researched argument draft essay
Reeree
whatameicansthinkonguncontrol.pdf
What Americans Think About Gun Control: Evidence from the General Social Survey, 1972–2016
Steven V. Miller, Clemson University
Objective. Gun control is a classic case of policy gridlock and we commonly assume public opinion is at the foundation of this gridlock. However, public opinion analyses of attitudes about gun control often say little about the topic itself and do not fully leverage our long-running survey data to assess partisan, regional, and temporal trends in attitudes toward gun control. Methods. I use over 26 waves of General Social Survey data from 1972 to 2016 to analyze the main public opinion cleavages (partisanship, urban/rural distinctions, and Census regions) of gun control. Results. I find that partisanship and ruralness are not robust predictors of attitudes about gun control and that partisan polarization is only partial and recent. Further assumptions about regional variation in attitudes toward gun control need reevaluation. Conclusion. Gun control policy gridlock says more about polarization at the elite level than at the mass level. Future research can also do well to assess issue-linkage concerns on specific gun control policy measures.
Gun control is one of the most controversial public policy issues in the United States. Its salience has only increased after recent mass shootings in places like the Pulse night club and the Inland Regional Center in San Bernardino. The San Bernardino incident in particular led to a strong reaction from some leading newspapers trying to push this debate into legislative action. The New York Daily News ran a front-page cover that engaged in deliberate issue linkage of the San Bernardino incident with other semiautomatic rampages in Aurora, Colorado Springs, and Newtown, labeling all perpetrators as “terrorists” (New York Daily News, 2015). The New York Times even dusted off a device it had not used in 95 years—the front-page editorial—to argue that legislative unwillingness to push sensible gun control legislation into law is a “moral outrage and national disgrace” (The Editorial Board, 2015). Successive incidents since, prominently the October 2017 mass shooting in Las Vegas that now ranks as the deadliest individual act of mass murder in U.S. history, only raise more questions about our gun laws and the salience of this controversial public policy debate. The demonstrations by the New York Daily News and New York Times still resonate in light of more current events.
The New York Daily News and New York Times made these demonstrations because gun control legislation is a nonstarter in Congress. The day after the San Bernardino incident, Senate Republicans voted down legislation that would have closed background check loopholes and banned subjects on the terror watch list from purchasing firearms (Joseph, 2015). Republican opposition was almost unanimous, which conforms to the preferences of the National Rifle Association (NRA) on this issue. The implication here is that Republican opposition to any gun control legislation follows the direct preferences of its core interest
∗Direct correspondence to Steven V. Miller, Clemson University, Department of Political Science, 232 Brackett Hall, Clemson, SC 29634-1354 〈[email protected]〉. SOCIAL SCIENCE QUARTERLY, Volume 100, Number 1, February 2019 C© 2018 by the Southwestern Social Science Association DOI: 10.1111/ssqu.12555
What Americans Think About Gun Control 273
groups and its constituents. This speaks well to political science scholarship that says public opinion (e.g., Page and Shapiro, 1983), especially on salient issues (Burstein, 2003), drives public policy in the United States. Gridlock may best define gun control policy of the past several decades and public opinion is the foundation of this gridlock. However, much of what we know about the public opinion of gun control is incomplete. The studies that most inform us about the public opinion of gun control rely on weak data with limited spatial or temporal scope or are more interested in the policy issue as an extension of other topics in political science (e.g., elite messaging, information environments). None leverage our longest-running American survey data sets to fully explore the trends in public opinion of gun control that otherwise inform our intuitions.
In this article, I report an analysis of a 44-year span of General Social Survey (GSS) data to fully explore the public opinion of gun control. I take care to relate the findings to the important cleavages that a review of the literature will emphasize: partisanship, urban/rural divides in America, and regional variation across the Midwest, Northeast, South, and West. My findings tell a mixed story about these cleavages and offer clarifications of conventional wisdom. I find no robust differences between rural and urban respondents in their opposition to gun control. I also find that increasing Republican partisanship does not robustly reduce support for gun control and that most Republicans actually support gun control. Gun control is not necessarily a partisan issue over time even if it has quickly become one. I also find that the regional variation in attitudes about gun control belie simple statements that emphasize opposition in the South and West. I conclude that there is still substantial support for gun control legislation and a lot of conventional wisdom about gun control merits reevaluation.
Gridlock and What We Know About the Public Opinion of Gun Control
Academic research about gun control policy in the United States spans social science disciplines as diverse as criminal justice, political science, and sociology, but shares two recurring themes. The first is an identification that gun control policy is a classic case of gridlock. Spitzer’s (2015a) summary notes that gun control policy, for the most part, still hews to maximalist interpretations of the Second Amendment to the U.S. Constitution as it applied to the immediate postindependence of the country. The 13 newly independent states, which all hugged the Atlantic Coast and were surrounded by territories held by hostile foreign powers (prominently Britain), saw this amendment to its new constitutional charter as useful for the mobilization of a militia in the event of another invasion of foreign forces. This does not suggest the United States has never had substantively important gun control movements succeed in the past. Indeed, gun control regulations are as old as the colonial United States itself. The exigencies of Prohibition-era gang violence saw a slew of gun control initiatives passed by state governments and Congress that especially targeted automatic weapons these gangs used. However, these historical facts, and even the state- by-state variation in states as diverse as New York and Wyoming, belie that gun control policy and interpretations of the constitutionality of gun control are still tightly tethered to the country’s independence even as the United States today barely resembles the United States of 1787 in almost every other respect.
Spitzer (2015b) explains what the cycle of gridlock has resembled of late. It starts with a dramatic event that generates widespread public outrage, for which examples that highlight the gridlock problem span from the 1989 Cleveland Elementary School shooting in California to more recent mass shootings in places such as Newtown, Colorado Springs,
274 Social Science Quarterly
San Bernardino, Orlando, and Parkland (among several others). Popular outrage leads to calls from gun control proponents to pass stronger gun control laws, ultimately articulating economic language that cheap gun supply decreases the costs associated with perpetrating large-scale violence and gun control intervention would raise the costs (and decrease the likelihood) of these events (cf. Wintemute, 2005; Bartley and Williams, 2018). However, the reaction from anti-gun-control groups, prominently the NRA, blocks almost all of these initiatives. Even the first Bush Administration found it difficult to reconcile how banning military-style rifles, such as the one used by Patrick Purdy in the 1989 Cleveland Elementary School shooting, would spare sporting rifles of interest to the members of the NRA and the Republican voting bloc. The Bureau of Alcohol, Tobacco, and Firearms (ATF) later imposed a ban on “assault weapons” after this incident, which Congress codified into the 1994 Federal Assault Weapons Ban. This had a 10-year window that expired in 2004 and Congress has yet to pass another similar policy in light of more recent shootings from Columbine to Majory Stoneman Douglas. The policy gridlock story is ultimately one of outrage, action, and a reaction that blocks changes to the gun control status quo.
The theme of gridlock implies a second shared theme of gun control research. Gridlock may define American gun policy but mass-level public opinion is at the foundation of this policy inertia. Spitzer’s (2015b) gridlock cycle begins with a dramatic event that stokes mass-level public opinion for increased gun control measures to prevent the recurrence of these events. The reaction from gun control opponents siphons the public opinion of gun rights activists through groups, such as the NRA, to lobby state and national politicians to oppose gun control measures. Policy gridlock follows in an American political system with a myriad of institutional barriers to legislation, but mass-level public opinion on gun control is the foundation on which this gridlock operates. A review of research that speaks to this will identify three main cleavages by which we can understand how public opinion influences policy gridlock on gun control (Patterson and Eakins, 1998).
Partisanship might be the biggest and most obvious component of the public opinion foundation to gun control policy gridlock. Gun rights activists today have found a home in the Republican Party. They serve as a core component of the Republican base given the extent to which the NRA has successfully connected gun rights issues to the overall “conservative” movement that unifies a Republican (GOP) voting base that may otherwise have opposing interests on other core GOP policies (Horwitz and Anderson, 2009). Likewise, gun control advocates have increasingly found a sympathetic audience with the Democratic Party after gun owners and gun rights activists sorted into the GOP. The partisan divide on gun control is evident when Senate Democrats react to events such as the San Bernardino shooting by introducing legislation that would have closed background check loopholes, which Senate Republicans unanimously rejected. The implication is that partisanship at the mass level is also a major public opinion barrier to gun control legislation. Existing studies on the public opinion of gun control routinely find partisan differences between self-identified Republicans and Democrats that provide further intuition to this argument (e.g., Haider-Markel and Joslyn, 2001; Pearson-Merkowitz and Dyck, 2017).
Other public opinion components of gun control policy gridlock speak more to the distribution of gun control attitudes than the ideological or partisan attitudes that may influence them. Here, research has generally found two different but related components of gun control public opinion. The first is a clear urban/rural divide in which rural Americans oppose gun control measures, while respondents in cities are more open to gun control measures, all things considered (e.g., Erskine, 1972; Boor and Bair, 1990). Intuitively, rural Americans, especially those with farm or livestock, are more likely to have firearms
What Americans Think About Gun Control 275
than city dwellers and are more likely to have practical uses for them (e.g., pest control or euthanizing sick livestock). They are thus more likely to oppose gun control measures that they interpret as infringing on their interests (Wolpert and Gimpel, 1998).
Regional variation constitutes the third cleavage in the public opinion on gun control and the policy gridlock that follows. Opposition to gun control clusters more in the South and West, certainly in relation to the Northeast. Arguments for why this regional variation emerges highlight the symbolic importance of guns to “southernness” and southern culture, even connecting the proliferation of guns in the region and opposition to gun control to traditions that emerged proximate to the Civil War (e.g., Dixon and Lizotte, 1987; Brennan, Lizotte, and McDowall, 1993). Arguments about opposition to gun control in the West say less about history and culture and tend to focus more on guns as recreational activity in an area of the United States more rural than most of the Northeast (cf. Gimpel, 1998; Wolpert and Gimpel, 1998). Despite major metropolitan areas in states on the Pacific Ocean, the West Census region features states more like Idaho in the Mountain division than California in the Pacific division. The implications are that gridlock follows in a political system that emphasizes rural votes and rural states more than votes in urban areas and more populated states (Page and Gillens, 2017) with the U.S. Senate being the clearest illustration of this disproportionality. Thus, the public opinion of gun control policy gridlock is as much a function of the distribution of gun control opposition in a political system like the United States.
Gun control policy is a case of policy gridlock and a review of scholarship from fields such as criminal justice, political science, and sociology will do well to capture the public opinion cleavages behind this gridlock. However, two limitations emerge for a general public and academic community that wants to know more about the public opinion of gun control. First, political scientists interested in this topic will note that analyses that focus on gun control are not much interested in the problem of gun control itself. Instead, analyses on the public opinion of gun control in political science treat this issue as an extension or new testing grounds of other hypotheses or topics. Haider-Markel and Joslyn (2001) and Pearson-Merkowitz and Dyck (2017) use the topic of gun control as the context for original surveys on broader topics of issue frames and how partisanship distorts information received from a respondent’s geographic or social environment. Filindra and Kaplan (2016) do well to examine attitudes about gun control but are mostly interested in a hypothesized effect of racial resentment on the gun policy preferences of white Americans. The results that follow are informative and help us locate some of the public opinion cleavages of gun control, but the analyses speak to other topics and concerns in political science. Gimpel (1998) and Wolpert and Gimpel (1998) also help us understand the fault lines in the public opinion on gun control but the main topic of interest in both analyses is the intersection of gun control between self-interest calculations and symbolic politics (cf. Sears et al., 1980). This comment is not intended to downplay these contributions to what we know about the public opinion of gun control. However, these works are speaking to different concerns. The analyses they provide are not exhaustive in the broad correlates and trends of the public opinion of gun control.
Second, the data used to illuminate the public opinion of gun control are often weak, relying on peculiar samples that are limited in spatial or temporal scope. It is not uncommon for scholarship, especially in criminal justice or sociology, to rely on convenience samples of college students (e.g., Payne and Riedel, 2002), opinion surveys from the NRA that may be as old as 20 years before the publication of the analysis (e.g., Brennan, Lizotte, and McDowall, 1993), cross-sectional samples that do not speak to temporal trends (e.g., Gimpel, 1998; Pearson-Merkowitz and Dyck, 2017; Wozniak, 2017), or samples from
276 Social Science Quarterly
the author’s state from which it may be difficult to generalize nationally (e.g., Lizotte and Bordua, 1980; Haider-Markel and Joslyn, 2001). Few leverage our long-running nationally representative survey data sets that do feature questions on gun control. Erskine’s (1972) analysis was impressive in scope for the time, but her trends end in the year of publication and well before gun control emerged as a major partisan issue. Likewise, Wolpert and Gimpel (1998) do well to leverage multiple samples at multiple points in time, but it should be unsurprising that their analysis, rooted largely in survey data from the 1980s, does not find strong evidence of the partisan rifts on gun control that emerged at the time of publication in 1998.
There is real value in the empirical analyses that I propose to conduct, which incorporate GSS data from 1972 to 2016 to uncover and elaborate the broad correlates of public opinion of gun control in the United States. Such an analysis would clarify much of our conventional wisdom about the public opinion of gun control at an important time in which the issue is one of the most salient policy issues in the United States. The next section will briefly describe the data and the methods I use before proceeding to full analyses of the data.
Research Design and Methods
I use GSS data for all available years from 1972 to 2016. The primary question of interest—whether the respondent favors or opposes a law that would require a person to obtain a police permit before he or she could buy a gun—is the most ubiquitous gun control item in the GSS. Its utility for the sake of this analysis is multiple. One, it appears in almost all of the available GSS waves from 1972 to 2016. Other gun control items that appear in GSS are one-time questions. Two, the substance of the gun control question that GSS asks is subtly aggressive. It prompts the respondent to ask whether state approval for a gun of any kind is necessary when gun control opponents tend to cite precisely this kind of state overreach as justification for a maximalist interpretation of the Second Amendment to the U.S. Constitution. Thus, gun control advocates and interested lawmakers in Washington almost never express a gun control proposal like this, instead preferring to massage objections from gun control opponents with policies that target gun show loopholes, terror watch lists, or military-style rifles like the AR-15. GSS respondents that favor this form of gun control are signaling their willingness to accept other aggressive gun control measures. I keep the binary nature of this variable where a 1 indicates support for this form gun control.
I also select data from the 2006 wave that features more granular questions about gun control policies. These include background checks for private sales, tougher penalties for illegal gun sales than drug sales, limiting semiautomatic weapons to the police and military, making it illegal to carry a gun while drunk, and support for tougher gun control legislation as a function of the 9/11 terrorist attacks. These are binary variables that assume a value of 1 for respondents that favor a more aggressive gun control proposal.
I am testing three main explanatory factors of interest. We commonly assume that the public opinion of gun control policy gridlock is a function of partisan fault lines and the distribution of public opinion against gun control in rural areas and the South and West. Stated more formally, I am evaluating the following hypotheses.
Hypothesis 1: Respondents who self-identify as more Republican are less likely to support gun control than respondents who self-identify as more independent or as more Democratic.
What Americans Think About Gun Control 277
Hypothesis 2: Respondents who live in rural areas are less likely to support gun control than respondents who live in urban areas.
Hypothesis 3: Respondents who live in the South or West are less likely to support gun control than respondents who live in the Midwest or Northeast.
I code these variables in the following way. First, I code partisanship on the familiar seven-point scale, where a respondent places her/his political affinities from a continuum of “strong Democrat” to “strong Republican.” Higher values communicate more partisanship with the (Republican Party). I code whether the respondent lives in a rural area by whether the GSS data code the respondent as living in a town of fewer than 2,500 people or whether he or she lives in an open country. Respondents who live in a town of more than 2,500 people are coded as not rural or “urban” by this understanding , which squares with how the U.S. Census Bureau outlined these terms in the 2010 Census.1 Finally, the GSS codes the condensed Census region—Midwest, Northeast, South, or West—in which the respondent lives. I include this in my analysis, though I will model these as random effects in the mixed effects modeling framework to better assess differences among regions. This has the benefit of allowing for more flexibility in modeling variation of different coefficients by these categories (i.e., allowing varying slopes for partisanship by region) and avoids the awkward implementation of selectively choosing a region to be baseline comparison in a fixed effects framework. Figures 4–6 will highlight the benefit of this approach.
I include additional control variables for the respondent’s age (in years), sex (a dummy for women), whether the respondent has a college education, fixed effects for race (leaving whites as the baseline), and whether the respondent has a gun in the household.
All dependent variables are binary, making the use of a logistic regression appropriate. However, there is important spatial and temporal heterogeneity across the 26 waves of GSS data that I use in my analysis. I account for this with mixed effects models with random effects for the condensed Census region in which the respondent resides and, for the primary dependent variable, the year of the survey. I allow the effect of partisan- ship to vary by year and condensed Census region. I also scale all nonbinary indepen- dent variables by two standard deviations to ensure rough comparability of coefficients (Gelman, 2008).
Results
My analyses of 26 waves of GSS data over 44 years suggest four interesting findings about public opinion and the gun control debate. First, partisanship with the Republican Party and ruralness do not robustly reduce support for gun control. Second, most Republicans actually support gun control, even in the current polarized political climate. Third, gun control is not necessarily a partisan issue but it is quickly becoming one. Finally, simple assumptions about regional variation mask important and more intriguing variation in attitudes toward gun control in the data.
Partisanship and Ruralness Do Not Robustly Reduce Support for Gun Control
Table 1 summarizes the results of six mixed effects regression models on the dependent variables I coded from the GSS data. The results yield few predictors with consistent effects
1〈https://www.census.gov/geo/reference/ua/uafaq.html 〉.
278 Social Science Quarterly
TA B
L E
1
M ix
e d
E ff e
ct s
M o
d e
ls o
f A
tt itu
d e
s To
w a
rd G
u n
C o
n tr
o l
R e
q u
ir e
P o
lic e
P e
rm it
B a
c kg
ro u
n d
C h
e c
k fo
r P
ri va
te S
a le
s To
u g
h e
r P
e n
a lti
e s
th a
n D
ru g
s L im
it S
e m
ia u
to m
a tic
s Ill
e g
a lt
o C
a rr
y G
u n
W h
ile D
ru n
k To
u g
h e
r G
u n
C o
n tr
o l
A ft
e r
9 /1
1
M o
d e
l1 M
o d
e l2
M o
d e
l3 M
o d
e l4
M o
d e
l5 M
o d
e l6
A g
e 0
.1 1
3 ∗
−0 .0
5 5
0 .3
5 8
∗ 1
.2 8
7 ∗
0 .1
7 7
0 .0
7 6
(0 .0
2 7
) (0
.1 8
9 )
(0 .1
5 2
) (0
.2 6
0 )
(0 .2
6 9
) (0
.2 2
4 )
C o
lle g
e e
d u
c a
te d
0 .3
1 5
∗ 0
.5 4
6 ∗
−0 .2
2 3
0 .3
7 9
0 .2
6 3
0 .0
7 7
(0 .0
3 5
) (0
.2 3
4 )
(0 .1
7 0
) (0
.2 8
7 )
(0 .3
2 3
) (0
.2 5
6 )
F e
m a
le 0
.6 5
0 ∗
0 .9
6 2
∗ 0
.1 1
3 2
.0 7
2 ∗
0 .6
3 7
∗ 1 .2
2 6
∗
(0 .0
2 7
) (0
.1 8
6 )
(0 .1
5 0
) (0
.2 6
6 )
(0 .2
5 7
) (0
.2 1
8 )
B la
c k
0 .0
8 4
−0 .2
8 9
0 .8
5 5
∗ 0
.5 1
6 −0
.5 9
2 −0
.2 1
4 (0
.0 4
4 )
(0 .2
9 5
) (0
.2 4
4 )
(0 .4
4 2
) (0
.3 9
7 )
(0 .3
8 1
) O
th e
r ra
c e
(n o
t w
h ite
) 0
.3 4
5 ∗
−0 .2
5 8
0 .8
2 6
∗ 0
.9 8
2 ∗
−0 .4
7 9
0 .3
1 0
(0 .0
7 1
) (0
.3 3
4 )
(0 .2
7 9
) (0
.4 7
3 )
(0 .4
2 4
) (0
.4 2
7 )
R u
ra l
−0 .3
9 4
∗ −0
.3 9
9 −0
.4 8
4 ∗
−0 .4
9 8
−0 .6
6 1
−0 .2
3 5
(0 .0
3 6
) (0
.2 4
5 )
(0 .2
1 6
) (0
.3 0
0 )
(0 .3
4 4
) (0
.2 8
8 )
P a
rt y
ID (D
to R
) −0
.4 0
5 ∗
−0 .3
0 7
0 .0
6 7
−0 .3
8 4
−0 .3
1 4
−0 .5
8 5
∗
(0 .0
6 5
) (0
.2 0
7 )
(0 .1
7 1
) (0
.2 5
3 )
(0 .3
3 9
) (0
.2 7
4 )
G u
n in
th e
h o
u se
h o
ld −0
.9 1
6 ∗
−0 .4
7 6
∗ −0
.5 4
8 ∗
−1 .2
2 5
∗ 0
.0 1
7 −1
.1 7
3 ∗
(0 .0
2 8
) (0
.1 9
4 )
(0 .1
6 2
) (0
.2 4
8 )
(0 .2
8 2
) (0
.2 2
3 )
N 3
5 ,9
7 2
8 3
8 8
3 5
8 3
2 8
4 2
7 9
0
∗ p <
0 .0
5 .
R a n d
o m
e ff e c t
p a ra
m e te
rs (i .e
., sl
o p
e a n d
in te
rc e p
t st
a n d
a rd
d e vi
a tio
n s)
o m
itt e d
fo r
p re
se n ta
tio n .
What Americans Think About Gun Control 279
across most estimations. The effect of having a gun in the household has a negative and significant effect in five of six estimations, with attitudes toward making it illegal to carry a gun while drunk as the lone exception. Generally, respondents with guns in their household (i.e., gun owners) are less likely to support gun control measures they see as infringing on their material interest, consistent with intuition from Wolpert and Gimpel (1998). Further, women are more likely to support gun control measures than men. Imposing stiffer penalties for illegal gun sales than illegal drug sales is the only gun control policy measure for which there was no discernible difference between men and women.
However, our go-to explanations for how public opinion at the mass level explains policy gridlock at the elite level do not enjoy similar levels of support. The partisanship–gun control relationship is not as robust at the individual level as we infer it to be at the elite level. Partisan differences appear in just two of the six estimations in Table 1. Increasing partisanship with the GOP coincides with less support for requiring a police permit before purchasing a gun and for making gun control laws tougher as a function of the 9/11 terror attacks. Likewise, the effect of a respondent living in a rural area is negative and significant in just two estimations, decreasing support for requiring a police permit to purchase guns and decreasing support for imposing stiffer penalties for illegal gun sales than illegal drug sales.
A “secret weapon” plot will communicate a similar story. “Secret weapon” is shorthand for a dot-and-whisker plot that summarizes multiple models that can better visualize time trends that would be lost if the data were pooled or even partially pooled (i.e., as a mixed effects model) (Gelman, 2005). Here, this means running a series of models for each available year in the GSS data on the measure about requiring police permits for gun purchases and reporting the results as faceted plots.2
Figure 1 reports these results and annotates each facet with the percentage of all regressions in which the particular coefficient had a statistically significant result. The results show clear and discernible effects for women and having a gun in the household for all 26 analyses. Every regression produced statistically significant effects for these two coefficients. Note that the effects of a respondent living in a rural area and increasing GOP partisanship produce negative and statistically significant results in just over half of all estimations. In the case of GOP partisanship, negative and discernible effects start to emerge only in the 1990s. The effect of living in a rural area is more reliably negative and significant in earlier survey waves than in later survey waves, especially in the surveys in the 1990s.
Partisanship and urban/rural divides constitute the two main cleavages in the public opinion of gun control that purport to explain gun policy gridlock at the elite level. However, Table 1 and Figure 1 show that the effects are not robust across multiple forms of gun control policies, nor are the effects consistent across time in the GSS data.
Most Republicans Actually Support Gun Control
Perhaps the most counterintuitive finding from survey data on attitudes toward gun control is that Republicans generally favor gun control legislation. This is not an anomaly unique to a specific year, nor necessarily a recent trend. There is not a year in the GSS data
2The variable for other nonwhite and nonblack races had separation problems in the analyses in 1974, 1977, and 1980. Intuitively, there were only 7, 15, and 10 nonwhite/nonblack respondents in these respective years, creating estimation problems. I dropped the variable from those years. The question about having a gun in the home first appears in 1973, which explains why estimations that include that covariate must drop the 1972 wave from the analysis.
280 Social Science Quarterly
FIGURE 1
A “Secret Weapon” Plot of the Correlates of Gun Control Support (GSS, 1972–2016)
% p < .05: 22.22%
% p < .05: 3.7%
% p < .05: 51.85%
% p < .05: 44.44%
% p < .05: 16.67%
% p < .05: 100%
% p < .05: 100%
% p < .05: 55.56%
Party ID (D to R) Gun in the Household
Black Other Race (Not White) Rural
Age College Educated Female
19 72 19
73 19
74 19
76 19
77 19
80 19
82 19
84 19
85 19
87 19
88 19
89 19
90 19
91 19
93 19
94 19
96 19
98 20
00 20
02 20
04 20
06 20
08 20
10 20
12 20
14 20
16 19
72 19
73 19
74 19
76 19
77 19
80 19
82 19
84 19
85 19
87 19
88 19
89 19
90 19
91 19
93 19
94 19
96 19
98 20
00 20
02 20
04 20
06 20
08 20
10 20
12 20
14 20
16
19 72 19
73 19
74 19
76 19
77 19
80 19
82 19
84 19
85 19
87 19
88 19
89 19
90 19
91 19
93 19
94 19
96 19
98 20
00 20
02 20
04 20
06 20
08 20
10 20
12 20
14 20
16
−2
−1
0
1
2
−2
−1
0
1
2
−2
−1
0
1
2
C o e ff ic
ie n t E
st im
a te
DATA: General Social Survey, 1972−2016.
NOTES: Other race (not white) variable dropped from 1974, 1977, and 1980 analyses for separation concerns. Gun ownership question was not asked in 1972. Annotations in each facet communicate percentage
of regressions with significant effects for each unique coefficient. Random effect parameters (i.e., slope and intercept standard deviations for condensed Census region) omitted for presentation.
in which support among strong Republicans for a law that would require a police permit for a gun purchase is below 50 percent.
Figure 2 is a time-series line chart of the percentage of responses that favor a law requiring a police permit before a person could buy a gun by partisan identification over all waves from 1972 to 2016 in which that item appears. For clarity, I subset the line chart to just the strong Democrats, the independents who do not lean toward either party, and the strong Republicans. The line chart shows that the nadirs in support for gun control among strong Republicans emerge in the most recent waves; 54.62 percent of strong Republicans favored requiring a police permit for a gun purchase in the 2012 wave. Support among the strong Republicans actually increased almost 5 percentage points to 59.54 percent in the two years after the 2012 survey, though fell again to 50.5 percent in 2016.
Figure 2 does show some other interesting trends over the years. Strong Democrats have become only a little more supportive of gun control over time; 75.16 percent of strong Democrats favored this gun control measure in 1972, which increased to 83.44 percent in 2016. Independents who do not lean toward one party have not changed over time. In 1972, support for this form of gun control was 69.23 percent among those independents, relative to 70.57 percent in 2016. Interestingly, strong Republicans tracked quite closely with strong Democrats before 1994. The year of the “Republican Revolution,” in which
What Americans Think About Gun Control 281
FIGURE 2
Support for Gun Control Among Partisans Over Time
83.44
70.57
50.56
0
10
20
30
40
50
60
70
80
90
100
1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 2016
Year
P e
rc e
n t
S u
p p
o rt
f o
r P
o lic
e P
e rm
it B
e fo
re O
w n
in g
a G
u n
Party ID a a aStrong Republican Independent Strong Democrat
DATA: General Social Survey (1972−2016).
TABLE 2
Expected Values of Support for Police Permits for Gun Purchases Among Strong Republicans
Category Expected Value 95% Interval
Female, does not own gun 0.843 (0.798, 0.882) Male, does not own gun 0.738 (0.672, 0.796) Female, gun owner 0.684 (0.611, 0.750) Male, gun owner 0.531 (0.451, 0.610) Male, gun owner, college educated 0.608 (0.529, 0.681)
the GOP took both chambers of Congress for the first time since 1953, saw support for this form of gun control among strong Republicans fall from 82.4 percent in 1993 to 67.77 percent in 1994.
Figure 2 clarifies the results from Model 1, telling us that strong Republicans are less likely to support gun control but this should not be confused as equivalent to a statement that strong Republicans are likely to “oppose” gun control. Table 2 tells a similar story with simulations of the regression from a multivariate normal distribution to generate quantities of interest. I set the explanatory variables at their typical values (i.e., urban white respondents around age 51 without a college education), set the party ID variable at its maximum (i.e., strong Republicans), and allow the gun ownership and gender variable to vary. The simulations show the point estimates for expected value of support for police permits for gun purchases, given the explanatory variables (e.g., women/men, strong GOP, owns gun/does not own a gun) are above 0.500 in every application. Further, the 95 percent interval for all expected values from the simulations are above 0.500 in all but one category: men with strong GOP affiliation, no college education, and with a gun in the home. However, one adjustment to that class of respondent—the presence of a college
282 Social Science Quarterly
FIGURE 3
Support for Gun Control by Year and Partisanship
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
Year
S im
u la
te d
P ro
b a b
ili ty
o f S
u p
p o rt
f o r
R e q
u ir
in g
a P
o lic
e P
e rm
it B
e fo
re O
b ta
in in
g a
G u n
Party ID Strong Republican Independent Strong Democrat
DATA: General Social Survey (1973−2016).
diploma—increases support for police permits before purchasing a gun to 0.608 with a 95 percent distribution of expected values that are above 0.500.
This is an important pedagogical lesson for pundits, scholars, and interested citizens who want to understand variation in public opinion about the gun control debate. Model 1 in Table 1 shows a negative relationship between increasing GOP partisanship and support for a police permit prior to owning a gun that is discernible from a zero relationship. In other words, increasing partisanship with the Republican Party decreases support for this form of gun control relative to those with lesser affinity for the GOP (including strong Democrats). The reader can mostly see this story in Figure 2. However, it would be a mistake, both statistical and substantive, to assume that statement is equivalent to “increasing GOP partisanship makes a respondent oppose this form of gun control.” This would not be true in this case. Strong Republicans still support this form of gun control, though strong Democrats support it more.
Gun Control Is Not Necessarily a Partisan Issue, But it Is Becoming One
Most Republicans are on board with gun control measures. However, we are observing a growing rift between Democrats and Republicans on this issue that seems to have increased during the Obama Administration.
Figure 3 shows more simulations of the dependent variable from a draw of a multivariate normal distribution for three different values of partisanship—strong Republicans, strong Democrats, and pure independents—across all survey waves with all other explanatory variables held at their typical value. The results show partisan differences first emerging as a result of the Republican Revolution of 1994, but show increasing polarization on this issue that seems to occur almost entirely during the Obama Administration. This coincides with prominent events like the spike in gun sales after Obama’s 2008 election (Bohn, 2008) and cues from the NRA that Obama was the most “anti-gun candidate ever” who will “take
What Americans Think About Gun Control 283
your guns away” (Smith, 2008). On average, Republicans are less likely to support this gun control measure than those whose political affinities gravitate more to the Democratic Party. However, Figure 3 shows that the effect of GOP partisanship is more pronounced in recent years than it is overall, underscoring similar trends in Figure 1 on how partisanship emerges as a negative and significant predictor of attitudes against gun control in waves starting after the Republican Revolution of 1994.
Most Republicans actually support gun control measures and the issue at stake is not necessarily as polarizing at the mass level as it is at the elite level. However, it is quickly becoming a partisan issue among the electorate.
The Partisan, Regional, and Temporal Variation in Support for Gun Control
Explanations for attitudes about gun control across regions in the United States tend to focus on the South and West. After all, guns purportedly play an outsized role in “southernness” and southern culture (e.g., Brennan, Lizotte, and McDowall, 1993) and the rural nature of most of the West region of the United States leads citizens in this region to view guns as essential for farm and livestock matters or for recreation (e.g., Wolpert and Gimpel, 1998). This may explain part of the partisan sorting since the 1990s; Republicans are better represented in most state-level and national-level offices in both regions as a whole. Further, the interest is invariably about these two regions relative to the Midwest and Northeast, which do not command as much attention by comparison. Figure 4 underscores how we commonly think about this regional variation in gun control, using statistical simulations to show that the typical white male in the 2016 survey wave of average age, partisanship, and education levels, who does not live in a rural area, and does not have a gun is less likely to favor gun control if he lives in the South or West than if he lives in the Northeast. There is no difference between the Midwest and the other three regions, but clear differences emerge comparing the South and West to the baseline of the Northeast.
Simple assessments of regional variation like this would belie the more interesting varia- tion in attitudes toward gun control across these four regions and by partisanship. Figure 5 summarizes simulations from a split of Model 1 in Table 1, subsetting the analyses of Model 1 in Table 1 to the survey waves before the “Republican Revolution” (1973–1993) and after it (1994–2016). Thereafter, I generate simulations with confidence intervals for varying
FIGURE 4
Support for Gun Control by Region in 2016
Midwest
Northeast
South
West
0.60 0.65 0.70 0.75 0.80
Simulated Probability (with 95% Intervals) of Support for a Police Permit to Purchase a Gun
NOTE: Quantities of interest are for white men living in urban areas, without a college education or gun in the home, with the average age and partisanship.
284 Social Science Quarterly
FIGURE 5
Support for Gun Control by Partisanship, Before and After the Republican Revolution
Strong Democrat
Weak Democrat
Ind., lean Democrat
Independent Ind., Lean Republican
Weak Republican
Strong Republican
Strong Democrat
Weak Democrat
Ind., lean Democrat
Independent Ind., Lean Republican
Weak Republican
Strong Republican
0.6
0.7
0.8
0.9
0.6
0.7
0.8
0.9
S im
u la
te d
P ro
b a b
ili ty
o f S
u p
p o rt
f o r
R e q
u ir
in g
a P
o lic
e P
e rm
it B
e fo
re O
b ta
in in
g a
G u n
Period Post−Republican Revolution Pre−Republican Revolution
DATA: General Social Survey (1973−2016).
levels of partisanship across the four Census regions in the model with all other explanatory variables held at their typical value. Figure 5 does not suggest any meaningful variation in attitudes toward gun control in the Midwest or the West as a result of the sorting of the electorate in 1994. Instead, we see two interesting and discernible differences in the Northeast and South. Republicans in the Northeast are less likely to support gun control after the Republican Revolution than Republicans in the Northeast before it. Further, we see a small but discernible difference in the South, though it involves the strong Democrats and not the strong Republicans. Strong Democrats in the South became more likely to support gun control after the Republican Revolution than strong Democrats before it, which would be consistent with a sorting hypothesis that emphasizes that the Republican Revolution changed the composition of the Democratic voting bloc as much as it changed the composition of the GOP (cf. Levendusky, 2009).
Figure 6 generates simulations of the likelihood of supporting mandatory police permits before obtaining a gun by condensed Census region and three values of partisanship: strong Democrats, strong Republicans, and pure independents who say they do not lean either way. The results are consistent with the previous discussion that suggests even strong Republicans support gun control and that gun control’s polarization on partisan lines is only recent. The South stands out because the overlap among strong Republicans, pure independents, and strong Democrats has the region better resemble the Northeast in more recent survey waves. Differences among partisans in both regions are not as discernible as simulations for the Midwest and West in the most recent survey waves. Those simulations in the Midwest and West suggest differences among all three partisan groups, with the widest differences among groups appearing in the West.
What Americans Think About Gun Control 285
FIGURE 6
Support for Gun Control by Partisanship and Region
1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 2016 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 2016
0.5
0.6
0.7
0.8
0.9
0.5
0.6
0.7
0.8
0.9
Year
S im
u la
te d
P ro
b a
b ili
ty o
f S
u p
p o
rt f o
r R
e q
u ir
in g
a P
o lic
e P
e rm
it B
e fo
re O
b ta
in in
g a
G u
n
Party ID Strong Republican Independent Strong Democrat
DATA: General Social Survey (1973−2016).
All told, there is clear regional variation in attitudes toward gun control by different values of partisanship and over time. However, the variation does not fit a simple pattern we otherwise casually assume. Strong Democrats in the South—and not the region’s strong Republicans, per se—observe a change in attitudes toward gun control after the Republican Revolution. Changes in attitudes toward gun control in the Northeast after that sorting of the electorate were more about changes among Republicans than Democrats. More recent waves have seen significant partisan sorting on attitudes toward gun control, though less so in the Northeast and South.
Conclusion
Gun control policy is arguably the most salient public policy issue in the United States today. More recent mass shootings from Newtown, Connecticut to Parkland, Florida only raise the stakes associated with this policy issue and have even led to unconventional advocacy from national news outlets like the New York Daily News and the New York Times. This increasing salience and these demonstrations reflect an apparent divisiveness of this issue that has only increased over time, coinciding with several decades of gridlock on gun control policy (Spitzer, 2015b) for which our conventional wisdom puts mass-level public
286 Social Science Quarterly
opinion cleavages at the foundation of this policy gridlock. However, the analyses that social scientists in fields as diverse as criminal justice, political science, and sociology have conducted to inform our conventional wisdom do not fully tell us a story that can help us understand the current political climate or the evolution in attitudes on gun control. The data that these analyses use are either weak with limited spatial and temporal scope or treat gun control as an extension of some other topic in political science. Few analyses since Erskine (1972) have adequately leveraged our longest-running nationally representative data sets to fully explore the trends and broad correlates in the public opinion of gun control for a wide audience.
This study informs an academic and lay audience about Americans’ attitudes toward gun control over 26 waves of the GSS data, resulting in several important findings. First, the two public opinion fault lines on gun control that we routinely identify as explaining gun control policy gridlock—increasing GOP partisanship and the urban/rural divide— do not robustly reduce support for gun control. The coefficients for partisanship and a respondent living in a rural area are significant in just two of the six models in Table 1. Figure 1, which summarizes individual regressions from Model 1 in Table 1 for every survey year, shows that the coefficients are significant in a little more than half of all the models. Second, most “strong Republicans” actually support a subtly aggressive form of gun control. The implication is that a negative relationship between partisanship and support for gun control should not be interpreted as equivalent to a misleading statement that self-identified strong Republicans oppose gun control. Third, polarization in the gun control debate as a function of partisanship is only recent and partial, with stark differences starting to emerge after the Republican Revolution of 1994 and emerging even more during the Obama Administration. Fourth, simple statements of regional variation on attitudes about gun control (i.e., about the South and West) mask important and more interesting variation we observe with actual data.
The findings I report should clarify what we think to be true about why gun control legislation is a nonstarter in Congress. Congress is more polarized now than it has ever been (e.g., Andris et al., 2015), which might explain why a divided government, in which the legislature is controlled by Republicans, would fail to pass gun control legislation during the Obama Administration. It is convenient to think Americans are polarized as well, especially if partisans adopt elite-level cues (cf. Zaller, 1992). My analysis suggests this is not true, which conforms well to general arguments that Americans are not nearly as polarized at the mass level as they are at the elite level (e.g., Hill and Tausanovitch, 2015). More than half of those identifying the most with the GOP support a subtly aggressive form of gun control that appears regularly in the GSS data.
It is important for the sake of an honest conversation that discussion of the gun control debate be mindful of these basic facts about where Americans stand on this issue. Consider a recent New York Times article from Bui and Sanger-Katz (2017), which attempted to map more than two dozen potential gun control policies on a two-dimensional space by reference to whether Americans support them and whether experts would say they are effective. The crux of the article focused on squaring mass-level preferences with expert- level opinion, but their article buried an interesting lead. Only one of the 29 policies they measured—requiring an individual to demonstrate a “need” for a gun—had less than 50 percent of support of Americans. Even then, support was 49 percent among the general public in their poll. That said, this true statement about the distribution of attitudes toward gun control in the data belie the trends we are seeing emerge among citizens. Republican partisans are gradually adopting elite cues on the topic of gun control. Right now, gun policy gridlock says more about partisan polarization at the elite level than at the mass level,
What Americans Think About Gun Control 287
but the trends in the data suggest partisans at the mass level will become a bigger part of the gun control policy gridlock in the near future.
This New York Times analysis, combined with my analyses, offers interesting hypotheses about issue framing on gun control for future research. My analyses show gun control has broad support over time, even among those with the strongest affinities for the GOP. Likewise, a battery of 29 proposed gun control policies of varying levels of effectiveness also enjoy wide support. Thus, it might be unsurprising that the broader questions about gun control that frame it as a constitutional issue or pitch the issue as the protection of the rights of gun owners find more lukewarm support for gun control (e.g., Pew Research Center, 2017). This could explain some of the messaging success for the NRA on this issue (Horwitz and Anderson, 2009) and more specific prompts about exact policy proposals may elicit more support, as we see in the GSS data. Further, the gun control policy proposal in the GSS is unique because it also engages in deliberate issue linkage (i.e., implicitly connecting gun control to confidence and trust in the police). Citizens routinely list the police as one of the most trusted institutions in the country, behind only the military and small businesses (Gallup, 2017). Gun control advocates may better advance their cause by linking police security with overall public security in advocating for gun control and framing the issue like this may have discernible effects in survey experiment prompts.
REFERENCES
Andris, Clio, David Lee, Marcus J. Hamilton, Mauro Martino, Christian E. Gunning, and John Armistead Selden. 2015. “The Rise of Partisanship and Super-Cooperators in the U.S. House of Representatives.” PLoS ONE 10(4):e0123507.
Bartley, Wm. Alan, and Geffrey Fain Williams. 2018. “The Role of Gun Supply I: 1980s and 1990s Youth Violence.” Unpublished Manuscript, Transylvania University.
Bohn, Kevin. 2008. “Gun Sales Surge After Obama’s Election.” CNN.com November 11. Available at 〈http://www.cnn.com/2008/CRIME/11/11/obama.gun.sales/〉. Boor, Myron, and Jeffrey H. Bair. 1990. “Suicide Rates, Handgun Control Laws, and Sociodemographic Variables.” Psychological Reports 66(3):923–30.
Brennan, Pauline Gasdow, Alan J. Lizotte, and David McDowall. 1993. “Guns, Southernness, and Gun Control.” Journal of Quantitative Criminology 9(3):289–307.
Bui, Quoctrung, and Margo Sanger-Katz. 2017. “How to Prevent Gun Deaths? Where Experts and the Public Agree.” New York Times January 10. Available at 〈http://www.nytimes.com/interactive/2017/ 01/10/upshot/How-to-Prevent-Gun-Deaths-The-Views-of-Experts-and-the-Public.html〉. Burstein, Paul. 2003. “The Impact of Public Opinion on Public Policy: A Review and an Agenda.” Political Research Quarterly 56(1):29–40.
Dixon, Jo, and Alan J. Lizotte. 1987. “Gun Ownership and the ’Southern Subculture of Violence’.” American Journal of Sociology 93(2):383–405.
Erskine, Hazel. 1972. “The Polls: Gun Control.” Public Opinion Quarterly 36(3):455–69.
Filindra, Alexandra, and Noah J. Kaplan. 2016. “Racial Resentment and Whites’ Gun Policy Preferences in Contemporary America.” Political Behavior 38(2):255–75.
Gallup. 2017. “Confidence in Institutions.” Available at 〈http://news.gallup.com/poll/1597/confidence- institutions.aspx〉. Gelman, Andrew. 2005. “The Secret Weapon.” Available at 〈http://andrewgelman.com/2005/03/07/the/_ secret/_weap/〉. ———. 2008. “Scaling Regression Inputs by Dividing by Two Standard Deviations.” Statistics in Medicine 27(15):2865–73.
288 Social Science Quarterly
Gimpel, James G. 1998. “Packing Heat at the Polls: Gun Ownership, Interest Group Endorsements, and Voting Behavior in Gubernatorial Elections.” Social Science Quarterly 79(3):634–48.
Haider-Markel, Donald P., and Mark R. Joslyn. 2001. “Gun Policy, Opinion, Tragedy, and Blame Attribution: The Conditional Influence of Issue Frame.” Journal of Politics 63(2):520–43.
Hill, Seth J., and Chris Tausanovitch. 2015. “A Disconnect in Representation? Comparison of Trends in Congressional and Public Polarization.” Journal of Politics 77(4):1058–75.
Horwitz, Joshua, and Casey Anderson. 2009. Guns, Democracy, and the Insurrectionist Idea. Ann Arbor, MI: University of Michigan Press.
Joseph, Cameron. 2015. “Senate Republicans Vote Down Legislation That Would Block Suspected Terrorists, Felons, Mentally Ill from Buying Guns.” New York Daily News December 4. Available at 〈http://www.nydailynews.com/news/politics/senate-gop-votes-terrorist-gun-bill-article-1.2454448〉. Levendusky, Matthew. 2009. The Partisan Sort: How Liberals Became Democrats and Conservatives Became Republicans. Chicago, IL: University of Chicago Press.
Lizotte, Alan J., and David J. Bordua. 1980. “Firearms Ownership for Sport and Protection: Two Divergent Models.” American Sociological Review 45(2):229–44.
New York Daily News. 2015. “He’s a Terrorist.” New York Daily News December 4. Avail- able at 〈http://www.nydailynews.com/news/top-new-york-daily-news-front-pages-2015-gallery-1.2474213? pmSlide=1.2474176〉. Page, Benjamin I., and Martin Gillens. 2017. Democracy in America? What Has Gone Wrong and What We Can Do About It. Chicago, IL: University of Chicago Press.
Page, Benjamin I., and Robert Y. Shapiro. 1983. “Effects of Public Opinion on Policy.” American Political Science Review 77(1):175–90.
Patterson, Samuel C., and Keith R. Eakins. 1998. “Congress and Gun Control.” Pp. 45–73 in John M. Bruce and Clyde Wilcox, eds., The Changing Politics of Gun Control. Lanham, MD: Rowman; Littlefield.
Payne, Brian K., and Robert Riedel, Jr. 2002. “Gun Control Attitudes and the Criminal Justice Student: Do Differences Exist?” College Student Journal 36(2):314–22.
Pearson-Merkowitz, Shanna, and Joshua J. Dyck. 2017. “Crime and Partisanship: How Party ID Muddles Reality, Perception, and Policy Attitudes on Crime and Guns.” Social Science Quarterly 98(2):443–54.
Pew Research Center. 2017. Public Views About Guns. Available at 〈http://www.people- press.org/2017/06/22/public-views-about-guns〉. Sears, David O., Richard R. Lau, Tom R. Tyler, and Harris M. Allen, Jr. 1980. “Self-Interest vs. Symbolic Politics in Policy Attitudes and Presidential Voting.” American Political Science Review 74(4):670–84.
Smith, Ben. 2008. “NRA: Obama Most Anti-Gun Candidate Ever, Will Ban Guns.” Ben Smith’s Blog (Politico) August 6. Available at 〈http://www.politico.com/blogs/ben-smith/2008/08/nra-obama-most-anti- gun-candidate-ever-will-ban-guns-010821〉. Spitzer, Robert J. 2015a. Guns Across America: Reconciling Gun Rules and Rights. New York, NY: Oxford University Press.
———. 2015b. The Politics of Gun Control, 6th ed. Boulder, CO: Paradigm Publishers.
The Editorial Board. 2015. “End the Gun Epidemic in America.” New York Times December 4. Available at 〈http://www.nytimes.com/2015/12/05/opinion/end-the-gun-epidemic-in-america.html〉. Wintemute, Garen. 2005. “Guns and Gun Violence.” Pp. 45–96 in Alfred Blumstein, ed., The Crime Drop in America, 2nd ed. New York, NY: Cambridge University Press.
Wolpert, Robin M., and James G. Gimpel. 1998. “Self-Interest, Symbolic Politics, and Public Attitudes Toward Gun Control.” Political Behavior 20(3):241–62.
Wozniak, Kevin H. 2017. “Public Opinion About Gun Control Post-Sandy Hook.” Criminal Justice Policy Review 28(3):255–78.
Zaller, John R. 1992. The Nature and Origins of Mass Opinion. New York, NY: Cambridge University Press.
Copyright of Social Science Quarterly (Wiley-Blackwell) is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
