Drug Use among U.S. Workers, Work Place Safety, and Drug Testing Policies
» Project Details
There are two dominant issues to consider when examining drug use among US, workers. First, what effect does drug use have
on outcomes that affect workplace safety and productivity? Embedded within this question is whether drug use affects
workplace accidents, absenteeism, and turnover. Second, how successful are workplace drug testing programs in deterring drug
use among US workers? An important interrelationship between these two issues is evident: If drug use has a negative impact
on workplace safety and productivity, and drug use is deterred by drug testing programs, then one might conclude that drug
testing programs are an efficient policy approach for minimizing drug use and increasing economic productivity in the United
We approach the two issues mentioned above through an analysis of data from the 1994 National Household Survey on Drug
Abuse (NHSDA). The 1994 NHSDA included a number of relevant questions about workplace accidents, absences, turnover,
drug testing programs, and perceptions of drug testing programs. Overall, we find that 12 percent of full and part-time workers
reported past year illicit drug use, including 10 percent who reported marijuana use and 2 percent who reported cocaine use.
Moreover, about five percent of workers had a work-related accident in the past year, six percent reported an unexcused
absence from work in the past month, and almost two percent were fired from a job in the past year. Finally, about 34 percent of
workers said that their employer tests employees for drugs at hiring while 20 percent said that their employer tests employees
for drugs on a random basis.
Multivariate models designed to predict workplace outcomes indicate that illicit drug use has no measurable effect on
work-related accidents and inconsistent effects on unexcused absences from work or being fired from a job in the past year. For
example, workers who used marijuana 50 or more times in the past year were about twice as likely as those who had never
used marijuana to report an unexcused absence from work, but these groups were not significantly different in terms of
work-related accidents or the likelihood of being fired from a job in the past year. Models using a comparable measure of past
year cocaine use yielded similar results.
Evidence for the deterrent effect of drug testing policies is mixed. The relationship between drug use and working for an
employer with a drug testing program is inconsistent, although there appears to be a cut-off point for marijuana use. Workers
who used marijuana 12 or more times in the past year were about one-and-a-half times as likely as those who had never used
marijuana to say they worked for an employer who had no drug testing program of any kind. The results for cocaine use were
mercurial, however, employee perceptions of drug testing policies are more closely linked to the level of marijuana and cocaine
use than are workers' reports of actual drug testing programs. Multivariate models indicate a clear dose-response effect
between frequency of marijuana use and perceptions of drug testing programs. That is, as marijuana use becomes more,
frequent, workers report a greater aversion to working for an employer who tests employees for drug use at hiring or on a
random basis. Models examining cocaine use failed to show a similar dose-response effect; there was no clear differentiation
between heavy and light cocaine users in terms of perceptions of drug testing programs.