REASON MAGAZINE

November 2002

Urine -or You're Out

Drug testing is invasive, insulting, and generally irrelevant to job performance. Why do so many companies insist on it?

By Jacob Sullum  jsullum@reason.com

"I ain't gonna pee-pee in no cup, unless Nancy Reagan's gonna drink it up."

-- from the 1987 song "I Ain't Gonna Piss in No Jar," by Mojo Nixon

In 1989 the U.S. Supreme Court upheld a drug test requirement for people seeking Customs Service positions that involved carrying a gun, handling classified material, or participating in drug interdiction. Justice Antonin Scalia dissented, calling the testing program an "immolation of privacy and human dignity in symbolic opposition to drug use." Scalia noted that the Customs Service policy required people to perform "an excretory function traditionally shielded by great privacy" while a monitor stood by, listening for "the normal sounds," after which "the excretion so produced [would] be turned over to the Government for chemical analysis." He deemed this "a type of search particularly destructive of privacy and offensive to personal dignity."

Six years later, Scalia considered a case involving much the same procedure, this time imposed on randomly selected athletes at a public high school. Writing for the majority, he said "the privacy interests compromised by the process of obtaining the urine sample are in our view negligible."

Last March the Supreme Court heard a challenge to a broader testing program at another public high school, covering students involved in any sort of competitive extracurricular activity, including chess, debate, band, choir, and cooking. "If your argument is good for this case," Justice David Souter told the school district's lawyer, "then your argument is a fortiori good for testing everyone in school." Scalia, who three months later would join the majority opinion upholding the drug test policy, did not seem troubled by that suggestion. "You're dealing with minors," he noted.

That factor helps explain Scalia's apparent equanimity at the prospect of subjecting every high school student to a ritual he had thought too degrading for customs agents. But his nonchalance also reflects the establishment of drug testing as an enduring fact of American life. What was once the "immolation of privacy and human dignity" is now business as usual.

While the government has led the way, the normalization of drug testing has occurred mainly in the private sector, where there are no constitutional barriers to the practice. Today about half of all U.S. employers require applicants, workers, or both to demonstrate the purity of their bodily fluids by peeing into a cup on demand. For defenders of liberty, this situation arouses mixed feelings.

On the one hand, freedom of contract means that businesses should be allowed to set whatever conditions they like for employment. People who don't want to let Home Depot or Wal-Mart sample their urine can take their labor elsewhere. The fact that drug testing is widespread suggests either that applicants and employees do not mind it much or that it enhances profits enough to justify the extra cost of finding and keeping workers, along with the direct expense of conducting the tests.

On the other hand, the profit motive is clearly not the only factor driving the use of drug testing. Through mandates and exhortation, the government has conscripted and enlisted employers to enforce the drug laws, just as it has compelled them to enforce the immigration laws. In 1989 William Bennett, then director of the Office of National Drug Control Policy, cited drug testing by employers as an important element of the government's crackdown on recreational users. "Because anyone using drugs stands a very good chance of being discovered, with disqualification from employment as a possible consequence," he said, "many will decide that the price of using drugs is just too high." The Institute for a Drug-Free Workplace, a coalition that includes companies that supply drug testing services as well as their customers, echoes this line. "Employers and employees have a large stake and legitimate role to play in the 'war on drugs,'" the institute argues. "A high level of user accountability...is the key to winning the 'war on drugs.' "

Why Test?

Federal policies requiring or encouraging drug testing by private employers include transportation regulations, conditions attached to government contracts, and propaganda aimed at convincing companies that good corporate citizens need to take an interest in their workers' urine. From the government's perspective, it does not matter whether this urological fixation is good for a company's bottom line. And given the meagerness of the evidence that drug testing makes economic sense, it probably would be much less popular with employers if it were purely a business practice rather than a weapon of prohibition. If it weren't for the war on drugs, it seems likely that employers would treat marijuana and other currently illegal intoxicants the way they treat alcohol, which they view as a problem only when it interferes with work.

Civilian drug testing got a big boost in 1986, when President Reagan issued an executive order declaring that "drugs will not be tolerated in the Federal workplace." The order asserted that "the use of illegal drugs, on or off duty," undermines productivity, health, safety, public confidence, and national security.

In addition to drug testing based on "reasonable suspicion" and following accidents, Reagan authorized testing applicants for government jobs and federal employees in "sensitive positions." Significantly, the order was based on the premise that "the Federal government, as the largest employer in the Nation, can and should show the way towards achieving drug-free workplaces." Two years later, Congress approved the Drug-Free Workplace Act of 1988, which demanded that all federal grant recipients and many contractors "maintain a drug-free workplace." Although the law did not explicitly require drug testing, in practice this was the surest way to demonstrate compliance.

Private employers, especially big companies with high profiles and lucrative government contracts (or hopes of getting them), soon followed the government's lead. In its surveys of large employers, the American Management Association found that the share with drug testing programs increased from 21 percent in 1987 to 81 percent in 1996. A 1988 survey by the Bureau of Labor Statistics estimated that drug testing was required by 16 percent of work sites nationwide.

Four years later, according to a survey by the statistician Tyler Hartwell and his colleagues, the share had increased to nearly half. In the 1997 National Household Survey on Drug Abuse (the source of the most recent nationwide data), 49 percent of respondents said their employers required some kind of drug testing.

As many as 50 million drug tests are performed each year in this country, generating revenue in the neighborhood of $1.5 billion. That's in addition to the money earned by specialists, such as consultants and medical review officers, who provide related services. Drug testing mainly affects pot smokers, because marijuana is much more popular than other illegal drugs and has the longest detection window. Traces of marijuana can be detected in urine for three or more days after a single dose, so someone who smoked a joint on Friday night could test positive on Monday morning. Daily marijuana smokers can test positive for weeks after their last puff. Because traces linger long after the drug's effects have worn off, a positive result does not indicate intoxication or impairment. (See sidebar.)

The relevance of such test results to job performance is by no means clear. But in the late 1980s and early '90s, government propaganda and alarmist press coverage combined to persuade employers that they could no longer rely on traditional methods for distinguishing between good and bad workers. "When employers read in Time and Newsweek and U.S. News & World Report that there was an epidemic of drug abuse in America, they got scared like everyone else," says Lewis Maltby, president of the National Workrights Institute and a leading critic of drug testing. "They didn't want some pothead in their company causing a catastrophe and killing someone. Drug testing was the only answer that anyone presented to them, so they took it." Because drug testing was seen as an emergency measure, its costs and benefits were never carefully evaluated. "Most firms are understandably rigorous about making major investment decisions," Maltby says, "but drug testing was treated as an exception."

My interviews with officials of companies that do drug testing -- all members of the Institute for a Drug-Free Workplace -- tended to confirm this assessment. They all seemed to feel that drug testing was worthwhile, but they offered little evidence to back up that impression.

Link Staffing Services, a Houston-based temp agency, has been testing applicants since the late 1980s. "In the industry that we are in," says Amy Maxwell, Link's marketing manager, "a lot of times we get people with undesirable traits, and drug testing can screen them out real quick." In addition to conducting interviews and looking at references, the company does background checks, gives applicants a variety of aptitude tests, and administers the Link Occupational Pre-employment Evaluation, a screening program that "helps identify an applicant's tendency towards characteristics such as absenteeism, theft and dishonesty, low productivity, poor attitude, hostility, and drug use or violence." Although the drug testing requirement may help impress Link's customers, it seems unlikely that urinalysis adds something useful to the information from these other screening tools. Asked if drug testing has affected accident rates or some other performance indicator, Maxwell says, "We probably don't track that, because we have other things that [applicants] have to pass."

Eastman Kodak, which makes photographic supplies and equipment, tests all applicants in the U.S. but tests employees (except for those covered by Department of Transportation regulations) only when there's cause for suspicion of drug-related impairment. Wayne Lednar, Eastman Kodak's corporate medical director, says safety was the company's main concern when it started doing drug testing in the 1980s. "Our safety performance has substantially improved in the last 10 years on a worldwide basis, not just in the United States," Lednar says. "That improvement, however, is not one [for which] the drug testing approach in the U.S. can be the major explanation. A very large worldwide corporation initiative driven by line management is really what I think has made the difference in terms of our safety performance."

David Spratt, vice president for medical services at Crown Cork & Seal, a Philadelphia-based packaging manufacturer, says that when the company started doing drug testing in the early 1990s, "there was a concern that employees who used drugs were more likely to have problems in the workplace, be either the perpetrators or the victims of more accidents or more likely to be less productive." But like Eastman Kodak, Crown Cork & Seal does not randomly test employees; once they're hired, workers can use drugs without getting into trouble, as long as they do their jobs well. "What drives our concern is work performance," Spratt says. "If there is such a thing [as] 'recreational use,' we would probably not find that out."

Asked if the company has any evidence that drug testing has been effective, Spratt says: "That's not typically the way these things start out. They typically start out with, "We gotta do drug testing, because the guy up the street is doing drug testing, and the people who walk in and see his sign will come down and sign up with us for a job." We're going to get the skewed....They will be a different group who may be less than desirable."

Margot Brown, senior director of communications and public affairs at Motorola, which makes semiconductors, cell phones, and two-way radios, says that when the company started doing drug testing in 1988, "They were trying to control the quality of their products and the safety of their work force." Asked whether the goals were accomplished, she says: "Our productivity per employee did go up substantially....Who knows if that was coincidental or not? Those were good years for Motorola."

Phantom Figures

As those remarks suggest, drug testing became broadly accepted without any firm evidence that it does what it's supposed to do: improve safety, reduce costs, and boost productivity. "Despite beliefs to the contrary," concluded a comprehensive 1994 review of the scientific literature by the National Academy of Sciences, "the preventive effects of drug-testing programs have never been adequately demonstrated." While allowing for the possibility that drug testing could make sense for a particular employer, the academy's panel of experts cautioned that little was known about the impact of drug use on work performance. "The data obtained in worker population studies," it said, "do not provide clear evidence of the deleterious effects of drugs other than alcohol on safety and other job performance indicators."

It is clear from the concessions occasionally made by supporters of drug testing that their case remains shaky. "Only limited information is available about the actual effects of illicit drug use in the workplace," admits the Drug-Free America Foundation on its Web site. "We do not have reliable data on the relative cost-effectiveness of various types of interventions within specific industries, much less across industries. Indeed, only a relatively few studies have attempted true cost/benefit evaluations of actual interventions, and these studies reflect that we are in only the very early stages of learning how to apply econometrics to these evaluations."

Lacking solid data, advocates of drug testing tend to rely on weak studies and bogus numbers. The Office of National Drug Control Policy, for example, claims a 1995 study by Houston's Drug-Free Business Initiative "demonstrated that workplace drug testing reduces injuries and worker's compensation claims." Yet the study's authors noted that the "findings concerning organizational performance indicators are based on numbers of cases too small to be statistically meaningful. While they are informative and provide basis for speculation, they are not in any way definitive or conclusive, and should be regarded as hypotheses for future research."

Sometimes the "studies" cited by promoters of drug testing do not even exist. Quest Diagnostics, a leading drug testing company, asserts on its Web site that "substance abusers" are "3.6 times more likely to be involved in on-the-job accidents" and "5 times more likely to file a worker's compensation claim." As Queens College sociologist Lynn Zimmer has shown, the original source of these numbers, sometimes identified as "the Firestone Study," was a 1972 speech to Firestone Tire executives in which an advocate of employee assistance programs compared workers with "medical-behavioral problems" to other employees. He focused on alcoholism, mentioning illegal drugs only in passing, and he cited no research to support his seemingly precise figures. Another number from the Firestone speech appears on the Web site of Roche Diagnostics, which claims "substance abusers utilize their medical benefits 300 percent more often than do their non-using co-workers."

Roche also tells employers that "the federal government estimates" that "the percentage of your workforce that has a substance abuse problem" is "about 17 percent." This claim appears to be a distortion of survey data collected by the National Institute of Mental Health (NIMH). As summarized by the American Psychiatric Association, the data indicate that "nearly 17 percent of the U.S. population 18 years old and over will fulfill criteria for alcohol or drug abuse in their lifetimes." By contrast, Roche is telling employers that 17 percent of the population meets the criteria at any given time. Furthermore, the vast majority of the drug abusers identified by the NIMH were alcoholics, so the number does not bolster the case for urinalysis aimed at catching illegal drug users.

According to a study published last February in the Archives of General Psychiatry, less than 8 percent of the adult population meets the criteria for "any substance use disorder" in a given year, and 86 percent of those cases involve alcohol. The study, based on data from the National Comorbidity Survey, found that 2.4 percent of respondents had a "substance use disorder" involving a drug other than alcohol in the previous year. So Roche's figure -- which is also cited by other companies that profit from drug testing, such as RapidCup and eVeriTest -- appears to be off by a factor of at least two and perhaps seven, depending upon whether "substance abuse problem" is understood to include alcohol.

Drinking Problems

This ambiguity seems to be deliberate. To magnify the size of the problem facing employers, the government and the drug testing industry routinely conflate illegal drugs with alcohol. But it's clear that employers are not expected to treat drinkers the way they treat illegal drug users. Although drinking is generally not allowed on company time, few employers do random tests to enforce that policy. In 1995, according to survey data collected by Tyler Hartwell and his colleagues, less than 14 percent of work sites randomly tested employees for alcohol. And while 22 percent tested applicants for alcohol, such tests do not indicate whether someone had a drink, say, the night before. In any case, it's a rare employer who refuses to hire drinkers.

When it comes to illegal drugs, by contrast, the rule is zero tolerance: Any use, light or heavy, on duty or off, renders an applicant or worker unfit for employment. "With alcohol, the question has always been not 'Do you consume?' but 'How much?' notes Ted Shults, chairman of the American Association of Medical Review Officers, which trains and certifies physicians who specialize in drug testing. "With the illegal drugs, it's always, 'Did you use it?'

The double standard is especially striking because irresponsible drinking is by far the biggest drug problem affecting the workplace. "Alcohol is the most widely abused drug among working adults," the U.S. Department of Labor notes. It cites an estimate from the Substance Abuse and Mental Health Services Administration that alcohol accounts for 86 percent of the costs imposed on businesses by drug abuse.

In part, the inconsistency reflects the belief that illegal drug users are more likely than drinkers to become addicted and to be intoxicated on the job. There is no evidence to support either assumption. The vast majority of pot smokers, like the vast majority of drinkers, are occasional or moderate users. About 12 percent of the people who use marijuana in a given year, and about 3 percent of those who have ever tried it, report smoking it on 300 or more days in the previous year. A 1994 study based on data from the National Comorbidity Survey estimated that 9 percent of marijuana users have ever met the American Psychiatric Association's criteria for "substance dependence." The comparable figure for alcohol was 15 percent.

According to the testing industry, however, any use of an illegal drug inevitably leads to abuse. "Can employees who use drugs be good workers?" Roche asks in one of its promotional documents. Its answer: "Perhaps, for awhile. Then, with extended use and abuse of drugs and alcohol, their performance begins to deteriorate. They lose their edge. They're late for work more often or they miss work all together....Suddenly, one person's drug problem becomes everyone's problem." This equation of use with abuse is a staple of prohibitionist propaganda. "It is simply not true," says the Drug-Free America Foundation, "that a drug user or alcohol abuser leaves his habit at the factory gate or the office door." The message is that a weekend pot smoker should be as big a worry as an employee who comes to work drunk every day.

Employers respond to the distinctions drawn by the government. Under the Americans With Disabilities Act, for example, alcoholics cannot be penalized or fired without evidence that their drinking is hurting their job performance. With illegal drugs, however, any evidence of use is sufficient grounds for disciplinary action or dismissal.

A Crude Tool

A more obvious reason government policy shapes employers' practices is that many do not want to hire people who break the law. A positive urinalysis "proves someone has engaged in illegal behavior," observes drug testing consultant Michael Walsh, who headed the task force that developed the federal government's drug testing guidelines. "All companies have rules, and this is a way of screening out people who are not going to play by the rules." He concedes that "you are going to rule out some people who would have made really good employees, and you are going to let in some people who make lousy employees."

Still, he says, "in a broad way, it's a fairly decent screening device."

Perhaps the strongest evidence in support of drug testing as a screening device comes from research involving postal workers conducted in the late 1980s. A study reported in The Journal of the American Medical Association in 1990 found that postal workers who tested positive for marijuana when they were hired were more prone to accidents, injuries, absences, disciplinary action, and turnover.

The differences in these rates were relatively small, however, ranging from 55 percent to 85 percent. By contrast, previous estimates had ranged from 200 percent for accidents to 1,500 percent for sick leave. "The findings of this study suggest that many of the claims cited to justify pre-employment drug screening have been exaggerated," the researchers concluded.

Even these comparatively modest results may be misleading. The study's methodology was criticized on several grounds, including an accident measure that gave extra weight to mishaps that occurred soon after hiring. A larger study of postal workers, reported the same year in the Journal of Applied Psychology, confirmed the finding regarding absenteeism but found no association between a positive pre-employment drug test and accidents or injuries. On the other hand, workers who had tested positive were more likely to be fired, although their overall turnover rate was not significantly higher.

It's hard to know what to make of such findings. As the National Academy of Sciences noted, "drug use may be just one among many characteristics of a more deviant lifestyle, and associations between use and degraded performance may be due not to drug-related impairment but to general deviance or other factors." On average, people who use illegal drugs may be less risk-averse or less respectful of authority, for example, although any such tendencies could simply be artifacts of the drug laws.

In any case, pre-employment tests, the most common kind, do not catch most drug users. Since people looking for a job know they may have to undergo a drug test, and since the tests themselves are announced in advance, drug users can simply abstain until after they've passed. For light users of marijuana, the drug whose traces linger the longest, a week or two of abstinence is probably enough.

Pot smokers short on time can use a variety of methods to avoid testing positive, such as diluting their urine by drinking a lot of water, substituting someone else's urine, or adulterating their sample with masking agents. "Employers are very concerned that there's always a way to cheat on a drug test," says Bill Current, a Florida-based drug testing consultant. "The various validity testing methods that are available are always one step behind the efforts of the drug test cheaters."

Generally speaking, then, drug users applying for jobs can avoid detection without much difficulty. "The reality is that a pre-employment drug test is an intelligence test," says Walsh. The people who test positive are "either addicted to drugs, and can't stay away for two or three days, or just plain stupid....Employers don't want either of those." Alternatively, applicants who fail a drug screen may be especially reckless or lazy. In short, it's not safe to draw conclusions about drug users in general from the sample identified by pre-employment tests. By the same token, however, such tests may indirectly measure characteristics of concern to employers.

The upshot of all this is something that neither supporters nor opponents of drug testing like to admit: Even if drug use itself has little or no impact on job performance -- perhaps because it generally occurs outside the workplace -- pre-employment testing still might help improve the quality of new hires. If so, however, it's a crude tool. As an index of undesirable traits, testing positive on a drug test could be likened to having a tattoo. Refusing to hire people with tattoos might, on balance, give a company better employees, but not because tattoos make people less productive or more prone to accidents.

How Much?

Maltby, president of the National Workrights Institute, argues that such benefits are too speculative to justify drug testing, and he believes employers are starting to realize that. "Times are tougher than they were 15 years ago," he says. "Money is tighter, and employers are scrutinizing all of their expenditures to see if they are really necessary. Initially, in the late '80s or early '90s, employers looked at drug testing and said, 'Why not? ' Now employers look at drug testing like everything else and say, 'Where's the payoff' And if nobody sees a payoff, programs get cut -- or, more often, cut back."

One example is Motorola, which has seen its profits slide recently and plans to eliminate a third of its work force by the end of the year. When Motorola started doing drug testing, the company's communications director says, "The cost wasn't really a factor because we really felt like it was something we should attend to at the time." But Motorola recently scaled back its urinalysis program, which for a decade included random testing of employees; now it tests only applicants.

Motorola's decision may be part of a trend. The share of companies reporting drug testing programs in the American Management Association's surveys of large employers dropped from a peak of 81 percent in 1996 to 67 percent last year. Some of that drop may reflect a new questionnaire the organization started using in 1997. The new survey is less focused on testing, which could have changed the mix of companies that chose to participate. But the downward trend continued after 1997.

Once drug testing became common, it acquired a certain inertia: Employers who didn't do it worried that they might be at a disadvantage in attracting qualified workers or maintaining a positive public image. Employers who did it worried that stopping would hurt their recruitment or reputations. Yet without abandoning drug testing completely, a company can save money by giving up random tests. Even if it keeps random tests, it can save money by testing less frequently -- the sort of change that would not be widely noticed.

Still, one reason drug testing endures is that it does not cost very much, especially from the perspective of a large employer. Eastman Kodak, which has more than 100,000 employees worldwide, pays just $12 to $15 per test. Even considering additional expenses (such as the medical review officer's time), and even with thousands of applicants a year, the total cost is a drop in the bucket. Drug tests cost Cork Crown & Seal, which has nearly 40,000 employees worldwide, $25 to $30 per applicant, for a total of less than $100,000 a year. Motorola, which will have about 100,000 employees after this year's cutbacks, spent something like $1 million a year when it was doing random testing of employees -- still not a significant concern to a corporation with billions of dollars in revenue (at least, not until profits took a dive).

Small companies, which have always been less inclined to do drug testing, have to pay more per test and are less able to afford it. They also have lower profiles. "If G.M. were to be on the front page of The Wall Street Journal, announcing that they dropped their drug testing program, I wouldn't want to own their stock," Maltby says. He recalls a conversation in which the president of a Fortune 500 company told him that a few million dollars a year was a small price to pay for the reassurance that drug testing gives stockholders.

The direct costs of drug testing are not the whole story, however. Wayne Sanders, CEO of the paper products giant Kimberly-Clark, has to keep shareholders in mind, but he also worries about the message that drug testing sends to employees. In 1986, when Sanders was the company's head of human resources, managers pressured him to start doing drug testing, arguing that otherwise Kimberly-Clark would get all the addicts rejected by other employers. According to The Dallas Morning News, Sanders, "who wasn't about to pee in a bottle," thought the notion was "utter bunk." He successfully argued that "the idea of urine testing was demeaning and completely alien in a culture based on trust and respect."

There is some evidence that the atmosphere created by drug testing can put employers at a disadvantage. A 1998 Working USA study of 63 high-tech companies found that pre-employment and random drug testing were both associated with lower productivity. The researchers, economists at Le Moyne College in Syracuse, speculated that drug testing programs may create a "negative work environment" that repels qualified applicants and damages employee morale.

The Familiarity Factor

Yet survey data suggest that most Americans have gotten used to the idea that their urine may be part of the price they pay to get or keep a job. In the National Household Survey on Drug Abuse, the share of employees who said they would be less likely to work for a business that tested applicants fell from 8 percent in 1994 to 5 percent in 1997. Random testing of employees was somewhat less popular, with 8 percent saying it would be a negative factor in 1997, compared to 14 percent in 1994. Even among current users of illegal drugs, only 22 percent said pre-employment testing would make a job less appealing in 1997 (down from 30 percent in 1994), while 29 percent said random testing would (down from 40 percent in 1994) -- which suggests how ineffective testing is at identifying drug users.

For those who object to drug testing, the natural tendency is to give in and take the test, on the assumption that a few protests are not likely to change a well-established business practice. But in jobs that require a high level of training or experience, even one person's objection can make a difference. An executive with a global management consulting company says he discussed his use of psychedelics with senior management early on "because I didn't want any negative repercussions later." When the company considered starting a drug testing program, he recalls, "I said, 'I'm not going to subject myself to mandatory testing because I don't have a problem. You know I don't have a problem, so testing me is not going to fly. And I think testing a bunch of people you pay upper five figures to mid to upper six figures is silly.'...The idea was dropped. I like to think I had some impact on that."

CONTINUATION OF ARTICLE CAN BE FOUND AT http://www.reason.com/0211/fe.js.urine.shtml