The analytic sample included 78 878 purchasers after exclusions. A total of 136 purchasers with no follow-up information, 635 with no census tract information for the entire follow-up period, and 29 with missing race or ethnicity have been excluded from the figure. Those arrested but not convicted for the crimes of interest at baseline are not shown. Purchasers may appear in both the DUI (driving under the influence) and non-DUI conviction categories.
The figure excludes purchasers arrested but not convicted for DUI (driving under the influence) or non-DUI crimes at baseline.
eAppendix 1. Linking Methods
eAppendix 2. Socioeconomic Index
eFigure 1. Graphs of Schoenfeld Residuals for Each Exposure Combination and Each Outcome
eFigure 2. Flow Chart for Analytic Sample
eFigure 3. Purchaser Locations [A] and Purchaser Density per 1000 by Census Tract [B] in 2001
eTable 1. Comparison of Eligible Study Population and Analytic Sample Across a Range of Variables
eTable 2. Absolute Differences in Probability of Arrest at 5 and 12 Years Associated With DUI Conviction at Purchase
eTable 3. Hazard Ratios for Arrest for Violent Crime Index Crimes Among First-time and Repeat Purchasers
eTable 4. Hazard Ratios for Arrest for Violent Crime Index Crime by Number of DUI Convictions Prior to Index Purchase in 2001
eTable 5. Risk of Arrest or Conviction Associated With DUI and Alcohol-Related Arrests or Convictions
eTable 6. Risk of Arrest Associated With DUI and Alcohol-Related Arrests or Convictions, Including 249 Individuals With Missing Criminal History Data
eTable 7. Risk of Arrest Associated With DUI and Alcohol-related Arrests or Convictions, Excluding Purged and Procedural Crimes
eTable 8. Risk of Subsequent Arrest Associated With DUI Conviction Prior to Handgun Purchase: Partially Adjusted
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Kagawa RMC, Stewart S, Wright MA, et al. Association of Prior Convictions for Driving Under the Influence With Risk of Subsequent Arrest for Violent Crimes Among Handgun Purchasers. JAMA Intern Med. 2020;180(1):35–43. doi:10.1001/jamainternmed.2019.4491
Are prior convictions for driving under the influence associated with the risk of subsequent arrest for a violent crime among legal purchasers of handguns?
In this longitudinal cohort study of 79 678 handgun purchasers in California, 9% of the purchasers with prior convictions for driving under the influence and 2% of the purchasers with no prior criminal history were subsequently arrested for murder, rape, robbery, or aggravated assault during the 13 years of follow-up.
This study’s findings suggest that prior convictions for driving under the influence may be associated with the risk of subsequent arrest for a violent crime among legal purchasers of handguns.
Alcohol use is a risk factor for firearm-related violence, and firearm owners are more likely than others to report risky drinking behaviors.
To study the association between prior convictions for driving under the influence (DUI) and risk of subsequent arrest for violent crimes among handgun purchasers.
In this retrospective, longitudinal cohort study, 79 678 individuals were followed up from their first handgun purchase in 2001 through 2013. The study cohort included all legally authorized handgun purchasers in California aged 21 to 49 years at the time of purchase in 2001. Individuals were identified using the California Department of Justice (CA DOJ) Dealer’s Record of Sale (DROS) database, which retains information on all legal handgun transfers in the state.
The primary exposure was DUI conviction prior to the first handgun purchase in 2001, as recorded in the CA DOJ Criminal History Information System.
Main Outcomes and Measures
Prespecified outcomes included arrests for violent crimes listed in the Crime Index published by the Federal Bureau of Investigation (murder, rape, robbery, and aggravated assault), firearm-related violent crimes, and any violent crimes.
Of the study population (N=79 678), 91.0% were males and 68.9% were white individuals; the median age was 34 (range, 21-49) years. The analytic sample for multivariable models included 78 878 purchasers after exclusions. Compared with purchasers who had no prior criminal history, those with prior DUI convictions and no other criminal history were at increased risk of arrest for a Crime Index–listed violent crime (adjusted hazard ratio [AHR], 2.6; 95% CI, 1.7-4.1), a firearm-related violent crime (AHR, 2.8; 95% CI, 1.3-6.4), and any violent crime (AHR, 3.3; 95% CI, 2.4-4.5). Among purchasers with a history of arrests or convictions for crimes other than DUI, associations specifically with DUI conviction remained.
Conclusions and Relevance
This study’s findings suggest that prior DUI convictions may be associated with the risk of subsequent violence, including firearm-related violence, among legal purchasers of handguns. Although the magnitude was diminished, the risk associated with DUI conviction remained elevated even among those with a history of arrests or convictions for crimes of other types.
Firearm violence is a public health and safety problem in the United States, with 14 542 homicides and more than 400 000 violent victimizations involving a firearm in 2017.1,2 Alcohol consumption is a known risk factor for violence perpetration, including violence that involves firearms, and firearm owners are more likely than others to report risky drinking behaviors.3-6 In a meta-analysis of studies involving homicide offenders, an estimated 34% of people who committed homicide with a firearm were under the influence of alcohol at the time of the crime.7 Hypothesized mechanisms of action include the direct neurocognitive effects related to alcohol use, which facilitate aggression; an underlying propensity to engage in harmful activities, including both risky alcohol use and violent behavior; and interactions between these factors.5,8
Given the association between alcohol use and violence, including firearm violence, many states have adopted statutes limiting firearm purchase or possession by persons with a history of risky alcohol use.9 However, the standards are often vaguely defined and difficult to implement. For example, Ohio prohibits firearm possession by anyone who is a “habitual drunkard,”10 and Tennessee prohibits the sale of firearms to those “who are addicted to alcohol.”11 Only Indiana, Maryland, Pennsylvania, and Washington, DC, have well-defined standards for prohibiting firearm possession among those with a history of risky alcohol use, such as accruing 3 or more convictions for driving under the influence (DUI) during a specified period.4
California has enacted restrictions for firearm access beyond those in federal laws. For example, in 1990, a state law prohibited firearm purchase or possession for a period of 5 years among those admitted to an inpatient psychiatric facility for an evaluation of dangerousness to self or others.12 In 1991, the restrictions in California were expanded to prohibit firearm purchase and possession for 10 years among people with convictions for specific misdemeanor violent crimes.13 In 2013, the California Legislature enacted a 10-year prohibition on firearm purchase and possession by people who accrued 2 or more convictions for DUI or certain other alcohol-related misdemeanor crimes within a 3-year period.14 Jerry Brown, California’s then-governor, vetoed the bill, citing a lack of evidence linking “crimes that are non-felonies, non-violent, and do not involve misuse of firearms”15 to violent behavior or improper firearm use.
A prior study from our research group assessed the association between risky alcohol use and the risk of future violence among firearm owners.16 The study, published in 2018, found that handgun purchasers in California with histories of alcohol-related crimes (particularly DUI crimes) were at increased risk of subsequent arrest for violent or firearm-related crimes. The prior study was limited; for example, it relied on old data (handgun purchases were made in 1977 and follow-up was through 1991), and the sample size was small, including 4066 purchasers, of whom 1272 had prior convictions for alcohol-related crimes. Of those with prior alcohol-related convictions, 32.8% were subsequently arrested for a violent or firearm-related crime compared with 5.7% of those with no prior criminal history.
In this retrospective longitudinal study, we estimated the association between prior convictions for DUI and the risk of subsequent arrest for violent crimes listed in the Crime Index published by the Federal Bureau of Investigation (murder, rape, robbery, and aggravated assault)17 among legally authorized purchasers of handguns in California, using a larger study population and more recent data. We also estimated associations using a broader category of alcohol-related crimes and for the following 2 secondary outcomes: arrest for firearm-related violent crimes and for any violent crimes.
The University of California, Davis Institutional Review Board approved the study protocol18 and waived the need for obtaining informed consent.
The study population included all legal purchasers of handguns in California in 2001 who were aged 21 to 49 years as of the 10th postpurchase day, the first day on which firearms could be acquired (California requires a 10-day waiting period). For the individuals who made multiple purchases in 2001, follow-up began after the first purchase. The selected age range reflects the facts that handgun purchasers in California must be aged at least 21 years, and violent crime rates among persons 50 years or older are relatively low.19 The study cohort was identified using the California Department of Justice (CA DOJ) Dealer’s Record of Sale (DROS) database, which retains information on all legal handgun transfers in the state.
The study participants were followed up from their 10th postpurchase day in 2001 until December 31, 2013, or a prior date when they could no longer be affirmatively identified as living California residents using public records. These records initially included the California Death Statistical Master File, California voter registration records, and DROS records. If a period of 3 years or more for an individual remained unaccounted for (53.0% of the study sample), we queried LexisNexis Public Records (see eAppendix 1 in the Supplement).
Our main exposure was DUI conviction prior to the index purchase. We tested associations for any prior DUI conviction and for the number of DUI convictions (1, ≥2). Conviction for other alcohol-related crimes (eg, public drunkenness) served as a secondary exposure. Crimes not involving alcohol consumption, such as having an open container in the vehicle, were excluded from this category. We linked the participants to the Criminal History Information System records by using deterministic and probabilistic matching to identify the exposure status.20
Our primary outcome was an arrest charge for a Crime Index–listed violent crime, which includes murder, rape, robbery, and aggravated assault.17 We used arrest charges because they are recorded more rapidly and completely than are convictions.21 Although not all criminal charges corresponded to an arrest event, we have referred to these as arrests throughout for simplicity. Secondary outcomes were arrests for firearm-related violent crimes and an umbrella category of all violent crimes, defined following the Federal Bureau of Investigation and World Health Organization recommendations.17,22,23 Secondary outcomes also included convictions for each outcome.
Individual-level control variables included sex, age, and self-reported race/ethnicity (DROS categories were collapsed to include American Indian, Asian, black, Latino, white, and other individuals) recorded in DROS; the number of handguns purchased from 1985 (the earliest year for which records were available) until the index purchase (0, 1-3, or ≥4); and time elapsed between the most recent DUI and non–DUI-related arrests and the index purchase for individuals with prior criminal histories (0-2, >2 years). Community characteristics included census tract demographics (population size and density, proportion of persons aged 20 to 44 years who were aged 20 to 24 years, and percentages of males, Latino individuals, and black individuals) from the American Community Survey with annual interpolations by Geolytics24,25; census tract alcohol outlet densities using 4 separate license types (bar/pub/tavern, restaurant [beer, wine], restaurant [spirits], and off premises)26; county population and violent and property crime rates27; and county firearm suicides as a proportion of total suicides (2001-2013), a proxy for firearm ownership prevalence.28,29 We adjusted for socioeconomic status using a census tract–level index that combined standard indicators (eAppendix 2 in the Supplement).
The data analysis was conducted from December 2017 to August 2019. We used Cox proportional hazards regression to estimate adjusted hazard ratios (AHRs) with 95% CIs. We included DUI history as a 3-level variable to categorize purchasers as having convictions at the time of purchase, arrests but no convictions, or neither (the referent group). We hypothesized that having a history of arrests or convictions for crimes other than DUI may affect the associations of interest. Therefore, our models included interactions between DUI history and a variable that distinguished participants who had prior arrests or convictions for non-DUI crimes from those who did not. We present separate HRs for associations with DUI convictions in the presence or absence of non-DUI criminal history as well as associations with non-DUI criminal history alone and the combination of DUI convictions and non-DUI criminal history. In separate models, we included additional interaction terms with race/ethnicity and sex to test whether associations differed across subgroups.
All covariates described above were hypothesized to have an association with risk of DUI conviction and future arrest for a violent crime, or simply with future arrest for a violent crime, and were included in all adjusted models. We also explored more parsimonious models that included all individual level variables, community characteristics with P < .25 in the model of arrest for any violent crime, and, for conceptual reasons, county violent crime rate. We included community variables as time-varying covariates (except for firearm ownership prevalence), updated in 2005 and 2010 to account for participants relocating and changes within a community over time. We used graphs of Schoenfeld residuals to assess the proportional hazards assumption (eFigure 1 in the Supplement). The plotted residuals remained moderately flat over time, suggesting that this assumption was met.
We graphed Kaplan-Meier curves of our primary and 2 secondary outcomes stratified by conviction history at the time of purchase. We used Kaplan-Meier analyses to calculate absolute differences in the probability of arrest for our 3 outcomes at 5 and 12 years after purchase, comparing those with and without a DUI conviction at the time of first purchase in 2001.
Statistical significance was assessed at the 0.05 level (2-sided). Analyses were conducted using SAS software, versions 9.3 and 9.4 (SAS Institute Inc).
We tested the sensitivity of our results to changes in criminal history event coding by excluding crimes that were regularly removed from criminal records by the CA DOJ (eg, pre-1976 possession of marijuana) and procedural charges (eg, probation violations, contempt of court) in defining the presence or absence of criminal history. Separately, we included 249 participants with no recent criminal history for whom paper records possibly containing early (pre-1976) criminal events could not be located as having no criminal history. We also compared results using annual updates for the time-varying covariates. Finally, we tested for whether associations differed between first-time and repeat purchasers, and for a dose-response relationship between the number of DUI convictions purchasers had prior to handgun purchase and each outcome.
Data from DROS suggested that 2001 was a typical year. The sales volume of handguns in 2001 (160 619) was close to the mean volume during 1996 to 2005 (181 877 [range, 126 233-244 569]).
The study population comprised 79 678 handgun purchasers (eFigure 2 in the Supplement). In all, 13 292 handgun purchasers (16.7%) had prior arrests or convictions at the time of purchase and 1511 (1.9%) had prior DUI convictions (1.6% had 1 and 0.3% had ≥2) (Table 1). Ninety-five purchasers (0.1%) had no DUI convictions but had convictions for other alcohol-related crimes, and 7663 (9.6%) had convictions unrelated to alcohol. After exclusions, the analytic sample included 78 878 purchasers with nonmissing race/ethnicity and community-level data for their residence in 2001, 2005, or 2010. Of them, 65 387 (82.9%) resided in California through the end of the observation period. The purchasers were under observation for 10 738 253 person-months, with those residing in California through December 2013 accounting for 9 742 573 person-months and those censored earlier for 995 680 person-months. The median age of the study participants was 34 (range, 21-49) years. The participants included 72 522 males (91.0%) and 7156 females (9.0%), were predominantly white individuals (54 868 [68.9%]) (Table 1), and resided largely in urban areas (eFigure 3 in the Supplement). Distributions of measured variables did not differ meaningfully between the eligible study population and the analytic sample (eTable 1 in the Supplement).
Of 1495 purchasers with prior DUI convictions, 131 (8.8%) were subsequently arrested for Crime Index–listed violent crimes (20 of 404 [5.0%] with DUI convictions and no non-DUI arrests or convictions and 111 of 1091 [10.2%] who had both DUI and non-DUI arrests or convictions) (Figure 1). Among 65 700 purchasers with no prior criminal history, 1188 (1.8%) were subsequently arrested for Crime Index–listed violent crimes, and among 10 771 with only non-DUI arrests or convictions, 816 (7.6%) were.
In adjusted comparisons, purchasers with DUI convictions and no non-DUI arrests or convictions had more than double the risk of arrest for a Crime Index–listed crime compared with purchasers who had no criminal history (AHR, 2.6; 95% CI, 1.7-4.1) (Table 2). Among purchasers with prior non-DUI arrests or convictions, DUI convictions remained associated with subsequent arrest for Crime Index–listed violent crimes (AHR, 1.4; 95% CI, 1.1-1.7).
Most participants (75 [57.3%]) with a DUI conviction and a subsequent arrest for a Crime Index–listed violent crime also had prepurchase convictions for non-DUI crimes, and an additional 36 (27.5%) had arrests without convictions. The combination of preexisting DUI convictions and non-DUI criminal history was associated with the greatest risk of subsequent arrest for a Crime Index–listed violent crime (AHR, 5.9; 95% CI, 4.9-7.2) (Table 2).
A history of DUI conviction was also associated with an increased risk of arrest for a firearm-related violent crime and for any violent crime irrespective of whether the participants had preexisting arrests or convictions for non-DUI crimes. Compared with purchasers who had no prior criminal history, those with prior DUI convictions and no other criminal history were at increased risk of arrest for a firearm-related violent crime (AHR, 2.8; 95% CI, 1.3-6.4) and any violent crime (AHR, 3.3; 95% CI, 2.4-4.5) (Table 2).
Kaplan-Meier curves for the primary and 2 secondary outcomes stratified by conviction history at purchase are shown in Figure 2; absolute differences calculated from our Kaplan-Meier analyses are shown in eTable 2 in the Supplement. As an example, comparing those with and without a DUI conviction at the time of purchase, the difference in the probability of arrest for a Crime Index–listed violent crime after 12 years was 0.07 (95% CI, 0.05-0.08).
We found no statistically significant interactions between DUI history and race/ethnicity or sex (Table 3). A statistically significant interaction would suggest that the association between DUI history and arrest for a violent crime differs, on average, depending on the race/ethnicity or sex of the individual.
We found no evidence of a difference in the associations between first-time and repeat purchasers or of a dose-response relationship between the number of prior DUI convictions and the risk of subsequent arrest for a Crime Index–listed violent crime (eTables 3 and 4 in the Supplement). The results of analyses examining the slightly broader category of alcohol-related crimes as the exposure or exchanging convictions for arrests as the outcome were largely consistent with those in the primary analyses (eTable 5 in the Supplement). The sensitivity analyses had consistent results (eTables 6 and 7 in the Supplement) as did more parsimonious models that included fewer variables (eTable 8 in the Supplement).
In this population of authorized purchasers of handguns in California, a history of DUI conviction at the time of purchase was found to be associated with a significant increase in the risk of subsequent arrest for a Crime Index–listed violent crime, a firearm-related violent crime, and any violent crime independent of a range of individual and community-level measures. Participants with both a history of DUI conviction and other preexisting non-DUI arrests or convictions were at greatest risk of subsequent arrest for a violent crime. Among those who had arrests or convictions for non-DUI crimes, the HRs associated specifically with DUI convictions were smaller. This finding suggests that a DUI conviction prior to the time of purchase offers a relatively modest increase in explanatory power over a history of arrests or convictions.
Our findings are consistent with those of studies among more general populations, which show that alcohol use is associated with an increased risk of violence perpetration30-32; our study extends those findings specifically to legal purchasers of handguns. Our previous study involved handgun purchasers from 1977, prior to the extension of California’s prohibiting criteria for firearm ownership to those with convictions for violent misdemeanors in 1991. The basic findings of the 2 studies, however, are similar, and we have now analyzed more recent data in a much larger study population.
As in the prior study,16 we did not find a dose-response relationship between the number of prior DUI convictions and the risk of subsequent arrest for a Crime Index–listed violent crime. This result contrasts with previous research on criminal activity among DUI offenders.33-38 Prior studies, however, often compared the prior criminal histories of first-time and repeat DUI offenders. Our study sample was distinct in that participants with prior DUI convictions could not also have been convicted of a felony or, within 10 years of purchase, of certain violent misdemeanors; those convictions would have prohibited them from purchasing firearms legally in California. Among DUI offenders in general, the presence of other criminal history may be associated with an increased risk for future criminal activity39; selection criteria associated with legal handgun purchase in California are a possible reason why we did not find a dose-response relationship.
Alternatively, the absence of a dose-response relationship may reflect the finding that even 1 DUI conviction suggests a sustained pattern of risky behavior; DUI offenders are estimated to drive while impaired from 200 to 2000 times before their first arrest.37 In a longitudinal study, the DUI recidivism rate of first-time offenders more closely resembled that of repeat offenders than the first-offense rate of nonoffenders,37 a finding similar to ours that is aligned with this alternative hypothesis.
The bar for firearm ownership is particularly high in California; firearm purchasers must have no convictions for a wide range of violent misdemeanors in the 10 years prior to purchase. As a result, many individuals with a higher likelihood of violence perpetration were likely excluded from our sample. Assuming that DUI conviction history and violent misdemeanor conviction history are associated, the risk for future violence among firearm owners with DUI convictions may be even greater in the 25 states where individuals with violent or firearm-related misdemeanor convictions remain eligible to purchase firearms.40
Finally, of note, the risk of arrest and conviction given criminal behavior is not uniformly distributed across the population, and some subgroups, such as specific racial groups, may be over- or underrepresented in the criminal history data for reasons other than their propensity to drive while impaired, such as enforcement behavior.41,42 In the present study, we found no evidence to suggest that DUI conviction is a nonuniform marker of risk for future violence across specific races/ethnicities or by sex.
Our study relied on criminal records to identify risky alcohol use and violent behavior. Violence prevention policy often relies on such administrative records, making this practice a useful approach. However, convictions are several steps removed from behavior, and, as noted previously, DUI convictions are rare compared with self-reported rates of alcohol-impaired driving. We were unable to study associations between other forms of risky alcohol use, such as binge drinking, and violence in our study population. The exclusion of purchasers older than 50 years limits the generalizability of our results. California’s violent misdemeanor prohibition discussed above may also limit the generalizability of our results to states without this prohibition, as may other differences in firearm policy (eg, states without comprehensive background check policies). Finally, between 1959 and 2014, California’s definition of DUI included impairment by alcohol, drugs, or any combination of the two.43 However, annual reports on California DUIs beginning in 1991 show that at least 97.0% of DUI crimes involved blood alcohol concentrations above 0.08% (80 ng/dL).
This longitudinal study of handgun purchasers in California, who were followed up from 10 days after purchase in 2001 through 2013, found that a history of DUI conviction is associated with a substantially higher risk of subsequent arrest for a range of violent crimes. Although the elevated risk was diminished, it remained even among participants with a history of arrests or convictions for crimes of other types.
The federal government and many states have restricted the purchase and possession of firearms by members of high-risk groups, including persons convicted of felonies, domestic violence misdemeanors, and other violent misdemeanors. The available evidence suggests that such policies may reduce the incidence of violent criminal activity among persons whose access to firearms is restricted.44-48 Our findings suggest that comparable benefits may arise from similar restrictions on persons convicted of DUI crimes.
Accepted for Publication: August 13, 2019.
Corresponding Author: Rose M. C. Kagawa, PhD, MPH, Violence Prevention Research Program, Department of Emergency Medicine, School of Medicine, University of California, Davis, UC Davis Medical Center, 2315 Stockton Blvd, Sacramento, CA 95817 (email@example.com).
Published Online: September 30, 2019. doi:10.1001/jamainternmed.2019.4491
Author Contributions: Drs Kagawa and Wintemute had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Stewart, Kass, Cerdá, Gruenewald, Wintemute.
Acquisition, analysis, or interpretation of data: Kagawa, Stewart, Wright, Shev, Pear, McCort, Pallin, Asif-Sattar, Sohl, Gruenewald, Studdert, Wintemute.
Drafting of the manuscript: Kagawa, Stewart, Shev, Asif-Sattar.
Critical revision of the manuscript for important intellectual content: Kagawa, Wright, Shev, Pear, McCort, Pallin, Asif-Sattar, Sohl, Kass, Cerdá, Gruenewald, Studdert, Wintemute.
Statistical analysis: Stewart, Wright, Shev, McCort, Gruenewald.
Obtained funding: Gruenewald, Wintemute.
Administrative, technical, or material support: Kagawa, Pallin, Asif-Sattar, Sohl, Kass, Gruenewald, Studdert.
Supervision: Kagawa, Gruenewald, Wintemute.
Conflict of Interest Disclosures: Dr Kagawa reported receiving grants from the National Institute on Alcohol Abuse and Alcoholism, the California Wellness Foundation, the Heising-Simons Foundation, and the Robertson Fellowship in Violence Prevention Research during the conduct of the study. Drs Stewart, Kass, Cerdá, Gruenewald, and Studdert and Mss Asif-Sattar and Sohl reported receiving grants from the National Institute on Alcohol Abuse and Alcoholism during the conduct of the study. Mss Wright, Pear, and Pallin, Drs Shev and Wintermute, and Mr McCort reported receiving grants from the National Institute on Alcohol Abuse and Alcoholism, the California Wellness Foundation, and the Heising-Simons Foundation during the conduct of the study.
Funding/Support: This study was funded by grants 1R01AA023551-01A1 from the National Institute on Alcohol Abuse and Alcoholism, 2014-255 from the California Wellness Foundation, and 2016-219 from the Heising-Simons Foundation and by the Robertson Fellowship in Violence Prevention Research. Funding from the National Institutes of Health was not used in developing and describing the policy implications of the findings from this research.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Kirthana Srikanth (an undergraduate student at University of California, Davis) scanned paper criminal rap sheets and abstracted data from these records for use in the analysis. She was compensated for her work. Andrew Gaidus and Bill Ponicki (GIS specialist and research scientist at the Prevention Research Center) geocoded addresses and linked them to census and alcohol outlet information. They were compensated for their work.
Create a personal account or sign in to: