Table 1.

Articles reporting data on age of onset for firearm carriage and patterns of firearm carriage

Article	Design	Sample size	Study sample and setting	Unique attributes	Key results for patterns of carriage	Limitations	Nos score
Arria et al. (1995)	Cohort Surveyed annually for 5 years	1714	Students from 43 classrooms within 19 public schools from 5 districts in Baltimore, MD	M age (w1) 9 M age (w5) 13 50% Male 73% Black 26% White	Past 1-year carriage = 10% (male); 1% (female). Rates of gun carriage increased across waves of data, starting at < 2% (w1) and increasing to 10% (male) and 1% (female) by wave 5 Less lethal weapon carriage (stick, knife) was associated with initiating carrying a gun at later waves across multiple waves	Limited generalizability Single city Younger adolescent sample Missed students who may have dropped out at later waves Self-report data	6
Ash et al. (1996)	Qualitative Semi-structured interviews	63	Youth offenders recruited from 5 detention centers in Atlanta, GA	Ages 13–18 M age 15.7 67% Male 66% Black 34% White	Patterns of carriage: Identified 4 distinct carrier types Non-carriers (n = 14): 71% of this group owned a gun Intermittent carriers (n = 14): Defined as having weeks and/or months without carrying; each owned ≥ 1 gun Part-time carriers (n = 11): Carried more frequently than Intermittent carriers and had increasing frequency over time in response to stressors (e.g., friend was shot) Constant carriers (n = 24): Defined as carrying all/most of the time; had carried constantly since acquiring a gun Small minority of carriers decreased carriage over time (n = 6) Purposeful acquisition of first firearm positively correlated with becoming frequent/constant carriers (p < 0.05)	Limited generalizability Nonrandom convenience sample of youth offenders from single city Self-report data Small sample	6
Beardslee et al. (2018a)	Cohort Pittsburgh Youth Study Analysis from 6-month surveys (3-year; grade 2–4) and 12-month surveys (age 10–17)	485	Random sub-sample of male participants from youngest cohort of large school-based longitudinal study in Pittsburgh, PA	100% male 56% Black 41% White Excluded non-White/Black participants	Lifetime carriage = 27% (Black); 12% (White): Black youth more likely to carry in adolescence than White youth (OR = 3.3); Youth with higher levels of conduct problems and peer delinquency at earlier childhood waves, as well as those with increases in conduct problems across early childhood, were more likely to initiate gun carriage prior to age 18 Examining whether racial differences in carrying were due to differential exposure to risk factors or differential sensitivity, study found more support for differential exposure, with 60% of the race effect on carriage being attributable to either initial peer delinquency levels or initial levels of conduct problems	Limited generalizability Limited to male Black/White youth sample Single city sample Self-report data. Small sample size may limit ability to detect group differences in gun carriage predictors	6
Beardslee et al. (2018b)	Cohort Pathway to Desistance Study: Youth surveyed every 6-month for 3 years then every 12-month for 4 years	1170	Male youth offenders recruited from court system in two counties; one in AZ (Maricopa) and one in PA (Philadelphia)	Ages 14—19 at baseline 100% male. 42% Black 34% Hispanic 19% White 100% offenders 70% on active probation at baseline	Across 10 waves, carriage rates ranged from 15% (w1) to 10% (w10) with non-linear decrease; peak gun carriage was 17% at w7 data collection (i.e., 4 years after baseline; mean age 21) Youth with gun violence exposure (witness/victim of gun violence) were 43% more likely to engage in gun carriage at the next wave after controlling for time-stable and time-varying (exposure to peers who carried; exposure to peers engaged in other criminal acts, developmental changes, changes in gun carrying from incarceration or institutionalization) covariates. No evidence that non-gun violence exposure conferred same risk	Limited generalizability Male sample of juvenile offenders in two states Self-report data Youth offenders likely have fluctuations in carrying over time	6
Dong and Wiebe (2018)	Cohort NLSY97 administered annually between 1997 to 2011 and then biennially (16 waves)	1585	Youth living in urban areas (as defined by Census criteria) at wave 1 data collection	Data from the 1997 Nat. Longitudinal Survey of Youth (NLSY97) Age 14.27 (SD1.49) 76% Male 29% Black 26% Hispanic 44% White	27.1% lifetime handgun carriage incidence 4–6% past 12-month handgun carriage (at each wave) Mean age of initiation of carriage = 18.2 years old Temporary peak of carriage at age 15; decline during late adolescence until a steady increase in carriage at age 22 Participants carried for an average of 2.7 years Urban Black/Hispanic youth were more likely to carry during late adolescence/early adulthood and less likely to carry than White youth during adulthood Duration of carriage was associated with higher odds of violence perpetration (OR = 1.2) and selling drugs (OR = 1.2) Persistent carriers (those carrying before and after age 18) had higher likelihood of arrest than adolescent carriers (IRR = 2.6)	Data could not distinguish between legal and illegal handgun carriage Self-report data Measures of adult drug sale and violent offending restricted to respondents reporting prior arrest; small comparison group	6
Lizotte et al. (1997)	Cohort Rochester Youth Survey administered at 6-month intervals (w1–9) then 2.5 years intervals (w 9–10)	615	Male youth recruited from Rochester public schools (7th/8th grade); High-crime areas (defined by arrest rates) oversampled	M Age 15 100% male 63% Black Analysis limited to male youth in the sample for w4 through w10	22% carried a gun at some point during any wave of the study; 6% carried a gun at w4 (age 15) 10% carried at w10 (age 20) Overall carriage rate consistent across waves Patterns of Carriage Only 1/3 of subjects carried from one wave to another, suggesting intermittent carriage among participants 53% who reported carrying only carried in one wave 21% who reported carrying carried in 2 waves 26% who reported carrying carried for ≥ 3 waves	Self-report data Sample limited to male youth Oversampled high-crime areas, not involving suburban/rural sample	6
Loeber et al. (2004)	Cohort Developmental Trends Study with dyad (parent and their child) cohort administered at 12-month intervals for 13 years	177	Male children living with ≥ 1 biological parent referred from primary care clinics in either PA or GA	Ages 7–12) (baseline); followed annually to 19 30% Black 70% White 53% Urban 57% not living with biological father 41% from low SES	20% carried a gun at least once during the study 1% reported carrying a gun at age 12 12% carried carrying a gun at age 17 Earliest age of carriage was 10 years old (< 2%) Patterns of Carriage 61% reported carrying for only 1 year 31% carried for 2-years 8% carried for 3–4 years	Limited generalizability Small convenience sample referred from healthcare clinic Self-report data. Sample limited to male youth	7
Reid et al. (2017)	Cohort Pathways to Desistance Study, administered every 6 months for 2 years	1170	Convicted male youth offenders in PA and AZ (recruited from court setting). Proportion charged with drug offenses capped at 15%	Ages 14–19 M age 16 100% male 42% Black 34% Hispanic 19% White	Among the 51% of youth who reported lifetime carriage: 57% carried during 1 period 25% during 2 periods 11% during 3 periods 6% during 4 periods 2% during all 5 periods	Limited generalizability Sample limited to male offenders from two cities Self-report data	6
Sheley (1994)	Cross-sectional	758	Male students from 10 urban schools in 5 cities in CA, IL, LA, and NJ. All schools had prior firearm incidents	M age 16 Range (age 15–17) 72% Black 3% White 19% Hispanic	12% of students carried a gun routinely (all/most of the time) with 23% reporting carriage intermittently (now/then) Youth who used and sold drugs or only sold drugs but did not use drugs were more likely to carry routinely (19% vs. 5%, p < 0.05)	Cross-sectional data Non-representative sample of urban youth (non-random) Self-report data	4
Spano and Bolland (2011)	Cohort Mobile Youth Survey administered annually for 2 years	1049	Adolescents from 12 high-poverty neighborhoods in Mobile, Alabama were recruited from school, homes, community and church locations	Age 9–19 M age 13 (T1) 42% male 100% Black (2000–2001) Sample limited to those who were included in T2 and had not carried at T1	8% initiated gun carrying at T2 (past 90-day); Results support the nexus hypothesis (carriage occurs among a small number of gang youth at increased risk for violence behavior and exposure) over the diffusion hypothesis (gun carriage results directly from violence exposure) for initiation of gun carriage in urban youth Youth gang members who were exposed to violence at T1 were more likely to initiate gun carriage at T2 (OR = 6.5) Gang members engaged in violent behavior at T1 were more likely to initiate gun carriage at T2 (OR = 5.8) Gang members who were engaged in violent behavior and had exposure to violence at T1 were more likely to initiate gun carriage at T2 (OR = 7.7)	High attrition rate (36%) Focused on T1 factors predicting carriage initiation (T2) Single 1-year follow-up Limited generalizability 100% Black sample Single city Sample of at-risk youth	6
Spano et al. (2012)	Cohort Mobile Youth Survey administered annually for 2 years	1049	Adolescents from 12 high-poverty neighborhoods in Mobile, Alabama were recruited from school, homes, community and church locations	Age 9–19 M age 13 (T1) 42% male. 100% Black (2000–2001) Sample limited to those who were included in T2 and had not carried at T1	8% initiated gun carrying at T2 (past 90-day); Key findings: Consistent with the stepping stone model, youth engaged in violent behavior (T1) were more likely to initiate carriage at T2 after controlling for violence exposure (AOR = 1.8) Consistent with the cumulative risk model, youth engaged in violent behavior and exposed to violence at T1 were more likely to initiate carriage at T2 compared to youth who had neither (AOR = 2.5)	High attrition rate (36%) Focused on T1 factors predicting carriage initiation (T2) Single one-year follow-up Limited generalizability 100% Black sample Single city, Sample of at-risk youth	6
Spano and Bolland (2013)	Cohort Mobile Youth Survey administered annually for 2 years	1049	Adolescents from 12 high-poverty neighborhoods in Mobile, Alabama were recruited from school, homes, community and church locations	Age 9–19 M age 13 (T1) 42% male. 100% Black (2000–2001) Sample limited to those who were included in T2 & had not carried at T1	8% initiated gun carrying at T2. Key findings: Youth experiencing violent victimization (OR = 2.3) and violent behavior (OR = 1.9) at T1 increased the likelihood of initiating gun carriage at T2 when examined separately, but only violent victimization (OR = 2.1) was significant when both were included in the model There was no difference between the likelihood of youth who initiated gun carriage at T2 for offensive (i.e., violent behavior only at T1) versus defensive (i.e., violent victimization only at T1) purposes	High attrition rate (36%) Focused on T1 factors predicting carriage initiation (T2) Single one-year follow-up Limited generalizability 100% Black sample Single city, Sample of at-risk youth	7
Steinman and Zimmerman (2003)	Cohort Flint Adolescent Study administered annually for 4 years during high-school (9th-12th grade)	705	African-American public high school students with a GPA ≤ 3.0 at risk for school dropout	M age 15 49% male 100% Black Sample limited to Black students given low base rates of gun carriage among White students Included youth who had left school	20% of youth reported carrying a gun (80% had never carried) Patterns of carriage 15% carried episodically (during one or two waves) 5% carried persistently (during three or four waves) Compared to non-carriers, episodic carriers were more likely male (AOR = 3.6), selling drugs (AOR = 3.2), engaging with adults who carry (AOR = 1.6), engaging in fighting behaviors (AOR = 1.6) and using marijuana (AOR = 1.03) Compared to persistent carriers, episodic carriers were more likely engaging in fighting behaviors (AOR = 1.6) and selling drugs (AOR = 3.3)	Self-report measures. Limited generalizability Sample limited to Black youth Study did not extend earlier than 9th grade so earlier risk and protective factors could not be examined	7
Vaughn et al. (2017)	Cross-Sectional Analysis uses pooled data from the National Survey on Drug Use and Health	7872	Analysis of adolescents age 12–17 from the NSDUH national sample who reported past year gun carriage (2002–2013)	Age 12–14: 40% 15–17: 60% 83% Male 15% Black 60% White 20% low SES 78%Urban	Latent class analysis (LCA) of handgun carrying (1 + times) adolescents identified 4-class solution: “Low-Risk” (48%): Low substance use, violence, delinquency “Alcohol/MJ users” (20%): High substance use, low levels of violence and delinquency “Fighters” (20%): High levels of violence and delinquency, but low levels of MJ and other drug use “Severe” (12%): High levels of alcohol, MJ, other drugs, violence and delinquency behaviors Socio-demographics: Compared to other classes, the low risk class was made up of more rural White youth from high SES households with fathers. Alcohol/MJ users were more likely older (15–17) Black youth from urban settings. “Fighter” class was more likely younger (12–14) Black youth from low SES homes in urban settings. “Severe” class had highest % females (males still more common), and were from urban settings. Behavioral: Compared to the low-risk group, youth in classes 2–4 were more likely to report greater risk propensity, greater parental conflict, and lower school engagement. This was most pronounced in the “severe” group. The Alcohol/MJ group and Severe group had lower levels of parental-limit setting Frequency of gun carriage: Severe subset youth were more likely to carry frequently compared to the low risk (RR 1.5), alcohol/MJ class (RR 1.5), and fighter class (RR 1.4). No difference between first 3 classes for frequency. Class 4 (severe) had higher likelihood of lifetime and past year arrest	Cross sectional data Self-report Motivations not included	7
Watkins et al. (2008)	Cross-sectional	967	338 adolescent males (age < 17) in detention; 629 adult male arrestees from Correctional Facilities in St. Louis, MO	Adolescents M age 15 100% Male 94% Black Adults M age 31 100% Male 87% Black	Adolescents: 46% reported carrying a gun outside of the home most/all of the time, 41% seldom carried (once per month), 13% had not carried in last year Adults: 12% carried frequently (most/all of the time), 19% seldom (once per month), 69% not at all	Limited generalizability Small sample of juvenile arrestees from a single city Mostly Black sample Self-report measures Cross-sectional data	6