H. Saner, Robert MacCoun and P. Reuter


Graduate School of Public Policy University of California· Berkeley


H. Saner, Robert MacCoun and P. Reuter

October 1994

The judgments and conclusions of this paper are solely those of the author(s), and are not necessarily endorsed by the Graduate School of Public Policy, by the University of California, or by any other agency.

On the Ubiquity of Drug Selling Among Youthful Offenders, 1985-1991: Age, Period or Cohort Effect?

H. Saner l , R. MacCoun2, and P. Reuter3

Running Head: Drug Selling Among Youthful Offenders Key Words: drug selling, cohort analysis, youth

Acknowledgments : The authors wish to thank Jay Carver, head of the DC Pretrial Services Agency, for providing the data used in this paper, and the PSA staff for the assistance they have provided in using these data. We also thank D. McCaffrey for helpful statistical comments and advice.

Mailing Address of First Author: Dr . H. Saner The RAND Corporation P.O . Box 2138 Santa Monica, CA 90407-2138

IThe RAND Corporation, Santa Monica, California 2 Gra duate School of Public Policy, University of California at Berkeley 3Un iversity of Maryland


We present a multiple cohort analysis of rates of participation in drug offenses versus other crime in an urban sample, based on official charge data on young adults from the Pretrial Services Agency in the District of Columbia for the years 1985 to 1991.

We make lower-bound

estimates of how many individuals from particular population groups resident in the District are involved in drug-related criminal activities, examine trends in drug and non-drug charges in Washington, D.C., and disentangle the age, cohort and period effects in the variation in participation in drug offenses across multiple birth cohorts in the city.

We estimate that up to 30 percent of the young,

black male population of the District of Columbia were charged with drug distribution during this time.

Charge rates for drug distribution

activities appear to peak around age 24, decreasing slowly thereafter. Large and nonlinear period effects were observed for all drug-related charge rates, while increasing linear period effects were found for nondrug rnisdrneanors. observed.

Cohort effects in drug-related charge rates were also

Levels of participation in drug distribution charge rates

were lower for older cohorts, while the cohort share with a drug possession charge declined for younger cohorts.

However, when age and

period effects are included in the models, these cohort effects are muted or disappear, except in the case of non-drug misdemeanors.


1. INTRODUCTION As the drug problem has become more visible, the apparent explosion in young urban males' involvement in the drug distribution business has received increasing attention.

Drug dealing is perceived

as being enormously lucrative, offering unique prospects for financial mobility; the media regularly report earnings of hundreds of dollars a day for young, poorly educated sellers (e.g. Time, May 9, 1988, pp. 2133; Finnegan, 1990).

These financial attractions are seen as one of the

primary reasons for the intractability of the drug problem.

While the

anecdotal and qualitative evidence pointing towards increasing involvement in drug crime by younger groups is quite compelling (e.g. Finnegan, 1990; Williams, 1989), systematic examination of the extent and dynamics of participation over time in drug distribution activities in a community, or the share of particular population groups involved in this type of offense has been rare.

In an earlier study of the drug-

selling population of the District of Columbia, using data from 1985 to 1987, Reuter, MacCoun and Murphy (1990) found that a substantial fraction of the young black male population in the District was involved in drug distribution activities.

When participation was examined across

cohorts, these authors found that while older cohorts were also heavily involved with drugs, they appeared to play less of a role in the drug trade.

However, they could not determine whether this pattern was truly

an effect of age (that is, that it would hold for all future comparisons across birth cohorts) or whether it represented a change in behavior of incoming cohorts.


Although the consistency of the age-crime and period-crime relationships have been well-documented for different types of personal and property crime (Greenberg and Larkin, 1985; Hirschi and Gottfredson, 1983), the effects of age, disentangled from those of cohort and period, have not been examined systematically for drug distribution and possession offenses. different.

Yet the dynamics of these crimes may be very

Other researchers have found that ages of peak activity vary

by type of crime, particularly in the distributions of person and property offenses (Greenberg and Larkin, 1985; Hirschi and Gottfredson, 1983).

The peak years for property offenses (based on official data)

are the mid-to-late teens.

On the other hand, offenses against persons

in official data tend to peak in the late teens or early twenties. Involvement in the drug distribution trade has become increasingly prevalent among juveniles and young adults, particularly in major urban areas in the U.S.

(Deschenes and Greenwood, 1993).

Traditionally, the

primary initiates to drug selling have been addicted users, who turn to dealing as a way to finance their habit (Johnson et al., 1990).


recent studies (Dernbo, Williams, et al., 1991; Deschenses and Greenwood, 1993; Reuter, MacCoun. and Murphy, 1990) indicate that many adolescent crack sellers are not users, at least not initially. do become heavy users over time (Reuter et al., 1990).

Nevertheless, many Though the

temporal sequence may have reversed, it is likely that heavy users continue to rely on drug sales to support their crack habits. Everingham and Rydell (1994, p.30) estimate that about half of all cocaine initiates remain users after two years, and about a third remain users after a decade.

Studies of heroin addicts

(e.g., Hanlon, Nurco,

Kinlock, and Duszynski, 1990; Hser, Anglin, and Powers, 1993) suggest


that many remain involved with both heroin use and acquisitive crimes for much of their adult lives.

This leads us to expect that the

downturn in drug offense-specific rates will be slower than for other offenses.

There may also be another important and complicating

difference between drug offenses and other criminal offenses, in terms of the stability of the age relationship.

Initiation into drug use is

seen as an epidemic phenomenon, with a strong cohort element (see, for example, Golub and Johnson (1994), and Johnston, O'Malley and Bachman (1991)).

If drug selling is related to drug use, we would expect the

age-drug selling relationship to be less stable than that for other offenses.

In this article, we present the first mUltiple cohort analysis of rates of participation in drug offending and other crime in an urban sample.

We also use data across a seven year span to examine trends in

drug and non-drug charges, and to disentangle the age, period and coho=t effects in the variation in participation in offending across diffe=ent birth cohorts in a single city.

1.1 Methodological Background Age effects, or patterns of variation in crime rates with age, a=e associated with chronological aging via the dimensions of biological, psychological and social maturation.

As people age through their

twenties and beyond, their involvement in different types of criminal activity tends to decline.

Thus examination of issues regarding the

prevention or control of crime often concentrates particularly on the high rates of criminal involvement among adolescents and young adults, attempting to unscramble the observed relationships between offending


and such key constructs as the controls exerted by bonds to social institutions (Hirschi, 1969), peer networks and co-offending (Tonry, Ohlin and Farrington, 1991), and breakdowns in community structure (Reiss, 1988; Simcha-Fagan and Schwartz, 1986).

General societal changes which affect people of all ages and in all cohorts are considered to be period influences (Glenn, 1981; Lab, 1988; Ryder, 1965).

By contrast, cohort effects are characterized as

the consequences associated with membership in different birth cohorts. These effects are group-specific, and thus are an issue only during the lifetime of the cohort (Lab, 1988).

The potential for confounding

between age and cohort effects is succinctly stated by Greenberg:


from one year to the next each subsequent cohort of fifteen-year-olds has a higher crime rate that the one before it - a rate which also remains constant as the cohort grows older - then in any given year, older cohorts will be committing crimes at lower rates than younger cohorts.

This creates the false impression that involvement in crime

diminishes with age"

(Greenberg, 1983, p.31).1

1 Easterlin's hypothesis posits population composition, or relative cohort size, as an explanation for changes in crime rates via a series of economic and institutional overload mechanisms (Easterlin, 1987). More generally, resources are scarcer for larger cohorts than for smaller ones.

A number of researchers have examined this hypothesis,

but the results have not been conclusive (Easterlin and Shapiro, 1979; Maxim, 1985; O'Brien, 1989; Steffensmeier et al., 1987).


It is difficult to distinguish between changes due to age and changes due to period by following only one cohort over time.


to some degree, the contributing influences of these factors can be clarified by examining several different cohorts over time (Farrington, Ohlin, and Wilson, 1986).

Age, period and cohort effects and the

distinctions among them have been studied in a variety of criminological contexts.

In their highly influential study based on self-report and

interview data, Elliot, Huizinga and Menard (1989) examined age, period and cohort trends in delinquency, mental health problems and drug use among seven cohorts from the National Youth Survey.

They found that,

unlike other forms of criminal or delinquent behavior in their sample, substance abuse appears to peak at the end of adolescence and the beginning of adulthood.

They also discovered a consistent, increasing

period trend in substance use from 1976 to 1980, and a positive trend after 1980 in alcohol and polydrug use.

In a more recent study, Golub

and Johnson (1994) examine trends in illicit drug use among youthful arrestees in Manhattan from 1987 through 1993, finding strong cohort effects in the decline in detected cocaine use in that population.

l.2 Limitations and Caveats

This study is based on official charge data on adults from the Pretrial Services Agency in the District of Columbia for the years 1985 to 1991.

This is not a random sample of the dealing population, and we

can make no inferences about the total number or demographic composition of the dealer population in general in the District.

Not all drug

sellers get arrested in the course of one year's (or even seven years') dealing activity, nor are all sellers equally at risk of getting



Those who sell in highly visible and notorious markets are

more at risk than those who sell in more discreet settings.

Thus, our

data are skewed towards individuals over the age of 18 at the low end of the distribution system who operate in relatively exposed settings.


assume that the poor or near poor among dealers are more likely to be found in these settings seems reasonable.

Thus these data form a biased

sample of the adult drug-retailing population in the District; the total drug-selling population is likely to be less young, black and male than these data suggest.

However, we can use such data to make lower-bound

estimates of how many individuals from particular population groups resident in the District are involved in drug-related criminal activities .

The charged adult population in washington D. C. over these years was overwhelmingly male, black and young.

Given the demographics of the

city, this is hardly surprising, although relative to the general population in the District, males, youths and blacks are over represented in the arrested population.

The 1990 census indicated that

about two-thirds of the city's population was black, and that 20% of the black population had incomes below the poverty line, more than three times the proportion of whites living in poverty in the city (u . S. Bureau of the Census, 1991).

The high percentage of black District

males in the arrested population may reflect the much higher poverty rates of blacks in the District, as well as the high fraction of blacks among males in the high-risk years.

This may be particularly true for

those arrested on drug distribution charges.

It should also be

recognized that the findings presented here are specific to a time and place, and should certainly not be generalized to other locations or



This point is illustrated by a recent study of individuals

who reported selling an illicit drug within the last 12 months in the 1991 National Household Survey of Drug Abuse.

Caulkins and McCaffrey

(1993) showed that of the approximately 1.8 million members of the household population who reported having sold drugs in the last year, only 19% are black, even though 64% of the subset reporting arrest for a drug offense are black.

Caulkins and McCaffrey (1993, p13) argue that

"Inasmuch as black dealers (and users) are more likely to participate in visible, urban street markets and policy directs enforcement towards those markets, the result is that blacks suffer an undue fraction of drug arrests."

The stringency of enforcement in the city during the mid-1980s also suggests that, with respect to the rates for specific coho=ts, ou= findings cannot be generalized to other cities. District is exceptionally stringent.

By some measures the

The ratio of drug arrests to

population is higher in Washington than in other major cities - 14.8 arrests per thousand residents, whereas most of the comparison cities had rates of less than 8 per thousand (District of Columbia Office of Justice Plans and Analysis, 1987).

The city also has a very high rate

of felony drug convictions per capita - about 5 per thousand, as compared with 2 per thousand in Manhattan and 1 per thousand in Los Angeles (Reuter et al., 1990). consumption patterns.

In addition, Washington has unusual drug

First, PCP has been a much more important drug in

the District than in any other U. S. city .

One indication is that more

than one-third of all persons arrested in the District in 1988 tested positive for PCP, whereas in most other cities the figure was less than 5% (National Institute of Justice, 1989).

Second, several indicators


suggest that Washington's suburban drug problems in the late 1980s were much more severe than those of other suburban areas (Reuter, Haaga, Murphy and Praskac, 1988).

Third, the city ranks among the leaders with

respect to heroin problems (Reuter,et al., 1990, p11).

We present data on the trends in drug and non-drug charges over the period 1985 to 1991.

In the context formed by these trends, we then

focus on a demographic group which has been the subject of considerable media attention, and which is over-represented in our data - young, black males who are resident in the District.

We follow multiple

cohorts over the 7 years to examine a number of hypotheses about the variation in patterns of drug distribution in the District.

First, we

expect that the downturn in the age-crime relationship for distribution and possession offenses in this population will be slower than those found for other crimes.

Second, as suggested in the earlier analyses of

these data for the years 1985 to 1987 (Reuter, et al., 1990), the rate of participation in drug offending will vary across older and younger cohorts, with younger cohorts becoming more involved in drug distribution activities. During the years spanned by our data, the District experienced an unprecedented explosion in hard drug use, drug selling and drug-related violence.

For example, the percentage of arrestees testing positive for

cocaine rose from 15 percent in March 1984, when testing began, to over 60 percent in 1988 (Reuter, Ebener, and McCaffrey, 1994).

It is no

coincidence that this is the period in which crack cocaine emerged on the scene. enforcement.

There was a corresponding dramatic increase in drug law This is suggested by the astounding growth in recent years

in the number of juvenile arrests involving drug sales.

In 1981 only


343 arrests for drug offenses occurred; that number rose to a peak of 1913 by 1988 and declined to 935 by 1991.

By 1987, drug cases accounted

for more than half of all felony indictments by the local prosecutor. Thereafter, the intensity of drug enforcement declined from its peak, though it remained well above the level of the early 1980s.


conditions lead us to expect that participation in drug distribution activities in the District (as measured by official data) will be subject to increasing period effects through to the mid-1980s, declining again by the beginning of the 1990s.


2.1 The Pretrial Services Agency Data from Washington D.C. The Pretrial Services Agency (PSA) is the District of Columbia agency responsible for providing information to District judges on persons charged and awaiting decisions about pretrial detention.


PSA report includes the arrestee's criminal history, information about his/her current status in the community, his/her education, and the results of a drug test. record at the time.

Not every charge in the District produces a PSA

Arrests the MPD disposes of at the precinct level,

without any referral to the prosecutor's office, will generally not go through the PSA.

However, such charges are dominated by minor offenses.

The PSA data has two types of records: a case record which includes information about an individual case, and a person record which tracks an individual (with pointers back to the case record for some kinds of information).

The data are maintained for successive arrests,

and contain relatively rich information about the individual's

-10demographic characteristics and life circumstances, including previous contacts with the criminal justice system, schooling level, current residence and employment.

Because the PSA data includes individual

identifiers, it can be used to create birth cohorts of individuals resident in DC (for at least one year) between 1985 and 1991.

The crime

rates for these birth cohorts can be examined to estimate the effects of age and time period on rates of drug offending.



of the data

It is important to note some major limitations of the PSA data as a description of the offending population in Washington D.C.

First, the

PSA data do not include data on juvenile (younger than 18) arrestees, so no analysis of this group is possible.

Nor do the data include the

juvenile arrests of the adult arrestees, although it does indicate the year in which they were first arrested.

Second, not all offenders, and,

more specifically, drug distributors, get arrested in the course of one year's illegal activity, nor are all offenders equally at risk for getting arrested.

In general, standard arrest-based analyses assume

equal probabilities of detection of offenses in each year.

Also, in the

absence of specific information about the relationship between age and risk of arrest, it is generally assumed that risk of arrest is not correlated with age.

Of course, offenders may learn from experience how

to avoid arrest, although as they age they may become better known to the police and thus more likely to be arrested (Greenberg, 1983).

2.3 Defining District of Columbia Residency Many arrestees record multiple addresses; one may be the current residence, another the address to which mail should be sent, still


another the one through which the arrestee is most likely to be reached quickly.

The PSA attempts, with the arrestee's consent, to verify

current residential information.

A person received a District resident

code if (1) his/her latest recorded address was in the District, and (2) the record indicated that he/she had lived in the District for at least one year.

This is a conservative definition of residency from a

demographer's point of view (Shryock and Siegel, 1973).

We discuss

issues of cohort assembly and migration later in this section.

2.4 Defining Drug Charges A lexicographic ordering scheme was used as an index of degree of involvement in illegal drug markets.

We assigned each person charged to

the first possible category in the following ordered list:

1. Drug sale or distribution; 2. Drug possession; 3. Other drug charge; 4. Felony; 5. Misdemeanor.

We classified "Possession with Intent to Distribute" as a distribution charge.

We further collapsed the categories by including

any felonies in the "Other drug charge" category in the "Drug sale or distribution"; otherwise "Other drug charge" was classified as a "Drug possession" charge.

Thus we classified an individual as having a drug

distribution charge if he had such a charge, though he may have had any number of other charges.

We classified an individual as a drug

possession charge if he had at least one such charge, but no distribution charge, and so on.

Note that this hierarchy rule is


different from that used by the Metropolitan Police Department in its arrest reports;

the MPD would list an arrest that included a violent

felony charge and a drug distribution charge as an arrest for the violent offense.

2.5 Methods of Cohort Analysis

We present age-period-cohort analyses for each category of offense as described above - drug distribution, drug possession, non-drug felony, and non-drug misdemeanor.

The dependent variable in each

analysis is the age-period-cohort-specific arrest rates for single years of age between 18 and 29, for each year from 1985 to 1991.

The rates

displayed considerable heteroskedasticity, and so were transformed to a more variance-stable form using the natural logarithm function. Age is one of the most robust predictors of rates of many types of crime and delinquency. these analyses.

Thus we expect the age effect to predominate in

The period effect, which captures the impact of social

and historical events which cast their influence across all ages at a given time, includes changes in enforcement patterns in evidence in the District between 1985 and 1991.

Other demographic research (Hobcraft,

Menken, and Preston 1982) suggests that the period effect may take precedence over the cohort effect.

Thus we present our models in a

nested fashion, with age entered first

(A), period (A+P) or cohort


second, and finally a full model including all three effects (A+P+C). We estimate age, period and cohort effects on offense rates using a conventional linear model approach (Hobcraft, Menken, and Preston 1982), with notation similar to that in Steffensmeier, Streifel and Shihadeh (1992), so that


In(YAPe) = bo + Lai~ +

LpjP + Lc.C., j


where In(YAPC) is the logarithmic transformation of the age-periodcohort-specific offense rates; bO, ai, Pj, and ck are standard regression coefficients; and Ai, Pj, and Ck represent the ith, kth age, period or cohort respectively .

jth and

Ai, Pj, and Ck are included as

indicator variables via 0/1 coding. A key problem in the analysis of cohort data is the logically confounded relationship among age, period and cohort.

As a number of

researchers have emphasized (Glenn, 1976, 1981; Greenberg and Larkin, 1985), it is not possible to distinguish among these effects using statistical techniques alone (except under unusual nonlinearity conditions).

Rather, a priori knowledge or extra-statistical evidence

must be brought to bear in the interpretation of these data. researchers have challenged this view.


In a seminal paper, Mason et al.

(1973) set forth a model which attempts to separate the effects confounded in cohort data.

However, estimation of their model requires

that the relationship of at least one of the variables - age, period or cohort - to the dependent variable be constrained to be nonlinear. Other researchers have sidestepped this collinearity problem by imposing constraints on the effects to be estimated.

For example, Greenberg and

Larkin (1985) examine differences in the age-crime relationship for different types of offenses, using models based on an assumption of equal period effects.

Steffensmeier et al.

(1992) estimate age, period

and cohort effects on age-period-specific arrest rates using models in which the effects of the two oldest cohorts are set equal.


We also break the impasse of perfect collinearity by imposing a number of constraints on our models.

First, as is typical in analyses

using indicator variables, the oldest age (age 29) is the contrast category for Age, and the final period (year 1991) is the contrast category for Period.

We further impose a series of restrictions on the

Cohort variable by placing the six cohorts (defined fairly arbitrarily by year of birth) into three two-year categories - 1962 and 1963, 1964 and 1965, and 1966 and 1967.

This provides us with an estimable model,

while still allowing for comparisons between older and younger cohorts.

2.6 Computing Cohort Size and the Risk of Arrest One way of expressing the degree of involvement in illegal and/or drug distribution activities of a particular cohort is to compute the fraction of the population at risk that commits least one crime during some observation period.

(or is arrested for)


The risk calculations

presented here use artificially created cohort sizes, based on estimates by the District government, and the 1980 and 1990 census counts.

In a

series of sensitivity analyses, we also include some adjustments for undercounts in those two census years.

The U.S. Census Bureau has

estimated that black adult males are systematically undercounted in census figures.

It estimates that the 1980 census data undercount 18-

year-old black males by 1% and the level of undercount increases linearly for each year of age, so that 29-year-old black males are undercounted by 12% 1987).

(Passel, Siegel, and Robinson, 1982; Tillman,

The local District figures include the institutional popUlation

- such groups as soldiers living in barracks, and college students living in dormitories.

(The inmates of the prison for the District,

Lorton Prison located in Virginia, are not included in these estimates.)


These estimates also have the advantage over


Bureau of the Census

counts of incorporating knowledge about the local population specifically about local migration patterns - in order to make reasonable population estimates (Reuter, MacCoun, and Murphy, 1990, pll 7) .

The calculation of the size of the population at risk for each cohort is not based on an actual age cohort, since the composition of this at-risk population changes over time due to in- and outmigration in the District.

The Census Bureau notes that at the State level,

"[Ilnternal migration is the most important and complex component of population change.

Unfortunately, migration data are often the least

timely and least comprehensive of the population data sets" (Wet rogan, 1988).

In the District of Columbia, there has been a net out-migration

of approximately 2 percent, based on the total beginning population, from 1980 to 1990;

approximately 60 percent of this outmigration has

been to the adjacent states of Maryland and Virginia.

The black

population outmigration rate was considerably higher, at about 6 percent (Wetrogan, 1988). In a discussion of the biases introduced to cohort studies of crime by in- and outmigration, Gordon (1976) notes that the size and direction of bias depends on the size of the pool of migrants, and on how representative of the cohort they are.

Based on a comparison of

rates of arrest for delinquency in Philadelphia, Gordon (1976)


that black juvenile migrants residing in a "high delinquency" neighborhood acquired court records at rates similar to those found for black males in the city as a whole.

The close agreement in average

rates indicated that if the migrants in this one neighborhood were


typical of migrants in Philadelphia, there was no significant difference in delinquency rates between migrants and residents.



result lends some credence to the expectation of insignificant differences in offense rates between migrants and nonrnigrants; however, this may have changed in the last twenty years, as migration patterns have changed.

If drug dealing is a uniquely attractive offense and

large cities are better places in which to deal drugs, then the immigration to Washington over the past decade for young black males may include a disproportionate number of dealers.

While we believe that the

biases caused by migration effects would have to be unrealistically high to affect the calculations provided here, this potential error should be recognized. We calculate the risk of arrest for each cohort as the ratio of the number of people arrested from the population of interest to an estimate of the total population at risk.

For our purposes, this is

considered to be the mid-year population of the number of black males of a particular age in the particular year.

When cumulative arrests are

considered over the 7-year period for which our data is available, the denominator is taken to be the average of the 7 mid-year population estimates.

Intercensal estimates of this denominator are obtained


an exponential interpolation technique (Pebley, 1993: Palmore, 1983):

where Pt2 = the size of the cohort in 1990, Ptl = the size of the cohort in 1980, r=rate of population growth, and t=each year between the 2 census years.


Our methods of computing estimates of cohort size for particular population groups (i.e. young black male residents) in the District involve an assumption of even rates of migration across the narrow range (18-29) of focus in this study, as well as different ethnic groups, cohorts and years.

Nevertheless, given the paucity of current data

about these populations in the District, we believe that our standard demographer's estimation procedure, in conjunction with the sensitivity analyses taking the estimated undercount in account, allows us to arrive at reasonable and conservative (via the incorporation of the undercount adjustments) estimates of cohort size for use in the calculations of risk of arrest for each of the cohorts in this analysis.


3.1 Trends in Drug and Non-drug Charges As seen in Table I, over the seven years from 1985 to 1991, some 98,104 different people were charged with at least one criminal offense in the District, and after 1985 the year-by-year totals remained at fairly stable levels, with a median of about 21,000.

[Table I here) - '

A total of 45,392 different people, or 46.3 percent of all charged, were charged with some drug offense.


Furthermore, of those

persons charged with drug offenses, two thirds were charged with the more serious distribution, sale or manufacturing offense.

The relative

share of drug charges to other offenses peaked at about 49 percent in 1987, decreasing steadily in subsequent years to 35 percent in 1991. However, the types of drug charges contributing to that share changed


quite dramatically over time: in 1985 the ratio of people charged with drug-selling to those charged with drug possession was 1.2 - that is, almost equal numbers of people were charged with each type of crime.


1991 this ratio had almost tripled to 3.1 - almost three times as many persons charged with drug-selling as those charged with possession. This may reflect changes in both enforcement policy and participation in the drug trade.

While more people may have been engaged in drug selling

activities, the police may also have been pushing to arrest and charge offenders found in possession of illegal substances with the more serious charge of uPossession with intent to distribute H

Once again,

1987 represented a peak in numbers of people charged with drug distribution offenses.

However, counts of persons charged with other

felonies or misdemeanors continued to rise over this period, reaching a high in 1991.

These figures provide a context for our subsequent

findings, by providing an estimate of the number of persons who were involved, at least occasionally, in drug distribution in the District over the seven-year period covered by our data. Since city residents are primarily subject to city enforcement, whereas residents from elsewhere are primarily in the hands of other jurisdictions, we separated persons resident in the city from the total population of persons in the PSA data.

Furthermore, the District has

come to dominate drug arrests in the metropolitan area.

For example, in

1986 the District arrested 21.1 individuals per 1,000 residents compared with 2.8 for the Maryland suburbs, and 2.1 for Northern Virginia (Reuter, Haaga, Murphy and Praskac, 1988).

In the present analysis,

residents accounted for 75 percent of those charged with drug selling, compared with 54 percent of those charged with drug possession and 60


percent of those charged with a nondrug felony.

For all groups, the

population was heavily male, black, and young.

Those people charged

with drug distribution were substantially more likely to have each of those characteristics; the non-drug misdemeanor group, substantially less likely.

The characteristics of all D.C . residents charged with

drug and other offenses are presented in Table II.

[Table I I here)

Given that young black males account for the


share of

all persons charged with drug distribution, we turn now to focus our subsequent discussion on a series of analyses examining the effects of age, cohort and period on participation rates for this group - black males aged 18 to 29, who were resident in the District between 1985 and 1991.

We follow 6 birth cohorts - the oldest born in 1962 and the

youngest in 1967.

These cohorts are further collapsed into 3 two-year

cohorts in some analyses and figures, both for model and clarity of presentation.



The grid in Table I I I allows for easy

identification of age and year of arrest for each


[Table I I I here)

3.2 Age, Period and Cohort Effects A~e

Effects We begin by considering the distribution of drug and other

offenses across age groups in this population .

The percentage of each

cohort of young black males charged for drug distribution offenses, drug

-20possession, non-drug felonies and non-drug misdemeanors, averaged over the seven years of our data, are presented by age in Figure 1.2

[Figure 1 here]

Here we see some surprising patterns.

The charge rates for drug

distribution and possession offenses are highest for people in their early twenties, and decline slowly thereafter.

The charge rates for

non-drug felonies increase slightly to a peak at age 24 and decrease as age increases through the mid to late twenties.

However, the non-drug

misdemeanor charge rate actually increases between ages 19 and 28.


trend is contrary to a large body of literature showing an almost immutable decreasing trend in crime rates with age (at least after the teens and early twenties - see for example, Hirschi and Gottfredson, 1983).

In fact, when the non-drug misdemeanors and the less serious

(usually misdemeanor) drug possession charges are combined, the curve does exhibit the usually observed slow decrease with age.

We are not

sure why the disaggregated non-drug misdemeanor age curve behaves in this uncharacteristic fashion, but hypothesize that other unmeasured changes in law enforcement practice may be contributing to these effects.

It is possible, for example, that a number of young offenders

who would have been arrested for misdemeanors during other periods were charged with drug offenses instead (Klein et al., 1991).

2 Although our data include ages 18 and 29, they are not shown in Figure 1 because the estimates of charge rates for these ages were very imprecise.


Period Effects The percentage of 2-year cohorts charged for drug and non-drug offenses are shown in Figures 2a and 2b.

As we anticipated, large and

nonlinear period effects are apparent in the drug distribution curves, and similar but muted effects occur in the possession curves for the three 2-year cohorts.

The percentage arrested on drug distribution

charges for each cohort increased dramatically after 1985, peaked in 1987, and had plateaued at a lower level by the beginning of the 1990s. (Cohort effects are also evident in these curves.

We return to a

discussion of these and other cohort effects in the following section.) These period effects are to be expected, given the intensification of law enforcement against drug offenders in the District during those years.

More important than the rising total of arrests for drug-related

crimes is the dramatic increase in the number of arrests for the more serious offense, sale and manufacture of drugs, as opposed to drug possession. The same nonlinear period effects are not observed in other categories of non-drug felony crimes or misdemeanors.

Here, the

percentage of charges per cohort rise steadily in a linear trend between 1985 and 1991. [Figures 2a and 2b here)

Cohort Effects We explore cohort effects for young criminal offenders in two ways, using different definitions of "population at risko.

First, we

focus on participation in drug distribution activities among all those charged with a criminal offense .

That is, for each successive birth

cohort from 1962 to 1967, we ask what percentage of residents charged


with a criminal offense had at least one drug distribution charge between 1985 and 1991.

Second, we calculate the risk of arrest for drug

and other offenses for each entire cohort, using general population estimates in the denominator. Figure 3 shows that, for people charged with at least one offense during 1985 to 1991, the rate of drug distribution offenses rises across successive birth cohorts.

That is, for persons born in 1966 or 1967 who

had a criminal charge, about 53 percent had at least one drug distribution charge during the years 1985 to 1991; for persons born in 1962 or 1963, the figure was 44 percent.

By contrast, note that the

cohort share with a drug possession charge declined for later cohorts 11 percent of the 1966/1967 charged cohort had such a charge, compared with about 16 percent of the 1962/1963 charged cohort.

Thus it appears

that older charged cohorts are also involved with drugs, although to a lesser degree than the younger cohorts, but play somewhat less of a role in drug distribution.

[Figure 3 here] We express the significance of drug distribution to a particular birth cohort by calculating the risk that an individual in that group will be arrested early in his adult life for selling or manufacturing drugs.

In Table IV we present the basic data for the risk calculation.

The second through fifth columns show figures on the number of persons in each cohort charged with different kinds of offenses (using our lexicographic ordering rule) between 1985 and 1991; the sixth column shows the number charged with at least one criminal offense.

The final

two columns contain average estimates of the cohort's total size - that is, the number of black males resident in the District of Columbia who


were born in a particular year.

As a way of assessing the sensitivity

of our results to the procedures used to estimate these "population-atrisk" denominators, we have included undercount adjustments for each cohort, based on the Census Bureau's estimates.

In Table V we convert

these numbers to risks of arrest.

[Tables IV and V here]

The risk of arrest on any type of charge for every cohort is extremely high, ranging from 55 to almost 63 percent.

The slightly

lower rates for the youngest cohorts appear to be due to a small decline in the charge rates for non-drug offenses in those groups.

For all 6

cohorts, the risk of arrest on drug distribution charges was substantial, ranging between 1 in 4 and 1 in 3.

Of black males



1962, 26.4 percent were charged with a drug selling offense at least once in the period 1985-1991. The risk for successive cohorts rises to a peak of 31.2 percent for the 1965 cohort, and plateaus at 30 percent for the youngest cohorts, a level significantly higher (Z=283.3, p
Although the 1968

cohort was not eligible for adult arrest in 1985, its risk of being arrested and charged for drug selling remains high, at 28 percent for years 1986 to 1991.

Incorporating the appropriate undercount

adjustments into the denominators does not change this pattern (although of course it does lower the absolute levels of risk for each cohort) . The increased share of drug distribution arrests among younger cohorts is also noteworthy given their declining shares of arrests on possession and non-drug charges.


These figures are cumulative over the 7-year period covered by our data.

Looking back to Figures 2a and 2b which show changes over time,

we see that, as expected, levels of participation in drug distribution crimes is lower for older cohorts - that is, the oldest cohort (1962/1963) has lower levels of participation than the younger cohorts, and this trend holds almost uniformly over the years.

However, for all

cohorts, risk of arrest on any drug-related offense peaked in 1987 and decreased to a plateau in the early 1990's.

A Complete Model

The above discussion of complicated effects highlights the need for an analysis which accounts for age, period and cohort effects simultaneously.

In Table VI we present the proportion of variation in

arrest rates accounted for several indicator variable models for each of the four offense categories. 3

The full age-period-cohort models produce

R2 values ranging from 0.748 for non-drug felonies to 0.967 for possession offenses, indicating that the fully specified models fit the data very well indeed. 4

3 A separate ANOVA of charges aggregated across categories revealed highly significant interactions of the charge categories with age and period, but not cohort.

For ease of presentation, we present analyses

separately by category in Table IV.

4 We also examined whether relative cohort size, as opposed to cohort membership, explained any variation in arrest rates (see Footnote 2 above for a brief discussion of Easterlin's hypothesis), but these


[Table VI here]

Age alone accounts for between 52% (Drug possession) and 68%


drug Misdemeanor) of the variation observed in these age-period-cohortspecific arrest rates.

The offense rates for the drug-related charges

tend to peak in the early twenties, after which they slowly decline.


contrast, the non-drug misdemeanor and the non-drug felony rates are highest among the oldest arrestees, and significantly lower for those in their late teens and early twenties. When Period or Cohort are added to the Age only model

(see columns

3 and 4 in Table VI), significantly more variation in the offense rates is explained in all but two instances - the Age/Cohort models for distribution and non-drug misdemeanor arrest rates.

However, when all

three effects are included in the fully specified models, significant cohort effects occur only in the case of the non-drug misdemeanor category, where the 1966/1967 cohort's rates of arrest are lower on average than those for earlier cohorts. Nonlinear period effects in both drug arrest categories indicate rates of arrest at their peak during 1986 and 1987, declining the rest of the decade.


Period effects are also discernible in the non-

drug misdemeanor category, but unlike the drug offenses, the rates of arrest were highest in 1990 and 1991, and lowest in 1985.

In the full

model for non-drug felonies, no significant period effects were found. We hypothesize that this may be the carryover from enforcement practices during these years.

Two possibilities suggest themselves. First, non-

effects were not significant when age and period were included in the models.


drug offenses may be on the rise in the later years because nonspecialist offenders who were being charged for drug-related crimes are now being caught for other types of offenses.

Second, as the emphasis

on drug charging decreased in the later part of the 1980s, more resources may have been available for police departments to pursue other types of crimes. Figure 4 provides a graphical depiction of our key results: A consistent nonlinear age effect for drug distribution offenses, moderated by a nonlinear period effect, particularly among the younger ages. [Figure 4 here]

4. CONCLUSIONS We have used official charge data spanning seven years to examine the criminal activities of six cohorts of young, adult, African-American males resident in

Washington, D.C.

Particular attention was given to

participation in the drug trade and to separating out age, period and cohort effects.

We believe that three findings deserve attention.

First, confirming an earlier study with just 3 years of data (Reuter, MacCoun and Murphy, 1990), we found extremely high rates of participation in the drug trade for these cohorts.

For the 1967 cohort,

our lower bound estimate was 29 percent, and it was no lower than 25 percent for any of the six cohorts.

Given that not all participants

were arrested and charged, the true figure may have been substantially higher.

Clearly, involvement in the drug trade was not an unusual

experience for 1960s birth cohorts of black male Washington residents.


For all cohorts a majority of those with some criminal charge were charged with drug offenses.

Second, we found substantial nonlinear period effects in the drug distribution curves. These findings confirm our expectations since, as we have noted earlier, the period from 1985 to 1987 was the peak of an extended crackdown on drug distribution in the District. Cohort effects were also evident in the drug distribution charge rates, in that the offense rate rose steadily across successive birth cohorts. However, when age, period and cohort variables were included in the complete models, most observed cohort effects were swamped by the effects due to age and period, and remained significant only in the case of non-drug misdemeanors, where the youngest two-year cohort's rates of arrest


lower on average than those for earlier cohorts.

Third, we found age-crime profiles for non-drug felonies and misdemeanors that were different from that found in numerous earlier studies of property and personal crimes.

Instead of

peaking around 20 and then declining sharply, the Washington data for non-drug felonies showed prevalence rates increasing slightly from age 19 through 28; for non-drug misdemeanors there is clear upward trend from 19 to 29.

We conjecture that this pattern may

reflect the consequences of the epidemic of drug initiation in the 1980s and offer the following account.

Frequent drug users show slow desistance rates.

That is well

established for heroin (Hser, Anglin and Powers, 1993) and appears to be true also for cocaine.

Everingham and Rydell (1994) estimate that 30

percent of cocaine initiates would still be using 10 years later; the


figure might well be higher for those initiates from high drug-use areas.

The Drug Use Forecasting data show that those arrested for non-

drug offenses have very high drug use rates; in the District of Columbia 59% tested positive for some drug in 1991. decline in the percentage by age.

Moreover, there is little

The highest age group for cocaine in

1991 for Washington was 31-35; 72 percent tested positive for cocaine and for the group over 35 it was still 59%.

Continued use of expensive drugs is likely to extend criminal careers.

That would help explain our findings.

Given that there

is abundant evidence that drug initiation rates have been falling for since the mid-1980s, this effect may be specific to a small set of birth cohorts.

Nonetheless, it suggests that projections

of offense rates based on sharp declines of crime rates by the mid-20s will underestimate property and violent crime for the next decade.

Enforcement efforts against street drug markets are largely premised on two control strategies, deterrence and incapacitation.


analyses do not directly test the efficacy of these strategies in the District during the period under study.

Nevertheless, our findings

certainly suggest the possibility that deterrent or incapacitative effects were quite modest at best.

First, our data indicate dramatic

levels of initiation to drug selling at a time when enforcement efforts were quite high, suggesting that sanction risks had relatively weak deterrent effects (cf. MacCoun, 1993).

Second, our findings appear

compatible with the replacement hypothesis (e.g., Blumstein, 1993), which holds that incapacitation is of limited effectiveness for economic crimes


such as drug selling, because economic incentives will entice others to quickly replace incapacitated offenders.

Such an effect is doubly

pernicious: it expands the number of drug offenders in a community--and perhaps the proportion of a cohort with criminal records--and it may well increase the number of territorial disputes, a major source of marketrelated violence.




Blumstein, A. (1993). Making Rationality Relevant--The American Society of Criminology 1992 Presidential Address. Criminology, Vol. 31, 1-16. Caulkins, J. and McCaffrey, D. (1993) Drug dealers in the Household population. (Unpublished manuscript) . Collins, J.J. (1981) Alcohol careers and criminal careers. In J.J. Collins (Ed.) Drinking and Crime, New York: Guildford Press. Dembo, R., L. Williams, A. Getreu, L. Genung, J. Schrneidler, E. Berry, E. Wish, and L. LaVoie. (1991). A longitudinal study of the relationships among marijuana/hashish use, cocaine use and delinquency in a cohort of high risk youths. Journal of Drug Issues, Vol. 21, 2, 271-312. Deschenes, E.P., and P.W. Greenwood, Offender.


Tr€ating the Juvenile Drug

District of Columbia Office of Planning, Demography Division (1993). Personal Communication. Elliot, D.S., Huizinga, D. and Menard S. New York: Springer-Verlag.

(1989) Multiple Problem Youth.

Easterlin, R. (1987) Birth and Fortune: The Impact of Numbers on Personal Welfare. 2n ed. New York: Basic Books. Easterlin, R. and Shapiro, M. o. (1979). Homicide and fertility rates in the United States: A co~~ent. Social Biology, 26, 341-343. Everingham, S. S. and Rydell, C. P. (1994) Modeling the demand for cocaine. The RAND Corporation, MR-332-0NDCP/A/DPRC Farrington, D.P. (1983) Offending from 10 to 25 years of age. In K.T. Van Dusen and S.A. Mednick (Eds). Prospective studies in crime and delinquency. Boston: Kluwer-Nijhoff. Farrington, D.P., Ohlin, L.E. and Wilson, J.Q. controlling crime. New York: Springer-Verlag. Finnegan, W.

(1986) Understanding and

(1990) New Yorker, Sept 9, 51-84.

Glenn, N. D. (1976) Cohort analysts' futile quest: statistical attempts to separate age, period and cohort effects. American Sociological Review, 41, 5, 900-904. Glenn, N.D. (1981) The utility and logic of cohort analysis. Journal of Applied Behavioral Science, 17, 247-257. Golub, A. and Johnson, B. D. (1994) A recent decline in cocaine use among youthful arrestees in Manhattan, 1987 through 1993. American Journal of Public Health, 84, No.8, 1250-1254.


Gordon, R. (1976) Prevalence: The rare datum im delinquency measurement and its implications for the theory of delinquency. In M. Klein (ed), The Juvenile Justice System. Beverly Hills: Sage. Greenberg, D.A. (1983) Age and Crime. In Encyclopedia of Crime and Justice, S.H. Kadish (Ed), The Free Press, New York. Greenberg, D.A. and Larkin, N. J. (1985) Age-cohort analysis of arrest rates. Journal of Quantitative Criminology, 1, 227-240. Hanlon, T.E., Nurco, D.N., Kinlock, T.W., and Duszynski, K.R. (1990). Trends in criminal activity and drug use over an addiction career. American Journal of Drug and Alcohol Abuse, 16, 223-238. Hirschi, T. (1969) Causes of delinquency. University of California Press, Berkeley, CA. Hirschi, T. and Gottfredson, M. (1983) Age and the explanation of crime. American Journal of Sociology, 89, 552-584. Hobcraft, J., Menken, J., and Preston, S. (1982). Age, period and cohort effects in demography: A review. Population Index, 48, 4-43. Hser, Y., Anglin, M.D., and Powers, K. (1993). A 24-year follow-up of California narcotics addicts. Archives of General Psychiatry, 50, 577584. Johnson, abuse in M. Tonry Justice:

B.D., Williams, T., Dei, K.A., and Sanabria, H. (1990). Drug the inner city: IMpact on hard drug users and the community. In and J.Q. Wilson (eds.), Drugs and Crime (Vol. 13 of Crime and A Review of Research). Chicago: University of Chicago.

Johnston, L.D., O'Malley, P.M., and Bachman, J.G. (1991). Drug use among American high school seniors, college students and young adults, 19751990. (Volume 1: High School Seniors), Washington D.C.: National Institute of Drug Abuse. Klein, S.P., Ebener, P. Abrahamse, A., and Fitzgerald, N. (1991) Predicting criminal justice outcomes: What matters? The RAND Corporation, R-3972. Lab, S.P. (1988) Analyzing change in crime and delinquency rates: The case for cohort analysis. Criminal Justice Research Bulletin, 3, 1-7.

MacCoun, R. J. (1993). "Drugs and the Law: A Psychological Analysis of Drug Prohibition," Psychological Bulletin, Vol. 113, 497-512.

Mason, K.O., Mason, W. M., winsborough, H.H., and Poole, W.K. (1973) Some methodological issues in cohort analysis of archival data. American Sociological Review, 38, 242-258. Maxim, P. (1985). Cohort size and juvenile delinquency: A test of the easterlin hypothesis. Social Forces, 63:661-679~


National Institute of Justice (1989) Drug Urinalysis Forecasting. Research in Action, Washington D.C. O'Brien, R.M. (1989). Relative cohort size anbd age-specific crime rates: An age-period-relative-cohort size model. Criminology, 27:57-78. Palmore, J. (1983) Measuring mortality, fertility and natural increase: A self-teaching guide to elementary measures. Honolulu: East-West Population Institute. Passel, J., Siegel, J. and Robinson, G. (1982) Coverage of the National Population in 1980 Census, by Age, Sex and Race. Bureau of the Census. Washington D.C.: U.S. Government Printing Office. Pebley, A.

(1993) Personal Communication.

Reiss, A.J. (1988). Co-oofending and Criminal Careers. In "Crime and Justice: An Annual Review of Research", vol 10, M. Tonry and N. Morris (Eds), University of Chicago Press, Chicago. Reuter, P., Ebener, P. and McCaffrey, D. (1994). Drug use patterns. In Besharov, D. (ed.) When drug addicts have children, Washington, American Enterprise Institute. Reuter, P., Haaga, J., Murphy, P., and Praskac, A. (1988). Drug Use and Drug Programs in the Washginton Metropolitan Area. The RAND Corporation, R-3655-GWRC. Reuter, P., MacCoun, R. and Murphy, P. Corporation, R-3894-RF.

(1990) Money from Crime, The RAND

Ryder, N.B. (1965) The cohort in the study of social change. American Sociological Review, 30, 843-861. Shannon, L.W. (1981) Assessing the relationship of adult criminal careers to juvenile careers. Iowa City, IA: Iowa Urban Community Research Center. Shryock, H.S. and Siegel, J. S. Demography.

(1973) The methods and materials of

Simcha-Fagan, o. and Schwartz, J. (1986). Neighborhood and delinquency: An assessment of contextual effects. Criminology, 24, 667-703. Steffensmeier, D., Streifel, C., and Harer, M.D. (1987). Relative size and youth crime in the United States, 1953-1984. American Sociological Review, 52:702-710.


Steffensmeier, D., Streifel, C., and Shihadeh, E.S. (1992). Cohort size and arrest rates over the life course: The Easterlin hypothesis reconsidered. American Sociological Review, 57, 306-314. Tillman, R. (19B7) The size of the "criminal population": The prevalence and incidence of adult arrest. Criminology, 25, 3, 561-580. Tonry, M., Ohlin, L.E., and Farrington, D.P. (1991) Human Development and Criminal Behavior. Springer-Verlag: New York.



Wetrogan, S. I. (1988). Projections of the populations of states, by age, sex and race: 1988 to 2010. Current Population Reports, P-25, No. 1017. Williams, T. (1989) The cocaine kids. Reading, Mass: Addison-Wesley.



Table I Persons Charged With Criminal Offenses. 1985-1991

1985 1986 1987 1988 1989 1990 1991

Drug Dist'n 3807 5842 6922 6443 6081 5580 6093

Drug Poss'n 3186 3896 4208 3261 2313 1858 1964

Other Felony 5615 5866 6468 6747 7383 7672 7888

Other Misdem 4470 4936 4956 4877 4973 5980 6684

17078 20540 22554 21328 20750 21090 22629

1985-1991 5








5 The totals count different individuals over the seven years of data, excluding multiple charges.


Table II Characteristics of all P C. residents charged with drug and other offenses. 1985-1991 (Percent) Distribution Male Age 18-29 Black Total number:


Non-drug felony

Non-drug rnisdem

85.1 59.9 98.3

80.2 51. 5 94.5

81. 7 53.4 92.6

71. 3 53 . 7 84.9





Grand Total Number: 60496


Table III Aqe by Year of Birth and Year of Arrest Birth/Arrest 1962 1963 1964 1965 1966 1967

1985 23 22 21 20 19 18

1986 24 23 22 21 20 19

1987 25 24 23 22 21 20

1988 26 25 24 23 22 21

1989 27 26 25 24 23 22

1990 28 27 26 25 24 23

1991 29 28 27 26 25 24


Table IV bla~k male Qi~tti~t te~id.ent~ in 19S5-l99l. b:i :ieat Qf bitth and. t:i~e Qf ~hatg:e

Number of ~hatg:ed.





Year Born

Drug Distribution

Drug Possession

Nondrug Felon:i

Nondrug Misd.

All Charges

Estim. Total Population

19621967 1967 1966 1965 1964 1963 1962

4990 967 1000 1034 1030 960 899

1461 187 224 232 245 288 285

3001 417 478 491 546 545 524

1537 201 237 270 252 309 268

11889 1772 1939 2027 2073 2102 1976

3211 3305 3316 3337 3353 3404

Total Pop. with Undercount 3339 3470 3515 3571 3621 3710


Table y Percentage of Black Male Residents Charged in 1985-1991, by Year of Birth and Type of Charge










Any charge (Under-count)

58.0 (53.3)

62.7 (58.1)

62.1 (58.1)

61.1 (57.7)

58.7 (55.6)

55.2 (53.1)

Any Drug Charge (Under-count)

34.8 (31.9)

37.2 (34.5)

38.2 (35 . 5)

38.1 (36.0)

37.0 (35.2)

35.9 (34.6)

























23.2 (21.3)

25.5 (23.4)

23.9 (22.3)

23.0 (21.7)

21.7 (20 . 6)

19.3 (18.5)

Type of charge

Drug Distribution Charge (Under-count) Drug Possession Charge (Under-count) Non-drug Charge (Under-count)


Table VI Proportion of variance in Arrest Rates as Explained by Different Models (p-yalue for full model in parentheses below)

Offense Distribution


Non-drug Felony

Non-drug Misdemeanor






































Figure 1 Cbarne R ___ of =~~¥~~~a~t~~e~S;:b~~~~~~~~~~~~~LrV 1-y Of s e Category cobort cb n _ year b i_r tAge h fand e __ a yeraned ... __ over 1985-1991) arged. IP ercentaae












CD 01 ~



,, ,

... ' ••••• 4t'


........... .



•.... .. '

... _.#1"'.".-






.. '.


- 0 - - Possession

.s:. 4.0 o


••••. Non·Drug Felony


o ~


.A. •






. .•• "/!r ••••••• i!£ •• ... • .A· •


- - 6- •. Non-Drug Misdemeanor



0.0 L-~--~--~~--~--~~--~~ I 19














Figure 2a Percentage of Cohort Charged. by Year. for Drug Offenses











,.. ,.. , ........ ' ... -... ~ " .... --"


.... .



I ,"




0 "t=

•... "" ""



GI 01 «I

~ ...... .


' ,,

•......... "







.... ......... .



............ ..................



- ...J.

.. GI U GI


Distribution: 1962-63 cohort

- .. - Distribution: 1964-65 cohort - ...... Distribution: 1966-67 cohort Possession: 1962-63 cohort


Possession: 1964-65 cohort •• Possession: 1966-67 cohort

- -6. -





3.0 _Ir'

.... ,.. -

........ _0--_.... 2.0

,.."" ...



, ...... ~'"

•••••• ~. • •

-.............&.., .............

-<> •• -- ............ ...... ~



--_. ......

t- :.....


.. ..........


0.0 1985











Figure 2b Percentage of Cohort Charged. by Year. for Non-Drug Offenses






10. 0~--------------------------------------------------------~

9.0 8.0











o .c o





1: CD



...- - - - - - .... - - - ...-~.:.:.. .'

- - Felony: 1962·63 cohort - .. - Felony: 1964·65 cohort • ..... Felony: 1966·67 cohort

a-- Misdemeanor: 1962·63 cohort

............... .' -~ ................. .

.. ....

Misdemeanor: 1964·65 cohort •. Misdemeanor: 1966·67 cohort

- -6 -






3.0 2.0

--~- ---- .....----

~~ ~~~ ~~- -


•••••••• _ .. _ ••••••••• -0 . . - 0 ••• -.

- ... _-.. '

. ' ;:,_



0.0 + 1 - - - - - + - - - - - - + - - - - - 1 - - - - - - + - - - - - - + - - - - - - 1 1988 1989 1990 1991 1987 1985 1986 Year




Figure 3 Distribution of Charges by Cohort for Black Male Residents aged 18-29.










40 0


-.c 0







• Distribution

(. ",::::




• Possession

mNon-Drug Felony IJ Non-Drug Misdemeanor


o 1962-63



Birth Cohort




Figure 4 Number Charged with Drug Distribution by Age and Year. 1985-1991

( I..

-----"0 CD




























Fig4 9/26/94

(Revised 1017/94)

WORKING PAPER SERIES Graduate School of Public Policy University of California Papers 1-184 are no longer active in series


Scotchmer, Suzanne and Eddie Dekel, "On The Evolution of Optimizing Behavior." March 1991.


Ellwood, John W. and James P. Walsh, "Mergers, Acquisitions, and The Pruning of Managerial Deadwood." May 1991.


Schrag, Joel and Suzanne Scotchmer, "Character Evidence in Hierarchical Justice." May 1991.


Trow, Martin, "American Higher Education: 'Exceptional' or Just Different?" May 1991.


Quigley, John M. and Daniel L. Rubinfeld, "Public Choices in Public Higher Education." August 1991.


Friedman, Lee S. and Christopher Weare, "The Two-Part Tariff, Practically Speaking." January 1992.


Kirp, David L., "Fetal Hazards, Gender Justice and The Justices: The Limits of Equality." January 1992.


Banks, Dwayne A., "Voluntary and Proprietary Hospital Behavioral Response to Socioeconomic Stimuli." April 1992.


Smolensky, Eugene and Robert Plotnick, "Inequality and Poverty in the United States: 1900 to 1990. n July 1992.


Kirp, David L., "A Suburb in Search of Its Soul. " July 1992


Kirp, David L., "Among Schoolchildren: The Equities in The Schoolhouse. n July 1992.


O'Regan, Katherine M. and John M. Quigley, "Family Networks and Youth Access to Jobs." August 1992.


Mauldon, Jane, "Now I'll Tell You, Now I Won't: An Investigation of The Internal Consistency of Survey Responses." September 1992.


Smolensky, E., R. Plotnick, E. Evenhouse and S. Reilly, "Give Up on Trickle Down?" January 1993.


Kirp, David L. and Ronald Bayer, "The Politics of Needle Exchange for AIDS Prevention." March 1993.


Quigley, John M., "Testimony on Housing and Urban Policy." June 1993.


Kirp, David L. and Cyrus Driver, "The Aspiration of Educational Mandates and The Realities of Localism" or "How rules Generated in Washington and Sacramento are Reconfigured in a California Suburb." August 1993.



Kirp, David L, "Doomed By Design: The Demise of 'The Biggest Little City in The Nation'." September 1993.


Jarvinen, Denis, "On-Site and Off-Site Waste Management in California: The Analysis of Markets and Environment Liability." November 1993.


Smolensky, Eugene, Eirik Evenhouse, and Siobhan Reily, "A Social Safety Net for the Negev." December 1993.


Kirp, David L., "After The Band Stopped Playing; AIDS in Our Time and, Maybe, AIDS Forever." March 1994.


Mauldon, Jane and Kristin Luker, "Contraception at First Sex: The Effects of Sex Education." May 1994.


Quigley, John M. and Katherine M. O'Regan, "Teenage Employment and the Spatial Isolation of Minority and Poverty Households." June 1994.


Kirp, David L., "Rethinking The Legacy of Tuskegee in The Era of AIDS." July 1994.


Mauldon, Jane and Susan Miller, "Child-Bearing Desires and Sterilization Among United States Women: Patterns by Income and AFDC Recipiency." August 1994.


MacCoun, Robert, Peter Reuter and Thomas Schelling, "Assessing Alternative Drug Control Regimes." October 1994.


MacCoun, Robert, "Differential Treatment of Corporate Defendants by Juries: Are 'Deep Pockets' The Cause?" October 1994.


MacCoun, Robert and Jonathan Caulkins, "Examining The Behavioral Assumptions of The National Drug Control Strategy." October 1994.


MacCoun, Robert, H. Saner, and P. Reuter, "On The Ubiquity of Drug Selling Among Youthful Offenders, 1985-1991: Age, Period or Cohort Effect?" October 1994.

To request copies or list ofpapers, write or call: Working Paper Series Coordinator Graduate School of Public Policy University of California 2607 Hearst Avenue, Berkeley, CA 94720 (510) 642-4670. The cost of each working paper is $5.00 (except to libraries). Make check or money order payable to:

UC Regents.

On the Ubiquity of Drug Selling Among Youthful Offenders, 1985 ...

MacCoun_Peter Reuter_1995.pdf. On the Ubiquity of Drug Selling Among Youthful Offe ... t Effect__H. Saner_R. MacCoun_Peter Reuter_1995.pdf. Open. Extract. Open with. Sign In. Main menu. Displaying On the Ubiquity of Drug Selling Among Youthful Offenders, 1985-1991. Age, Period or Cohort Effect__H. Saner_R.

17MB Sizes 0 Downloads 49 Views

Recommend Documents

Drug Offenders Receiving Mandatory Minimums in Each Drug Type
19,787. 10,982. 55.5. 4,094. 20.7. 4,711. 23.8. Powder Cocaine. 3,891. 1,570. 40.3. 935. 24.0. 1,386. 35.6. Crack Cocaine. 1,582. 860. 54.4. 441. 27.9. 281. 17.8. Heroin. 2,830. 1,646. 58.2. 587. 20.7. 597. 21.1. Marijuana. 3,534. 2,610. 73.9. 772. 2

[PDF] Selling Among Wolves: Without Joining the Pack ...
... Pack by Michael Q Pink The International Christian Business Institute Biblical ... the pack revolutionary deists answers pdf http vpnews co uk top notch PDF Book ... Online Free, Read Best Book Online Selling Among Wolves: Without Joining ...

inside-of-a-selling-gratified-expect-selling-on-steroid-1499499562359 ...
Home; OnEvery Side Us; This Steroid Was Specially. Formulated To Give Your BodyALeaner ... Page 2 of 2. inside-of-a-s ... 499562359.pdf. inside-of-a-se .

Sex Offenders 4-2015.pdf
Page 1 of 2. Oskaloosa Public. Library Policy. Sex Offenders Policy. Adopted: June 2009 Last Revision: April 2015. In accordance with Chapter 692A.113(1.f) of ...

On Differences among Elementary Theories of Finite ...
there is a non-trivial definable filter in D2. ... possible candidate of definable filters is the collection of non-cappable ... Ann. Pure Appl. Logic, 55(2):125–151,.

[PDF] Download Selling Sickness: How Drug Companies are Turning ...
... health decisions on the highest quality scientific data openly and objectively ... Drug Companies are Turning Us All Into Patients Top Ebook Download Best ...

eBook Download Selling Sickness: How Drug ...
... billion in private cash has Diamond Reynolds girlfriend of Philando Castile speaks to a crowd outside of J J Hill Montessori School during a memorial on July 7 2016 in St Paul Minnesota Get up to the minute entertainment news celebrity interviews

[PDF] Download Special Needs Offenders in the ...
... DONWLOAD NOW http online ebooksales top Download community based vocational with constant exposure to serious offenders and special education and ...

PDF[EPUB] Ubiquity: The Science of History . . . or Why ...
InformationWeek com News analysis and research for business technology .... Experts have never been able to explain the causes of any of these disasters.

Leslie Parrish's Youthful Genesis PDF Book Free Download.pdf ...
It was rediscovered by the scientists from the. Christian Health Science Institute. Using this method, you can reclaim your skin's. natural God given, radiant glow ...

Effects of Prohibition, Enforcement and Interdiction on Drug Use.pdf ...
Effects of Prohibition, Enforcement and Interdiction on Drug Use.pdf. Effects of Prohibition, Enforcement and Interdiction on Drug Use.pdf. Open. Extract.

There was a problem previewing this document. Retrying... Download. Connect more apps... Try one of the apps below to open or edit this item. 1985.pdf. 1985.

inside-of-a-selling-gratified-expect-selling-on-steroid ...
inside-of-a-selling-gratified-expect-selling-on-steroid-1499499562359.pdf. inside-of-a-selling-gratified-expect-selling-on-steroid-1499499562359.pdf. Open.

pdf-12108\oversight-hearing-on-office-of-national-drug-control ...
... apps below to open or edit this item. pdf-12108\oversight-hearing-on-office-of-national-drug ... licy-anti-drug-media-campaign-from-brand-bibliogov.pdf.

Standing on an Apple Box The Story of a Girl among ...
Standing on an Apple Box The Story of a Girl among the Stars PDF Download Full Online, epub free Standing on an Apple Box The Story of a Girl among the ...

Watch Starchaser The Legend of Orin (1985) Full Movie Online.pdf ...
Sign in. Loading… Whoops! There was a problem loading more pages. Retrying... Whoops! There was a problem previewing this document. Retrying.

[PDF] Download Standing on an Apple Box: The Story of a Girl among ...
... Woman Walking from master of suspense Sharon Bolton the sole survivor of a ... an Apple Box: The Story of a Girl among the Stars Best Book, Download Best ...

Studies on genetic variability among the derived lines of B x B, B x R ...
derived from B x B, B x R and R x R crosses in Rabi sorghum at RARS, Bijapur .... out for using in statistical analysis. The mean of five plants in each replication.

national facility on drug development - Anna University
Applications are invited for the various research positions (Purely ... Delhi sponsored National Facility on Drug Development for Academic ... Mobile No:.

Flying Blind on Drug Control Policy_National Academies_The ...
Page 2 of 4. MARK A. R. KLEIMAN. Flying Blind on Drug. Control Policy. Not knowing about the actual pat- terns of illicit drug abuse and drug. distribution cripples policymak- ing. As the subtitle of a National. Academies report put it four years. ag

Consideration on Sentencing for Drug Offenders_The Sentencing ...
Consideration on Sentencing for Drug Offenders_The S ... overnor's Commission on Certainty in Sentencing_.pdf. Consideration on Sentencing for Drug ...

PdF Drug Interaction Facts 2014: The Authority on Drug ...
Interactions Full Read ePub Online ... pharmaceutical manufacturer’s marketing strategy First a spillover arises when a How to Handle ... Enjoy proficient essay writing and custom writing services provided by professional academic writers ...

The breakfast club 1985
Convergencespeed force.Claire Dames AshliOrionLondonKeyes.58237291744. Dresden sub nl. ... Datastructures pdf.Angularjs pluralsight.Tributeto elliott. smith.3 ResearchTopic QuestionHowhas technology changed and affected shopping trendsand thesocial b

PDF/ePUB Affiliate Selling: Building Revenue on the ...
Download and Read Affiliate Selling Building Revenue On The Web ebooks epub easy books to ... The Web Reading this book with the PDF affiliate selling revenue on the web is yours Popular ..... to build a virtual storefront, and how to add.