relationship between employee drug use and the characteristics of workplace drug testing. Hoffman and Lavison (1999) found that adults who use marijuana or cocaine at least weekly are more likely than other employed adults to work for companies with no testing program and less likely to work for companies with preemployment testing or a random postemployment program. For example, compared with persons who never used marijuana or who used it more than three years earlier, people who used marijuana at least weekly within the year prior to the survey were about one-third as likely to work for employers with postemployment drug testing. However, because the study is cross-sectional, it cannot shed any light on whether this association is attributable to self-selection (drug testing deters current drug users from working for firms with testing programs) or to deterrence (the threat of drug testing leads drug users to terminate or reduce their drug use).
Some proponents of employment drug testing point to the fact that the percentage of employees who test positive has declined significantly (from 18 percent in 1987 to 5 percent in 1997) in tandem with a significant increase in the proportion of large employers requiring drug testing (from 21 percent in 1987 to 81 percent in 1997). It is clear, however, that there need be no causal link between these two trends. First, the decline in percentage of positive tests reflects the overall decline in prevalence of drug use in the young adult population, which began in the early 1980s and extended into the 1990s. Second, a greater proportion of drug tests are based on random selection and therefore the pool of employees being tested has been expanded to include more nonusers.
The committee recommends that the Bureau of Labor Statistics monitor the measures taken by employers to discourage use of illegal drugs by their employees, including drug testing, and that the National Institute on Drug Abuse support rigorous research on the preventive effects and cost-effectiveness of workplace drug testing.
There is very little systematic evidence regarding the application of school disciplinary sanctions to student use of illegal drugs (or alcohol or tobacco). Two recent surveys shed some light on the subject. In 1997 the Department of Education commissioned a survey of school principals in a nationally representative sample of regular public elementary, middle and secondary schools in the United States (National Center for Education Statistics, 1998). According to their reports, a substantial majority of schools purport to have a zero-tolerance policy toward student use of