This short note documents how drug use impacts how many days an employee misses work and whether drug testing is effective in reducing the number of drug users. The first step is to show that those who use drugs in the past month are more likely to miss one or more days of work than those who did not use drugs in the past month. Figure 1, below, shows the percentage of people who did and did not use drugs in the past month across three scenarios–no drug test at work, drug test at work, and random drug test at work.
The baseline percentage of employees who missed one or more days of work in the past month and who also did not use drugs in the past month is around 11% to 12% across all test scenarios. For those who did use drugs in the past month, the percentage of those who missed one or more days at work increases. 39.72% of those who used drugs in a “No Drug Test at Work” scenario missed one or more days at work.
Something else is apparent in this figure. Those who used drugs in a “Drug Test at Work” scenario and “Random Drug Test at Work” scenario have an even higher percentage of absenteeism from work than those in a “No Drug Test at Work” scenario. This is counter-intuitive because drug testing should reduce the percentage who are likely to miss work. However, my interpretation is that those who use drugs in a scenario with drug tests are hardcore users and even more likely to miss work.
| Figure 1. Average Percentage Miss One or More Days of Work for Those Who Did or Did Not Use Drugs in the Past Month | ||
| Drug Testing Scenario | Did Not Use Drugs in Past Month | Did Use Drugs in Past Month |
| No Drug Test at Work | 12.06% | 39.72% |
| Drug Test at Work | 11.55% | 56.85% |
| Random Drug Test at Work | 12.72% | 54.38% |
The next step is to look at the percentage of total respondents who used drugs in the past month across three testing scenarios–no drug test, drug test, and random drug test–to see if drug testing decreases the percentage of respondents who used drugs relative to “no drug test.” The results are in Figure 2, below. The headline result is that almost all drugs experience a decline in the percentage of respondents who used drugs in the last month, except for crack, heroine, and PCP. These three drugs are the most resilient to testing.
| Figure 2. Percentage of Total Respondents who Used Drugs in the Past Month | |||
| Drug | No Drug Test at Work | Drug Test at Work | Random Drug Test at Work |
| Marijuana | 14.10% | 7.96% | 6.23% |
| Cocaine | 1.15% | 0.77% | 0.58% |
| Crack | 0.09% | 0.16% | 0.16% |
| Heroine | 0.14% | 0.16% | 0.23% |
| Hallucinogens | 0.90% | 0.37% | 0.33% |
| LSD | 0.33% | 0.12% | 0.10% |
| PCP | 0.01% | 0.01% | 0.02% |
| Ecstacy | 0.34% | 0.19% | 0.21% |
| DMT/AMT/Foxy | 0.06% | 0.02% | 0.03% |
| Ketamin | 0.02% | 0.02% | 0.02% |
| Salvia | 0.04% | 0.01% | 0.00% |
| Inhalants | 0.24% | 0.15% | 0.10% |
| Methamphetamine | 0.22% | 0.18% | 0.20% |
To continue the discussion on whether drug testing reduces drug users, see Figure 3, below. This figure shows drug users in a “test” scenario as a percentage of drug users in a “no test” scenario. It also shows drug users in a “random test” scenario as a percentage of drug users in a “no test” scenario. This captures how much drug testing impacts the number of drug users.
All drugs in a “test” scenario and “random test” scenario experience a decline in users compared with the “no test” scenario–except for crack, heroine, and PCP. Ketamin, in the “random test” scenario, did not decline, either. However, PCP seems to be the most resilient drug. 100% of PCP users in the past month missed one or more days of work in the past month. This is true in a “no test” setting, a “test” setting, and a “random test” setting.
| Figure 3. Does Drug Testing Reduce the Percentage of Drug Users? | ||
| Drug | Drug Users in a Test Scenario as a Percentage of Drug Users in a No Test Scenario | Drug Users in a Random Test Scenario as a Percentage of Drug Users in a No Test Scenario |
| Marijuana | 56.45% | 44.22% |
| Cocaine | 67.05% | 50.03% |
| Crack | 189.09% | 179.59% |
| Heroine | 114.23% | 169.77% |
| Hallucinogens | 41.73% | 37.14% |
| LSD | 34.95% | 31.61% |
| PCP | 208.92% | 347.48% |
| Ecstacy | 55.23% | 62.96% |
| DMT/AMT/Foxy | 25.92% | 42.18% |
| Ketamin | 86.14% | 109.35% |
| Salvia | 26.70% | 0.00% |
| Inhalants | 62.97% | 43.37% |
| Methamphetamine | 81.82% | 91.66% |
The next question is whether random drug testing produces better results than general testing, and the answer is no. I ran a t-test, comparing drug users in a “test” scenario as a percentage of drug users in a “no test” scenario and drug users in a “random test” scenario as a percentage of drug users in a “no test” scenario. The t-statistic was 0.6931, which says that general testing and random testing are not significantly different.
This short note documents that drug testing does reduce the number of drug users, except for users of crack, heroine, and PCP. These are the drugs most resilient to testing. Furthermore, 100% of PCP users missed at least one day of work in all settings–“no test,” “test,” and “random test.” Finally, random testing does not seem to provide more power in reducing the number of drug users than in general testing.
Source: http://datafiles.samhsa.gov/