PUBLISHED
UNITED STATES COURT OF APPEALS
FOR THE FOURTH CIRCUIT
No. 12-5045
UNITED STATES OF AMERICA,
Plaintiff - Appellee,
v.
BENJAMIN TOD CARTER,
Defendant - Appellant.
Appeal from the United States District Court for the Southern
District of West Virginia, at Charleston. John T. Copenhaver,
Jr., District Judge. (2:09-cr-00055-1)
Argued: March 20, 2014 Decided: April 30, 2014
Before NIEMEYER and DIAZ, Circuit Judges, and HAMILTON, Senior
Circuit Judge.
Affirmed by published opinion. Judge Niemeyer wrote the
opinion, in which Judge Diaz and Senior Judge Hamilton joined.
ARGUED: Lex A. Coleman, OFFICE OF THE FEDERAL PUBLIC DEFENDER,
Charleston, West Virginia, for Appellant. Philip Henry Wright,
OFFICE OF THE UNITED STATES ATTORNEY, Charleston, West Virginia,
for Appellee. ON BRIEF: Mary Lou Newberger, Federal Public
Defender, Jonathan D. Byrne, OFFICE OF THE FEDERAL PUBLIC
DEFENDER, Charleston, West Virginia, for Appellant. R. Booth
Goodwin, II, United States Attorney, Joshua C. Hanks, Assistant
United States Attorney, OFFICE OF THE UNITED STATES ATTORNEY,
Charleston, West Virginia, for Appellee.
NIEMEYER, Circuit Judge:
Following his conviction and sentencing for possessing two
firearms while being an unlawful user of and addicted to a
controlled substance (marijuana), in violation of 18 U.S.C. §
922(g)(3), Benjamin Carter appealed, contending that § 922(g)(3)
infringed on his right to bear arms, in violation of the Second
Amendment. We vacated the judgment and remanded the case to the
district court to allow the government to substantiate the fit
between § 922(g)(3) and the government’s important interest in
protecting the community from gun violence. See United States
v. Carter (“Carter I”), 669 F.3d 411 (4th Cir. 2012). After
taking evidence from both sides, the district court held that
the government had carried its burden in justifying the
regulation of guns under § 922(g)(3), and Carter filed this
second appeal.
Because we agree with the district court that the
government adequately demonstrated a reasonable fit between its
important interest in protecting the community from gun violence
and § 922(g)(3), which disarms unlawful drug users and addicts,
we now affirm.
I
In Carter I, we recited the facts:
Responding to complaints of suspected drug activity at
735 Central Avenue, Charleston, West Virginia, a two-
2
unit apartment building where Carter was living at the
time, Charleston police investigated by knocking on
doors and talking with persons who answered. After
finding evidence of marijuana use in the first unit,
the officers proceeded to knock on Carter’s door.
Carter answered and allowed the officers to enter his
apartment. Upon smelling marijuana, the officers
questioned Carter, who acknowledged that he had been
smoking marijuana and indeed that he had been using
the drug for 15 years. The officers recovered from
the apartment 12 grams of loose marijuana, 15 grams of
partially smoked blunts, a digital scale, $1,000 in
larger bills, and $122 in smaller denominations.
Carter also informed the officers about two firearms
in his closet -- a semi-automatic pistol and a
revolver -- and disclosed that he had purchased the
weapons from a friend a week earlier for his defense.
He later explained in more detail that he had
purchased the guns because he lived in “a bad
neighborhood” and needed weapons to protect himself
and his nephew, who also lived with him in the
apartment. Indeed, at sentencing, Carter’s attorney
represented to the court that one month after Carter’s
arrest in this case, the other unit in the apartment
building was burglarized, and his neighbor was shot
eight times.
Carter I, 669 F.3d at 413.
After Carter was indicted for violating 18 U.S.C. §
922(g)(3), he filed a motion to dismiss the indictment, arguing
that the statute violated his Second Amendment rights. When the
district court denied his motion, Carter entered a conditional
guilty plea that preserved his right to appeal the court’s
ruling on the motion. After accepting Carter’s guilty plea, the
court sentenced Carter to three years’ probation.
On appeal, we vacated the judgment and remanded the case to
the district court for further consideration of Carter’s Second
3
Amendment challenge. We assumed that Carter’s circumstances
implicated the Second Amendment but held that, because he could
not claim to be a law-abiding citizen, any infringement of his
right to bear arms would not have implicated a “core” Second
Amendment right. Carter I, 669 F.3d at 416; see also District
of Columbia v. Heller, 554 U.S. 570, 635 (2008). We therefore
applied intermediate scrutiny to review Carter’s challenge.
Carter I, 669 F.3d at 417. Under intermediate scrutiny, the
question thus became whether there was “a reasonable fit”
between § 922(g)(3) and “a substantial [or important] government
objective.” Id. (quoting United States v. Chester, 628 F.3d
673, 683 (4th Cir. 2010)) (internal quotation marks omitted).
We readily concluded that the government had advanced an
important governmental interest in protecting the community from
crime and, in particular, from gun violence. Carter I, 669 F.3d
at 417. On whether disarming drug users and addicts through §
922(g)(3) reasonably served that interest -- whether there was
“a reasonable fit between the important goal of reducing gun
violence and the prohibition in § 922(g)(3)” -- we noted that
the government could “resort to a wide range of sources, such as
legislative text and history, empirical evidence, case law, and
common sense, as circumstances and context require[d].” Id. at
418. We found that while the government had made plausible
commonsense arguments about the risks of mixing drugs and guns,
4
it had “presented no empirical evidence or data to substantiate
them.” Id. at 419-20. Therefore, in light of Chester and
United States v. Staten, 666 F.3d 154 (4th Cir. 2011), we
remanded the case to the district court to “allow the government
to develop a record sufficient to justify its argument that drug
users and addicts possessing firearms are sufficiently dangerous
to require disarming them.” Carter I, 669 F.3d at 419.
On remand, both the government and Carter submitted a
number of publications and studies to the district court about
the behavioral tendencies of drug users. After considering the
evidence, the court concluded that the government had carried
its burden, finding that the data indicated “a correlation
between violent crime . . . and drug use.” While the court
acknowledged that the government’s studies did not prove “a
strict causal nexus” between drug usage and violence, it found
that “the two factors frequently coincide.” In addition, it
pointed to “common-sense notions” that supported the fit between
drug users and violence, noting (1) that drug users are more
likely to encounter law enforcement; (2) that their criminal
associations increase the risk of violence; (3) that the high
price of drugs is likely to lead to violent property crimes; and
(4) that drug use impairs judgment. The court then concluded:
Based upon the narrowed design of the statute, the
empirical and scholarly evidence relied upon, the
weight of precedent nationwide, and common sense, the
5
United States has shouldered its burden of
establishing that section 922(g)(3) is reasonably
fitted to achieve the substantial governmental
objective of protecting the community from crime by
keeping guns out of the hands of those impaired by
their use of controlled substances. The court,
accordingly, concludes that section 922(g)(3) is
constitutional as applied to Mr. Carter.
From the district court’s judgment on remand, Carter filed
this second appeal.
II
Carter contends that, on remand, the government still
failed to prove that a regulation disarming drug users
reasonably serves the important governmental interest of
protecting the community from gun violence. 1
The government was required to show that the fit between §
922(g)(3) and the government’s important goal is “reasonable,
not perfect.” Carter I, 669 F.3d at 417 (quoting United States
v. Marzzarella, 614 F.3d 85, 98 (3d Cir. 2010)). It was not
required to prove that “the regulation is the ‘least intrusive
means of achieving the relevant government objective, or that
there be no burden whatsoever on the individual right in
1
Carter also presents arguments in his brief that we
previously resolved in Carter I, presumably to preserve them for
further review. He again argues that we should employ strict
scrutiny in reviewing his claim that § 922(g)(3) infringes on
his Second Amendment rights; that § 922(g)(3) is overly broad;
and that § 922(g)(3) is underinclusive. Because we disposed of
these issues in Carter I, we discuss them no further in this
opinion. See Carter I, 669 F.3d at 416-17, 420-21.
6
question.’” Staten, 666 F.3d at 159 (quoting United States v.
Masciandaro, 638 F.3d 458, 474 (4th Cir. 2011)). Moreover, its
burden in this case was lower than in other § 922(g) Second
Amendment cases because of § 922(g)(3)’s “limited temporal
reach” -- i.e., the fact that § 922(g)(3)’s prohibition lasts
only as long as the individual remains an unlawful drug user or
addict. Carter I, 669 F.3d at 419.
Carter argues that the district court, in concluding that
the government carried its burden, erred in two respects: (1)
it improperly relied on factors other than empirical evidence in
evaluating the soundness of § 922(g)(3); and (2) it failed to
recognize that the studies submitted by the government were
inadequate because they related to drug use generally rather
than marijuana use specifically and they failed to prove a
causal link between marijuana use and violence. He maintains
that the studies he submitted demonstrate that, in fact,
“marijuana users are not prone to violent behavior.” (Emphasis
added). We address these points seriatim.
A
On the scope of the district court’s consideration on
remand, Carter contends that the court improperly relied on
factors other than empirical evidence in evaluating the validity
of § 922(g)(3). He asserts that in Carter I, we rejected the
7
government’s use of non-evidentiary support, such as its
reliance on common sense, and that therefore the court was
required to consider only evidence “presented in the crucible of
an adversary proceeding.” While he acknowledges that the
district court did in fact receive empirical studies into
evidence, he notes that its determination “included a heavy
reliance on other factors,” such as the “design of the statute,”
the “weight of precedent nationwide,” and “common sense.”
Without these other factors, he argues, the government’s showing
was insufficient.
Carter’s argument misreads our prior opinion in this case.
In Carter I, we held that, in establishing the “fit between a
regulation and a governmental interest,” the government “may
resort to a wide range of sources, such as legislative text and
history, empirical evidence, case law, and common sense.”
Carter I, 669 F.3d at 418 (emphasis added). While it is true
that we found the government’s commonsense arguments, standing
alone, insufficient to justify § 922(g)(3), that did not imply
that legislative text and history, case law, and common sense
could play no role in justifying Congress’s enactment. To the
contrary, we noted that the government’s commonsense arguments
in this case were plausible and therefore supported §
922(g)(3)’s constitutionality, observing that the government’s
remaining burden “should not be difficult to satisfy.” Id. at
8
419. In short, our holding in Carter I clearly did not preclude
the district court from considering factors other than empirical
evidence, and, as such, the district court did not err in
upholding § 922(g)(3) by relying on “the narrowed design of the
statute, the empirical and scholarly evidence[,] . . . the
weight of precedent nationwide, and common sense.”
B
Focusing on the substance of the studies presented by the
government to the district court, Carter contends that the data
were inadequate because they related to drug use generally
rather than marijuana use specifically and because they failed
to prove a causal relationship between marijuana use and
violence. He maintains that the studies he submitted, by
contrast, demonstrated that “marijuana users are not prone to
violent behavior.” (Emphasis added).
We have little trouble concluding that the studies
presented to the district court by both the government and
Carter indicate a strong link between drug use and violence. A
study by Carrie Oser and colleagues, offered by the government,
found that probationers who had perpetrated violence in the past
were significantly more likely to have used a host of drugs --
marijuana, hallucinogens, sedatives, and heroin -- than
9
probationers who had never been involved in a violent episode. 2
A 2004 survey of prisoners by the Bureau of Justice, again
offered by the government, found that almost 50% of all state
and federal prisoners who had committed violent felonies were
drug abusers or addicts in the year before their arrest, as
compared to only 2% of the general population. 3 That survey also
found that inmates who were dependent on drugs or abusing them
were much more likely to have a criminal history. 4 The
government also presented a study by Lana Harrison and Joseph
Gfroerer, which found that individuals who used marijuana or
marijuana and cocaine, in addition to alcohol, were
significantly more likely to engage in violent crime than
individuals who only used alcohol. 5 And finally, the government
presented a study by Virginia McCoy and colleagues, which found
2
Carrie B. Oser et al., The Drugs-Violence Nexus Among
Rural Felony Probationers, 24 J. Interpersonal Violence 1285,
1293 tbl.1 (2009) (hereinafter Oser et al., Nexus).
3
Bureau of Justice Statistics, U.S. Department of Justice,
Drug Use and Dependence, State and Federal Prisons, 2004, at 7 &
tbl.6 (2007) (hereinafter BJS Survey).
4
BJS Survey, at 7 & tbl.7.
5
Lana Harrison & Joseph Gfroerer, The Intersection of Drug
Use and Criminal Behavior: Results from the National Household
Survey on Drug Abuse, 38 Crime & Delinquency 422, 433 tbl.4
(1992) (hereinafter Harrison & Gfroerer, Intersection).
10
that chronic cocaine and opiate users were more likely than
nonusers to engage in robbery and violence. 6
Carter seeks to marginalize these studies, arguing first
that they are too broad and discuss only “general categories of
offenders, including those who abuse a range of controlled
substances.” He contends that, even if there is a link between
“harder” controlled substances and violence, the government’s
evidence does not indicate that marijuana users are prone to
violence. To the contrary, he claims that the evidence he
submitted disproves such a link. Yet, even if such a
particularized demonstration is necessary -- an issue we need
not reach -- the studies presented by the government amply
demonstrate a connection between marijuana use specifically and
violence. The Harrison and Gfroerer study, for instance, found
that, “[e]ven after controlling for other variables[,] such as
age, race, income, education, and marital status, . . . using
marijuana in the past year . . . [was] significantly related to
criminal behavior.” 7 Also, the study by Oser and colleagues
6
H. Virginia McCoy et al., Perpetrators, Victims, and
Observers of Violence: Chronic and Non-Chronic Drug Users, 16 J.
Interpersonal Violence 890, 900 (2001).
7
Harrison & Gfroerer, Intersection, at 432-35 & tbl. 6.
The study used logistic regression and found that individuals
who used marijuana in the past year were more than twice as
11
found that, among probationers, individuals who had been
involved in violence were more likely to have used marijuana. 8
Finally, the 2005 National Survey on Drug Use and Health found
that individuals arrested for a serious violent or property
offense in the last year were much more likely than non-
arrestees to have used marijuana. 9
Moreover, the evidence that Carter offered to refute the
link between marijuana use and violence -- a study by Evelyn Wei
and colleagues 10 -- actually provides additional evidence that
marijuana use and violence coincide. 11 The Wei study tracked the
likely to report both committing and being booked for violent
crimes.
8
Oser et al., Nexus, at 1293 tbl.1.
9
Office of Applied Studies, Substance Abuse and Mental
Health Services Administration, National Survey on Drug Use and
Health: Illicit Drug Use Among Persons Arrested for Serious
Crimes (2005). This survey found that 46.5% of individuals
arrested for a violent offense (murder, rape, robbery, or
aggravated assault) or a property offense (burglary, theft,
motor vehicle theft, or arson) had used marijuana in the past
year, compared to 10.0% of those not arrested for any serious
offense.
10
Evelyn H. Wei et al., Teasing Apart the Developmental
Associations Between Alcohol and Marijuana Use and Violence, 20
J. Contemp. Crim. Just. 166 (2004) (hereinafter Wei et al.,
Teasing Apart).
11
Carter also presented a 2003 “West Virginia Drug Threat
Assessment” report to cast doubt on the link between marijuana
use and violence based on its statement that “[m]arijuana
distributors in West Virginia occasionally commit violent crimes
12
behavioral development of “inner-city adolescent males” for ten
years and found that, “at age 18, frequent marijuana users were
11 times more likely than nonfrequent users to . . . engage in
violence.” 12 The study also found that marijuana use in one year
frequently predicted violence in the subsequent year. 13 Carter
argues nonetheless that the Wei study militates in his favor
because, when it controlled for “risk factors,” the correlation
between marijuana use in adolescence and violence in young
adulthood was not statistically significant. 14 In this instance,
we do not think that the Wei study’s failure to identify a
to protect their product and turf; however, marijuana abusers
rarely commit violent crimes.” National Drug Intelligence
Center, U.S. Department of Justice, West Virginia Drug Threat
Assessment 13 (2003). This conclusory statement, however, lacks
any empirical or even anecdotal support, and therefore we accord
it no weight.
12
Wei et al., Teasing Apart, at 171, 176.
13
Wei et al., Teasing Apart, at 177 & tbl.3.
14
Wei et al., Teasing Apart, at 177-178 & tbl.3. To provide
a bit more detail: Wei and colleagues used logistic regression
to measure whether marijuana use in adolescents aged 11 to 14
was correlated with their engaging in violence when they were
aged 15 to 20. To isolate the effect of marijuana use, Wei and
colleagues controlled for various “risk factors”: self-reported
property crime, low academic achievement, poor communication
with caretaker, caretaker perception of bad neighborhood,
African-American ethnicity, and hard drug use. After
controlling for those variables, they found that adolescents who
used marijuana were still 1.91 times more likely to engage in
violence later in young adulthood. However, Wei and colleagues
called this relationship “spurious” because the p-value was only
0.068. Id. at 178.
13
statistically significant correlation is particularly relevant. 15
Indeed, we note that the study, even when controlling for risk
factors, still found that adolescents who used marijuana were
almost twice as likely to engage in violence when they became
young adults. Thus, the Wei study, far from undercutting the
government’s position, provides it with strong support.
Carter also objects to the government’s evidence on the
grounds that it demonstrated, at most, a correlation between
marijuana use and violence and not a causal relationship.
Quoting the Wei study, he argues that “[t]he relationship
between marijuana use and violence ‘is due to the selection
effects whereby these behaviors tend to co-occur in certain
individuals, not because one behavior causes the other.’”
(Emphasis added) (quoting Wei et al., Teasing Apart, at 166).
This argument is flawed, however, because it assumes,
incorrectly, that Congress may not regulate based on
correlational evidence. We conclude that it may and that the
15
First, we think it rather irrelevant to § 922(g)(3) --
which concerns active unlawful drug users -- whether marijuana
use among adolescents predicts violence years later. Second,
one of Wei’s “risk factors” was hard drug use. But Congress
would be well within its rights in disarming marijuana users if
such users were more likely to engage in violence because of
their hard drug use. Controlling for hard drug use improperly
weakened the correlation. Third, and most critically, a p-value
of 0.068 indicates that there was only a 6.8% chance that the
correlation was due to chance. Scientists may insist on p-
values of 0.05, but Congress is not so constrained.
14
government need not prove a causal link between drug use and
violence in order to carry its burden of demonstrating that
there is a reasonable fit between § 922(g)(3) and an important
government objective. See Staten, 666 F.3d at 164-67 (upholding
§ 922(g)(9)’s disarmament of those convicted of a misdemeanor of
domestic violence in large part based on correlational evidence
about recidivism rates). Indeed, the studies put forward by
both Carter and the government in this case illustrate just how
powerful correlational evidence can be. The Harrison and
Gfroerer study and the Wei study both used logistic regression
to show that individuals who used marijuana were much more
likely to engage in violence, even controlling for multiple
demographic and behavioral variables including age, race,
economic status, marital status, and educational level. While
eliminating these potentially confounding variables does not
prove that marijuana use causes violence, it substantially
bolsters the link and helps to justify regulating gun possession
by marijuana users. We have emphasized that, under intermediate
scrutiny, the fit between the regulation and the harm need only
be reasonable, not perfect. Carter I, 669 F.3d at 417. The
correlational evidence put forward by the parties in the present
case easily clears that bar.
While the empirical data alone are sufficient to justify
the constitutionality of § 922(g)(3), we find that common sense
15
provides further support. In Carter I, we noted the
government’s argument that “due to the illegal nature of their
activities, drug users and addicts would be more likely than
other citizens to have hostile run-ins with law enforcement
officers, which would threaten the safety of the law enforcement
officers when guns are involved.” 669 F.3d at 419. The
government also warned that “the inflated price of illegal drugs
on the black market could drive many addicts into financial
desperation, with the common result that the addict would be
‘forced to obtain the wherewithal with which to purchase drugs
through criminal acts either against the person or property of
another or through acts of vice such as prostitution or sale of
narcotics.’” 16 Id. Finally, the government suggested that drugs
“impair [users’] mental function . . . and thus subject others
(and themselves) to irrational and unpredictable behavior.” 17
Id. at 420 (omission in original) (internal quotation marks
omitted); see also United States v. Dugan, 657 F.3d 998, 999
(9th Cir. 2011) (“Habitual drug users . . . more likely will
16
This hypothesis finds support in the evidence submitted
by the government, which indicates that approximately 18% of
federal and state prisoners committed their crimes in order to
obtain money for drugs. BJS Survey, at 6. This figure rises to
30% for state prisoners arrested for property offenses. Id.
17
This suggestion was also borne out by the government’s
evidence, which reports that 32% and 26% of state and federal
prisoners, respectively, were using drugs at the time of their
offense. BJS survey, at 5.
16
have difficulty exercising self-control, particularly when they
are under the influence of controlled substances”). We find all
three of these observations convincing, and Carter has provided
no argument grounded in either logic or evidence to undercut
them.
Finally, we observe that every court to have considered the
issue has affirmed the constitutionality of § 922(g)(3) under
the Second Amendment. See, e.g., Dugan, 657 F.3d at 999; United
States v. Yancey, 621 F.3d 681, 682 (7th Cir. 2010) (per
curiam); United States v. Seay, 620 F.3d 919, 925 (8th Cir.
2010); United States v. Richard, 350 F. App’x 252, 260 (10th
Cir. 2009). Indeed, the majority of these courts found the
statute constitutional without relying on any empirical studies.
See Dugan, 657 F.3d at 999; Seay, 620 F.3d at 925; Richard, 350
F. App’x at 260.
At bottom, we conclude that the empirical evidence and
common sense support the government’s contention that drug use,
including marijuana use, frequently coincides with violence.
Carter has failed to present any convincing evidence that would
call this conclusion into question. Accordingly, we join our
sister circuits in holding that § 922(g)(3) proportionally
advances the government’s legitimate goal of preventing gun
17
violence and is therefore constitutional under the Second
Amendment. The judgment of the district court is
AFFIRMED.
18