The federal government of the United States has demonstrated increasing interest in campus
violence over the past decade, beginning with the Student Right To Know and Campus Security Act
of 1990.1 Since 1992, federal law has required that universities publish statistics about criminal victimization on campus. Unfortunately, these data reveal little about gun-related injury and fatality rates, whether on or off campus, and nothing about suicide rates at colleges.
Despite this dearth of reliable data, some authors claim that guns at college have become a public health problem. Recent articles report that guns are a growing menace on college campuses2 and that the prevalence of alcohol and firearms on and around college campuses has had deadly effects.3 Responding to such reports, the Association for Student Judicial Affairs unanimously adopted a resolution in 1994 urging colleges to support tough rules and laws to keep guns off campuses. Yet, nothing is known for certain about guns and gun use at college.2
The authors of Violence on Campus, published in 1998, noted that despite widespread concern about violence on campus, "numbers and percentages of students who possess weapons on their campuses are . . . unknown."1(p117) General weapon carrying at college4-6 was examined in three previous studies; all reported that from 6% to 7% of students carry some form of a weapon on campus. Only one of the studies5 asked specifically about guns, which accounted for 14% of the weapons carried. Our survey is the first to report on-campus firearm possession with information on associated demographic and
behavioral characteristics. The data come from a national random sample of students at accredited 4-year colleges in the United States.
One hundred thirty (93%) of the 140 colleges that participated in the original 1993 College Alcohol Study were surveyed again in 1997. Ten of the 140 colleges that participated in the initial survey did not participate in
1997 because they could not provide a satisfactory random sample of students and their addresses in the time required. For the present analysis, these 130 colleges constitute the study population.
In the original 1993 College Alcohol Study, researchers drew a random national sample from the American Council on Education’s list of accredited 4-year colleges by using probability proportionate to size of enrollment (the sample is described elsewhere).7 This sample contained few colleges for women only and few with less than 1,000 students. To adjust for this problem, the investigators added an oversample of 15 additional colleges with enrollments of fewer than 1,000 students and 10 all-women institutions. Nine colleges, including
seminaries and military and allied health schools, were subsequently dropped from the study because they were considered inappropriate.
In 1993, 140 out of 195 randomly selected colleges agreed to participate. The primary reason for nonparticipation, according to college administrators, was the inability to provide a random sample of the students and their home addresses in the time required for the study. The 55 nonparticipating schools differed from the 140 schools in the study only in enrollment size—colleges with fewer than 1,000 students were underrepresented as a proportion. Because small colleges had originally been deliberately oversampled, sufficient numbers were present for our statistical analysis.
The 130 colleges that participated in the 1997 survey were located in 40 states and the District of Columbia; they represent a cross-section of American higher education. In the sample, White students (77%) predominated, and more women (60%) than men were included, partially because of the inclusion of 6 colleges with enrollments limited to women.
By comparison, the proportions in a 1995 national sample of 4-year undergraduate institutions were 78% White students and 54% women.8
Administrators at each college were asked to provide a random sample of undergraduates drawn from the total enrollment of full-time students. Each of the 130 participating colleges provided a sample of 230 students. We sent a questionnaire entitled College Alcohol Study to students at the end of February 1997. We conducted three separate mailings over a 3-week period; first a questionnaire, then a reminder postcard, followed by a second questionnaire. We timed the mailing to avoid the periods immediately before and after spring break to assure that student responses were based on a time when they were at school. Responses were voluntary and anonymous.
To encourage students to respond, we offered the following cash awards: one $1,000 bonus to a student whose name was drawn from among students responding within 1 week; one $500 award; and 10 $100 awards to students selected from all those who responded.
Of the 29,717 students to whom questionnaires were mailed, 2,797 were eliminated from the sample because schools had given us incorrect addresses or the students had withdrawn from school or were on leave of absence. This reduced
the sample size to 26,920. A total of 15,685 students returned questionnaires, yielding an overall response rate of 58%. The response rate may be an underestimate because it does not take into account all of the students who may not have
Response rates from colleges varied from 26% to 88%, with 12 colleges having response rates below 45%. Response rates were not significantly associated with college or student characteristics, including gun ownership rates. Analyses from the 12 schools with response rates of less than 45% did not differ from analyses that excluded these schools. Results reported in this article include all 130 schools.
The key question in this study was "Do you have a working firearm with you at college?" Respondents chose among the following three options: no; yes, a handgun; and yes, a semiautomatic. Because the term semiautomatic generally refers to handguns, we considered a yes response to either gun question as having a handgun at school. The questionnaire also asked for detailed information about students’ drinking behavior, the extent to which students experienced specific problems as a consequence of drinking, and other behavioral and health-related issues.
To examine how much more likely college gun owners were to have certain demographic and behavioral characteristics compared with students who did not have a gun, we used bivariate and logistic regression analyses. The independent variables in the model are listed in Table 1.
We formed a composite variable, "external effects," by combining three alcohol-related independent variables. External effects were considered present when the student responded positively to having experienced at least one of the following: (a) being arrested for driving under the influence (DUI), (b) damaging property as a result of alcohol ingestion, or (c) driving an automobile shortly after consuming five or more drinks in a row within the previous month.
In all, 15,531 students responded to the gun possession question; of those, 15,040 provided complete data on all of the variables. Missing responses to any of the questions used in the multivariate logistic model resulted in exclusion of the students’ information from the data analyses. The risk factor odds ratios (ORs) presented in the table. are from the multivariate model.
Approximately 3.5% of the college student respondents (6% of the men and 1.5% of the women; OR = 4.0, p < .001) reported they had a working firearm at college (Table 1). Gun ownership was most prevalent among Native Americans, 12% of whom had a gun (OR = 2.4, p < .01); Whites had the next highest rate (4%). Having a gun was more common among students aged 21 years or older (OR = 1.9, p < .001).
Student gun owners were more likely to attend a public rather than a private college (OR = 1.7, p < .001) and to attend a school in the South (OR = 3.1, p < .001) or in the West (OR = 2.0, p < .01). In bivariate but not multivariate analysis, we found that gun owners were more likely to attend a rural than an urban college.
Specific housing and school-related affiliations were also associated with having a gun. Two thirds of the students with guns lived off campus. Among the 53% of students living off campus, 5.3% had guns, compared with 1.3% of the students who lived on campus (OR = 2.7, p < .001). Students who had guns were significantly more likely to be members of fraternities or sororities (OR = 1.6, p < .01) and to live with a spouse or significant other (OR = 1.8, p < .001).
Approximately 40% of the college students in the survey consumed five or more alcoholic drinks in a row over the 2 weeks before the survey (a level classified as binge drinking). Among those students, 4.3% owned a gun, compared with 2.9% of those who were not binge drinkers. Of the 274 students (2%) who responded that they needed an "alcoholic drink first thing in the morning" as an eye-opener, 12.4% had a gun (OR = 2.0, p < .001). Cocaine or crack use since the beginning of the term was more common among gun owners (7%) than among those who did not own a gun (3%).
Having a working firearm at college was more likely if the student had, since the term began, been arrested for DUI, damaged property as a result of alcohol ingestion, or driven an automobile after consuming five or more alcoholic
drinks in the month before the survey. Seven percent of the students who engaged in any of these alcohol-related behaviors had guns, compared with 3% of those who had not had an externality-related driving/drinking episode. Among owners of guns, 36% engaged in at least one of these behaviors, compared with 19% of the student respondents who did not own a gun.
A possible dose-response relationship was suggested by the data: that is, as the frequency of gun ownership increased, the number of the students’ externality-related behaviors increased. Of the 1,961 students who were involved in one such activity, 6% owned a handgun; of the 409 students who engaged in two activities, 9% were gun owners; and of the 7 students who engaged in all three of these alcohol-related behaviors, approximately 43% owned guns (not shown). When each of these three variables (DUI, binge and drive, and damage property) was separately exchanged for the composite externality variable
in the logistic regression (not shown), the multivariate odds ratios were all significant at the p < .05 level (OR for DUI = 2.3; OR for binge and drive = 1.4; OR for damaging property = 1.4).
Among the 2% of college students who indicated that they had consumed alcohol within 6 hours of sustaining an injury severe enough to require medical attention in the past year, 9% reported having a gun with them at school, compared with 3% of those who were not injured (OR = 2.5, p < .001). In the group of 276 students who sustained injuries in driving after a drinking episode, 17 (6%) were injured in a car crash, 57 (20%) in a fight, 68 (25%) in a sports or recreation-related incident, and 134 (49%) in miscellaneous activities. Gun ownership was significantly greater among the students who were injured in alcohol-related fights, car accidents, or sport activities, compared with those not injured as well as those injured in other activities (not shown).
Sixteen percent of students injured in fights owned guns at school, 18% of those injured in a car crash owned guns, 12% of those injured in sports/recreational activities owned guns, and 4% of those injured in other activities owned guns (not shown). At least 1 respondent from each of the 113 schools reported having a gun at school (87%). Eliminating the 5 schools with the highest gun ownership rates did not alter our results.
College campuses are populated with individuals at high risk for unintentional and violent injury (people 18–22 years old), the vast majority of whom are single and experiencing freedom from home and parental supervision for the first time. Studies of this population indicate that consuming alcohol is commonplace and that binge drinking puts both alcohol drinkers and nondrinkers at risk for motor vehicle injuries and fatalities, property damage, trouble with campus or local police and, especially for women, unwanted or unprotected sex.7 Our survey contained many questions about
alcohol and alcohol-related behaviors and is also the first survey to include information on firearm possession in a national sample of college students.
College students who have guns are, as are gun owners generally9,10 more
likely to be male and White. Gun owners in the United States are more likely to live in southern and western regions and in rural areas.9 Although we do not know each student’s permanent residence, we do know
that students who attend schools in the South, the West, and in rural areas are more likely to have guns at school.
Previous studies of the adult population in the United States have examined the gun–alcohol connection and found that increased alcohol consumption, including binge drinking, is associated with questionable gun storage practices
and ownership of semiautomatic firearms.11,12 We found that, among college students, several alcohol and alcohol-related variables were also associated with having a gun. Students were more likely to have a gun if they had damaged property as a result of alcohol ingestion, been arrested for DUI, or driven an automobile after consuming five or more alcoholic drinks.
These alcohol-related behaviors suggest that college gun owners are more likely than those who do not own guns to engage in activities that put themselves and others at risk for severe or life-threatening injuries. Damaging property when intoxicated suggests an inability to contain aggressive impulses. Driving after binge drinking and being arrested for drunk driving also suggest poor judgment and indifference to the effect one’s actions have on the well-being and safety of others.
Our study does not address whether students who drink and drive feel less prone to fatal car crashes than the average individual. Nor does it consider whether gun owners, who are overrepresented in the population of drunk drivers, feel similarly immune to the connection linking gun ownership to increased risk of suicide, homicide, and unintentional gun fatalities.13-16
In 1994, a national review concluded that alcohol is involved in two thirds of college student suicides, in 90% of campus rapes, and in 95% of the violent crime on campus.17 Current evidence suggests a strong association between alcohol consumption and both depression and violent behavior, including rape, domestic violence, homicide, and suicide.18-28 Furthermore, binge drinking is a risk factor for fighting and for carrying weapons as well as for other risky behaviors among high school students.29,30
Approximately 40% of college students reported they had consumed five or more alcoholic drinks in a row over the 2 weeks preceding the survey, a finding consistent with the findings of other
studies.4-6,31 Among these students, 4.3% owned a gun, compared with 2.9% of those who were not binge drinkers. Binge drinking did not retain statistical significance in the multivariate
model because of colinearity with other alcohol-related variables included in the model. When other alcohol-related variables were excluded from the model, however, binge drinking remained associated with gun possession (OR = 1.4, p < .01).
We also found a robust association between gun ownership and a marker for more chronic alcohol dependence—needing an alcoholic drink (an eye opener) to get the day started. Of the 274 students (2%) who responded that they needed an eye opener in the morning, 12.4% owned a gun. This relationship remained statistically significant in the multivariate model despite the presence of other alcohol-related terms.
Our study has several limitations. First, it is a cross-sectional analysis and can only describe associations; it cannot show causation. Second, the data come from a survey, and sample surveys are subject to sampling error; in a sample of 15,000, the results are subject to an error margin of +1% for each question because of chance variation in the model. Third, although students at each school were randomly sampled, the schools that participated in the survey were a nonrandom subset of the original random sample. However, the extent of distortion resulting from working in our nonrandom subset is likely to be slight. The main reason for nonparticipation among the one third of the schools that did not participate was their inability to provide a random sample of students and their addresses within the time required for the study. Such a reason is not a priori thought to be associated with gun ownership.
Finally, self-report data may be subject to inaccuracies because of social desirability responses, recall bias, intentional distortions, or noncandid responses.32 For example, although registered gun owners provide generally valid responses to questions about gun ownership,33 individuals who own guns illegally may be reluctant to admit ownership. With respect to drinking and substance abuse, important independent variables in our study, research has tended to support the validity of self-report data.34-36
Roughly 25% of the households in the United States contain handguns.9 By comparison, our finding that 3.5% of college students have working handguns at school appears quite low. Three and one-half percent
undoubtedly underrepresents the percentage of college students who have a handgun in their permanent residence (usually their parents’ house). Many students probably choose not to bring their guns with them to college, just as they choose not to bring such durable goods as toasters, baseball bats, and dinnerware. In addition, most colleges do not permit gun possession on campus. Why some students choose to bring their guns and others choose not to is unclear.
Unfortunately, our study included only one question about firearms; it did not examine what motivates students to have guns with them at school.
The only previous study to attempt to understand why college students carry weapons (types of weapons were not differentiated in that study) showed that armed students, both men and women, reported that they felt less safe than
students who did not carry weapons.37 We do not know to what extent this perception was based on actual experiences that differentiated armed from unarmed students. Because many guns are owned for protection,9 it is possible that similar fears play a role in the gun ownership we discovered in our study. Unfortunately, we know of no data to corroborate or refute such speculation.
In addition, we do not know why gun ownership and alcohol-related behavioral externalities are related or why gun ownership and alcohol-related injury in car accidents and sports activities are associated. We assume that unobserved
variables explain these associations. It seems doubtful that driving while drunk or damaging property while intoxicated causes an individual to purchase a firearm or that owning a firearm causes one to drink heavily, drive while intoxicated, or become involved in automobile accidents. Perhaps the root cause is an attitude toward risk that leads individuals to
risk losing control of their behavior and leads them to purchase firearms.38,39
Whether a student who gets into fights buys a gun as a consequence of trauma or fear of trauma or, alternatively, is emboldened by possession to act more aggressively is not known. Our study cannot resolve such issues.
Possessing guns on college campuses is usually prohibited by institutional policy.37 Although
reported firearm prevalence in our study is small (3.5%), almost no college seems exempt from some students having working firearms. Efforts aimed at reducing gun possession at college should take into account both the relative risk and the absolute risk of gun possession associated with identifiable characteristics. For example, although the odds ratio associated with needing alcohol in the morning is greater than 2, only 274 students, 34 of whom owned guns, reported such a need. Focusing exclusively on these 274 students misses 94% of the guns. Interventions aimed at behaviors not as strongly associated with gun ownership but more prevalent, such as binge drinking, can potentially affect more people with guns.
From our study, we cannot determine whether "guns are a growing menace on college campuses"2(p33) or whether "the prevalence of alcohol and firearms on and around college campuses has had deadly effects."3(p2) Our study contains no data on whether guns cause problems at colleges or the percentage of students whose gun possession violates either the law or campus policies. However, our study does provide information for the first time on the number and characteristics of students with working firearms at college; the evidence indicates a worrisome association between gun possession and serious alcohol problems.
Support for this study was provided, in part, by the Robert Wood Johnson Foundation (Dr Wechsler), a Health Services Research Award (Dr Miller), an investigator award from the Robert Wood Johnson Foundation (Dr Hemenway), and a grant from the Centers for Disease Control and Prevention (CDC) to the Harvard Injury Control Research Center.
1. Violence on Campus: Defining the Problems, Strategies for Action. Gaithersburg, MD: Aspen; 1998.
2. Lederman D. Weapons on campus? Officials warn that colleges are not immune from the scourge of handguns. Chron Higher Educ. March 9, 1994:A33,34.
3. Nichols W. Violence on campus: The intruded sanctuary. FBI Law Enforcement Bulletin. June 1995:1–5.
4. Presley C, Meilman P, Lyerla R. Alcohol and Drugs on College Campuses: Use, Consequence, and Perceptions of the Campus Environment. Carbondale, IL: The Core Institute; 1993.
5. Patrick K, Covin J, Fulop M, Calfas K, Lovato C. Health risk behaviors among California College students. J Am Coll Health. 1997;45:265–272.
6. Lewis D, Goodhart F, Burns W. New Jersey college students’ high-risk behavior: Will we meet the health objectives for the year 2000? J Am Coll Health. 1986;45:119–126.
7. Wechsler H, Davenport A, Dowdall G, Moeykens B, Castillo S. Health and behavioral consequences of binge drinking in college: A national survey of students at 140 campuses. JAMA. 1994;272:1672–1677.
8. US Department of Education. Digest of Education Statistics. Washington, DC: National Center of Educational Statistics; 1997.
9. Blendon R, Young J, Hemenway D. The American public and the gun control debate. JAMA. 1996;275:1719–1722.
10. Davis JA, Smith TW. General Social Surveys, 1972–1994. Chicago, IL: National Opinion Research Center; 1994. Machine-readable data file.
11. Nelson DC, Grant-Worley JA, Powell K, Mercy J, Holtzman D. Population estimates of household firearm storage practices and firearm carrying in Oregon. JAMA. 1996; 275:1744–1748.
12. Hemenway D, Richardson E. Characteristics of automatic or semiautomatic firearm ownership in the United States. Am J Public Health. 1997;87:286–288.
13. Kellermann A, Reay DT. An analysis of firearm-related deaths in the home. N Engl J Med. 1986;314:1557–1560.
14. 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.
15. Bailey JE, Kellerman 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.
16. Kellermann AL, Rivara FP, Rushforth NB, Banton JE, et al. Gun ownership as a risk for homicide in the home. N Engl J Med. 1993;329:1084–1091.
17. CASA Commission on Substance Abuse at Colleges and Universities. Rethinking rites of passage: Substance abuse on America’s campuses. New York: Columbia University; June 1994.
18. Gustafson R. Threat as a determinant of alcohol-related aggression. Psychol Rep. 1986;58:287–297.
19. Leonard KE, Bromet EJ, Parkinson EK, Day NL, Ryan CM. et al. Patterns of alcohol use and physically aggressive behavior in men. J Studies Alcohol. 1985;46:279–282.
20. Gerson LW. Alcohol-related acts of violence: Who was drinking and the acts that occurred. J Studies Alcohol. 1978; 39:1294–1296.
21. Rada RT, Kellner R, Laws DR, Wilson WW. Drinking, alcoholism and the mentally disordered sex offender. Bulletin Am Acad Psychiatry and Law. 1978;6:296–300.
22. Murphy CM. Treating perpetrators of adult domestic violence. Md Med J. 1994;43:877–883.
23. Cook PJ, Moore MJ. Violence reduction through restriction on alcohol availability. Alcohol Health Res World. 1993;17:151–157.
24. Goodman RA, Mercy JA, Loya F, et al. Alcohol use and interpersonal violence: Alcohol detected in homicide victims. Am J Public Health. 1986;76:144–149.
25. Brent DA, Perper JA, Allman CJ. Alcohol firearms and suicide among youth: Temporal trends in Allegheny County, Pennsylvania 1960–1983. JAMA. 1987;257:3369–3372.
26. Welte JW, Abel EL, Wieczorek W. The role of alcohol in suicides in Erie County NY, 1972–1984. Public Health Reports. 1988;103:648–652.
27. Meyers HB, Zepeda SG, Murdock MA. Alcohol and trauma: An epidemic syndrome. Western J Med. 1990;153:149–153.
28. Shepherd JP, Robinson L, Levers BGH. Roots of urban violence. Injury. 1990;21:139–141.
29. Valois RF, Vincent ML, McKeown RE, Garrison CZ, Kirby SD. Adolescent risk behaviors and the potential for violence: A look at what’s coming to campus. J Am Coll Health. 1993;41:141–147.
30. Wechsler H, Dowdall GW, Davenport A, Castillo S. Correlates of college student binge drinking. Am J Public Health. 1995;85:921–926.
31. Johnston LD, O’Malley PM, Bachman JG. Drug Use Among American High School Seniors, College Students, and Young Adults, 1975–1990. Vol 2. Washington, DC: Government Printing Office; 1991. US Department of Health and Human Services Publication ADM 91-1835.
32. Aday LA. Designing and Conduction of Health Surveys: A Comprehensive Guide. San Francisco: Jossey-Bass; 1989.
33. Kellermann AL, Rivara FP, Banton J, Reay DT, Flinger CL. Validating survey responses about gun ownership among gun owners of registered handguns. Am J Epidemiol. 1991:131:1080–1084.
34. Freier MC, Bell RM, Ellickson PL. Do Teens Tell the Truth? The Validity of Self-Report Tobacco Use by Adolescents. Santa Monica, Calif: RAND; 1991, RAND publication N-3291-CHF.
35. Cooper AM, Sobell MB, Sobell LC, Maisto SA. Validity of alcoholics’ self-reports: Duration data. Int J Addict. 1981; 16:401–406.
36. Midanik L. Validity of self report alcohol use: A literature review and assessment. Brit J Addict. 1988;83:1019–1030.
37. Presley CA, Meilman PW, Cashin JR. Weapon carrying and substance abuse among college students. J Am Coll Health. 1997;46:3–8.
38. Jessor R, Donovan J, Costa FM. Beyond adolescence: Problem behavior and young adult development. New York: Cambridge University Press; 1991.
39. Elliott DS, Huizinga D, Menard S. Multiple problem youth: Delinquency, substance abuse and mental health problems [abstract]. New York: Springer-Verlag; 1991.
All of the authors are affiliated with the Harvard School of Public Health. Matthew Miller and David Hemenway
are with the Department of Health Policy and Management, and Henry Wechsler is with the Department of Health and Social