Examining Changes in Attitudes on the Death Penalty: 40 Years of Public Opinion

In 1608 the first recorded execution in what is now the United States occurred in the Jamestown colony of Virginia [1]. Since that time, few social issues have fueled debate and pendulum swings as capital punishment. In 2016 [2], support for the death penalty was the same as it was in 1937, 60% [3]. This longview is deceiving, however. From 1937 to 1953, public support for the death penalty for murder steadily increased, peaking at 68%. Then, from 1953 to 1966, the trend reversed, declining to 42%, the lowest in polling history. The pendulum swung again, reaching 80%, the highest level of public support, in 1994. Since that year, support has declined, except for a few years, to its current level [3]. The most recent data from the Pew Research Center [3] reveals that support for the death penalty is strongest among Republicans, men, Whites, those 30 to 64 years, those with some college education, and White Evangelicals. Other than reporting demographic breakdowns of survey respondents, Pew provides no analysis of the relationship between individuals with different backgrounds and their views on the death penalty.


Introduction
In 1608 the first recorded execution in what is now the United States occurred in the Jamestown colony of Virginia [1]. Since that time, few social issues have fueled debate and pendulum swings as capital punishment. In 2016 [2], support for the death penalty was the same as it was in 1937, 60% [3]. This longview is deceiving, however. From 1937 to 1953, public support for the death penalty for murder steadily increased, peaking at 68%. Then, from 1953 to 1966, the trend reversed, declining to 42%, the lowest in polling history. The pendulum swung again, reaching 80%, the highest level of public support, in 1994. Since that year, support has declined, except for a few years, to its current level [3]. The most recent data from the Pew Research Center [3] reveals that support for the death penalty is strongest among Republicans, men, Whites, those 30 to 64 years, those with some college education, and White Evangelicals. Other than reporting demographic breakdowns of survey respondents, Pew provides no analysis of the relationship between individuals with different backgrounds and their views on the death penalty.
The research literature on the factors that help explain opinions regarding the death penalty is relatively sparse. In writing a concurring opinion to the majority in Furman v. Georgia [4], Justice Marshall concluded that "I believe that the great mass of citizens would conclude on the basis of the material already considered that the death penalty is immoral, and therefore unconstitutional: [4]. Justice Marshall believed that if average citizens were fully informed about the death penalty they would find it abhorrent. Following this landmark case which found that the death penalty violated the U.S. Constitution 8th Amendment prohibiting "cruel and unusual punishment," researchers set out to test the "Marshall hypothesis" that information can affect the public's decision on the death penalty [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19].

Methods
This research examines 40 years of data (1974-2014) from the General Social Survey (GSS) to understand and explain changes in attitudes towards capital punishment. The GSS is the most comprehensive, highly validated, and long-standing survey of demographic, attitudinal, and behavioral characteristics of American residents including thousands of different variables. Since 1972, the GSS has utilized increasingly sophisticated multi-stage national sampling strategies to interview an average of nearly 2000 adults in each year of the 30 years in which the survey was conducted. The cumulative number of interviews during this timeframe totals 59,599 [44]. This research is descriptive and attempts to explain changes in attitudes towards capital punishment related to a variety of socio-cultural factors.

Results
From 1972 to 2014, the General Social Survey included nearly 5600 variables and utilized five questions that measure some aspect of capital punishment: Does respondent favor capital punishment? (Only asked in 1972 and 1973) Favor or oppose death penalty for murder? Importance of death penalty issue to respondent? How much information does respondent have of the death penalty? And how firm is respondent's opinion of the death penalty? For the purposes of this research, the second question, Favor or oppose the death penalty for murder? is the dependent variable. This question has been tracked for the study period -1974 to 2014 -and is the most foundational question regarding capital punishment. Interestingly, basic opposition to capital punishment is significantly related to the importance of the death penalty issue (χ2[3, N=462] = 25.599, p=.000) and how much information the respondent has about the death penalty ((χ2[3, N=460] = 12.7454, p=.005). In the former instance, those who said that the death penalty was one of the most important issues were significantly more likely than expected (49.9 v. 55, respectively) to support the death penalty for murder while those who said that the death penalty was not an important issue were less likely than expected (9 v. 18.9) to support the death penalty.
In the later instance, those who said that they had all the information they needed about the death penalty were significantly more likely than expected (41 v. 37.7) to support the death penalty than those who said they had very little information (76 v. 89.8, respectively). Attitudes on the death penalty changed significantly over the 40 year period under review (χ2[4, N=9211] = 173.712, p=.000, for 10-year intervals). In 1974, 66.5% of those surveyed supported the death penalty. Support peaked in 1985 (79.5%), 1990 (79.4%), and 1994 (79.2%); since 1994, support has steadily declined to 64.3% in 2014. Figure 1, below, illustrates these trends. (Figure 1 favor death penalty for murder by years) Table 1, below, summarizes the relationship between the dependent variable and each of the independent variables explored in this research from 1974 to 2014. (Table 1 relationship between independent variables and dependent variable (favor or oppose death penalty for murder) for All Years 1974-2014). While, overall, age is significantly related to opinions about the death penalty (those under 30 and 70 and older were less likely than expected to favor the death penalty [11,369 v. 11,674.4, respectively], and those 30-69 were more likely than expected to favor the death penalty [23,563 v   The actual and expected cell counts in each of the study years (1974,1984,1994,2004, and 2014) followed the same pattern as the overall trend but were significantly different in only two years, 1974 and 2014. Support for all age groups has declined since 1994 except for those under 30, which has declined since 1984. (Figure 2 favor death penalty by age by survey years). The sex of the respondent was significantly related to attitudes towards the death penalty in each decade.
In each year of the study period, men were more likely that expected to favor the death penalty for murder (3139 v. 2952.9, respectively, for the total) while women were more likely than expected to oppose (1643 v. 1456.9, respectively, for the total). Support among both males and females have declined since 1994. (Figure 3 favor death penalty by sex by survey years). In each of the study years, marital status was significantly related to attitudes towards the death penalty. In each decade, respondents who were married were more likely to favor the death penalty for murder than expected (3724 v. 3546.4, respectively, for the total), while those who had never been married were less likely to favor the death penalty than expected (1242 v. 1379.3, respectively, for the total), except in 1984 when there was little difference between actual and expected counts.  There was very little difference between actual and expected counts for those who were widowed, divorced, or separated. Support for the death penalty for murder has declined for all marital statuses since 1994 except those who are separated and for whom support has increased between 2004 and 2014. (Figure 4 favor death penalty by marital status by survey years). Differences in views on the death penalty based on education level (highest degree earned) varied significantly in every year except 1974 when actual and expected cell counts were similar. In each of the other decades, those with less than a high school education, a bachelor's degree, or a graduate degree were less likely than expected to favor the death penalty (866 v. 941.6, 902 v. 923.9, and 405 v. 477.7, respectively, for 1984-2014). Those with a high school or associate's degree were more likely than expected to favor the death penalty (3064 v. 3290.0 and 383 v. 353.0, respectively, for 1984-2014). Opposition to the death penalty has increased among all education levels since 1994 except among those with a bachelor's degree, which started to increase in 1984, and among those with an associate's degree, which experienced an increase from 2004 to 2014. (Figure 5 favor death penalty by education level by survey years). Support for the death penalty varied significantly by the race of the respondent in each of the years under review. In each decade, Whites were more likely to support the death penalty than expected (5683 v. 5373.7, respectively, overall), though support has declined since 1994. Blacks were more likely to oppose than expected (613 v. 334.2, respectively, overall), though support has increased since 2004.  A third category used by the GSS, Other, was generally less likely to support the death penalty than expected (330 v. 360.5, respectively, overall). (Figure 6 favor death penalty by race by survey years). Overall, those earning less than $15,000 a year were less likely than expected to favor the death penalty (3619 v. 3713.2, respectively), while those earning $25,000 or more were more likely than expected to favor the death penalty (1687 v. 1642.3, respectively); those making between $15,000 and $25,000 showed very little difference between actual and expected counts. The differences were significant in 1974 (χ2 [ (Figure 7 favor death penalty by income by survey years). In each decade of the past 40 years, those who identify as Democrats have consistently and significantly opposed the death penalty more than expected (1246 v. 944.2, respectively, overall, p=.000), while Republicans have supported more than expected (1977 v. 1682.1, respectively, overall, p=.000).  Independents and Other political parties have not contributed to the level of significance, generally having actual and expected counts that are similar. Support for the death penalty among Democrats and Independents declined after 1994, and increased among Republicans and Other political parties after 2004. (Figure 8 favor death penalty by political party by survey years). Political views (Liberal, Moderate, and Conservative) are significantly related to political party (χ2[6, N=50,120] = 5350.485, p=.000), but weekly correlated (Φ=.394, p=.000). Those who consider themselves as liberal were less likely than expected to favor the death penalty (1415 v. 1702.3, respectively, overall), while conservatives were more likely than expected to favor the death penalty (2471 v. 2217.6, respectively, overall). Overall, moderates were slightly more likely to support the death penalty than expected (2483 v. 2441.9, respectively), but their support has declined since 1994. The significant differences between liberals and conservatives held true in each year of the study period. Overall, support for the death penalty decreased among liberals and moderates after 1994, but increased among Republicans after 2004. (Figure 9 favor death penalty by political views by survey years). Overall, religion is significantly related to opinions on the death penalty, though this was not the case in 1984 and 1994 when actual and expected counts were similar for all groups. In 1974, Catholics were more likely than expected to favor the death penalty, while Protestants, Jews, and None were less likely than expected.  . Support for the death penalty among the largest group of respondents, Baptists (32.8% of the total), declined after 1994 (when 78.6% were in favor, still fewer than expected) to 2014 (when only 67% favored, with actual and expected counts roughly equal). Those who said they belonged to other denomination (20.2% of the sample, second largest), were less likely than expected to consistently support the death penalty (749 v. 786.4, respectively, overall). Methodists, who accounted for 14.5% of respondents, also expressed less support for the death penalty after 1994, but were still more likely than expected to support the sanction overall (609 v. 569.1, respectively).   Both Methodists and Lutherans (10.0% of the total) were the most likely to support capital punished overall (76.2% and 76.1%, respectively). African Methodists, by far the smallest denomination (less than 1% of the total), had the lowest level of support for the death penalty (41.2% overall). (Figure 11 favor death penalty by denomination by survey years). Responses to the question "How fundamentalist are you currently?" were collapsed into three categories: fundamentalist, moderate, and liberal. Those who indicated that they were fundamentalist were overall more likely than expected to favor the death penalty (1911 v. 1898.9, respectively), as were moderates (2711 v. 2684.9, respectively). Liberals were more likely than expected to oppose the death penalty overall, 1722 v. 1760.2, respectively. And while fundamentalism is significantly related to attitudes towards the death penalty during the entire 40 years covered by this study, it was only significant in one of the sample years: 1974. In 1974, those who self-identified as moderates, were much more likely than expected to favor the death penalty (513 v. 471.1, respectively), while fundamentalists and liberals were less likely than expected for favor the death penalty. It is interesting to note that in 1974, 55.8% of respondents identified as moderate; in 2014, only 41.9% did so. Those identifying as liberal totaled 16   Another measure of religiosity is strength of one's affiliation to a particular religion, which is significantly related to attitudes on the death penalty in each decade. The level of support for the death penalty was strongest among those who said their affiliation was not very strong, and in each decade that support was greater than expected (2757 v. 2587.1, respectively). Those who have a strong affiliation and those who have no religious affiliation were less likely than expected to support the death penalty (2401 v. 2500.9 and 718 v. 797.0, respectively, overall). The actual and expected counts among those with a somewhat strong affiliation were close in value. Support for the death penalty declined since 1994 among all groups except those with no religious affiliation, which increased from 57.7% in 2004 to 61.3% in 2014. (Figure 13 favor death penalty by strength of affiliation by survey years). How often respondents attend church was significantly related to opinions on the death penalty in every year except 1974. Overall, there is a direct relationship between attendance, where those who attend most often are the most likely to oppose the death penalty; the one outlier is those who attend more than once a day have the highest level of support (though they only account for 0.7% of the total). Those who never attend, attend less than once a year to once a year, and those who attend several times a year to once a month are more likely than expected to favor the death penalty. Those who attend once a month to several times a month, and who attend nearly every week to every week are less likely than expected to favor the death penalty. (Figure 14 favor death penalty by church attendance by survey years).

Discussion
The death penalty is one of the most contentious social issues in the United States. While a majority of Americans continue to support the death penalty for murder, that support has steadily declined since 1994 except for a slight uptick in 2010. Since 1974, 19 states have abolished the death penalty, and 18 states are simply not applying the sanction (Death Penalty Information Center, 2017c). In addition, the number of people being sentenced to death is declining, from 211 in 2009-2010 to 83 in 2013 as are populations on death row [46]. In 2013, the average time between sentencing and execution was 186 months (15.5 years); that same year, 39 people were executed while 30 died of other causes awaiting execution [46]. This research filled two gaps in the existing literature on the relationship between various social factors and attitudes towards the death penalty. First, there is a lack of research that measures more than a few factors that influence opinions. This research examined 13 factors that are significantly related to attitudes on the death penalty. Second, there is a lack of longitudinal research on changes in attitudes and the factors that affect them.
This research evaluated those changes over a 40 year period using highly validated survey data that include approximately 50,000 respondents. Overall, the results of the multivariate analyses closely parallel those of the bivariate analysis of the trend in attitudes towards the death penalty from 1974-2014. After peaking in 1994, most of the values for the independent variables reflected the overall decline in support for the death penalty. There were, however, a few notable exceptions. From 2004 to 2014, support for the death penalty increased among Blacks, those who are separated from their spouse, those who completed junior college (1.5% of the sample in 1974, 6.1% in 2014), those who identify as Republican or Other, those who identify as conservative, those who identify as not religious (both in terms of affiliation and strength of affiliation), those who identify as either Episcopalian or Other denomination, and among those who pray more than once a day (less than 1% of the total sample). Future research will attempt to understand the degree to which these exceptions are interrelated and how they might affect future trends.

Conclusion
This research found that support for the death penalty is highest among Whites, males, respondents age 30-69, those who are married, have an associate's degree, and earn more than $25,000. In addition, support for the death penalty was strongest among those who identify as Republican, conservative, Protestant (Baptist in particular), fundamentalist, do not have a strong affiliation to their religion, and attend church infrequently. The relationship between religion and the death penalty is more complex than one might assume, however. While religion is often associated with conservative views on social issues support for the death penalty among most religions has declined since 1994. But another trend (for example, see Green & Kelso [45]) not examined in the current analysis is the overall decline in religiosity in the United States. For example, in 1974, 64.1% of respondents self-identified as Protestant and 25.7% as Catholic; in 2014, 44.6% of respondents self-identified as Protestant and 24% as Catholic. In 1974, 35.1% of respondents attended church once to twice a week; in 2014, that fell to 24%. While there is a close relationship between religion and attitudes towards capital punishment, the overall influence of religion in general, and on this issue in particular, has declined. Subsequent studies on this subject should examine the goodness of fit of the trends on attitudes on the death penalty and various social characteristics [47].