Where There Are Guns, There Are Suicides: Study

This has been published earlier this year. Here is the study:

Firearm Availability and Unintentional Firearm Deaths, Suicide, and Homicide among 5–14 Year Olds

Mathew Miller, MD, MPH, ScD; Deborah Azrael, PhD; David Hemenway, PhD

From the Harvard School of Public Health, Department of Health Policy and Management, Boston, Massachusetts.

JOURNAL OF TRAUMA 2002;52:267-275

Background: In the United States, only motor vehicle crashes and cancer claim more lives among children than do firearms. This national study attempts to determine whether firearm prevalence is related to rates of unintentional firearm deaths, suicides, and homicides among children.

Methods: Pooled cross-sectional time-series data (1988–1997) were used to estimate the association between the rate of violent death among 5–14 year olds and four proxies of firearm availability, across states and regions.

Results: A statistically significant association exists between gun availability and the rates of unintentional firearm deaths, homicides, and suicides. The elevated rates of suicide and homicide among children living in states with more guns is not entirely explained by a state’s poverty, education, or urbanization and is driven by lethal firearm violence, not by lethal nonfirearm violence.

Conclusion: A disproportionately high number of 5–14 year olds died from suicide, homicide, and unintentional firearm deaths in states and regions where guns were more prevalent.

Key Words: Firearms; Guns; Children; Homicide; Suicide; Unintentional death; Accidents; Violence

In the United States, only motor vehicle crashes and cancer claim more lives among children 5–14 years old than do firearms.1 Between 1988 and 1997, the last 10 years for which complete U.S. data are available, 6,817 children 5–14 years of age died from firearms.1

In contrast, children in other industrialized nations are not dying from guns. Compared with children 5–14 years old in other industrialized nations, the firearm-related homicide rate in the United States is 17 times higher, the firearm-related suicide rate 10 times higher, and the unintentional firearm-related death rate 9 times higher.2 Overall, before a child in the United States reaches 15 years of age, he or she is 5 times more likely than a child in the rest of the industrialized world to be murdered, 2 times as likely to commit suicide and 12 times more likely to die a firearm-related death.2,3

Within the United States, case-control studies have found that the purchase of a handgun4 and the presence of a handgun in the home are strongly associated with an increased risk of homicide and suicide among adults5–7 and an increased risk of suicide among adolescents.5–12 Cohort, cross-sectional, and interrupted time series studies suggest a strong link between the availability of guns and rates of homicide and suicide among adults13–20 and with the rate of unintentional firearm death among all age groups.21

Case-control and cohort studies of lethal firearm violence have several advantages over cross-sectional studies, but they have been geographically limited and have not focused on children. Nationally representative cross-sectional studies, on the other hand, have been hampered by the lack of direct measures of gun availability at levels smaller than the nine census regions. Our study extends previous findings by focusing on children and by using three different state-level proxies for gun availability (one survey-based measure available for a nonrandom 21/50 states; two nonsurvey measures derived for all 50 states) and an additional survey-based regional-level measure of gun availability to explore the relationship between firearm availability and violent death among 5–14 year olds. State-level analyses adjust for state urbanization, poverty, and education levels.

MATERIALS AND METHODS
We used pooled cross-sectional time-series data from the 50 states over a 10-year period (1988–1997) to examine the association between four different measures of the availability of firearms and the corresponding rates of suicide, homicide, and unintentional firearm deaths among children 5–14 years old.

State- and year-specific population figures and data for the number of suicides (International Classification of Diseases, Ninth Revision [ICD-9] codes E950.0–E959), homicides (E960.0–E969), suicides by firearm (E955.0–E955.4), homicides by firearm (E965.0–E965.4), and unintentional deaths caused by firearm (E922.0–E922.9) come from the National Center for Health Statistics (NCHS) mortality files. Deaths from firearms of undetermined intention (ICD-9 E985) constitute less than 3% (186/6,187) of all firearm deaths among 5–14 year olds and are excluded from analyses. Region-specific population and mortality figures were derived by aggregating the corresponding state-based figures.

Two of our measures of firearm availability come from published survey-based estimates of household firearm ownership, one at the regional level and the other at the state level. Our regional survey-based measure of firearm availability, collected at the level of the nine census regions, comes from the average of reported household gun ownership rates between 1988 and 1997, as reported in the General Social Surveys (GSS).22 At the state level, published data on reported household gun ownership rates are available for 21 states from the Behavioral Risk Factor Surveillance System (BRFSS) from the 1990s.23 The 21 states for which the BRFSS obtained data on household gun ownership are a nonrandom sample of those states that self-selected to ask questions about household firearms. However, these states did not appear to select on the basis of the rates of violent death among children or the purported relationship between gun levels and violent death rates. The implications of this nonrandom selection cannot be evaluated directly. Nevertheless, since results from the BRFSS states are substantively similar to those obtained from the nationally representative gold-standard survey based estimates of household firearm ownership (GSS), the BRFSS estimates can be viewed as corroborating rather than primary evidence of the validity of the association observed with GSS estimates. Furthermore, as both the BRFSS the GSS measures are very highly correlated with each other and with our independently derived proxies, these measures as a group supply consistent findings about the association between firearm availability and violent death among children.

Direct measures of household firearm ownership are not available at the state level for all 50 states. When all 50 states are analyzed (at the state level), proxies for firearm availability are two derived measures: (1) Cook’s Index, developed and previously validated at the city level;16 and (2) the fraction of all suicides that involve a gun. Both of these measures have been used in cross-sectional studies within the United States,24–27 and Cook’s Index has also been independently correlated with household gun ownership levels across 14 industrialized nations.28

Cook’s Index for a given state in a given year is calculated by averaging (over all age groups) the percentage of all suicides committed with a firearm and the percentage of all homicides committed with a firearm. That is, Cooks Index = (fraction of suicides with guns + fraction of homicides with guns)/2. In our investigation of childhood mortality rates, to avoid having violent deaths among children appear in both the dependent and the key independent variables, we used a modified Cook’s Index that excluded suicides and homicides among children 0–19 years of age from the calculation. Since 5–14 year old children account for less than 2% of all suicides and homicides, it is not surprising that analyses using the modified index were qualitatively and quantitatively similar to those using a Cook’s Index derived from all age groups. All results presented are based on the modified index.

The second mortality-derived estimate of firearm availability is closely related to Cook’s Index: the fraction of suicides that are gun related. This proxy is referred to as FS/S in the text to indicate that it is the number of firearm suicides in a given state-year (among adults) divided by the total number of suicides in that state-year (among adults). The proxy FS/S is based on the assumption that firearms are likely to be more readily available in states where guns make up a larger fraction of all suicides than in states where guns make up a smaller fraction of all suicides—independent of the number or rate of suicides in a state. A state’s FS/S and Cook’s Index reflect the distribution of firearm vs. nonfirearm means of suicide (in the case of FS/S) and the distribution of firearm vs. nonfirearm means of suicide and homicide (in the case of Cook’s Index). Neither FS/S nor Cook’s Index inherently reflects the rate of suicides or homicides in a state and so do not bias our testing of the null hypothesis, i.e., that there is no relationship between gun availability and overall suicide rates. FS/S merely reflects the distribution of these means.

That FS/S does not by construction (i.e., in itself) bias the overall associations is corroborated by the results of a Monte Carlo-type simulation of 10,000 recursive loops in which positive random numbers are generated for the number of suicides, firearm suicides, and total population in the 50 states. These random numbers are generated so that FS < S. As expected, the correlation between overall suicide rates and FS/S is zero.

Qualitatively and statistically similar results were obtained whether Cook’s Index (and FS/S) assumed the average Cook’s Index (average FS/S) for each state over the 10-year study period, a 5-year rolling average, or a specific value for each state-year. We present results using 10-year averaged values for all our proxies because the gun stock in the United States is so high (over 200 million guns) that changes in a state’s stock are likely to be quite small from year to year, and because using a 5-year rolling average would require us to drop data from 1996 and 1997. The 10-year averaged measures yielded regressions that were qualitatively similar to regressions using 5-year rolling averages and also to regressions in which we used the specific Cook’s Index (FS/S) for each state-year.

To make the comparisons among our four proxies more intuitive, all were standardized so that each proxy has a mean of zero and a SD of 1. The raw average values (1988–1997) for Cook’s Index, FS/S, and the BRFSS survey-based gun ownership levels are presented out to 2 decimal points in Table 1, ranked according to Cook’s Index. The dependent variable used in our analyses is the number of deaths per population per state-year among 5–14 year olds. Distributions of death rates were skewed and variances were greater than the means. Consequently, negative binomial models were used (rather than Poisson).

Table 1. State-Level Proxies of Firearm Availability, Average (Nonstandardized) Values 1988–1997, Ranked by Cook’s Index

A state’s homicide, suicide, and unintentional firearm death rates in a given year are not independent from rates in that state in other years. To account for this nonindependence, standard errors in regressions were corrected by clustering observations (by state in the regressions presented in Table 2; by region in Table 3).

Table 2. Crude and Multivariate Adjusted Incidence Rate Ratios (IRRs) of State-Level Homicide, Suicide, and Unintentional Gun Deaths among 5–14 Year Olds in the United States, by State-Level Proxies of Firearm Availability (1988–1997)

Table 3. Crude Incidence Rate Ratios (IRRs) of Regional-Level Homicide, Suicide, and Unintentional Gun Deaths among 5–14 Year Olds in the United States, by Region-Level Proxies of Firearm Availability (1988–1997)

Primary analyses use incidence rate ratios (IRR), obtained by exponentiating beta coefficients in the negative binomial regressions, to express the magnitude of the association between a state’s suicide, homicide, and unintentional firearm death rate and that state’s standardized proxy for gun availability. Since the SD of each of the standardized proxies is, by construction, equal to 1, the reported IRRs represent the percentage change in the dependent variable (e.g., the suicide rate) for a unit change in the independent variable (i.e., for a change of 1 SD of the proxy under consideration). Because the proxies differ somewhat from each other in their ranges (and a given proxy will have a different range when considered at the state vs. the regional level), comparisons of IRRs must take into account the range of the particular proxy under the conditions specified. The relevant ranges are specified in the legends for each table.

In the state-based analyses, multivariable analyses adjust for other independent variables that have been found to be associated with violent death, including the percentage of the population living in poverty,29 the percentage of the adult population with at least a high school education,30 and the percentage of the state’s population living in urban areas.31,32 When data for these control variables were not available for all years, values for missing observations were interpolated from surrounding years. Whether interpolations were linear interpolations from the surrounding years or averages of the 4 years closest to the missing year did not materially affect results. Linear interpolations were used in the data presented. Data for control variables come from the Statistical Abstracts of the United States (education, poverty, and urbanicity). In the region-based analyses we do not control for other variables because of the small number of observations.

Mortality data were electronically available in 5-year age groupings. The age-groupings we could choose from in evaluating the effect of firearm availability on violent death among children were children aged 0–4, 5–9, 10–14, and 15–19. There are several reasons we did not include children under age 5 in our evaluation. Evaluating the relationship between firearm availability on suicide in 0–4 year olds is impossible, since death among children in this age group is never attributed to suicide. Similarly, homicide in this age group is largely attributable to shaken baby syndrome or similar trauma, a far less common means among older children. Nevertheless, when we evaluated the relationship between firearm availability and violent death among 0–14 year olds we obtained results very similar to those among 5–14 year olds. We did not include 15–19 year olds in our evaluation because the social and cultural forces that influence violence among this age group are so different from those that affect younger children.

RESULTS
Over the 10-year study period (1988–1997), 6,817 children 5–14 years of age were killed with firearms in the United States (3,447 firearm homicides, 1,588 firearm suicides, and 1,782 unintentional firearm deaths). An additional 1,889 children died from nonfirearm homicide; 1,328 from nonfirearm suicides.

Regardless of the proxy chosen, in the multivariate analyses at the state level, we found a positive and statistically significant association between gun availability and state-level rates of unintentional firearm deaths, homicides, firearm homicides, suicides, and firearm suicides among children (Table 2). The increased rate of homicide and suicide in states with high gun levels was accounted for by significantly elevated firearm (but not nonfirearm) suicide and homicide rates (Table 2). Results from the 21-state BRFSS sample are significant even though they do not include three out of five of the states with the highest gun levels and three out of five states with the lowest gun levels (as ranked by Cook’s Index).

At the regional level we still generally observe, regardless of the proxy chosen, a positive and statistically significant association between gun availability and the level of unintentional firearm deaths and suicides among children (Table 3). The only exception is for the homicide rate. Results with the GSS proxy are similar to those obtained with the other proxies, but the firearm availability–homicide relationship is not statistically significant (Table 3).

Table 4 compares the actual number of children 5–14 years old who died from unintentional firearm deaths, suicides, and homicides (1988–1997) in the five states with the highest Cook’s Index values to the corresponding rates in the five states with the lowest Cook’s Index values. These states are chosen on the basis of their extreme firearm ownership levels, not on the basis of their extreme violent death rates among children. Although the number of children in the two groups was similar, compared with children living in the low-gun states (Hawaii, Massachusetts, Rhode Island, New Jersey, and Delaware), children living in the high-gun states (Louisiana, Alabama, Mississippi, Arkansas, and West Virginia) were 16 times as likely to die from unintentional firearm injury, 7 times as likely to die from firearm suicide, 3 times as likely to die from firearm homicide and, overall, twice as likely to die from suicide and homicide.

Table 4. Homicide, Suicide, and Unintentional Gun Deaths among 5–14 Year Olds: The Five U.S. States with the Highest vs. the Lowest Average Cook’s Index of Gun Availability (1988–1997)a

Our firearm proxies give similar results, since they are highly correlated (Table 1). Not only are Cook’s Index and (its component) FS/S highly correlated, but these proxies are also highly correlated with survey-based measures. At the state level, the correlation coefficient for the BRFSS survey-based estimates of household firearm ownership (among the 21 states for which data are available) is 0.81 with Cook’s Index and 0.89 with FS/S. Considering the subgroup of 21 states for which BRFSS provides household firearm ownership levels, the five states with the highest household ownership levels according to the BRFSS are the same five states with the highest FS/S and constitute four out of five of the states with the highest Cook’s Index. Similarly, the five states with the lowest household gun levels according to the BRFSS correspond to the same five states with the lowest levels according to both FS/S and Cook’s Index. At the regional level, our modified Cook’s Index and FS/S are also highly correlated with household firearm ownership levels reported in the GSS (correlation coefficient = 0.91 and 0.96, respectively).

DISCUSSION
The present study examines the relationship between firearm availability and violent death among children within the United States. We found that each of our four proxies leads to the same conclusion: children 5–14 years old were more likely to die from unintentional firearm injuries, suicides and homicides if they lived in states (or regions) with more rather than fewer guns. In contrast, nonfirearm homicides and nonfirearm suicides were not significantly associated with the availability of guns. The relationship between guns and violent death among children remains statistically significant even after controlling for state-level poverty, education, and urbanization.

If, as has been suggested for adolescents and adults,33–35 suicides among children are commonly impulsive acts, the easier it is to find lethal means, such as firearms, the more suicides there might be. On the other hand, if the choice of firearm has less to do with the availability of the weapon than with the strength of the intent, persons determined to kill others or themselves will work harder to get a gun where guns are less available, or will substitute other lethal means. Consistent with some,13,15,17–20 but not all,36 previous studies among U.S. adults, we found that not only firearm-related but also overall suicide rates were significantly associated with state gun levels, suggesting that among 5–14 year olds substitution of equally lethal means for guns is incomplete.

Despite the strong and robust association between lethal firearm violence and state gun levels, we failed to find a statistically significant relationship between state gun levels and either nonfirearm homicide or nonfirearm suicide among 5–14 year olds (Table 2). To the extent that rates of nonfirearm lethal violence reflect violent tendencies, our study indicates that 5–14 year old children living in high-gun states are not significantly more lethally violent toward themselves than are children in low-gun states, nor are they significantly more likely to be victims of lethal nongun attacks. Rather, the disproportionately high level of overall lethal violence where guns are more available suggests that where there are more guns, violence is more likely to turn lethal.

Our proxies for firearm availability are even more strongly associated with the rate of unintentional firearm death than with the rate of intentional firearm deaths (Table 2, Table 3 and Table 4). Over the 10-year study interval, 6,817 children between 5 and 14 years of age died from firearms, 1,782 from unintentional firearm injury alone. We could find no data to suggest that where there are more guns parents care less about their children’s welfare. Yet, unintentional firearm deaths are an order of magnitude greater in high-gun compared with low-gun states (Table 4).

Our study includes both survey-based reports of household firearm ownership rates and estimates of firearm availability that rely on mortality data. We used Cook’s Index and FS/S because these proxies allow us to analyze data from all 50 states. In addition, using measures that rely on different estimating mechanisms may capture different (perhaps complementary) aspects of the relevant variable. The extent to which Cook’s Index or FS/S captures some aspects of firearm availability better (or less well) than do survey-based estimates of household gun ownership rates is unknown. In any event, household gun ownership levels (BRFSS and GSS measures) and our mortality-derived estimates (Cook’s Index and FS/S) are highly correlated, suggesting that they are providing information about the same construct, at least in this age group. We obtained substantively and statistically similar results with all four proxies.

Our findings are robust (Table 2, Table 3 and Table 4). The proxy chosen does not drive the regression results. Regressions are also not driven by either the largest states or the states most extreme in gun levels. Statistically significant and qualitatively consistent results were produced regardless of whether the data analyzed were all 50 states, the 40 largest, or the 40 smallest states; excluding the 5 states with the highest (or lowest) Cook’s Index (or FS/S) did not materially or statistically alter our findings. Even when we used the survey-based estimates of household firearm ownership rates among only 21 states (or among only nine regions), we obtained similar results. Including the 186 (2.6%) firearm deaths coded as firearm deaths of undetermined origin (ICD-9 E985) did not alter our findings, regardless of whether these deaths were included as firearm suicides, firearm homicides, or unintentional firearm deaths. Although we do not know exactly how measures of household firearm ownership relate to the availability of firearms to children, it is remarkable that despite the crudeness of our measures, survey based and proxy alike, we consistently see a relationship between these measures and the rate of violent death among children.

The use of FS/S as a proxy of firearm assumes nothing about the relative rate of overall suicides in a state. If in state A for every 100 suicides 90 are firearm suicides and in state B for every 100 suicides 10 are firearm suicides, use of the proxy FS/S assumes only that guns are more readily available in state A than in state B. The null hypothesis (which we set out to test) states that gun availability does not influence the overall suicide rate, i.e., that if people really want to commit suicide they will find the means. FS/S merely reflects the distribution of these means. Therefore, FS/S does not, per se, bias our testing—and ultimate rejection—of the null hypothesis. Even at the level of the nine census regions, the null hypothesis can be rejected: overall suicide rates are higher where there are more guns, whatever measure of firearm availability is chosen. These findings are consistent with those of others who have described incomplete substitution as an explanation for decreases in suicide rates when a particularly lethal means is restricted.37–40

A potentially more problematic issue is that of reverse causation, though only in the case of homicide (reverse causation is not a problem for suicide or for unintentional firearm deaths). It might be that where homicide rates are higher, individuals (most likely individuals older than our cohort) are more likely to obtain guns in the belief that they are protecting themselves and their families. In this case, the direction of any causal relationship between high gun levels and high homicide rates cannot be determined. In addition, homicide rates among children are not expected to be independent of homicide rates among adults in the same state; hence the removal of 5–14 year olds from the calculation of Cook’s Index and FS/S is only a partial solution to the problem of reverse causation. Nevertheless, our finding that all our proxies for gun availability are significantly related to a state’s rate of gun-homicide and overall homicide but not to the rate of nongun homicide is consistent with firearm availability playing some causal role in homicide rates among children.

Drawing causal inferences from group data to individual behaviors is generally referred to as the “ecological fallacy.”41–44 For example, although the poverty rate in a given state with a high unintentional gun death rate may be disproportionately high, that does not prove that the actual individuals in this state who are dying from guns are disproportionately poor. On the other hand, if a person dies from gunfire, that particular individual did come in contact with a bullet. The ecological fallacy is thus not likely to be a major issue with our analyses.

Another limitation of our study is that our analyses may not account for some reasons that states with higher household gun levels have higher violent death rates. Although we include some state-level confounders (poverty, urbanization, and education), these represent only a small number of the characteristics likely to affect suicides, homicides, or unintentional firearm deaths. We do not, for example, account for parenting practices, domestic abuse, or firearm storage patterns. It is not clear, however, whether accounting for these or other state-level characteristics would revise the magnitude of observed association upward or downward.

Many geographic U.S. studies find a positive and statistically significant relationship between gun density measures and overall homicide45 and suicide46 rates. Consistent with these studies, we find that of the 6,817 children 5–14 years killed with firearms between 1988 and 1997, a disproportionately large number, per population, died in states where guns were more prevalent, regardless of the proxy chosen to estimate firearm availability. Moreover, the elevated rates of suicide and homicide among children living in states with more guns appears to be driven by lethal firearm violence, not by nonfirearm violence. Our findings suggest that, on average, where there are more guns children are not protected from becoming, but are rather much more likely to become, victims of lethal violence.

REFERENCES

National Center for Injury Prevention and Control, Centers for Disease Control and Prevention: Injury Mortality Statistics. Available at http://wonder.cdc.gov/mortsql.shtml.
Centers for Disease Control and Prevention. Rates of homicide, suicide, and firearm-related deaths among children: 26 industrialized countries. MMWR Morb Mortal Wkly Rep. 1997; 46: 101–105.
Krug EG, Dahlberg LL, Powell KE. Childhood homicide, suicide, and firearm deaths: an international comparison [published erratum appears in World Health Stat Q. 1997;50:210]. World Health Stat Q. 1996; 49: 230–235.
Cummings P, Koepsell TD, Grossman DC, Savarino J, Thompson RS. The association between the purchase of a handgun and homicide or suicide. Am J Public Health. 1997; 87: 974–978.
Kellermann AL, Rivara FP, Somes G, et al. Suicide in the home in relation to gun ownership. N Engl J Med. 1992; 327: 467–472.
Kellermann AL, Rivara FP, Rushforth NB, et al. Gun ownership as a risk factor for homicide in the home. N Engl J Med. 1993; 329: 1084–1091.
Bailey JE, Kellermann AL, Somes GW, Banton JG, Rivara FP, Rushforth NP. Risk factors for violent death of women in the home. Arch Intern Med. 1997; 157: 777–782.
Brent DA, Perper JA, Goldstein CE, et al. Risk factors for adolescent suicide: a comparison of adolescent suicide victims with suicidal inpatients. Arch Gen Psychiatry. 1988; 45: 581–588.
Brent DA, Perper JA, Allman CJ, Moritz GM, Wartella ME, Zelenak JP. The presence and accessibility of firearms in the homes of adolescent suicides: a case-control study. JAMA. 1991; 266: 2989–2995.
Brent DA, Perper J, Moritz G, Baugher M, Allman C. Suicide in adolescents with no apparent psychopathology. J Am Acad Child Adolesc Psychiatry. 1993; 32: 494–500.
Brent DA, Perper JA, Moritz G, Baugher M, Schweers J, Roth C. Firearms and adolescent suicide: a community case-control study. Am J Dis Child. 1993; 147: 1066–1071.
Bukstein OG, Brent DA, Perper JA, et al. Risk factors for completed suicide among adolescents with a lifetime history of substance abuse: a case-control study. Acta Psychiatr Scand. 1993; 88: 403–408.
Loftin C, McDowall D, Wiersema B, Cottey TJ. Effects of restrictive licensing of handguns on homicide and suicide in the District of Columbia. N Engl J Med. 1991; 325: 1615–1620.
Sloan JH, Kellermann AL, Reay DT, et al. Handgun regulations, crime, assaults, and homicide: a tale of two cities. N Engl J Med. 1988; 319: 1256–1262.
Wintemute GJ, Parham CA, Beaumont JJ, Wright M, Drake C. Mortality among recent purchasers of handguns. N Engl J Med. 1999; 341: 1583–1589.
Cook P. The effect of gun availability on robbery and robber murder: a cross sectional study of 50 cities. Policy Stud Rev Annu. 1979; 3: 743–781.
Clark R, Jones P. Suicide and the increased availability of handguns in the United States. Soc Sci Med. 1989; 28: 805–809.
Boor M, Bair JH. Suicide rates, handgun control laws, and sociodemographic variables. Psychol Rep. 1990; 66 (3 pt 1):923–930.
Markush RE, Bartolucci AA. Firearms and suicide in the United States. Am J Public Health. 1984; 74: 123–127.
Marzuk PM, Leon AC, Tardiff K, Morgan EB, Stajic M, Mann JJ. The effect of access to lethal methods of injury on suicide rates. Arch Gen Psychiatry. 1992; 49: 451–458.
Miller M, Azrael D, Hemenway D. Firearm availability and unintentional firearm deaths. Accid Anal Prev. 2001; 33: 447–484.
Davis J, Smith T. General Social Surveys, 1972–1998 [machine-readable data file]. Principal Investigator James A. Davis; Director and Co-Principal Investigator, Tom W. Smith; Sponsored by National Science foundation—NORC ed.—Chicago: National Opinion Research Center [producer]; Storrs, CT: The Roper Center for Public Opinion Research, University of Connecticut [distributor]; 1998.
Powell KE, Jacklin BC, Nelson DE, Bland S. State estimates of household exposure to firearms, loaded firearms, and handguns, 1991 through 1995. Am J Public Health. 1998; 88: 969–972.
Lester D. Gun control, gun ownership, and suicide prevention. Suicide Life Threat Behav. 1988; 18: 176–180.
Lester D. Gun ownership and suicide in the United States. Psychol Med. 1989; 19: 519–521.
Lester D. Relationship between firearm availability and primary and secondary murder. Psychol Rep. 1990; 67: 490.
Lester D. The availability of firearms and the use of firearms for suicide: a study of 20 countries. Acta Psychiatr Scand. 1990; 81: 146–147.
Killias M. International correlations between gun ownership and rates of homicide and suicide. Can Med Assoc J. 1993; 148: 1721–1725.
Young TJ. Poverty, suicide, and homicide among Native Americans. Psychol Rep. 1990; 67 (3 pt 2):1153–1154.
Saucer PR. Education and suicide: the quality of life among modern Americans. Psychol Rep. 1993; 73: 637–638.
Fingerhut LA, Ingram DD, Feldman JJ. Homicide rates among US teenagers and young adults: differences by mechanism, level of urbanization, race, and sex, 1987 through 1995. JAMA. 1998; 280: 423–427.
Fingerhut LA, Jones C, Makuc DM. Firearm and motor vehicle injury mortality: variations by state, race, and ethnicity—United States, 1990–91. Adv Data. 1994; 242: 1–12.
Browning CH. Handguns and homicide: a public health problem. JAMA. 1976; 236: 2198–2200.
Baker SP. Without guns, do people kill people? [editorial]. Am J Public Health. 1985; 75: 587–588.
Adelson L. The gun and the sanctity of human life; or the bullet as pathogen. Pharos. 1980; 43: 15–25.
Sloan JH, Rivara FP, Reay DT, Ferris JA, Kellermann AL. Firearm regulations and rates of suicide: a comparison of two metropolitan areas. N Engl J Med. 1990; 322: 369–373.
Rich CL, Young JG, Fowler RC, et al. Guns and suicide: possible effects of some specific legislation. Am J Psychiatry. 1990; 147: 342–346.
Carrington PJ, Moyer S. Gun control and suicide in Ontario. Am J Psychiatry. 1994; 15: 606–608.
Gunnell D, Middleton N, Frankel S. Method availability and the prevention of suicide: a re-analysis of secular trends in England and Wales 1950–1975. Soc Psychiatry Psychiatr Epidemiol. 2000; 35: 437–443.
Hassall C, Trethowan WH. Suicide in Birmingham. BMJ. 1972; 1: 717–718.
Schwartz S. The fallacy of the ecological fallacy: the potential misuse of a concept and the consequences. Am J Public Health. 1994; 84: 819–824.
Piantadosi S, Byar DP, Green SB. The ecological fallacy. Am J Epidemiol. 1988; 127: 893–904.
Kalimo E, Bice TW. Causal analysis and ecological fallacy in cross-national epidemiological research. Scand J Soc Med. 1973; 1: 17–24.
Dutton DG. Patriarchy and wife assault: the ecological fallacy. Violence Vict. 1994; 9: 167–82.
Ohsfeldt R, Morrisey M. Firearms, firearms injury and gun control: a critical survey of the literature. Adv Health Econ Health Serv Res. 1992; 13: 65–82.
Miller M, Hemenway D. The relationship between firearms and suicide: a review of the literature. Aggression Violent Behav. 1999; 4: 59–75.
EDITORIAL COMMENT
Firearm Availability and Childhood Death
Violence is a significant problem that trauma surgeons and other health care professionals see literally everyday. When firearms mix with violence, the outcome is much more likely to be lethal.1 In the United States, we live in an excessively violent society2 that also happens to be the world’s largest market for civilian firearms.3 Yet, at the same time, we are committed to improving the health and welfare of our children. Are we really to believe then that plentiful firearms and healthy children can coexist?

The study by Miller and colleagues provides important information to help answer this question. These authors use four measures of firearm “availability” (two survey and two derived measures) to indicate firearm availability at both the state and regional level. As the authors indicate, these are rather crude measures of firearm availability and provide aggregate data only. The derived measures (a modified Cook Index and the FS/S) are estimates of firearm availability based on the fractions of adult homicides and/or adult suicides that are firearm-related in each state. The authors take appropriate care in describing the derivation, testing, and limitations of these measures, noting (although not in detail) that their high correlation with survey data on firearm availability support their use. The use of derived measures is clearly a limitation. However, these measures are some of the best available, since, unlike motor vehicles, civilian ownership of firearms is not archived for any serviceable length of time by federal or state data agencies.

During the 10-year study period, Miller and colleagues found that 6,817 children between the ages of 5 and 14 years of age were killed with firearms in the United States, outstripping those who were killed by non-firearm homicide or suicide. Even controlling for poverty, urbanization, and education (known correlates of firearm violence), a strong, positive association between firearm availability and firearm death was found in this age group at the state and regional levels. However, factors not controlled for in this study may have modified the relationship between firearm availability and firearm death, as indicated by the very high rates in Alaska, Idaho, and Montana, states not ranking in the top 10 high gun availability states. Future studies are needed to identify other risk factors that contribute to the exceptionally high firearm death rates for 5- to 14-year-olds in these states.

Although no conclusions about cause and effect can be made, this study provides compelling evidence that states with high firearm availability are states with high childhood firearm death rates. As trauma professionals we cannot and should not ignore the fact that children living in the five highest gun availability states are estimated to be 16 times more likely to die from unintentional firearm injury, 7 times more likely to die from firearm suicide, and 3 times more likely to die from homicide compared with those living in the five lowest gun availability states. These findings reinforce the belief that high levels of firearm availability place our children at risk and seriously undermine attempts to improve their health and welfare.

The lethality of firearm violence and its toll on the nation’s children should rouse trauma surgeons and other health care professionals to transfer some of the assertive spirit they display during trauma resuscitations to other, more proactive injury prevention activities. This study, along with lessons learned from successfully reducing the toll of motor vehicle crashes,4 point to three specific steps that can be taken. First, comprehensive data about all violent deaths is critical to building the scientific foundation for prevention. At the federal level, the National Violent Death Reporting System, modeled after the Fatality Analysis Reporting System (FARS), is being proposed as a critical step in providing detailed data about violent injury throughout the United States. We can support this endeavor through grassroots efforts and by working directly with our federal and state officials. Second, civilian registration of all firearms would provide valuable and accurate data on firearm availability. Registering a firearm should be no different from registering a car and should not infringe on legal firearm ownership. Finally, community-specific data applicable to our own “home towns” is crucial to dispelling local myths about firearms and firearm violence. Combining data from trauma centers, local medical examiners/coroners, police, and crime labs is vital to understanding and disseminating accurate profiles of firearm violence in our communities. Only then can we develop meaningful interventions. If we, as health care professionals are to safeguard and promote healthy lives for our children, these steps seem obvious. Their accomplishment await only our determination, support, and energy.

Therese S. Richmond, PhD, CRNP
C. William Schwab, MD
Charles C. Branas, PhD
Firearm Injury Center at PennUniversity of PennsylvaniaPhiladelphia, Pennsylvania

REFERENCES

Gotsch KE, Annest JL, Mercy JA, Ryan GW. Surveillance for fatal and nonfatal firearm-related injuries – United States, 1993–1998. MMWR Surveill Summ. 2001; 50 (SS-2):1–36.
Krug EG, Powell KE, Dahlberg LL. Firearm-related deaths in the United States and 35 other high- and upper-middle-income countries. Int J Epidemiology. 1998; 27: 214–221.
Diaz T. Making a Killing: The Business of Guns in America. New York, NY: New Press; 1999: 69–84.
Morbidity and Mortality Weekly Report. Motor-vehicle safety: A 20th century public health achievement. MMWR Morb Mortal Wkly Rep. 1999; 48: 369–374.
Submitted for publication December 12, 2000.

Accepted for publication July 24, 2001.

This research was supported in part by grants from the Centers for Disease Control and Prevention, the Joyce Foundation, the Robert Wood Johnson Foundation, the Packard Foundation, and the Center on Crimes, Communities and Culture of the Open Society Institute.

Address for reprints: Matthew Miller, MD, MPH, ScD, Harvard School of Public Health, Department of Health Policy and Management, 677 Huntington Avenue, Boston, MA 02115; email: mmiller@hsph.harvard.edu.

J Trauma 2002 February;52(2):267-275
Copyright © 2002 Lippincott Williams & Wilkins
All rights reserved

Previous Full Text | Full Text PDF (79 K) | References (42) | Next Full Text

Cooking the Numbers
Why the Harvard “Study” on Kids and Guns is Bunk

by Sean Oberle
Analysis@KeepAndBearArms.com

March 4, 2002

NOTE: If this looks too bunched up along the section where the charts are displayed, you can view this article outside of our normal site format here: http://www.KeepAndBearArms.com/Oberle/harvard.htm.

The factoid fry-cooks at Harvard School of Public Health have pulled a sleight of hand. You’ve probably seen the press reports[1] of their new “study,” in the Journal of Trauma which supposedly shows:

“The higher death rates in high gun states are due to differences in deaths from firearms” and “the strong and robust association between gun ownership and children’s violent death is compelling”[2] (see footnote for link to full-text article).”

They even (ahem) “show” that overall violence in kids is higher in the five most gun-prevalent states than in the five least. They do this by using a gun-prevalence proxy that two of them have recently acknowledged elsewhere is an inferior way to estimate gun prevalence — this comes in a paper for the Duke University Sanford Institute of Public Policy[3]. Not only that, but the creator of the proxy coauthored this Duke paper — even he, apparently, thinks it is inferior (see footnote for link to full-text paper).

What’s a proxy? Gun ownership levels can be difficult to determine, so the two ways to estimate it is to conduct surveys or to use a proxy — a formula based on something that involves guns and deduce ownership levels from that.

The proxy the Harvard crew used is called Cook’s index. Named for Philip Cook[4], it is based on the proportion of guns used in two types of violent death — homicide and suicide. It is the average of two fractions — firearm suicides (FS) over all suicides (S) and firearm homicides (FH) over all homicides (H). Its formula is (FS/S + FH/H)/2. The assumption behind it is that the more that firearms are represented in overall violent death, the more they are prevalent.

The Harvard trio also used a second proxy (sort of Cook’s lite) — simply FS/S. While the Duke paper identifies it as better, it has problems similar to Cook’s index, but to a lesser degree. As well, they looked at 21-state gun prevalence surveys done by the Centers for Disease Control (CDC) in the 1990s and a regional level survey[5]. However, they focused on Cook’s index in their article. It is from their Cook’s index calculations that they derive their pronouncements about the top and bottom five states, and it is on Cook’s index that I devote most attention.

In the Journal of Trauma article, the Harvard group claims there is no 50-state survey of gun prevalence. While perhaps technically true, there are 48-state surveys (no Hawaii and Alaska).

Could the Harvard gang simply have not known about them? Nope. In the Duke paper, two of them identify these 1996 and 1999 surveys. Indeed, the surveys were conducted for them — or at least for the Harvard Injury Control Research Center, which is their part of the Harvard School of Public Health. In any event, all of them have used this data in other articles, in 1998 and 2000 — prior to submitting the Journal of Trauma article (submitted in December 2000)[6].

This becomes important when reviewing the Harvard gang’s Journal of Trauma work — they imply that the 21-state survey is all they have to check their numbers against, and they note that their Cook's and FS/S calculations correspond with the 21-state survey — thus they must be relatively accurate, right?. Well, not exactly.

Chart 1 shows the relationships between four state-level measurements identified in the Duke and Trauma papers:

1) Harvard’s Cook’s index calculations
2) Harvard’s FS/S calculations
3) Harvard’s 48-state surveys
4) CDC’s 21-state surveys

(Note: see data sets by clicking charts, excepting chart 2)

If a state is red/blue and underlined, it shows variance but was excluded from the Harvard threesome’s check for their Cook’s index and FS/S calculations. In other words, Harvard’s Cook’s index and FS/S match survey-based estimates if you exclude most of the states that show variance with them.

Indeed, high variant states are the most likely to be excluded, and low variant states (black) are least likely to be excluded. Using the CDC 21-state survey as a check excludes 17 of 24 (70.8%) high-variance states and 9 of 17 (52.9%) moderate-variance states. But it excludes only 3 of 9 (33.3%) low-variance states.

Cook’s index, incidentally, does not measure quantity or rate of misuse. A town of 100,000 people, for example, with just one murder and two suicides in a year nonetheless could have a high Cook’s index if, for example, the homicide and one of the suicides were with guns. Its index would be 0.75 (maximum is 1.00). The formula: (1/1 + 1/2)/2

Cook’s index supposedly corrects for high rates of gun homicide and/or suicide in low-prevalence areas. For example, moving to the state level, Hawaii has a very low homicide rate, while New York has a higher rate. But both have low proportion of gun homicide/suicides, thus Hawaii ranks 1st and New York ranks 8th in low gun prevalence — 49 states have higher gun prevalence than Hawaii, and 42 have higher than New York.

OK, nothing odd so far. Now let’s look at two states, one with high violence and another with low violence — Maryland (high) and South Dakota (low). Maryland has a higher Cook’s index than South Dakota; therefore, Maryland must have more gun prevalence, right?

Yes, it must if you accept Cook’s index as valid. Indeed, in the Harvard analysis, South Dakota ranks as having less prevalence of guns than Maryland. In fact, South Dakota’s rates of gun homicide/suicide to overall homicide/suicide are so low that the state ranks 6th in the less prevalence scale. That’s right; the Harvard crew calculated that in 44 states — including New York (8th), Connecticut (11th), Illinois (12th), California (23rd) and Maryland (28th) — guns are more prevalent than in South Dakota.

In the FS/S proxy, however, South Dakota drops from 6th to 25th. In the 48-state survey, it drops from 6th to 44th. On the other hand, in the FS/S proxy, Maryland rises from 29th to 11th. In the 48-state survey, it rises from 29th to 13th.

Note that the all those high-homicide, high suicide states in the bottom using either Cook’s index or FS/S move out of the bottom — the bottom five states based on the 48-state survey are South Dakota (44, moves down 38 slots), Vermont (45, down 13), Montana (46, down 19), Idaho (47, down 11) and Wyoming (48, down 7).

The flaw (this is my analysis, not Duke’s) is that Cook’s index assumes that gun prevalence and proportion of use in homicide and suicide move together along one linear spectrum, when in reality their relationship must be looked at on a plane (see chart 2).

Cook’s wrongly assumes there are only two basic types of localities: (B) higher prevalence, higher proportion and (D) lower prevalence, lower proportion. However two additional types exist: (A) higher prevalence, lower proportion and (C) lower prevalence, higher proportion.

The problem caused by Cook’s linear flaw is that (A) and (C) locations get pushed towards the wrong ends of the prevalence spectrum.

Indeed, there is some circular reasoning going on: Overall homicides and suicides will be affected by the proportion of gun misuse because guns are the primary weapon of choice in both. Any changes in use-levels of the mechanism will affect the whole more than any changes in lesser-used mechanisms.

Thus, (A) states tend to have lower overall violence because they have lower proportions of gun misuse (despite their higher gun prevalence), and (C) states tend to have higher overall violence because they have higher proportions of gun misuse (despite their lower prevalence of guns).

In other words, the up and down movements of Cook’s index and violence levels are tied more to each other than to gun prevalence. This means that if you use Cook’s index as a proxy for firearm prevalence, it really is not much different from using overall violence as the proxy. Despite being hidden by omitting the first step from the thought process, the circular logic in the Harvard Journal of Trauma article is:

High violence = high Cook’s index (now, take a measurement of violence) = high violence.

The flaws in Cook’s index become more prevalent in other ways. Looking at Texas, for example, that state’s Cook’s index[7] falls from 1994’s 0.72 — (1,652/2,337 +1 ,568/2,113)/2 — to 1999’s 0.59 –- (1,224/2,005 + 794/1,391)/2. If Cook’s index is a valid measure of gun availability, then guns became 18% less available in Texas in the late 1990’s. Yep, Cook’s tells us that one in five guns left the state. Wow, those Texans must really hate their guns!

Warning:
Next 3 Charts Are Based on Bogus Cook’s Index Measurements
Take State Rankings with Grain of Salt

In any event, since Cook’s index and violence levels move up and down together — independently of gun prevalence — a Cook’s index ranking of states will have high or low violence clustered at the extreme ends[8].

This means that by looking at only the extreme ends (see Chart 3) — which the Harvard threesome emphasizes in the Journal of Trauma article — wrong impressions emerge, leading to inaccurate pronouncements.

The error, however, begins to become apparent when you look at all states (see Chart 4). The correlation exists only at the skewed ends. Notice that except for the ends, the trend is somewhat flat line. Why?

I can think of numerous hypotheses, including that the Cook’s index used by Harvard is derived from the whole population, but the yellow violence line refers to a sub-set of that population — children aged 5-14. The sub-set might have its own index to match its violence trend.

More likely is what we see by breaking “overall violence” out into its subgroups: homicide and violence (See chart 5).

By doing this — even though the order of supposed gun prevalence is bogus — we see that there seems to be some substitution — and thus balancing — in category of violence.

Excepting South Dakota, each marked state is noted twice (find matches by looking straight up or straight down). States with higher homicide rates have lower suicide rates for this age group and vice versa. The phenomenon isn’t limited to just a few states — it occurs in most states.

Why is that? I don’t know, and I don’t want to make half-cooked pronouncements like the Harvard crew does.

Now Let’s Look at Something Closer to Reality

What I do know is that when we look at the trend in violence based on the 48-state survey, rather than the faulty Cook’s Index (or FS/S), the lines do get rather flat — especially when you take into account two facts: the District of Columbia is not in these comparisons, and Idaho’s high violence is due to a single year aberration.

Note that on Chart 6 the trend shows rising violence from 0% to 30% gun prevalence. Then it flat lines. Then, around 70% gun prevalence it starts going down again. (Does it go down farther as ownership approaches 100%? I have no way of knowing.)

First Idaho: I’ve marked Idaho (ID) to the far right. Idaho’s violence is so high because, in 1990, it had an aberrantly high number of suicides.

Typically the state has 2-to-5 suicides per year for this age group, but in 1990 it had 11 (5 non-gun). In fact, that year’s suicides account for about a quarter of the entire 10-year tally — 11 out of 43. Remove 1990 (look at a different 10-year period) and Idaho’s overall violence level drops, and along with it drops the trend line — the high-prevalence end begins to look a lot like the low prevalence end.

There’s also a lesson here: Many other states have similarly low numbers year-to-year, so have similarly unstable numbers. While the 10-year pooling gets the numbers up, stabilizing them a bit, there’s enough instability that it’s risky to draw conclusions no matter what proxy or survey used.

Now D.C.: What do we do with the District of Columbia? It should be included. It has more people (about 572,000) than the entire state of Wyoming (about 494,000), and similar numbers to a few other states like North Dakota and Vermont[9].

On the other hand, Washington is a city, and simply throwing it into the mix is improper (as in apples and oranges). Cities generally have much higher violence rates than states because state rates moderate as high and low violence areas cancel each other out. D.C.’s violence rate for this period and age group is 9.66, more than three times the closest state — putting it in the mix would cause a huge spike, bending the trend way out of whack.

In my opinion, combining D.C. and Maryland into a single unit makes the most sense. Geographically and socio-economically, D.C. fits into Maryland. Indeed, the west half of the city has more in common with bordering Montgomery County, Md., than it does with the east half of the city, and the east half of the city has more in common with bordering Prince George’s County, Md., than it does with the west half.

Combining D.C. and Maryland this would do two things: move Maryland closer to the left of the gun-prevalence scale in much the way that other cities with extreme gun control — like New York City and Chicago — move their states in that direction. However Washington’s high violence rate would make Maryland’s violence rate go up (although the spike would not occur as it would with D.C. counted alone due to state-level moderation).

Farther to the left and higher up than it is now, Maryland would pull the trend upwards, further making either end of the trend similar looking.

We’ve got a soft bell curve with a slight dip in the middle.

Hawaii and Alaska: These states also are not included. Hawaii would pull the low-prevalence end towards low violence, and Alaska probably wouldn’t change the trend much at all.

Conclusion: I’m not going to make any conclusions about guns and kids. We are dealing with many low-quantity numbers. As the Idaho example illustrates, these numbers can be unstable, making conclusions risky.

What I will conclude is that the Harvard study used a faulty proxy with unstable numbers. It proves nothing.

Sean Oberle is a Featured Writer and gun control analyst for KeepAndBearArms.com. He can be reached at Analysis@KeepAndBearArms.com. View other articles from him at http://www.KeepAndBearArms.com/Oberle.

Footnotes

1 For example: Harvard School of Public Health Feb. 19 press release, available on the web at: http://www.hsph.harvard.edu/press/releases/press2192002.html, or Erica Nagourney, Feb. 26, New York Times, available on the web at: http://www.nytimes.com/2002/02/26/health/26SAFE.html.

2 Miller, Mathew; Azrael Debora, Hemenway, Dave. “Firearm Availability and Unintentional Firearms Deaths, Suicide, and Homicide among 5-14 Year Olds.” The Journal of Trauma, Feb. 2002, Vol. 52, No. 2, pp. 267-275. Full Text (with pdf reader): http://www.hsph.harvard.edu/press/releases/trauma_article.pdf.

3 Azrael, Deborah; Cook, Philip; Miller, Mathew. “State and Local Prevalence of Firearms Ownership: Measurements, Structure, and Trends.” Duke University Terry Sanford Institute of Public Policy, Working Paper Series (SAN01-24), Sept. 2001. Full Text (with pdf reader): http://www.pubpol.duke.edu/people/faculty/cook/SAN01-25.pdf.

4 Cook, Philip. “The Effect of Gun Availability on Robbery and Robber Murder: A Cross Sectional Study of 50 Cities.” Policy Studies Reviews Annual 3, 1979, pp. 743-781.

5 See footnote 2.

6 See footnote 3.

7 All violence data in this article is from the Centers for Disease Control Injury Mortality database (WISQARS), which can be accessed online at http://www.cdc.gov/ncipc/wisqars. Click on charts to see full data sets.

8 See footnote 7.

9 U.S. Census Bureau, Table: “Time Series of State Population Estimates,” available at: http://eire.census.gov/popest/data/states/populartables/table01.php.

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.