Firearm Prevalence and Social Capital

Firearm Prevalence and Social Capital DAVID HEMENWAY, PhD, BRUCE P. KENNEDY, EdD, ICHIRO KAWACHI, PhD, AND ROBERT D. PUTNAM, PhD

PURPOSE: Previous studies have shown a correlation between measures of social capital and morbidity, mortality, and violent crime. This article examines the association across U.S. states between social capital (as measured by mutual trust and civic engagement) and firearm availability. METHODS: The analysis uses OLS to determine degrees of association across U.S. states. Measures of mutual trust come from responses to questions on the U.S. General Social Survey that “you can’t be too careful in dealing with people,” and most people “would try to take advantage of you.” Measures of formal civic engagement come from responses to Lifestyle Survey questions concerning times volunteered, club meetings attended, community projects worked on, and church services attended. Informal civic engagement measures come from responses to number of times bowled, played cards, entertained at home, and gave or attended dinner parties, and number of greeting cards sent. The Lifestyle Survey also asked whether respondent believed whether “most people are honest.” The percentage of suicides from firearms, and the average percentage of suicides and homicides from firearms, are used as proxies for state firearm ownership rates. Control variables are the degree of urbanization, the rates of poverty, and the percentage of nonwhites in the state. RESULTS: Across the U.S. states, higher levels of firearm ownership are associated with significantly lower levels of mutual trust and civic engagement. CONCLUSION: While the analysis cannot show causation, states with heavily armed civilians are also states with low levels of social capital. Ann Epidemiol 2001;11:484–490. © 2001 Elsevier Science Inc. All rights reserved. KEY WORDS:

Social Capital, Social Network, Firearms, Guns, Mutual Trust, Civic Engagement.

INTRODUCTION Consider Robert Heinlein’s famous dictum that “An armed society is a polite society.” Knowing that one’s fellow citizens are armed, greater care is naturally taken not to give offense (1). We wonder, finally, about the quality of life in the kind of society where routine social order depends upon massive armament of the citizenry. . . . [An issue that] deserves to be raised, in our view, is the psychological effect on a community’s residents of the knowledge that many guns are in its homes, on its streets, and even in its schools. . . . [A] great deal of personal testimony, if not yet systematic evidence, indicates that fear is the dominant emotion inspired by the pervasiveness of guns and gun crime (2).

From the Department of Health Policy and Management (D.H.) and the Department of Health and Social Behavior (B.P.K., I.K.), Harvard School of Public Health, Boston, MA; and the John F. Kennedy School of Government, Harvard University, Cambridge, MA (R.D.P.). Address reprint requests to: David Hemenway, PhD, Department of Health Policy and Management, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115. Received June 23, 2000; revised March 7, 2001; accepted March 28, 2001. © 2001 Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010

The United States is the most heavily armed nation in the developed world (3). Many studies have examined the association of gun ownership levels, gun carrying, and guns in the home with unintentional injury, suicide and homicide, and other crimes. Although there are beliefs about the societal effects of an armed society—as the opening quotes indicate— no study seems to have examined the association of civilian firearm ownership levels with measures of social capital. Social capital is a broad sociologic concept describing the connection among individuals. It involves those social aspects of our daily lives that often matter the most—good will, fellowship, sympathy, and social intercourse. Social capital is typically measured by the amount of mutual trust and civic engagement among community members. Civic engagement includes connectedness, which is both formal (e.g., attending Rotary Club or Parent Teacher Association meetings) and informal (e.g., playing cards with friends or sending greeting cards). As a screwdriver (physical capital) or education (human capital) can increase both productivity and well being, so too can one’s social networks. The effect of social support, social contacts, and social networks on morbidity and mortality is receiving increasing attention in the medical literature. For example, recent studies have indicated that good social supports may protect against stress-related hypertension and the common 1047-2797/01/$–see front matter PII S1047-2797(01)00235-6

AEP Vol. 11, No. 7 October 2001: 484–490

Selected Abbreviations and Acronyms GSS  General Social Survey LS  DDB Needham Lifestyle Survey

cold (4, 5), and that church attendance and membership in social organizations may reduce the risk of mortality (6–10). Previous studies have examined the relationship between measures of social cohesion and firearm homicide and violent crime across U.S. states (11, 12). In this article we begin to explore the relationship between levels of firearm ownership and measures of social capital.

METHODS The study is a cross-sectional ecological study of U.S. states. The dependent variables are measures of mutual trust among community members and civic engagement. These are two core constructs of social capital, as presented by its principal theorists (13–16). Depending on whether there is sufficient sample size to include states with small populations, we analyze data for (a) the 39 most populous states, or (b) the 48 contiguous states. For the first analysis, we follow Putnam (15, 16) and use data from the General Social Survey (GSS) conducted by the National Opinion Research Center to estimate variations in levels of mutual trust. The GSS is a nationally representative survey of non-institutionalized adults over 18 years of age living in the United States. The surveys have been repeated many times over the last two decades and have included a set of questions on mutual trust. For this study, we average 5 years of cumulated data (1986–1990) from the GSS, representing 7679 individual observations. Of the 50 states, only 39 are represented due to the small population of some states. Because the sampling design of the GSS is intended to be representative of regions rather than states, we adjust individual responses using post-stratification weights to reflect the age, race/ethnic, and educational composition of each state. Detailed procedures for the poststratification weighting are described elsewhere (11, 17). Lack of trust in others is assessed from responses to two GSS items, “Do you think most people would try to take advantage of you if they got a chance, or would they try to be fair?” and “Generally speaking, would you say that most people can be trusted or that you can’t be too careful in dealing with people?” For each state we calculate the percentage of respondents who agreed that most people “would try to take advantage of you” and that “you can’t be too careful in dealing with people.” Higher percentages indicate lower levels of mutual trust. Civic engagement refers to the level of commitment of citizens to their communities, reflected by their involvement in community affairs. Typically this is measured by mem-

Hemenway et al. FIREARM PREVALENCE AND SOCIAL CAPITAL

485

bership in civic-related associations and groups that bring members of the community together around shared interests. Sufficient data on measures of involvement in the community are available to allow us to include all 48 contiguous states in the second analysis. We use data from the DDB Needham Lifestyle Survey (LS) conducted by Market Facts for the DDB Needham advertising firm. Participants for the LS mail panel survey are pre-recruited using commercially available lists generated from motor vehicle registries, credit card companies, telephone directories, and a variety of other sources. Responses to the pre-recruitment mailing vary from about 1% for inner city residents to 5–10% among middleclass Americans. From the pre-recruited mail panel (numbering perhaps 500,000 at any one time) a random, demographically balanced sample is drawn. The response rates for this sample range from 70–80%. The sample has been shown to yield results that are virtually identical to other surveys using conventional random samples (18). For example, compared to the GSS, the LS survey provides extremely similar estimates from comparable questions in the two surveys including both attitudes (e.g., toward feminism, legalization of marijuana or abortion, the Soviet Union, financial worries, basic social values, and mutual trust) and behaviors (e.g., smoking, video usage, hunting, gun ownership, church attendance, and leisure activities). Subgroup analyses of responses (e.g., by age, gender, education, region) also yield results within sampling error (16). The LS survey has been repeated annually since 1975 using the same methodology and contains data on respondents in the 48 contiguous states. In the present study, to obtain large enough samples for the least populous states, we average 23 years of cumulated data (1975–1997) from the LS, representing 65,387 individual observations from 48 states, ranging from 140 in Wyoming to 6570 in California, with an average of 1362 respondents in each state. Mutual trust is measured in the LS by the percentage of survey respondents in each state agreeing that “most people are honest.” Levels of formal civic engagement are measured by, in the past year: 1. 2. 3. 4.

the average number of times respondents volunteered, the average number of club meetings attended, the average number of community projects worked on, and the average number of church services attended.

In addition to these items that purport to capture social capital as it is more formally defined, we also include survey responses that may reflect informal social capital as evidenced by the extent of socializing with friends in such informal settings as at parties or in the home. The informal social capital variables are, in the past year: 5. the average number of times bowled, 6. the average number of times played cards,

486

Hemenway et al. FIREARM PREVALENCE AND SOCIAL CAPITAL

7. the average number of times entertained at home, 8. the average number of greeting cards sent, and 9. the average number of dinner parties given or attended. The main independent variable is a measure of the level of firearm ownership in the state. Survey data on firearm ownership are not available for many states, so surrogate measures are commonly used. One is the percentage of all completed suicides that are firearm suicides. That measure is highly correlated with the percentage of households reporting firearm ownership in surveys of 170 cities (r  .86) and across 16 developed nations (r  .91) (3). The measure has often been used as a proxy for firearm availability (3, 19–21). Another proxy measure has been the percentage of homicides committed with a firearm (22–25). For this study we use a second, related proxy for firearm ownership, which has been called “Cook’s index,” after Phil Cook (26, 27). Cook’s index is the average of the percent of completed suicides that use a firearm and the percent of homicides that use a firearm. Cook’s index has been validated for urban patterns of regional firearm ownership with data from surveys of residents (r  .91) (28). Cook’s index has more recently been validated at the regional level with GSS based estimates of handgun ownership (r  .96) and

AEP Vol. 11, No. 7 October 2001: 484–490

for 21 states that reported household firearm ownership levels on Behavioral Risk Factor Surveillance surveys (r  .83)(29). Guns are highly durable consumer products, and the stock of guns changes slowly over time. Our gun prevalence measures are stable across states over decades. For example, we calculated the correlation coefficient of the percentage of suicides that are firearm suicides across states for three 3-year time periods, 1979–1981, 1987–1989, and 1995– 1997. The correlation coefficients among these three periods were .94 between the first and the third time period, and .96, between the early and middle and middle and later time period. We use data from 1986–1990 for our proxy measures of firearm ownership. Suicide and homicide rates, and the rates attributable to firearms, were obtained for each state from the Compressed Mortality Files compiled by the National Center for Health Statistics for the Centers for Disease Control and Prevention. We use three control variables: percent urbanization (the percentage of state residents who live in urban areas), percent poverty (the percentage of households in a state that were living below the Federal poverty threshold), and percent nonwhite (the percentage of individuals in the state who were nonwhite).

TABLE 1. Variables Variables Formal social capital variables (n  48) Times volunteered past year Club meetings attended past year Community projects worked on past year Church services attended past year Informal social capital variables (n  48) Times bowled past year Times played cards past year Times entertained at home past year Number of greeting cards sent past year Dinner parties given or attended past year Mutual trust variables Trust: percent agreeing “most people are honest” Distrust 1: percent responding “Most people would try to take advantage of you if they had the chance.” GSS Survey (n  39) Distrust 2: percent responding “You can’t be too careful dealing with people” GSS Survey (n  39) Firearm variables Firearm ownership proxy 1 (48 states) Firearm ownership proxy 2 (Cooks Index, 48 states) Control variables Percent urban (n  48) Percent poverty (n  48) Percent nonwhite (n  48)

Mean

Standard Deviation

Maximum

Minimum

7.0 7.5 2.4 23.3

1.0 1.1 0.5 3.9

9.5 10.9 4.0 31.2

5.0 4.5 1.7 13.4

3.6 11.7 11.4 16.4 5.0

1.2 2.1 1.1 2.1 0.8

6.5 17.4 13.4 21.3 6.8

1.2 8.4 9.3 11.4 2.9

66.8

4.4

74.5

54.0

26.1

7.8

41.9

8.3

43.4

10.0

61.6

21.1

61.6 60.0

12.3 9.4

82.1 75.3

31.5 34.6

52.0 12.9 15.2

20.2 4.1 1.4

86.0 26.1 37.3

15.0 6.6 9.6

Firearm Ownership Proxy 1: Percent of completed suicides in the state that are firearm suicides. Firearm Ownership Proxy 2: (Cooks Index): Mean of percent of completed suicides that are firearm suicides, and percent of homicides which are firearm homicides.

0.10 0.21 0.14

0.48* 0.33* 0.09 0.46* 0.21 0.14

0.46* 0.21

0.07 0.20 0.23 0.24 0.30*

0.69* 0.57* 0.44*

0.60*

0.64*

0.14 0.37* 0.24 0.23 0.18

1.00

4

0.15

0.51* 0.76* 0.58* 0.56* 0.29*

0.47* 0.45* 0.53* 0.47* 0.63*

1.00 0.16

3

0.68* 0.62* 0.42*

1.00 0.49* 0.15

2

1.00 0.43* 0.48* 0.01

1

0.69*

1.00 0.63* 0.53* 0.21

6

b

0.62*

1.00 0.73* 0.46*

7

0.53*

1.00 0.51*

8

0.35*

1.00

9

1.00

10

11

12

0.67* 0.27 0.56* 0.20

13

14

0.73* 0.70* 0.61* 0.40* 0.44* 0.49* 1.00 0.57* 0.59* 0.54* 0.37* 0.45* 0.50* 0.90* 1.00

0.59* 0.69* 0.62* 0.54* 0.44* 0.78* 0.79* 1.00

0.71* 0.69* 0.52* 0.53* 0.52* 0.85* 1.00

0.65*

1.00 0.64* 0.64* 0.68* 0.42*

5

Correlations using GSS data are based on 39 states, all others on 48 excluding Alaska and Hawaii. Mean of (the percent of completed suicides that are firearm suicides and the percent of homicides that are firearm homicides). * p  0.05

a

Formal social capital (n  48) 1 Times volunteered past year 2 Club meetings attended past year 3 Community projects worked on past year 4 Church services attended past year Informal social capital (n  48) 5 Times bowled past year 6 Times played cards past year 7 Times entertained at home past year 8 Number of greetings cards sent past year 9 Dinner parties given or attended past year Mutual trust variables 10 Percent agreeing “Most people are honest” 11 Percent agreeing, “You can’t be too careful in dealing with people” (GSS)a 12 Percent agreeing, “Most people would try to take advantage of you if they got a chance” (GSS)a Firearm ownership variables 13 Percentage of all completed suicides in which a firearm was used 14 Cooks indexb

TABLE 2. Correlations among social capital and firearm variables

AEP Vol. 11, No. 7 October 2001: 484–490 Hemenway et al. FIREARM PREVALENCE AND SOCIAL CAPITAL

487

488

Hemenway et al. FIREARM PREVALENCE AND SOCIAL CAPITAL

AEP Vol. 11, No. 7 October 2001: 484–490

TABLE 3. Relationship of levels of gun ownership and social capital Gun ownership 1b Multivariatea

Univariate Dependent variables Formal social capital Times volunteered past year Club meetings attended past year Community projects worked on past year Church services attended last year Informal social capital Times bowled past year Times played cards past year Times entertained at home past year Number of greeting cards sent past year Dinner parties given or attended past year Mutual trust Percent agreeing “Most people are honest” Percent agreeing, “You can’t be too careful in dealing with people (GSS)c Percent agreeing, “Most people would try to take advantage of you if they got a chance” (GSS)c

Gun ownership 2b

B

p

B

p

0.04 0.03

.001 .022

.03 .07

.056 .000

0.05 0.03

.001 .145

.04 .05

.053 .029

0.01 0.07

.552 .147

.02 .04

.104 .562

0.01 0.05

.326 .358

.02 .08

.045 .311

0.07 0.05 0.07 0.12

.000 .064 .000 .000

.06 .11 .09 .11

.001 .001 .000 .000

0.07 0.05 0.07 0.13

.000 .165 .000 .000

.04 .09 .07 .09

.027 .017 .001 .008

0.04

.000

.02

.043

0.04

.000

.02

.112

0.14

.005

.13

.027

0.17

.009

.15

.029

0.33

.004

.18

.288

0.45

.004

.26

.191

0.28

.001

.24

.038

0.39

.001

.32

.018

B

p

Multivariatea

Univariate

B

p

a

Control variables in the multivariate analysis are % urbanization; % poverty; % nonwhite. b Gun Ownership Proxy 1: Percent of completed suicides in the state that are gun suicides. Gun Ownership Proxy 2: (Cooks Index): Mean of percent of completed suicides that are gun suicides, and percent of homicides that are gun homicides. c GSS data are based on 39 states, all others on 48 excluding Alaska and Hawaii.

Urbanization, poverty, and racial data for each state come from the 1990 United States Census Tape File STF 3A. The Federal poverty index is a wage-income based measure that does not include income from other sources such as public assistance programs. All analyses were conducted using Pearson’s correlation and ordinary least squares (OLS) regression.

RESULTS Table 1 shows the mean and standard deviations for each variable for the 48 contiguous states. On average, 60% of suicides in a state are firearm suicides, with a range among states of 38% to 82%. About two-thirds of people per state agree that most people are honest, with a range among states of 54% to 75%. In an average state, residents report attending approximately 23 church services per year, and 7.5 club meetings; they report on average playing cards 12 times per year, and going to or attending 5 dinner parties. In simple correlations over 39 states, both measures of firearm ownership are positively and significantly correlated with the percentage of the state’s residents who believe that “most people cannot be trusted” and “most people would try to take advantage of you” (Table 2). These results remain significant in the multivariate analysis for “most people would try to take advantage of you” (Table 3).

In simple correlations over 48 states, both measures of firearm ownership are negatively and significantly correlated with six of the measures of formal and informal civic engagement and of mutual trust (community volunteering, going bowling, entertaining at home, sending greeting cards, going to parties, and “most people are honest”). The absolute value of the correlation coefficients are often quite high—for example, one gun prevalence measure has a greater than .70 correlation with two of the measures of informal social capital (Table 2). By contrast, only one of the 10 social capital variables (church attendance) is positively associated with the measures of firearm ownership, and that correlation is small and is not statistically significant. In the multivariate analysis, all 10 social capital variables are negatively correlated with both measures of firearm ownership, and for six of the 10 variables the correlation is statistically significant. To provide a visual display of the results, a plot of Cook’s index versus one of the dependent variables, times volunteered, is given in Figure 1. Volunteering includes such things as visiting shut-in parishioners or coaching a little league team. Similar results were obtained for the other measures of social capital for which results were statistically significant (e.g., times bowled; times entertained at home; number of greeting cards sent; dinner parties given or attended; percent agreeing “most people are honest”; and for 39 states, the two GSS mutual trust variables).

AEP Vol. 11, No. 7 October 2001: 484–490

Hemenway et al. FIREARM PREVALENCE AND SOCIAL CAPITAL

489

FIGURE 1. Firearm ownership by community volunteering.

DISCUSSION This study provides evidence that in states where there are higher levels of firearm ownership, there are significantly lower levels of mutual trust and civic engagement. We cannot determine whether this result is due to cause and effect (or in which direction) or whether it is due to a third set of unmeasured variables that affect both firearm ownership rates and measures of social capital. We do know that the relationship is not explained by urbanization rates, poverty rates, or the percentage of state residents who are nonwhite. Our results are consistent with the notion that, in the United States, as more people have ready access to lethal weapons, the less the mutual trust and social interaction. Conversely, it could mean that where mutual trust is low, people are more likely to arm themselves. In either case, the outcome—areas of heavily armed civilians who do not trust each other—is not comforting. Surveys have shown that the large majority of U.S. adults say they would feel less safe if more people in their community acquired firearms (30, 31). Studies often find that states with higher levels of household firearm ownership have higher rates of homicide and suicide (22, 32–34). Many factors influence the levels social capital in a state. As Figure 1 illustrates, Utah and Nevada are polar opposites in terms of many measures of social capital, as they

are in terms of many measures of health status (35). Figure 1 also illustrates that many of the southern U.S. states have high levels of gun ownership and often low levels of various measures of social capital. Our cross-sectional study does not provide evidence about whether or not an armed society is a polite society. It does suggest, however, that an armed community may not be a very sociable or trusting one. U.S. states with heavily armed civilians are also states with low levels of social capital. Hemenway, Kawachi and Kennedy are recipients of the Robert Wood Johnson Foundation Investigator Awards in Health Policy Research. Hemenway also received support from the Center on Crime, Communities and Culture of the Open Society Institute.

REFERENCES 1. Snyder J. A nation of cowards. The Public Interest. Fall 1993. quoted by George Will, Newsweek, November 8, 1993. 2. Alba RD, Messner SF. Point Blank against itself: Evidence and inference about guns, crime and gun control. Journal of Quantitative Criminology. 1995; 11:391–410. 3. Kleck G. Targeting Guns: Firearms and their Control. New York: Aldine de Gruyter; 1997. 4. Strogatz DS, Croft JB, James SA, et al. Social support, stress, and blood pressure in black adults. Epidemiology. 1997;8:482–487. 5. Cohen S, Doyle WJ, Skoner DP, et al. Social ties and susceptibility to the common cold. JAMA. 1997;277:1940–1944.

490

Hemenway et al. FIREARM PREVALENCE AND SOCIAL CAPITAL

6. Seeman TE. Social ties and health: The benefits of social integration. Annals of Epidemiology. 1996;6:442–451. 7. Yasuda N, Zimmerman SI, Hawkes W, et al. Relation of social network characteristics to 5-year mortality among young-old versus oldold white women in an urban community. American Journal of Epidemiology. 1997;145:516–523. 8. Oman D, Reed D. Religion and mortality among the community-dwelling elderly. American Journal of Public Health. 1998;88:1469–75. 9. Rosengren A, Orth-Gomer K, Wilhelmsen L. Socioeconomic differences in health indices, social networks and mortality among Swedish men: A study of men born in 1933. Scandinavian Journals of Social Medicine. 1998;26:272–280. 10. Eriksson BG, Hessler RM, Sundh J, Steen B. Cross-cultural analysis of longevity among Swedish and American elders: The role of social networks in the Gothenburg and Missouri longitudinal studies compared. Archives of Gerontology and Geriatrics. 1999;28:131–148. 11. Kennedy BP, Kawachi I, Prothrow-Stith D, et al. Social capital, income inequality, and firearm violent crime. Social Science and Medicine. 1998;47:7–17. 12. Kawachi I, Kennedy BP, Wilkinson RG. Crime, social disorganization and relative deprivation. Social Science and Medicine. 1999;48:719–731. 13. Coleman JS. The Foundations of Social Theory. Cambridge, MA: Harvard University Press; 1926. 14. Putnam RD. The prosperous community: Social capital and economic growth. American Prospect. 1993;Spring:35–42. 15. Putnam RD. Making Democracy Work. Princeton, NJ: Princeton University Press; 1993. 16. Putnam RD. Bowling Alone: The Collapse and Revival of American Community. New York: Simon & Schuster; 2000. 17. Kawachi I, Kennedy BP, Lochner K, Prothrow-Stith D. Social capital, income inequality and mortality. American Journal of Public Health. 1997;87:1491–1498. 18. Groeneman S. Multi-purpose household panels and general samples: How similar and how different? Presented at the annual conference of the American Association for Public Opinion Research, May 12– 15, 1994. 19. Lester D. Firearm availability and the incidence of suicide and homicide. Acta Psychiatrica Belgica. 1988a;88:387–393.

AEP Vol. 11, No. 7 October 2001: 484–490

20. Lester D. Gun control, gun ownership and suicide prevention. Suicide and Life Threatening Behavior. 1988b;18:176–180. 21. Lester D. The availability of firearms and the use of firearms for suicide. Acta Psychiatrica Scandinavia. 1990;81:146–147 22. Brearly HC. Homicide in the United States. Chapel Hill: University of North Carolina Press; 1932 23. Cook PJ. The role of firearms in violent crime: An interpretive review of the literature with some new findings and suggestions for future research. In: Wolfgang M, Weiner N, eds. Criminal Violence. Beverly Hills, CA: Sage; 1982:236–289 24. Kleck G, Patterson EB. The impact of gun control and gun ownership levels on violence rates. Journal of Quantitative Criminology. 1993; 9:249–288. 25. Hellsten JJ. Motivation and opportunity: An ecological investigation of U.S. urban suicide, 1970–1990. PhD dissertation in Social Ecology, University of California, Irvine, 1995. 26. Cook PJ. The effect of gun availability on robbery and robbery murder: A cross section study of fifty cities. In: Haveman RH, Zellner BB, eds. Policy Studies Review Annual. Beverly Hills: Sage; 1979:743–781. 27. Cook PJ. The relationship between victim resistance and injury in noncommercial robbery. Journal of Legal Studies. 1986;15:405–416. 28. Cook PJ. The technology of personal violence. In: Tonry M, ed. Crime and Justice, Vol 14. Chicago: University of Chicago Press; 1991. 29. Miller M, Azrael D, Hemenway D. Firearm availability and unintentional firearm deaths. Accident Analysis and Prevention. 2001;33:477–484 30. Hemenway D, Solnick SJ, Azrael DR. Firearms and community feelings of safety. Journal of Criminal Law and Criminology. 1995; 86:121–132. 31. Miller M, Azrael DR, Hemenway D. Community firearms, community fear. Epidemiology. 2000;11:709–714. 32. Miller M, Hemenway D. The relationship between firearms and suicide: A review of the literature. Aggression and Violent Behavior: A Review Journal. 1999;4:59–75. 33. Seitz ST. Firearms, homicide, and gun control effectiveness. Law and Society Review. 1972;6:595–614. 34. Lester D. Relationship between firearm availability and primary and secondary murder. Psychological Reports. 1990;67:490 35. Fuchs VR. Who Shall Live? Health, Economics and Social Choice. New York: Basic Books; 1974.