Research and Practice

THE LIMITED RELEVANCE OF DRUG POLICY:

CANNABIS IN AMSTERDAM AND IN SAN FRANCISCO

[ABSTRACT]

Objectives. We tested the premise that punishment for cannabis use deters
use and thereby benefits public health.

Methods. We compared representative samples of experienced cannabis users
in similar cities with opposing cannabis policies--Amsterdam, the
Netherlands (decriminalization), and San Francisco, Calif
(criminalization). We compared age at onset, regular and maximum use,
frequency and quantity of use over time, intensity and duration of
intoxication, career use patterns, and other drug use.

Results. With the exception of higher drug use in San Francisco, we found
strong similarities across both cities. We found no evidence to support
claims that criminalization reduces use or that decriminalization increases
use.

Conclusions. Drug policies may have less impact on cannabis use than is
currently thought. (Am J Public Health. 2004;94:836 842)

[TEXT]

There is a trend among Western democracies toward liberalization of
cannabis laws. (Cannabis includes both marijuana and hashish.) In 1976, the
Netherlands adopted de facto decriminalization. Under Dutch law, possession
remains a crime, but the national policy of the Ministry of Justice is to
not enforce that law. After 1980, a system of "coffee shops" evolved in
which the purchase of small quantities of cannabis by adults was informally
tolerated and was then formally permitted in shops that were licensed.1 3
During the 1990s, Switzerland, Germany, Spain, Belgium, and Italy shifted
their drug policies in the Dutch direction. Portugal decriminalized
cannabis in 2001, and England similarly reclassified cannabis in 2004.
Canada and New Zealand are currently considering cannabis
decriminalization. These shifts constitute the first steps away from the
dominant drug policy paradigm advocated by the United States, which is
punishment based prohibition.4 6

Moving in the opposite direction, the United States has stiffened criminal
penalties for drug offenses and has increased arrests for cannabis
offenses. Since 1996, voters in 8 states and the District of Columbia have
passed medical-marijuana initiatives, but the federal government has
resisted implementation. In 2001, 723 627 people were arrested for
marijuana offenses.7 In 2002, the Drug Enforcement Administration began
raiding medical-marijuana organizations,8 and the White House Office of
National Drug Control Policy launched a campaign against marijuana.9,10

Such policies are designed to deter use. The core empirical claim made by
criminalization proponents is that, absent the threat of punishment, the
prevalence, frequency, and quantity of cannabis use will increase and will
threaten public health.11 16 The question of whether deterring use enhances
public health was beyond the scope of our study, but we did examine the
proposition that drug policies affect user behavior and deter use. It is
possible that the causal arrow points the other way--that user behavior
affects laws and policies, which has been the case with alcohol policies in
some countries.17 However, the Marijuana Tax Act of 1937, which first
criminalized cannabis, predated widespread cannabis use in the United
States and had clear political origins.18 21 In the Netherlands, de facto
decriminalization of cannabis was first forged in the late 1960s, when use
was spreading among the youth counterculture. But Dutch policymakers
decided that cannabis use was unlikely to lead to deeper deviance and that
criminalization could lead to greater harm to users than the drug itself.3
In neither country, then, was user behavior the effective cause of laws or
policies.

The presumed effects of cannabis policies have been explored by those who
are critical of criminalization in the United States22 and by those who are
skeptical of Dutch decriminalization. 23 However, until now there have been
no rigorously comparative studies of user behavior designed to assess
whether criminalization constrains use or whether decriminalization
increases it. Our study compared the career use patterns of representative
samples of experienced cannabis users in 2 cities with many similarities
but with different drug-control regimes--Amsterdam, the Netherlands
(decriminalization), and San Francisco, Calif (criminalization).

San Francisco was selected as the US comparison city not because it is
representative of the United States but because it is the US city most
comparable to Amsterdam. Both cities are large, highly urbanized port
cities with diverse populations of slightly more than 700000. They are
financial and entertainment hubs for larger regional conurbations, and they
have long been perceived within their home countries as cosmopolitan,
politically liberal, and culturally tolerant.

Law enforcement officials in San Francisco are not as zealous about
enforcing marijuana laws as law enforcement officials are in most other US
cities. Nonetheless, San Francisco is embedded in the drug policy context
of criminalization, which is a markedly different drug policy context than
that of Amsterdam. Buying and selling cannabis are permitted in Amsterdam
in 288 licensed "coffee shops,"24 and public use is permitted, whereas in
San Francisco, buying, selling, and public use of marijuana remain criminal
offenses. In Amsterdam, there is neither proactive nor reactive policing of
use or lowlevel sales, although police do enforce regulations against
coffee shops' advertising, selling to minors, and creating public
nuisances. In San Francisco, there is strong proactive and reactive
policing of sales, and there is moderate reactive policing of use.

These differences in drug policy context are palpable to users. San
Francisco students are suspended from schools and are placed in treatment
for marijuana use. San Francisco users risk citations, fines, and arrests
if they are detected buying, possessing, or using marijuana. In Amsterdam,
users face none of these risks. The use and sale of other illicit drugs
sometimes used by cannabis users is proactively policed in San Francisco.
In Amsterdam, police occasionally engage in reactive policing of complaints
about open use or sale of other drugs, but they do not proactively patrol
in search of these offenses.

Methods

We required not merely a random sample of cannabis users but a random
sample of users who had enough experience (defined as at least 25 episodes
of use during their lifetimes) to answer questions about career use
patterns. In Amsterdam, recruitment of users began as part of a drug-use
prevalence survey of the general population. This survey was administered
to a random sample that was obtained from Amsterdam's Municipal Population
Registry. The overall response rate was 50.2%, which yielded a sample of
4364.25 (The response rate was slightly below the 55% response rate of a
1990 iteration of the survey. Sampling details and an extensive response/
nonresponse study can be found in Sandwijk et al.24 or at
http://www.cedro-uva.org/lib .) Comparisons of responders with
nonresponders and with known city demographic data indicated no need for
weighting. All respondents who reported having used cannabis at least 25
times (nS5; 12.3%) were asked to participate in an in-depth interview about
their cannabis use. Of these 535 experienced users, 216 (40.5%) were
interviewed in 1996.26

This modest response rate necessitated a check of representativeness. We
compared the 216 users who responded with the 319 who did not on 12
demographic and drug-use prevalence variables. Respondents had slightly
higher levels of formal education and slightly higher past-year prevalence
of cannabis use,26,27 but otherwise, they showed no differences compared
with nonrespondents and thus were reasonably representative of experienced
cannabis users in the general population. Homeless and institutionalized
persons were not interviewed for either survey.

Beginning in 1997, the Amsterdam survey of experienced cannabis users was
replicated in San Francisco, where there is no population registry. To
remain consistent with Amsterdam, we first drew an area probability sample
by randomly selecting census tracts, blocks, buildings, households, and
adults within households. We administered a brief prevalence survey
containing demographic and drug-use prevalence questions. Unlike the
Amsterdam prevalence survey, which was an extensive study in its own right,
the brief prevalence survey in San Francisco was principally designed to
generate a random representative sample of experienced cannabis users.

The overall response rate of the San Francisco prevalence survey was 52.7%,
which yielded a sample of 891.28 Of these respondents, 349 reported that
they had used cannabis 25 or more times (39.2% of the population sample and
3 times the prevalence found in the Amsterdam sample) and were asked to
participate in the in-depth interview; 266 (76.2%) respondents were
ultimately interviewed in-depth about their career use patterns. As a check
on their representativeness, respondents were compared with nonrespondents
on 10 demographic and drug-use prevalence variables. No statistically
significant differences were found.

The Dutch questionnaire was translated for use in San Francisco.
Non-English-speaking Asian Americans were excluded because of the
prohibitive costs of translating instruments and training interviewers in
the many Chinese and other Asian languages found in San Francisco. This
exclusion was not consequential, because national prevalence studies show
that illicit drug use among Asian Americans is the lowest of any ethnic
group.29 Also, non English speakers are mostly elderly and are thus least
likely to be cannabis users. However, the instruments were translated into
Spanish, and bilingual interviewers conducted interviews when necessary.
Homeless and institutionalized persons were not interviewed.

Results

Age at Onset, First Regular Use, and Maximum Use

The mean age at onset of cannabis use was nearly identical in both cities:
16.95 years in Amsterdam and 16.43 years in San Francisco. The mean age at
which respondents commenced regular use (=once per month) also was nearly
identical: 19.11 years in Amsterdam and 18.81 years in San Francisco. The
mean age at which respondents in both cities began their periods of maximum
use was about 2 years after they began regular use: 21.46 years in
Amsterdam and 21.98 years in San Francisco. Clear majorities in both cities
reported periods of maximum use of 3 years or less.

Cannabis Use Patterns Over Time

We asked about the frequency and the quantity of use and the intensity and
the duration of intoxication. To assess how these dimensions of use may
have changed over time, we asked about each for 4 periods: first year of
regular use (=once per month), maximumuse period, past year (12 months
before the interview), and past 3 months (3 months before the interview).

Frequency of use. Figure 1 shows the frequency of reported marijuana use
for these 4 periods; the overall pattern is similar across both cities.
During first year of regular use, strong majorities reported use of
cannabis once per week or less, whereas small percentages reported daily
use. Frequency increased during the period of maximum use but declined
sharply thereafter. Amsterdam respondents reported more frequent use than
did San Francisco respondents during their first year of regular use
(t=4.019; dfG9; P= .000) and during their period of maximum use (t=2.979;
dfG9; I=.003). When the maximum-use period was compared with the past year,
daily use declined from 49% to 10% in Amsterdam and from 39% to 7% in San
Francisco. This decline was even greater for the past 3 months.

The basic trajectory of frequency of use across careers was parallel in
both cities. Most users reported a maximum-use period of 2 to 3 years,
after which the vast majority sharply reduced their frequency of use or
stopped altogether. Roughly three fourths of the respondents in each city
reported that they had used cannabis less than once per week or not at all
in the year before the interview.

Quantity of use. Figure 2 shows that in the first year of regular use, few
respondents in either city consumed large quantities of cannabis. Only 3%
in Amsterdam and 5% in San Francisco used 28 grams (approximately 1 ounce)
during an average month. Amsterdam respondents used significantly smaller
quantities than did San Francisco respondents during this period. When the
2 smallest categories were combined, two thirds in Amsterdam (66%) and
slightly less in San Francisco (59%) were found to have consumed 4 or fewer
grams per month during their first year of regular use. More than one third
used less than 2 grams per month during their first year of regular
use--38% in Amsterdam and 35% in San Francisco.

Quantities consumed during maximum-use periods were larger and very similar
across the cities. About two thirds of respondents consumed an average of
14 or fewer grams per month--69% in Amsterdam and 64% in San Francisco.
Less than 1 in 5 respondents in each city (18%) consumed an average of 28
grams per month or more during their maximum-use periods.

During the year before the interview, consumption among those who still
used cannabis declined sharply. Clear majorities used 4 or fewer grams per
month, although this proportion was smaller in Amsterdam (63%) than in San
Francisco (72%) (t= 2.207; df)7; P = .028). About 1 in 3 respondents in
each city reported no use. Overall, the patterns were parallel in both
cities; quantities used increased from first regular use through maximum
use but then quantities used declined steadily or use ceased altogether
over the course of the respondents' careers.

Intensity of intoxication. Respondents were asked to estimate "how high or
how stoned you generally got" when they consumed cannabis. Some recalled
this occurence with greater consistency than did others, but all of them
were able to make basic ordinal distinctions between more- and less-intense
highs. To increase reliability of respondents' estimates, we displayed a
6-point scale ranging from "light buzz" (1) to "very high" (6) and asked
them to select the number that best summarized their highs during each period.

Figure 3 shows that respondents in both cities generally increased the
intensity of their highs during periods of maximum use but moderated their
highs thereafter (past-year and past-month figures exclude those who had
quit). Amsterdam respondents were significantly more likely than San
Francisco respondents to report milder intoxication during the first year
of regular use and during maximum- use periods: mean scores for the first
year were 3.5 in Amsterdam and 3.9 in San Francisco (t=-3.180; dfG6;
P=.002), and these scores rose for maximum-use periods to 3.9 and 4.4,
respectively (t=-4.932, df= 413; P=.000).

The same pattern was found for the more recent periods, although the mean
scores declined. The proportion of respondents who chose 6 (very high)
remained small and was between 3% and 7% in both cities. For highs
experienced during the past year, Amsterdam respondents were again more
likely to report milder intoxication (t=-2.233; df10; P=.026). For the
past-3-month periods, majorities in both cities reported milder highs of 1
to 3 on the 6-point scale. In short, respondents in both cities reported
less intoxication with use over the course of their careers.

Duration of intoxication. We also asked "about how long" respondents were
high during a typical occasion of cannabis use. Reported durations were
correlated with frequency and with quantity but were not a function of
frequency and quantity alone. Here, too, we found a tendency toward
moderation over the course of users' careers in both cities. Figure 4 shows
that Amsterdam respondents reported highs of somewhat longer duration than
reported by respondents in San Francisco during the first year of regular
use (t=2.329; dfG6; P=.020). (One reviewer noted a divergence between San
Francisco respondents, who reported more intense highs during 3 of the 4
periods, and Amsterdam respondents, who reported highs of longer duration
during 1 period. Because we found no reason to suspect that either sample
played up or played down their responses to any of the questions, this
divergence may indicate culture-specific consumption styles or cultural
grammars of intoxication. 30,31) However, during the other 3 time periods
there were no significant differences: in each city, a clear majority of
users regulated their ingestion so that highs lasted 2 to 3 hours or less.
Substantial minorities in each city reported being high for 4 or more hours
during maximum-use periods, but these proportions dropped sharply after
those periods. Of those who used cannabis during the 3 months before the
interview, 89% in Amsterdam and 93% in San Francisco reported being high
for 2 to 3 hours or less.

Overall Career Use Patterns

We also asked respondents to characterize their overall career use
patterns. We presented a typology of trajectories32 and asked them to
identify the 1 that "best describes" their cannabis use over time (Table 1).

Two career use patterns were dominant in both cities. Pattern 4--gradual
increased use followed by sustained decline--was the most common (49.4% of
the combined sample). The second most common was Pattern 6 wide variation
over time (24.4% of the combined sample). Patterns 1, 2, 3, and 5 were each
selected by only 6% to 8% of the combined sample. Pattern 3--stable use
from the beginning onward--was selected significantly more often by
Amsterdam respondents (11.1%) than by San Francisco respondents (1.9%),
whereas Pattern 5--intermittent use (many starts and stops over time)--was
selected significantly more often by San Francisco respondents (9.5%) than
by Amsterdam respondents (3.2%).

These findings are consistent with findings on frequency and quantity of
use and intensity and duration of intoxication, and they have important
public health implications. Claims that cannabis produces addiction or
dependence13-15 lead one to expect that many experienced users would report
Pattern 2--escalation of use over time. But this pattern was reported by
only 6% in both cities, which means that 94% of respondents had overall
career use patterns that did not entail escalation across careers.

Other Illicit Drug Use

Another important question about the effects of drug policies concerns the
use of other illicit drugs. The "separation of markets," in which lawfully
regulated cannabis distribution reduces the likelihood that people seeking
cannabis will be drawn into deviant subcultures where "hard drugs" also are
sold is one public health objective of Dutch decriminalization. 1-3 The
reduction of cannabis use and thereby the reduction of the extent to which
it serves as a "gateway" to "harder" drugs is one public health objective
of US criminalization.11,12,14,16 Users who had ingested cannabis 25 times
or more during their lifetimes were far more prevalent in San Francisco
than in Amsterdam, and the same was true for users of other illicit drugs.
Table 2 shows a significantly lower lifetime prevalence of other illicit
drug use in Amsterdam than in San Francisco. During the 3 months before the
interview, prevalence of crack and opiate use also were significantly
higher in San Francisco, but cocaine, amphetamine, and ecstasy use were not
significantly different. Thus, rates of discontinuation the decline from
lifetime prevalence to prevalence during the past 3 months--were somewhat
higher in San Francisco for cocaine, amphetamine, and ecstasy; however,
rates of discontinuation were high (64%-98%) for all drugs in both cities.

Discussion

Proponents of criminalization attribute to their preferred drug-control
regime a special power to affect user behavior. Our findings cast doubt on
such attributions. Despite widespread lawful availability of cannabis in
Amsterdam, there were no differences between the 2 cities in age at onset
of use, age at first regular use, or age at the start of maximum use.
Either availability in San Francisco is equivalent to that in Amsterdam
despite policy differences, or availability per se does not strongly
influence onset or other career phases.

We also found consistent similarities in patterns of career use across the
different policy contexts. Although a few significant differences were
found in some dimensions of use during some career phases, the basic
trajectory was the same in both cities on all dimensions of use: increasing
use until a limited period of maximum use, followed by a sustained decrease
in use over time or by cessation. It is significant, from a public health
perspective, that clear majorities of experienced users in both cities
never used daily or used large amounts even during their peak periods, and
that use declined after those peak periods. Furthermore, both samples
reported similar steady declines in degree and duration of intoxication.
Only 6% in each city reported escalation of use over time.

We expected differences in drug policies to affect the duration of
cannabis-use careers and the rates of cessation. Criminalization is
designed to decrease availability, discourage use, and provide incentives
to quit. Decriminalization is said to increase availability, encourage use,
and provide disincentives to quit. Thus, we expected longer careers and
fewer quitters in Amsterdam, but our findings did not support these
expectations. Cannabis careers ranged from 1 to 38 years, and 95% of
respondents in both cities reported careers of 3 years or longer. The mean
career length was slightly greater in San Francisco (15 years) than in
Amsterdam (12 years), but this finding was mostly because of the somewhat
higher mean age in the San Francisco sample (34 years vs 31 years).
Similarly, nearly identical proportions of respondents in each city had
quit by the time they were interviewed--33.8% in Amsterdam and 34.3% in San
Francisco.

If drug policies are a potent influence on user behavior, there should not
be such strong similarities across such different drug control regimes. Our
findings do not support claims that criminalization reduces cannabis use
and that decriminalization increases cannabis use. Moreover, Dutch
decriminalization does not appear to be associated with greater use of
other illicit drugs relative to drug use in San Francisco, nor does
criminalization in San Francisco appear to be associated with less use of
other illicit drugs relative to their use in Amsterdam. Indeed, to judge
from the lifetime prevalence of other illicit drug use, the reverse may be
the case.

Our study has limitations and should be replicated in other cities over
longer periods. While our findings share the limitations of all self-report
studies (e.g., vague or selective memory, over- or understatements of
fact), we attempted to minimize these limitations by means of carefully
worded questions, extensive pretesting, and use of multiple measures. The
questionnaire is available under questionnaire at
http://www.cedro-uva.org/lib/cohen. canasd.html. Our comparable samples and
measures helped isolate the effects of drug policies, but "all else" is not
necessarily "equal." Cultural and social conditions in the United States
are different from in the Netherlands; therefore, cannabis use might
increase if the United States were to adopt a Dutch approach. Further
studies that examine prevalence before and after policy shifts would be
illuminating, although previous studies of the impact of marijuana
decriminalization among 11 US states during the 1970s found no increases.33-36

One hypothesis for future research is that with a widely used drug like
cannabis, the informal social controls that users develop as part of their
culture30,31,37-39 have more powerful regulatory effects on their behavior
than do formal social controls such as drug policies. This possibility
emerged from responses to questions about the circumstances respondents
found appropriate for cannabis use. In both cities, relaxation was the most
common purpose of use, and majorities from both cities reported that they
typically used cannabis with friends and at social gatherings. Majorities
in both cities most often mentioned work or study as situations in which
use was inappropriate. In both cities, 69% reported negative emotional
states as unsuitable for cannabis use, and 80% reported having advised
novices about the virtues of moderation.

Conclusions

These data suggest that most experienced users organize their use according
to their own subcultural etiquette--norms and rules about when, where, why,
with whom, and how to use--and less to laws or policies. When experienced
users abide by such etiquette, they appear to regulate their cannabis use
so as to minimize the risk that it will interfere with normal social
functioning. This pattern suggests that if formal drug policies are based
on the folk (informal) drug policies users themselves already practice,
drug policies may achieve greater relevance.

[SIDEBARS AND FOOTNOTES]

About the Authors

Craig Reinarman is with the Department of Sociology, University of
California, Santa Cruz, Calif. Peter D. A. Cohen and Hendrien L. Kaal are
with the Centre for Drug Research, University of Amsterdam, the Netherlands.

Requests for reprints should be sent to Craig Reinarman, Department of
Sociology, University of California, 1156 High Street, Santa Cruz, CA 95064
(e-mail: craigo@ucsc.edu).

This article was accepted June 7, 2003.

Contributors

C. Reinarman assisted in conceiving the study and analyzing the data. H. L.
Kaal composed the figures and the tables that form the core of the data
analysis. P. D. A. Cohen proposed the study, led its conceptualization,
designed and supervised the data analysis, and assisted with writing the
article.

Acknowledgments

The authors would like to acknowledge the generous financial support of the
Dutch Ministry of Health and the US National Institute on Drug Abuse (grant
1 R01 DA10501-01A1), which made this research possible. We thank Manja
Abraham, Ira Glasser, Harry G. Levine, Marsha Rosenbaum, Arjan Sas, and 3
anonymous American Journal of Public Health reviewers for their helpful
comments. An earlier version of this article was presented at the 97th
Annual Meeting of the American Sociological Association, Chicago, Ill,
August 15-19, 2002.

Human Participant Protection

This protocol was approved by all relevant institutional review boards and
funding agencies.

References

1. Engelsman EL. Dutch policy on the management of drug-related problems.
Br J Addiction. 1989;84: 211-218.

2. Leuw E, Marshall IH, eds. Between Prohibition and Legalization: The
Dutch Experiment in Drug Policy. Amsterdam, the Netherlands: Kugler
Publications; 1994.

3. Cohen PDA. The case of the two Dutch drug policy commissions: an
exercise in harm reduction, 1968- 1976. In: Erickson PG, Riley DM, Cheung
YW, O'Hare PA, eds. Harm Reduction: A New Direction for Drug Policies and
Programs. Toronto, Ontario, Canada: University of Toronto Press; 1997:17-31.

4. Musto D. The American Disease: Origins of Narcotics Control. New Haven,
Conn: Yale University Press; 1973.

5. Bruun K, Pan L, Rexed I. The Gentlemen's Club: International Control of
Drugs and Alcohol. Chicago, Ill: University of Chicago Press; 1975.

6. Bewley-Taylor D. The United States and International Drug Control,
1909-1997. London, England: Continuum; 2001.

7. Federal Bureau of Investigation. Uniform Crime Report: Crime in the
United States, 2001. Washington, DC: US Dept of Justice; 2002.

8. Murphy DE. California: medicinal marijuana raid. New York Times.
September 6, 2002:A20.

9. Office of National Drug Control Policy. An open letter to parents about
marijuana. New York Times. September 18, 2002:A23.

10. Office of National Drug Control Policy. My child is not an honor
student. New York Times. October 7, 2002:A5.

11. Kandel DB, Yamaguchi K, Chen K. Stages of progression in drug
involvement from adolescence to adulthood: further evidence for the gateway
theory. J Stud Alcohol. 1992;53:447-457.

12. Center on Addiction and Substance Abuse. Cigarettes, Alcohol, and
Marijuana: Gateways to Illicit Drugs. New York, NY: Center on Addiction and
Substance Abuse; 1995.

13. Hall W, Solowij N, Lemon J. The Health and Psychological Consequences
of Cannabis Use. Canberra, Australia: Australian Government Publishing
Service; 1994.

14. National Institute on Drug Use (NIDA). Marijuana: What Parents Need to
Know. Rockville, Md: NIDA; 1995.

15. Swan N. Marijuana antagonist reveals evidence of THC dependence in
rats. NIDA Notes. 1995;10:1-2.

16. Office of National Drug Control Policy. Drug Facts: Marijuana.
Available at http://www.whitehousedrugpolicy.gov. Accessed May 16, 2003.

17. Christie N. Scandinavian experience in legislation and control.
National Conference on Legal Issues in Alcoholism and Alcohol Usage,
Swampscott, Mass, 17-19 June 1965. Boston, Mass: Law-Medicine Institute,
Boston University; 1965.

18. Becker HS. Outsiders: Studies in the Sociology of Deviance. New York,
NY: Free Press; 1963.

19. Dickson DT. Bureaucracy and morality: an organizational perspective on
a moral crusade. Soc Problems. 1968;16:143-156.

20. Bonnie RJ, Whitebread CH. The Marihuana Conviction: A History of
Marihuana Prohibition in the United States. Charlottesville, Va: University
of Virginia Press; 1974.

21. Himmelstein JL. The Strange Career of Marihuana:Politics and Ideology
of Drug Control in America. Westport, Conn: Greenwood Press; 1983.

22. Nadelmann EA. Drug prohibition in the US: costs, consequences, and
alternatives. Science. 1989;245: 939-947.

23. MacCoun RJ, Reuter P. Interpreting Dutch cannabis policy: reasoning by
analogy in the legalization debate. Science. 1997;278:47-52.

24. Bieleman B, Goeree P. Coffee Shops Geteld: Aantallen Verkooppunten van
Cannabis in Nederland [Coffee Shops Counted: Numbers of Points of Sale of
Cannabis in the Netherlands]. Groningen, the Netherlands: Stichting
Intraval; 2000.

25. Sandwijk PJ, Cohen PDA, Musterd S, Langemeijer MPS. Licit and Illicit
Drug Use in Amsterdam Vol II. University of Amsterdam, Centre for Drug
Research, 1995.

26. Cohen PDA, Sas A. Patterns of Cannabis Use in Amsterdam Among
Experienced Users. Amsterdam, the Netherlands: University of Amsterdam,
Centre for Drug Research; 1997.

27. Cohen PDA, Kaal H. The Irrelevance of Drug Policy: Patterns and Careers
of Experienced Cannabis Use in the Populations of Amsterdam, San Francisco,
and Bremen. Amsterdam, the Netherlands: University of Amsterdam, Centre for
Drug Research; 2001.

28. Piazza, T, Cheng, Y. Sampling Methods and Field Results of the San
Francisco Drug Use Study. Berkeley, Calif: University of California, Survey
Research Center; 1999. Technical Report 44.

29. Substance Abuse and Mental Health Services Administration (SAMHSA).
Preliminary Estimates From the 1994 National Household Survey on Drug
Abuse. Rockville, Md: US Dept of Health and Human Services, SAMHSA, Office
of Applied Studies; 1995.

30. MacAndrew C, Edgerton R. Drunken Comportment: A Social Explanation.
Chicago, Ill: Aldine; 1969.

31. Alasutari P. Desire and Craving: A Cultural Theory of Alcoholism.
Albany, NY: State University of New York Press; 1992.

32. Morningstar P, Chitwood D. The Patterns of Cocaine Use. Rockville, Md:
National Institute on Drug Abuse; 1983.

33. Crusky WR, Berger LH, Richardson AH. The effects of marijuana
decriminalization on drug use patterns. Contempor Drug Problems.
1978;7:491-532.

34. Johnston LD, Bachman JG, O'Malley PM. Marijuana Decriminalization: The
Impact on Youth, 1975- 1980. Ann Arbor, Mich: University of Michigan,
Institute for Social Research; 1981. Monitoring the Future, Occasional
Paper 13.

35. Single EW. The impact of marijuana decriminalization. Res Adv Alcohol
Drug Problems. 1981;6: 405-424.

36. Maloff D. A review of the effects of the decriminalization of
marijuana. Contempor Drug Problems. 1981;10:307-322.

37. Becker HS. History, culture, and subjective experience: an exploration
of the social bases of drug-induced experiences. J Health Soc Behav.
1967;8:162-176.

38. Maloff D, Becker HS, Fonaroff A, Rodin J. Informal social controls and
their influence on substance use. In: Zinberg NE, Harding WM, eds. Control
Over Intoxicant Use: Pharmacological, Psychological, and Social
Considerations. New York, NY: Human Sciences Press; 1982:53-76.

39. Zinberg NH. Drug, Set, and Setting: The Basis for Controlled Intoxicant
Use. New Haven, Conn: Yale University Press; 1984