A Review of Drug Abuse Data Bases


This chapter will identify and describe briefly data sources which may be used to project nonmedical drug use among young adults in future years. A wide variety of sources have potential utility in this regard. They range from individual studies conducted by local school districts or States to major national surveys involving thousands of respondents.

Because of the proliferation of research on drug use in recent years, it was necessary to place some limitations on the data to be presented here. First, it was decided that since a complete review was conducted in 1974 (), it was unnecessary to duplicate those efforts. () The studies reviewed here are more recent and, with the exception of the earliest National Surveys, were not covered in the 1974 review.

Second, this review is limited to those data bases which are national in scope. State and local surveys have severe limitations for purposes of making national projections of nonmedical drug use. For example:

  1. 1. The definitions used for nonmedical use often vary from one local or State survey to another.
  2. 2. Various local and State surveys are conducted in different time periods, so that it is often difficult to piece together a national profile.
  3. 3. The use of different methodologies in developing estimates limits the possibility of developing comparable results from one geographic region to another.

For these reasons, it is virtually impossible to combine data from several independent studies at disparate locations in the U.S. to form national estimates of the number of nonmedical drug users.

Finally, wherever possible, this literature review focused on data sources describing the youth and young adult populations, since these are the groups most at risk for drug abuse.

There are three broad categories of data describing the prevalence of nonmedical drug use:

  1. 1. Treatment-oriented data systems, which compile statistics based on individuals who have sought care;
  2. 2. Surveys designed to determine the extent of various types of drug use among specific populations;
  3. 3. Federally sponsored surveys of the general population which are national in scope.

In order to support comparisons of different data sources, a standard format was adopted for describing each data source, in terms of the following parameters:

  1. 1. The purpose, date, and sponsor of the data source;
  2. 2. Respondents, sample size, and sample design;
  3. 3. Type of drug investigated;
  4. 4. Highlights of the results / findings; and
  5. 5. Limitations of the data base in terms of its potential usefulness for projecting nonmedical drug use among young adults.

A Review of Drug Abuse Data Bases: Treatment-Oriented Data Systems

Survey data

Because nonmedical drug use is of concern to individuals in many walks of life, a host of studies have been conducted to investigate the problem. Often these studies are undertaken by State or local agencies or by school systems. While data gathered by these studies can be extremely useful for monitoring the level of drug use in a particular locale, national estimates cannot be based on these figures. Therefore, this report will consider only surveys of national scope.

The National Institute on Drug Abuse has sponsored several nationwide surveys in the past decade. This section will discuss those surveys conducted by NIDA which have focused on special groups of respondents (e.g., young men or high school students). The series known as the National Household Surveys will be the topic of the section that follows.

1. Young Men and Drugs — A Nationwide Survey

Description. A landmark study was Young Men and Drugs — A Nationwide Survey (). Data on the nonmedical use of psychoactive drugs was collected in 1974 and 1975 on men aged 20 to 30 years. This study had three characteristics that no previous study combined:

  • The sample was representative of the general population rather than of clinical or other special populations.
  • All of the commonly used psychoactive drugs were studied in a standard framework, to allow comparisons between drugs in patterns and correlates of use.
  • Detailed information on the correlates and consequences of drug use were collected.

Respondents and Sampling. Data for this study were collected from October 1974 to May 1975 by personal interviews with 2,510 men out of an original sample of 3,024. The study was designed so that data would be representative of all men in the general population who were 20 to 30 years old, inclusive, in 1974. The survey utilized a multistage stratified random sample from Selective Service lists maintained by local Selective Service Boards. By this method, all young men in the U. S. registered with Selective Service had a known chance to be selected.

Drugs Investigated. Nine classes of drugs were investigated in this study: tobacco, alcohol, cannabis, psychedelics, stimulants, sedative-hypnotics, heroin, other opiates, and cocaine.

Information Collected. The core of the interviews conducted for this survey related to past and current drug use. The questionnaire included a series of screening questions to determine which drugs had been used, and how often. These were followed by detailed questions about drugs used ten times or more. The interview also covered a variety of demographic characteristics (e.g., education, religion, criminal behavior, occupation).

In addition, there were two brief self-administered questionnaires to obtain factual data and some indicators of attitudes and values.

Data / Results. The data from this survey show that:

  • Larger proportions of men in the younger cohorts than in the older used all drugs with the exception of alcohol and tobacco.
  • Median age at onset of use was lower in the younger than in the older cohorts.
  • Age is inversely related to drug use.

By the best estimates available,

  • More than 1,000,000 men in the 20-30 year age range had used heroin, over 2,500,000 had used cocaine, and more than 10,000,000 had used marijuana.
  • Of the men interviewed, 38.2 percent were currently (1974-1975) using marijuana.
  • Men were more likely to continue using alcohol and marijuana once they began using them than they were the other drugs.
  • Of the men interviewed, 49 percent of those born in 1954 were currently (1974-1975) using marijuana, while only 19 percent of those born in 1945 were currently using marijuana.

Limitations of the Data Base. This data base is limited for a number of reasons. First, it was intentionally limited to young men. Second, it was limited in its use of Selective Service lists; it excluded men who enlisted before age 18 and stayed in the service beyond age 26 (they were not required to register), as well as any men who simply failed to register. In addition, as the authors pointed out, much of the analysis of these data consisted of comparisons of one part of the sample with another, although there was no basis to assume that all parts of the sample were representative of the corresponding parts of the population because the sample had not been stratified.

2. Drugs and American Youth

Description. This study was begun in 1966 to examine the changing lifestyles, values and preferences of American youth on a continuous basis. A panel of 2,200 young men were followed for 3-1 / 2 years, from the fall of their tenth grade year to the spring of their first year out of high school. A second phase of the study involved surveys of male high school seniors between 1969 and 1974. Drug questions were included beginning in 1970. Then, in 1975, followup surveys were conducted with the entire class for five separate years after graduation.

Respondents and Sampling. 1,798 males were tested in 1970. This represented 71 percent of the original sample, which was drawn to be representative of the national population of boys who were starting tenth grade in public high schools in the continental United States in the fall of 1966.

Drugs Investigated. Drugs included in the survey were: alcohol, marijuana (including hashish), amphetamines, barbiturates, heroin, and hallucinogens.

Information Collected: For each drug class listed above, the following questions were asked:

  • How many of your friends would you estimate use the drug?
  • How often have you done this during part or all of the last year for other than medical reasons?
  • Previous to this past year (that is, before last summer), how often had you done this for other than medical reasons?

Additionally, opinions were gathered on the use of various drugs, and on the availability of various drugs.

Data / Results. Although 29 percent of the original sample were not interviewed in this study, the authors conclude that the “population estimates of such things as drug use were probably changed very little due to panel attrition.”

Pertinent results for the 1970 data are as follows:

  • Twenty-one percent had used marijuana sometime in the past.
  • For other drugs, usage at sometime in the past ranged from 1.8 percent for heroin to 7 percent for hallucinogens.
  • Thirty-four percent had used marijuana during the year just after high school graduation.
  • For other drugs, 2.2 percent had used heroin in the past year and 11.4 percent had used hallucinogens.

Limitations of the Data Base. Although these data appeared to be comprehensive, their usefulness for making national predictions was limited in several ways. First, there was the problem of attrition and its relation to drug usage. By comparing the remaining sample and the original sample the authors concluded this was not a problem. However, there was evidence to indicate that dropouts were underrepresented in the sample and it was difficult to know how their drug use patterns may have differed from the in-school sample.

The other limitations of the data base have been discussed earlier so only will be mentioned here. The age range was limited. Geographic dispersion was not assured. Finally, the drug data were not available for more than one year.

3. Drug Use Among American High School Students 1975-1977; Drugs and the Class of ‘78: Behaviors, Attitudes, and Recent National Trends; and Drugs and the Nation’s High School Students Five-Year National Trends

Description. All three of these reports (Johnston et al. 1977; Johnston et al. 1979a; Johnston et al. 1979b) are products of the project, Monitoring the Future: A Continuing Study of the Lifestyles and Values of Youth. This study was conducted by the University of Michigan, Institute for Social Research. These three reports present data, respectively, on the graduating classes of 1975 through 1977; the graduating classes of 1975 through 1978; and the graduating classes of 1975 through 1979. Each of these reports presents data on the following general topics

  • Current prevalence of drug use among high school seniors;
  • Trends in drug use since 1975;
  • Grade of first use;
  • Intensity of use;
  • Attitudes and beliefs regarding various types of drug use; and
  • Perceptions of certain relevant aspects of the social environment.

The general purpose of the Monitoring the Future study was to develop an accurate picture of the current situation and trends of illicit drug use among young Americans. This would, in turn, provide a rational basis for public debate and formulation of policy. In order to determine the extent of involvement with drugs, this study focused attention on higher frequency levels rather than on proportions of respondents who had “ever used” drugs.

Respondents and Sampling. The same data collection and sampling procedures were used for all five years. Data were collected from high school seniors in the spring of each year, beginning with the graduating class of 1975. Approximately 125 public and private high schools were selected to provide a cross-section of high school seniors.

It should be noted that each sampled school (except for half in 1975) was asked to participate in two data collections. This permitted replacement of half of the total sample of schools each year. The advantages of this were increased administrative efficiency and the ability to determine whether appreciable shifts in scores from one graduating class to the next were attributable to differences in the newly sampled schools.

Within each selected school, up to 400 seniors could be included in the data collection. In schools with fewer than 400 seniors the usual procedure was to include all seniors. In larger schools a subset of seniors was selected either by randomly sampling classrooms or by some other random method. Sample weights were assigned to each respondent to take into account variations in the size of the sample from school to school as well as variations in selection probability occurring at earlier sampling stages. The response rate over the five years averaged 79 percent.

Drugs Investigated. Eleven classes of drugs were distinguished for purposes of this study: marijuana, inhalants, hallucinogens, cocaine, heroin, natural and synthetic opiates other than heroin, stimulants, sedatives, tranquilizers, alcohol, and cigarettes. In making their responses, students were asked to exclude usage of these drugs under medical supervision.

Information Collected. Data were collected by means of five different questionnaire forms, distributed in an ordered sequence which ensured five virtually identical subsamples. One-third of each of the five questionnaire forms consisted of key or core variables common to all forms; all demographic variables and nearly all drug use variables were included in these core variables.

Data / Results. Data resulting from this study include the following:

  • In 1977, 61.6 percent of the seniors reported using illicit drugs at some time in their lives, as compared to 64.1 percent in 1978 and 65 percent in 1979. However, a substantial portion of all illicit users (42 percent in 1977 and 43 percent in both 1978 and 1979) have used only marijuana.
  • Roughly one-third of the seniors in 1977 through 1979 reported using an illicit drug other than marijuana at some time.
  • Marijuana was by far the most widely used illicit drug–47 percent of the class of 1975, 59 percent of the class of 1978, and 60 percent of the class of 1979 reported some use in their lifetime.
  • Heroin was the most infrequently used drug; only 1.6 percent of the sample in 1978 and 1.1 percent of the sample in 1979 admitted to ever having used it.

In general,

  • Higher proportions of males than females were involved in drug use, especially heavy drug use. However, nearly equal proportions of males and females reported at least some illicit use of drugs other than marijuana in the last year.
  • Between 1975 and 1979 there was only a small increase in the proportion who used some illicit drug besides marijuana, with lifetime prevalence rising only 1 percent (from 36 percent to 37 percent) between 1975 and 1979 and annual prevalence rising only 2 percent (from 26 percent to 28 percent).

Limitations of the Data Base. The two major biases inherent in this methodology stem from the use of high school seniors and the administration of the questionnaire in the classroom. As the authors note, this methodology excluded both high school dropouts (15-20 percent of each age cohort) and absentees. It is likely that both dropouts and students who were regularly absent were more likely to be involved in drug use than the remainder of the high school seniors. However, the authors note that this bias would quite probably remain the same from year to year, thus not affecting their trend data. In addition, it is important to note that the sampling methodology utilized in this study increased the likelihood of schools with larger senior classes being selected for data collection. However, weights were assigned to respondents to minimize any bias resulting from this sampling method.

A Review of Drug Abuse Data Bases: the National Survey

A Review of Drug Abuse Data Bases: Summary

The foregoing studies were reviewed to judge their adequacy as data bases for estimating the number of youthful drug users in the population. Several factors were considered in assessing each study’s potential usefulness for making projections, including: the respondents selected for study, the sampling methodology used, the drugs investigated, and the information collected. Based on these factors the following conclusions were derived.

Treatment-oriented data bases have one major advantage: the data are collected routinely every year. However, these data bases are not adequate for estimating youthful drug abuse for several reasons.

  • The respondents are self-selected and represent a unique subpopulation of drug users.
  • Geographical representation is not assured.
  • Generally, published data are not reported with the level of detail necessary for our estimates (e.g., age, race, or sex of user).

Therefore, although some of these data could possibly be used to corroborate drug use trends, they cannot be the basis from which the estimates will be derived.

Also considered as data sources for the drug use estimates were those high-quality surveys which investigated young adults and were national in scope. Those studies described in this review all contained sound sampling methodologies which made them warrant further attention. However, there were several reasons why, as a group, these studies were not suitable for prediction purposes:

  • Generally, the studies were one-shot endeavors. Therefore, it was impossible to use the data for developing trends.
  • The one study that was continued over a period of time did not investigate the full age range of interest.
  • The drugs investigated and the information collected were not consistent from study to study.

The five national surveys conducted by the National Commission on Marihuana and the National Institute on Drug Abuse shared certain characteristics which make them most suitable for predicting future drug use.

These surveys have maintained essentially the same format over the nine-year period that they cover. They have used similar sampling methodologies, drug classes, and wording of questions. The sampling methodology used has consistently been a sound one which has yielded reliable and valid results. Also, the sampling frame has adequately covered the major geographical areas of the United States. Another factor which makes these studies particularly useful is that generally, in the analyses, they highlight the age range of interest, 18-25 years.

There are some inconsistencies from year to year which make it impossible to compare all drug classes or all types of information for every year in which a study was conducted. Even with these problems, however, the 1976 and 1977 surveys, in particular, provide a wealth of data with which to estimate the number of 18-25 year-old drug users in the future.

The following chapter presents projections of drug abuse for young adults in 1985, 1990, and 1995.


Selections from the book: “Demographic Trends and Drug Abuse, 1980-1995”. Louise G. Richards, Ph.D. , ed. Estimates of probable extent and nature of nonmedical drug use, 1980-1995, based on age structure and other characteristics of U.S. population. National Institute on Drug Abuse Research Monograph 35, May 1981.