mainly, what influence can decisions of the Supreme Court have on the opinions of a contrarian public? In this paper, I will explore the positions of the Supreme Court and the American public towards issues of religious establishment. A brief analysis will be given of the study of Supreme Court influence on public attitudes. I will then present research to answer my central question: did the 2005 rulings in the Supreme Court cases of Van Orden v. Perry and McCreary County v. ACLU polarize public opinion toward the display of religious symbols in public spaces? First I will discuss how the Court analyzes religious establishment issues with the threepronged Lemon test. Supreme Court Jurisprudence and Public Attitudes The Establishment Clause of the First Amendment has been a source of great tension between the Supreme Court and the American public. Historically, constitutional law in this area has been separationist in nature prohibiting even the most general assistance to religion (Jelen & Wilcox 1995, Jelen 2010). At the center of Establishment Clause jurisprudence is the analytical framework set forth in Lemon v. Kurtzman (1971) (Kritzer & Richards 2003). This framework would invalidate government acts if it met one or more of three requirements. First, that it have the effect of advancing or inhibiting a religious practice. Second, that it would require an excessive entanglement of government and religion. Or third, that it would only serve a religious purpose (Jelen & Wilcox 1995). Since the 1980 s, several case outcomes reinforced the Court s separationist position and cemented the role of Lemon as Supreme Court doctrine. Stone v. Graham, in 1980, struck down the posting of the 10 Commandments in public schools. In 1987, the case of Edwards v. Aguillard found the mandatory teaching of creation science in public schools unconstitutional. Wallace v. Jaffree (1985), Lee v. Weisman (1992), and Santa Fe ISD v. Doe (2000) expanded on the prior school prayer decision in Engel v. Vitale (1962) to include bans on silent prayer, as well as prayers at graduation ceremonies and football games. In the few instances where the court ruled in a more accommodating manner, it has done so cautiously without making substantial changes to constitutional doctrine (Jelen & Wilcox 1995). American public opinion regarding the Establishment Clause has generally been contrary to the position of the Supreme Court. While at the abstract level public attitudes have supported a separationist position, when it comes to concrete applications public attitudes can best be described as accomodationist (Jelen 2010). This position would support an interpretation of the Establishment Clause which only prohibits preferential treatment to any particular religion (Jelen 2010). It seems that on the surface there exists a strong consensus on a high wall of separation between church and state. Just below that surface, the American public has traditionally been supportive of prayer in public schools, displays of religious symbols in the public sphere, military chaplains, and government mandated teaching of Judeo-Christian values (Jelen & Wilcox, 1995). The resulting division between the accomodationist position of the American public and the separationist position of the Supreme Court provides a framework through which we can analyze the influence of Court decisions on public attitudes. 2. Review of the Literature Vol. 18 (2016) Ryan Cannon p.2

Early explanations of the relationship between Supreme Court decisions and public opinion can generally be referred to as the positive response hypothesis (Franklin & Kosaki 1989). The idea being that the Supreme Court could confer legitimacy on the policy positions that it favors, pulling public opinion in line with the Court. Testing this hypothesis typically relied on analysis of aggregate opinion change, and was for the most part unsuccessful at finding evidence for a positive response relationship. Marshall (1987) looked at attitude shifts after eighteen Court decisions since the 1930 s, and was unable to identify movement in opinion. Similarly, research focused on change after Roe v. Wade (1973) found only slight differences in aggregate opinion (Johnson & Canon 1984). While measuring changes in aggregate opinion has not provided conclusive evidence for a relationship between Supreme Court decisions and public attitudes, Franklin and Kosaki (1989) were able to identify significant shifts in opinion at the individual and group level. Analysis of opinion change between groups following Roe showed strong evidence of the influence of the Court s actions on the structure of public opinion concerning abortion (Franklin & Kosaki 1989). However, the evidence did not point towards the positive response hypothesis. Franklin and Kosaki instead found evidence for what they called the structural response hypotheses, suggesting that the Court s decision had a polarizing effect on groups already occupying either side of the abortion debate. The micro-level social-psychological analysis contained in Franklin and Kosaki s work has since become the underlying principle of research on relationships between Supreme Court decisions and public attitudes. Hoekstra and Segel (1996) looked at changes in group opinion in local communities after the decision in Lamb s Chapel v. Center Moriches School District (1992). They identified a significant relationship between Supreme Court decisions and the geographical areas in which those legal battles ensued (Hoekstra & Segel 1996). This suggests that higher levels of knowledge about a particular case can increase support for the opinion of the Court among those who regard the decision as less important (Hoekstra & Segal 1996). Johnson and Martin (1998) found support for the structural response hypothesis by analyzing shifts in group public opinion data before and after abortion and capital punishment rulings. Beyond the structural response, they found that Supreme Court opinions which followed major landmark decisions in the same area had relatively little influence on public attitudes. This conditional response hypothesis has since been challenged by the work of Brickman and Peterson (2006). They studied the effects of two Supreme Court cases on abortion, which had a significant influence on individual level opinion, regardless of the number of previous actions. The analysis of opinion change on the individual level regarding gay civil rights in the wake of Bowers v. Hardwick (1986) and Lawrence v. Texas (2003) suggests a more nuanced relationship exists between Court decisions and public attitudes. While differences in opinion provide support for the structural response hypothesis, changes in support were significant only for specific demographic groups rather than those simply holding opposing positions (Allen, Haider- Markel, & Stoutenborough 2006). Women and liberals, for instance, were more likely to support gay civil rights after Lawrence, while Catholics and the wealthy were less likely to support gay civil rights after Bowers (Allen, Haider-Markel, & Stoutenborough 2006). This research also suggests that the influence of Court decisions is contingent on the level of knowledge the public has on the case. While the public s lack of knowledge about legal issues does not minimize the role of the Supreme Court, it does highlight just how conditional the potential influence of the Court actually is (Allen, Haider-Markel, & Vol. 18 (2016) Ryan Cannon p.3

Stoutenborough 2006). Individual level analysis has provided some support for the legitimation hypothesis. Research reanalyzing opinion changes in groups after Roe suggests that the public s move towards a broader acceptance of abortion was more significant than was previously thought. Hanley, Salamone, and Wright (2012) found that the positive response hypothesis is borne out by Roe, and that their analysis indicate[s] that the Court is able to sweep opinions to its side, even in highly controversial areas it is deciding on for the first time. Regardless of the hypothesis, testing the relationship between Supreme Court decisions and public attitudes has come to focus on individual level characteristics. Research concerning the influence of Establishment Clause cases is no different. Jelen and Wilcox (1995) found that positions on church-state separation in the abstract varied among groups. Education, income, age, sex, and race were all found to be significant predictors of support for a separationist or accomodationist position. Women as well as racial and religious minorities were generally found to be more separationist than men. Older people tended to take a highly accomodationist position usually associated with increased levels of conservatism (Jelen & Wilcox 1995). On the concrete level, individual characteristics also play an important role in measuring changes in support for accomodationist or separationist positions. Elifson and Hadaway (1985) identified significant divisions in support of prayer in public schools within the context of age, education level, and political affiliation. Conservative political affiliations were much more likely to garner support for school prayer, and age was expected to be a significant factor because of the tendency of older persons to be more politically conservative (Elifson & Hadaway 1985). Similarly, Elifson and Hadaway (1985) speculated that education provided a liberalizing effect, which broadened the context for analyzing the issue. The effects of both religious affiliation and race were statistically significant when analyzing support for school prayer, and are largely thought to be intertwined (Gash & Gonzalez 2008). African Americans, who tend to be more religiously observant, as well as evangelical Christians, showed a high likelihood of support for prayer in public schools compared to whites, Jews, and the nonreligious (Gash & Gonzalez 2008). 3. Central Hypothesis Within the scope of my research I will be testing the structural response hypothesis. The 2005 decisions in Perry and McCreary are unlikely to have moved public opinion on religious displays in public spaces in line with the court. The decisions underscored the court s preference for a separationist position; the display in McCreary was struck down because the court believed it served an overtly religious purpose, while the display in Perry was upheld only on the basis that it was secular enough. This stands in direct contrast with the American public s preference for an accomodationist approach. Rather than pulling the public in line with the court, the Court s decisions most likely polarized public opinion on the issue, effectively working to move those who only held mild opinions on the issue or no opinion at all into a strong position of either agreement or disagreement. As was discussed previously, the structural response hypothesis is only one possible explanation for the relationship between Supreme Court decisions and public opinion changes. It very well may be that the Court was able to pull public opinion more in line with its own position, as the positive response hypothesis suggests. There is also a distinct possibility that any changes observed in public opinion were unrelated to the Supreme Court decisions. For the purposes of my research, though, I will only be testing for the significance of a structural response. If there is any evidence for a positive response found, it will be left for further research to Vol. 18 (2016) Ryan Cannon p.4

determine if these changes are statistically significant. I will take as my null hypothesis that the Supreme Court s decisions in these two cases had no measurable effect on public opinion, with my alternative hypothesis being that the main effect of these two decisions was an increase in polarization. 4. Methodology Starting in 1997, the First Amendment Center has issued periodic surveys examining American attitudes towards the freedom of speech, press, assembly, and religion. The results from these surveys are published annually in the Center s State of the First Amendment report. Beginning in 2002, and continuing through 2005, this survey included a question about the public display of religious symbols. The question asked those surveyed to answer whether they agreed or disagreed with the following statement: Government officials should be allowed to post the 10 Commandments inside government buildings, using a scale of strongly agree, mildly agree, mildly disagree, strongly disagree, don t know, or refused. The surveys were typically administered during the beginning of June, with the 2005 survey being conducted a mere four days before the release of both the Perry and McCreary decisions. Using the data from this survey, I conducted group level analysis in order to identify opinion preferences within characteristic groups towards the display of the 10 Commandments inside government buildings. Three groups of demographic characteristics were used which were expected to have an effect on an individual s level of support for separationist or accomodationist positions. These three groups can be categorized as (I) immutable demographic characteristics, (II) long term social characteristics, and (III) values and political orientation. Group I is made up of sex, age, and race. Group II includes education level. Group III includes political and religious affiliation. The relationship between characteristics and support positions will then be used to predict a structural response. To measure changes in support following the Perry and McCreary decisions, a separate survey was needed. In the winter of 2005 Baylor University began wave one of its religion survey, which was followed up by wave two in 2007. Like the First Amendment Center Surveys, these also included a question about public displays of religious symbols. The Baylor survey asked participants in the survey to answer: To what extent do you agree or disagree that the federal government should allow the display of religious symbols in public places? The answer type, similar to the First Amendment Center surveys, was a scale of strongly agree, agree, disagree, strongly disagree, and undecided. The data from these two surveys was used to measure changes in public opinion after the release of the Perry and McCreary decisions. Conducting analysis of characteristic groups I, II, and III, I looked for significant changes in opinion between 2005 and 2007. Because of the answer type used in the Baylor survey, I was able to look at changes in polarization within each characteristic between those who simply agree or disagree, and those who strongly agree or strongly disagree. While differences in question type make it difficult if not impossible to compare directly the results of the First Amendment Center surveys and the Baylor University surveys, the former can at least serve as a basis for identifying the general support position and trend of polarization for specific attitudinal predictors pre Perry and McCreary. Once these groups have been identified, the latter survey can be used to look at trends of polarization for specific attitudinal predictors post Perry and McCreary. Public Opinion and Group Characteristics Prior to the Court Decisions: The State of the First Amendment Center Survey A key aspect of the structural response Vol. 18 (2016) Ryan Cannon p.5

hypothesis is that polarization will be most likely to occur among characteristic groups which hold strong positions either for or against the issue. Data from the State of the First Amendment Center survey was used to look at group positions, and any trends that appeared in the 3 years running up to the Supreme Court decisions. Groups that were more likely to hold accomodationist positions before the decisions were released will be more likely to see drops in mild positions and increases in strong support for the public display of religious symbols. Similarly, groups that were more likely to take a separationist position will, according to the structural response hypothesis, see drops in moderate positions and increases in strong disagreement. (I) Immutable Demographic Characteristics: Sex, Age, & Race When it comes to the public display of the 10 Commandments, female respondents were both more likely to hold an accomodationist position than men, as well as a position of strong agreement. By 2005, female respondents held a higher level of strong support over males, with both groups seeing slight decreases in moderate positions between 2003 and 2005 (See appendix Chart 3, Chart 5, and Chart 6). For the purposes of testing the structural response hypothesis, these results indicate that both males and females should see increases in strong support, with decreases in moderate positions following the Court decisions. Unlike the support found among male and female respondents, when broken down by age the State of the First Amendment Center survey data aligns with the relationship found by Jelen and Wilcox (1995). While all age groups were more likely to take a position of accomodationism, as age increased so did the likelihood of taking a strong position of agreement. By 2005, strong support for the public display of the 10 Commandments sat at 47% for 18-30 year olds, 53% for 31-44 year olds, 57% for 45-61 year olds, and 63% among those aged 62 and up. If a structural response occurs in this case, then polarization should become more evident among older groups after the Court rulings. These groups should see increases in strong support, with decreases in moderate positions. Finally, where Jelen and Wilcox (1995) found racial minorities more likely to take separationist positions than whites, black respondents in the State of the First Amendment Center survey were just as likely to take an accomodationist position as whites (See appendix Chart 1). In fact, black respondents were slightly more likely to hold a position of strong agreement than whites in 2003, 2004, and 2005 (See appendix Chart 3). Among both white and black respondents there was also a slight decrease in moderate positions between 2003 and 2005, along with increases in strong support (See appendix Chart 5, Chart 6, and Chart 3). Following the release of the Court decisions, then, we should see increases in strong support among both whites and blacks, with decreases in moderate positions if the structural response hypothesis holds. (II) Long Term Social Characteristics: Education Higher levels of education, which Elifson and Hadaway (1985) suspected had a liberalizing effect, seemed to correlate with an increase in the likelihood of a separationist position. As education increased, the level of strong support within each group decreased. In 2005, 63% of high school graduates held strong support, 6 among those with some college education, 5 among college graduates, and 39% among post graduates. However, all groups saw increases in strong support between 2003 and 2005 (See appendix Chart 3). Post graduates were more likely to take a separationist position over an accomodationist position in 2003 (58% versus 4, respectively). By 2005, though, post graduates along with all other education groups were more likely to hold accomodationist positions (See appendix Chart 1). Vol. 18 (2016) Ryan Cannon p.6

As the education level increased, there was also an increase in strong disagreement with the public display of the 10 Commandments. Strong disagreement in 2005 was at 13% among high school graduates, 21% among those with some college education, 23% among college graduates, and 33% among post graduates. Like the overall increasing trend in strong agreement, though, all groups except those with some college education saw decreases in strong disagreement between 2003 and 2005 (See appendix Chart 4). All of this indicates that following the release of the Supreme Court decisions, polarization should be most evident among high school graduates and post graduates; the former seeing increases in strong support, and the latter increases in strong disagreement. (III) Values and Political Orientation: Political & Religious Affiliation While Democrat, Independent, and Republican respondents were all likely to take accomodationist positions, it was much more likely among those holding conservative political affiliations. By 2005, 58% of Democratic respondents took a position of accomodationism, with Independents at 67%, and Republicans at 87%. Republican respondents were also much more likely to hold strong agreement than were independents and democrats (72%, 53%, and 46% respectively, by 2005).Democratic respondents were more likely to strongly disagree than either Independents or Republicans, with 31% taking a position of strong disagreement by 2005. Between 2003 and 2005, interestingly, all three groups saw significant increases in strong support for the public display of the 10 Commandments. All of this indicates that if a structural response occurs, polarization should be most evident in the most conservative political affiliations following the Court s decisions. Similar to what Gash and Gonzalez (2008) found, data from the State of the First Amendment Center survey indicated that the more religiously observant respondents, especially evangelical Christians, were the most likely to take an accomodationist position (See appendix Chart 1). Those identifying as Evangelical Christians were the most likely to take positions of accomodationism (91% by 2005). Protestants trailed slightly behind Evangelicals, with 81% taking an accomodationist position by 2005, and Catholics behind that with 74%. Strong agreement between these three groups had a similar trend, with 8 of Evangelicals strongly agreeing by 2005, 67% of Protestants, and 58% of Catholics. Those identifying as non-religious were more likely to take a position of separationism (58% by 2005), and held the highest levels of strong disagreement (42% by 2005). All of this data suggests that, following the Court s decisions, polarization should be most evident among the non-religious and Evangelical Christians. According to the structural response hypothesis, these two groups should see decreases in moderate positions with increases in strong disagreement for the non-religious and strong agreement for Evangelicals. Because of their already high levels of support for accomodationism and strong agreement, Catholics and Protestants are likely to also see polarization (although to a lesser degree), with increases in strong agreement. Public Opinion and Group Characteristics Post Court Decisions: The Baylor University Survey The characteristic groups and their positions on the public display of religious symbols garnered from the State of the First Amendment Center survey were identified within the Baylor University Survey, and then used to hypothesize a structural response. The predictions were derived not from the actual results of the State of the First Amendment Center survey, but from the theory underlying previous research on the structural response hypothesis. These predictions were compared against movement in the two years following the release of the decisions. If the structural response Vol. 18 (2016) Ryan Cannon p.7

hypothesis is correct, we should see polarization within each group toward positions which were strongly supported prior to the decisions. Within characteristic group I, we should see movement toward strong accomodationist agreement among males and females. As age increases, so should polarization towards strong accomodationist agreement. Strong accomodationism among all racial groups should also increase following the decisions. In characteristic group II, polarization towards accomodationism should be evident among those with the least amount of education. At the same time, those with the most amount of education should exhibit an increase in strong separationism. Finally, in characteristic group III, polarization towards accomodationism should be most evident among those with conservative political affiliations. This should also be the case among the more religiously observant, with polarization towards strong separationism among the non-religious 1 (I) Immutable Demographic Characteristics: Sex, Age, & Race In both 2005 and 2007, males and females were more likely to take accomodationist positions over separationism (See appendix Chart 7 and Chart 8). Contrary to the predictions made using the structural response hypothesis, though, between 2005 and 2007 males and females saw decreases in strong positions and increases in moderate ones (See Figures 1 & 2). Strong agreement for the public display of religious symbols among females dropped 3%, and strong disagreement fell 4%. At the same time, both agreement and disagreement increased 4% each. Males saw a similar drop in strong agreement (4%), and strong disagreement (5%), with a 2% increase in agree, and a 5% increase in disagree. This trend towards moderate positions was evident in most other characteristic groups as well. Among age groups, all were more likely to take positions of accomodationism over separationism (See appendix Chart 7 and Chart 8). However, among those 65 and older, there was an 8% decrease in strong agreement, with a 6% and 7% increase in agreement and disagreement (See Figure 3). Moderate positions also increased among those 31-44 years old, with decreases in both strong positions. Among the youngest respondents, those aged 18-30, strong disagreement decreased while disagreement and strong agreement rose. These shifts did not change the fact that even among the youngest respondents a position of accomodationism was more likely than one of separationism. Similar to this, support along racial lines did not seem to indicate any polarization. Like the State of the First Amendment Center survey data, both white and black respondents were more likely to hold accomodationist positions (See appendix Chart 7). While strong support among black respondents did not change between 2005 and 2007 (39% for both years), mild agreement increased by 3%, with disagreement and strong disagreement both remaining stable. Strong support decreased among white respondents by 4% between 2005 and 2007, with agreement increasing by 3%, and disagreement increasing by 5%. All of this taken together makes it clear that among characteristic group I, polarization was not occurring. If anything, there seemed to be a moderating effect, especially among groups where polarization should have been most evident. (II) Long Term Social Characteristics: Education Education, according to the structural response hypothesis, should provide a liberalizing effect, making polarization towards accomodationism more evident among those with the least education. At the same time, those with the most education should see the least amount of polarization towards accomodationism, and even some accompanying increases in separationism. Like the State of the First Amendment Center survey data, all education groups were more Vol. 18 (2016) Ryan Cannon p.8

likely to take accomodationist positions with the likelihood of this decreasing as education increased (See appendix Chart 7). Among high school graduates strong agreement dropped 9%, with agreement increasing by 5% and disagreement increasing by 4% (See Figure 4). Post graduates, who should have seen an increase in separationist positions, actually saw a slight increase in accomodationism (See appendix Chart 7). Strong disagreement also dropped between 2005 and 2007, with strong agreement decreasing by 2%, and both moderate positions increasing (agreement by 6%, and disagreement by 8%) (See Figure 5). This trend continued with college graduates, who saw strong disagreement and strong agreement decrease (3% and 8%, respectively), with a 7% increase in disagreement, and a 2% increase in agreement. While respondents with some college education and those with trade and vocational training did not experience the same moderating trend, they also did not show signs of polarization. Among those with some college strong agreement increased by only 1%, with agreement and disagreement both increasing by 2%. Strong agreement decreased among those with trade and vocational training by 1%, with agreement decreasing by 3%, and disagreement increasing by 8%. These results, like those found among characteristic group I, run contrary to the predicted structural response. Where polarization should have been most evident, among high school graduates and post graduates, there was a drop in strong positions and increases in both moderate ones. This would seem to align with what the positive response hypothesis may have predicted following the Court decisions. (III) Values and Political Orientation: Political & Religious Affiliation Political affiliation in the Baylor University survey, which was coded as Republican, Independent, and Democrat, also had respondents record their political affiliation for Republicans and Democrats as either strong, moderate, or leaning. This provides an interesting opportunity to look closer for signs of polarization. Movement toward strong positions of accomodationism should be most evident among conservative political affiliations, especially those identifying as strong Republicans. This movement toward strong accomodationism should also be least noticeable among liberal political affiliations, especially those identifying as strong Democrats. This group, if it does polarize, may even show movement towards separationism. This was, however, not the case. Among Republicans in general, between 2005 and 2007, strong agreement dropped 3%. At the same time agreement increased by 1%, and disagreement increased by 4% (See Figure 6). This shift towards moderate positions was even more pronounced among those identifying as strong Republicans. For this group, between 2005 and 2007, strong agreement dropped 8%, with agreement increasing by 5% and disagreement increasing by 4% (See Figure 7). At the same time, Democrats experienced a 9% decrease in strong disagreement, and a 2% decrease in strong agreement. This was accompanied by a 6% increase in agreement, and a 5% increase in disagreement (See Figure 8). Among strong Democrats, movement towards moderate positions was even bigger. Strong disagreement decreased by 8%, with strong agreement decreasing by 2%. Agreement increased by 7%, and disagreement increased by 6% (See Figure 9). This lack of polarization, even among those most likely to display it according to the structural response hypothesis, was also evident among religious affiliations. Like the structure of support found in the State of the First Amendment Center survey data, the non-religious were the most likely to take a position of separationism (69% in 2005), while support for accomodationism was highest among Catholics (73% in 2005), Protestants (7 in 2005), and Evangelicals (86% in 2005). This would indicate that if Vol. 18 (2016) Ryan Cannon p.9

polarization were to occur, it would be most evident among the non-religious and Evangelicals (although, with such high support, Catholics and Protestants should see polarization as well). Between 2005 and 2007, though, strong disagreement among the non-religious dropped 4% (from 42% to 38%). At the same time, both moderate positions increased by 2% (See Figure 10). Catholics and Protestants both had small decreases in strong support for the public display of religious symbols (2% and 1% respectively). Evangelicals, who held the highest level of strong support in 2005 (49%), saw a 9% decrease by 2007. Meanwhile, both moderate positions had increases between 2005 and 2007 (6% increase in agreement, and a 3% increase in disagreement) (See Figure 11). All of this indicates that polarization following the release of the Supreme Court decisions did not occur among the groups most likely to exhibit such an effect. Contrary to the structural response hypothesis, the data from the Baylor University religion survey shows a decrease in polarization among characteristic groups. The ramifications of this evidence, its limitations, as well as suggestions for follow up research will now be further discussed in light of these conclusions. 5. Conclusion & Discussion Results from the analysis of public opinion following Van Orden v. Perry and McCreary County v. ACLU indicates that polarization, at least among the groups identified in this experiment, did not occur. What did seem to happen was a sort of moderating effect, where groups which should have experienced a rise in strong opinions saw the opposite. These shifts in opinion towards more moderate positions, strangely enough, were even evident among the groups which held the strongest opinions prior to the Court s decisions. What, then, could be the reason for the lack of predicted polarization? While these results were contrary to the structural response hypothesis, polarization may still be occurring on a deeper level. Cross-sectional analysis between characteristic groups may yield evidence of polarization, as may investigation of public opinion on the individual level. If polarization was not found after further research, and if the shifts towards moderate positions evident in this experiment were found to be statistically significant (that is, outside of the possibility of random chance), then this could lend support to the positive response hypothesis. Similar to what Hanley, Salamone, and Wright (2012) found in the aftermath of Roe, the Court may have been able to pull public opinion in line with itself on this issue. The changes in opinion found following the decisions may have also been the natural progression of public opinion on the matter, completely devoid of influence by the Court. It very well could have been that most Americans were unaware of the two court cases and the Supreme Court s opinions. This would seem to be in line with the results found by Allen, Haider-Markel, and Stoutenborough (2006) who theorized that the level of influence the court has is contingent on the amount of knowledge the public has on the case. The surveys used in this study were also conducted on a national level, and as Hoekstra & Segel (1996) pointed out, the geographical areas where the cases originated are often associated with higher levels of knowledge on the issue. This suggests that while polarization did not occur on a national level, it may have still happened on the local level where the case would have been common knowledge. There are, of course, limitations to this study which should be taken into consideration when analyzing the findings. Unlike the State of the First Amendment Center survey, the survey used to analyze shifts in opinion following the Court s decisions asked respondents not about the 10 Commandments specifically, but about the public display of religious symbols in general. This is a very important distinction, especially when you take into consideration the Vol. 18 (2016) Ryan Cannon p.10

particular way in which Americans view the 10 Commandments. Support for displays in public courthouses, schools, and government buildings suggest that many Americans do not consider the 10 Commandments to be an explicitly religious symbol. Indeed, as was argued by lawyers in front of the Supreme Court, the 10 Commandments for many Americans represents a sort of moral code akin the justice system itself. Since many Americans view the 10 Commandments as a set of historical first laws, it would be unsurprising if support for their public display was higher than for religious displays in general. This may explain why, while there appeared to be a real conflict between public opinion and the position of the court, there did not seem to be any noticeable effect on the way the public views the display of explicitly religious symbols. If the Baylor University survey had asked specifically about the 10 Commandments, then there might have been evidence supporting the structural response hypothesis. Speculative theories such as these obviously require further research. Investigating the significance of the observed moderating effect may help to provide evidence for the positive response hypothesis. While my data is very suggestive of a positive response, it does not engage in multivariate analysis. This kind of control would provide independent assessment of the contribution of each characteristic to the observed support position. Given the nature of the dependent variable, some form of regression analysis would be the next logical step to take in trying to prove or disprove a positive response. Analysis of cross-sectional data, or movement in opinion on the individual level could also revive the possibility for a structural response. Likewise, if public opinion surveys were conducted following the Court s decisions asking about the 10 Commandments specifically, then analysis of this data would go a long way in clarifying opinion change in light of the way Americans view the 10 Commandments. What can be said is that among the groups identified within this study, polarization of opinion regarding religious displays in general did not occur. While the Court ruled in a manner contrary to the opinion of the public, it did not seem to exacerbate the tension between its own position of separationism, and the accomodationist position of the public. Caution should be taken when interpreting the findings of this study because a polarization effect could still be occurring on a number of different levels. With this in mind, though, what appears to have happened is a moderation of opinion where movement was generally away from strong positions and toward a middle ground. In a political climate rife with division, this study provides a glimmer of hope that polarization is not the inevitable outcome for every aspect of Supreme Court politics and public opinion. Acknowledgments First I would like to acknowledge and thank Professor John Gates for all of his guidance throughout the research, writing, and editing process. There is no doubt in my mind that his effort was instrumental to the completion and success of this paper. I will always be indebted to him for his generosity, time, and encouragement. I would also like to thank Professor Eric Engstrom for all of his time spent instructing the honors thesis seminar. His work made conducting research an enjoyable and exciting experience. Finally, I would like to thank my parents, who gave me the confidence to find my own connection to religion and spirituality. Finishing this thesis and the education which led to it would not have been possible without their endless support. Vol. 18 (2016) Ryan Cannon p.11

Figure 1. Male Support for the Public Display of Religious Symbols Figure 2. Female Support for the Public Display of Religious Symbols Percent of Male Respondents 45% 4 35% 3 25% 2 15% 5% Year Percent of Female Respondents 45% 4 35% 3 25% 2 15% 5% Year Strongly Agree Agree Strongly Agree Agree Strongly Strongly Percent of Respondents Aged 65+ 5 4 3 2 Figure 3. Support for the Public Display of Religious Symbols Among Respondents Age 65+ Year Strongly Agree Agree Strongly Vol. 18 (2016) Ryan Cannon p.12

Percent of High School Graduate Respondents Figure 4. Support for the Public Display of Religious Symbols Among High School Graduates 5 4 3 2 Strongly Agree Agree Strongly Support Response Percent of Postgraduate Respondents Figure 5. Support for the Public Display of Religious Symbols Among Postgraduates 4 35% 3 25% 2 15% 5% Strongly Agree Agree Strongly Support Response Percent of Republican Rspondents Figure 6. Support for the Public Display of Religious Symbols Among Republicans 5 4 3 2 2005 Year 2007 Strongly Agree Agree Percent of Strong Republican Respondents Figure 7. Support for the Public Display of Religious Symbols Among Strong Republicans 7 6 5 4 3 2 Year Strongly Agree Agree Strongly Strongly Vol. 18 (2016) Ryan Cannon p.13

Percent of Democratic Respondents 4 35% 3 25% 2 15% 5% Figure 8. Democratic Support for the Public Display of Religious Symbols Year Percent of Strong Democrat Respondents Figure 9. Support for the Public Display of Religious Symbols Among Strong Democrats 4 3 2 Year Strongly Agree Agree Strongly Agree Agree Strongly Strongly Percent of Non-Religious Respondents Figure 10. Non-Religious Support for the Public Display of Religious Symbols 45% 4 35% 3 25% 2 15% 5% Strongly Agree Agree Strongly Support Response Percent of Evangelical Protestant Respondents 6 5 4 3 2 Figure 11. Evangelical Protestant Support for the Public Display of Religious Symbols Strongly Agree Agree Strongly Support Response Vol. 18 (2016) Ryan Cannon p.14

Appendix State of the First Amendment Center Survey Please tell me whether you agree or disagree with the following statements. Government officials should be allowed to post the 10 Commandments inside government buildings. Strongly agree Mildly agree Mildly disagree Strongly disagree Don't Know Refused Chart 1. Support for Accomodationism Among Groups Chart 2. Support for Separationism Among Groups Variable Year Variable Year 2003 2004 2005 2003 2004 2005 Male 54% 63% 65% Male 43% 35% 33% Female 58% 67% 76% Female 38% 3 24% 18-30 58% 67% 67% 18-30 41% 31% 32% 31-44 57% 62% 71% 31-44 4 36% 29% 45-61 55% 65% 68% 45-61 42% 32% 3 62+ 6 66% 73% 62+ 37% 31% 24% White 58% 64% 71% White 4 34% 28% Black 57% 78% 67% Black 42% 19% 29% Hispanic 46% 7 75% Hispanic 51% 28% 25% Asian 7 71% 82% Asian 25% 29% 18% High School Graduate 71% 74% 81% High School Graduate 26% 23% 18% Some College 61% 68% 7 Some College 37% 3 28% College Graduate 51% 64% 67% College Graduate 46% 34% 31% Post Graduate 4 51% 51% Post Graduate 58% 47% 47% Democrat 48% 51% 58% Democrat 48% 47% 41% Independent 5 64% 67% Independent 48% 34% 32% Republican 7 82% 87% Republican 28% 15% 12% Catholic 53% 66% 74% Catholic 44% 32% 25% Protestant 66% 75% 81% Protestant 32% 23% 17% Fundamentalist/ Evangelical 77% 89% 91% Fundamentalist/ Evangelical 22% 9% 9% Jewish 32% 4 52% Jewish 64% 6 48% No Religious Affiliation 37% 43% 45% No Religious Affiliation 55% 55% 54% Other 63% 66% 45% Other 36% 31% 55% Vol. 18 (2016) Ryan Cannon p.15

Chart 3. Strong Agreement Among Groups Chart 4. Strong ment Among Groups Variable Year Variable Year 2003 2004 2005 2003 2004 2005 Male 37% 43% 51% Male 29% 22% 25% Female 39% 49% 61% Female 24% 21% 17% 18-30 3 37% 47% 18-30 23% 18% 2 31-44 39% 36% 53% 31-44 28% 22% 21% 45-61 39% 46% 57% 45-61 29% 21% 22% 62+ 45% 51% 63% 62+ 23% 22% 18% White 39% 45% 56% White 26% 22% 2 Black 39% 62% 59% Black 22% 14% 21% Hispanic 27% 43% 64% Hispanic 24% 15% 14% Asian 35% 41% 59% Asian 25% 18% 6% High School Graduate 5 58% 63% High School Graduate 16% 12% 13% Some College 44% 5 6 Some College 21% 21% 21% College Graduate 31% 43% 5 College Graduate 31% 23% 23% Post Graduate 24% 28% 39% Post Graduate 43% 34% 33% Democrat 28% 35% 46% Democrat 35% 36% 31% Independent 34% 4 53% Independent 31% 22% 22% Republican 52% 64% 72% Republican 15% 6% 7% Catholic 31% 37% 58% Catholic 24% 17% 19% Protestant 48% 58% 67% Protestant 19% 15% 11% Fundamentalist/ Evangelical 61% 74% 8 Fundamentalist/ Evangelical 11% 5% 6% Jewish 11% 25% 43% Jewish 46% 4 39% No Religious Affiliation 24% 25% 26% No Religious Affiliation 45% 39% 42% Other 46% 49% 41% Other 24% 22% 34% Vol. 18 (2016) Ryan Cannon p.16

Chart 5. Mild Agreement Among Groups Chart 6. Mild ment Among Groups Variable Year Variable Year 2003 2004 2005 2003 2004 2005 Male 17% 21% 14% Male 15% 13% 9% Female 19% 18% 14% Female 14% 9% 8% 18-30 27% 31% 2 18-30 18% 12% 11% 31-44 18% 27% 18% 31-44 12% 14% 7% 45-61 16% 19% 12% 45-61 14% 12% 9% 62+ 16% 15% 62+ 14% 8% 6% White 19% 19% 15% White 14% 12% 8% Black 18% 16% 9% Black 19% 5% 7% Hispanic 2 28% 11% Hispanic 27% 13% 11% Asian 35% 29% 24% Asian 12% 12% High School Graduate 21% 16% 19% High School Graduate 11% 5% Some College 17% 18% Some College 16% 9% 8% College Graduate 21% 21% 16% College Graduate 15% 11% 9% Post Graduate 17% 23% 13% Post Graduate 15% 14% 15% Democrat 21% 15% 12% Democrat 14% 11% Independent 16% 24% 14% Independent 16% 13% 9% Republican 18% 18% 15% Republican 13% 4% Catholic 22% 29% 15% Catholic 21% 14% 6% Protestant 18% 17% 14% Protestant 13% 7% 6% Fundamentalist/ Evangelical 16% 15% 11% Fundamentalist/ Evangelical 11% 4% 3% Jewish 21% 15% 9% Jewish 18% 2 9% No Religious Affiliation 13% 19% 18% No Religious Affiliation 15% 12% Other 18% 17% 3% Other 12% 9% 21% Vol. 18 (2016) Ryan Cannon p.17

Baylor University Religion Survey Results To what extent do you agree or disagree that the federal government should allow the display of religious symbols in public spaces? Strongly agree Agree Strongly disagree Undecided Chart 7. Support for Accomodationism Among Groups Chart 8. Support for Separationism Among Groups Variable Year Variable Year Male 65% 63% Male 3 3 Female 68% 69% Female 23% 23% 18-30 6 58% 18-30 32% 32% 31-44 61% 63% 31-44 3 31% 45-64 67% 69% 45-64 26% 14% 65+ 74% 72% 65+ 21% 22% White 67% 67% White 26% 27% Black 68% 71% Black 17% 16% Asian* 25% 38% Asian* 65% 54% High School Graduate 8 77% High School Graduate 16% Some College 72% 74% Some College 22% 2 Trade/Vocational 74% 7 Trade/Vocational 18% 23% College Graduate 67% 62% College Graduate 27% 3 Post Graduate 49% 51% Post Graduate 45% 43% Democrat 49% 53% Democrat 42% 38% Strong Democrat 41% 52% Strong Democrat 51% 4 Moderate Democrat 53% 56% Moderate Democrat 38% 36% Leaning Democrat 55% 49% Leaning Democrat 33% 4 Independent 63% 62% Independent 29% 29% Republican 86% 83% Republican 12% Leaning Republican 74% 76% Leaning Republican 21% 16% Moderate Republican 83% 82% Moderate Republican 11% 13% Strong Republican 94% 91% Strong Republican 5% 8% Catholic 73% 73% Catholic 19% 19% Black Protestant* 74% 77% Black Protestant* 17% 7% Protestant 7 69% Protestant 24% 26% Evangelical 86% 83% Evangelical 12% Jewish* 16% 19% Jewish* 73% 78% Non-Religious 2 22% Non-Religious 69% 67% Other* 47% 55% Other* 45% 36% *Characteristic groups omitted from analysis because of a low number of respondents. Vol. 18 (2016) Ryan Cannon p.18

Chart 9. Strong Agreement Among Groups Chart 10. Strong ment Among Groups Variable Year Variable Year Male 28% 24% Male 16% 11% Female 31% 29% Female 13% 9% 18-30 19% 26% 18-30 18% 13% 31-44 31% 23% 31-44 17% 12% 45-64 28% 29% 45-64 13% 65+ 35% 27% 65+ 11% 5% White 3 26% White 14% 9% Black 39% 39% Black 8% 9% Asian* 5% 23% Asian* 4 15% High School Graduate 44% 35% High School Graduate 3% 5% Some College 3 31% Some College 6% Trade/Vocational 29% 28% Trade/Vocational 9% 6% College Graduate 27% 19% College Graduate 16% 13% Post Graduate 19% 17% Post Graduate 26% 16% Democrat 18% 16% Democrat 25% 16% Strong Democrat 16% 18% Strong Democrat 32% 2 Moderate Democrat 19% 17% Moderate Democrat 22% 14% Leaning Democrat 16% 12% Leaning Democrat 16% 13% Independent 26% 24% Independent 14% 12% Republican 41% 38% Republican 4% 2% Leaning Republican 24% 28% Leaning Republican 11% 1% Moderate Republican 33% 34% Moderate Republican 4% 3% Strong Republican 6 52% Strong Republican 1% 1% Catholic 28% 26% Catholic 8% 4% Black Protestant* 51% 44% Black Protestant* 6% 4% Protestant 24% 23% Protestant 12% 8% Evangelical 49% 4 Evangelical 4% 3% Jewish* 2% 6% Jewish* 47% 25% Non-Religious 4% 3% Non-Religious 42% 38% Other* 18% 17% Other* 26% 18% *Characteristic groups omitted from analysis because of a low number of respondents. Vol. 18 (2016) Ryan Cannon p.19