dissenting:

Working on an oil rig is dangerous business. It requires total concentration, precise timing, a fair degree of coordination and a significant amount of speed. Rig accidents can have disastrous consequences, ranging from severed limbs and multiple deaths to massive despoliation of the environment. It goes without saying that drug abuse has no place on oil rigs and that a company operating oil rigs has the right — indeed, the obligation — to take decisive action when it obtains reliable information that some of its employees may be abusing drugs while on duty.

This is the unhappy tale of a company that did just that. Company officials reasonably believed that three employees had used drugs on the job, not once but repeatedly. Two eyewitnesses fingered the drug-using employees; the company pursued the matter promptly, but not precipitously, obtaining confirmation from yet a *205 third eyewitness before discharging the violators. The personnel action was taken in a balanced, detached, professional manner, free from any hint of rancor or personal animosity. Had the company acted less decisively, it would have betrayed its responsibility to other employees and the environment we all share. Yet when all is said and done, the fingered employees walk off with a cool third of a million dollars, while the company is left to pick up the tab, pay its lawyers and scratch its head wondering what it could have done differently. It is a question we all might ponder as we contemplate the bitter lesson of this cockeyed morality tale.

The majority characterizes this as a run-of-the-mill case, a cursory review for sufficient evidence. But there’s much more going on here. First, the majority violates a fundamental tenet of appellate review by upholding a verdict on a theory that was never presented to the jury. The jury was instructed to decide whether Parker Drilling discharged plaintiffs for just cause, defined by the court as a good faith belief that plaintiffs were using drugs on the job. Plaintiffs did not object to this instruction; accordingly, it is the law of the case. A careful review of the record discloses that no rational jury could have concluded that Parker failed to act in good faith. The majority nevertheless upholds the verdict against Parker, explaining that the jury was really deciding not whether company officials reasonably believed plaintiffs had violated the drug ban, but whether plaintiffs in fact did so. The majority thus upholds the verdict by ignoring the law of the case.

The majority next explains that Alaska law is clear: A good faith belief is not enough to constitute just cause. But it wasn’t clear to the district judge, who instructed the jury that good faith was enough; it wasn’t clear to plaintiffs, who never raised this argument at trial, in post-trial motions or on appeal: and it most certainly is not clear to me.

Worst of all, today’s decision makes us the first court in the country to require an employer to prove to the satisfaction of a jury that employees involved in an inherently hazardous activity, whom they reasonably suspected of using drugs on the job, were in fact guilty of doing so. The majority takes this action casually, without acknowledging the serious public policy consequences of its ruling. In so doing, my colleagues take a giant leap into a dangerous and heretofore uncharted no-man’s land, ill-serving the causes of environmental hygiene, industrial safety and worker privacy.

Facts

Plaintiffs Cantu, Howard and Sanders worked as floor hands on Rig 191, one of several oil rigs operated by Parker Drilling Company on Alaska’s North Slope. The floor area is probably the most dangerous part of the rig: In its center is a large drilling hole that is often overflowing with mud or has mud spraying out of it, making the floor slick and treacherous. As floor hands, plaintiffs’ job was to place in the hole and connect together 90-foot segments of steel pipe, each weighing 1500 lbs. Making matters more difficult, they had to perform this hard physical labor outdoors in temperatures of 30 degrees below zero, working with heavy iron tools, the lightest of which weighs more than 400 lbs. Worse still, the floor surrounding the hole is not stationary. Known as a rotary table, it quite frequently turns without warning, occasionally taking with it the leg of an inattentive floor hand. At its best, the work is hazardous — injuries are frequent, ranging from bruises and cuts to severed fingers and arms; at its worst, men lose their lives. Because of the enormous danger involved, Parker has a strict ban on the use or possession of drugs or alcohol on the rigs. 1 Cantu, Howard and Sanders testified that they were well aware of the ban.

During a routine inspection of the North Slope rigs on February 22, 1983, Safety *206 Director John Haynes was told by two employees — Billy Reynolds, a supervisor on Rig 191, and Joe Watkins, a rig employee who had formerly been a police officer— that plaintiffs were routinely smoking marijuana before and after shifts and on breaks during shifts. Immediately upon returning to the company’s main office in Anchorage the following afternoon, Haynes reported these accusations to Gary McCarrell, Parker’s Division Manager. The undisputed testimony discloses that McCarrell and his staff took the accusations very seriously, giving the matter their full attention. Hesitant to discipline employees on what might have been mere rumor, McCarrell telephoned the rig and asked that Reynolds and Watkins put their allegations in writing. Both men made out handwritten statements and faxed them to McCarrell the same evening. McCarrell and other Parker managers stayed in the office late that night to receive the faxes and formulate a reasoned strategy for dealing with the problem. None of the managers involved in making the decisions— McCarrell, Glenn Mosley and Faye Up-church — were personally acquainted with Cantu, Howard or Sanders.

Reynolds’ and Watkins’ statements disclosed that each man had seen the plaintiffs smoking and/or handling marijuana on the rig, and that other employees had mentioned smelling marijuana smoke coming from plaintiffs’ room. RT 3:84-86. Based on these statements, McCarrell decided to suspend plaintiffs, who at that time were off the rig on a normal two-week rotation. The suspension was to prevent plaintiffs from returning to the rig for their next shift, something McCarrell considered too dangerous in light of the serious allegations made against them. Still, in an effort to be entirely fair, McCarrell sought independent confirmation of the allegations before discharging plaintiffs.

Confirmation came on February 25, when McCarrell interviewed Covan Chapman, who had earlier roomed with plaintiffs. When McCarrell told Chapman that his former roommates were charged with on-the-job drug use, Chapman responded, “I told those guys this was going to get them.... [Tjhat is why I moved out of that room.... I told those guys this was going to happen to them.” RT 3:186-87; see also 3:25. On the basis of this information, McCarrell fired the plaintiffs. 2 So as not to ruin their chances for employment elsewhere, however, the discharge papers noted the reason for termination as violation of company policy.

On this record, no rational jury could have concluded that Parker failed to act in good faith. All the facts I have recited are wholly uncontroverted. Plaintiffs do not dispute that Reynolds and Watkins reported them to Haynes, that Haynes promptly informed McCarrell, that Reynolds and Watkins put their accusations in writing or that McCarrell sought and obtained further confirmation of the allegations from Chapman. Plaintiffs also admit that Parker Drilling had a strong policy against drug use on the job, that they were well acquainted with the policy and that they understood the reasons for it. Plaintiffs suggest no ulterior motive McCarrell, Mosley or Upchurch might have had for firing them.

Plaintiffs’ case at trial rested entirely on the theory that objective facts surrounding the terminations would prove the company had not acted in good faith. See pp. 207-09 & n. 4 infra. Plaintiffs never claimed *207 that the case turned on whether they in fact used drugs. Yet it is this point, conceded by plaintiffs, that the majority relies on to uphold the verdict. The majority never once suggests that the jury could have found a lack of good faith, and with good reason: The record simply would not support such an assertion. Rather, the majority changes the relevant inquiry. As I shall explain, the majority’s approach conflicts with an unbroken line of cases in this circuit which hold that jury instructions not objected to at trial may not be challenged on appeal. The majority also interprets Alaska law in a dangerous and unprecedented way, and wholly ignores serious public policy concerns.

Discussion

I. The Law of the Case

Federal Rule of Civil Procedure 51 provides: “No party may assign as error the giving or the failure to give an instruction unless that party objects thereto before the jury retires to consider its verdict, stating distinctly the matter objected to and the grounds of the objection.” Simply put, if you didn’t object to an instruction in the district court, you can’t challenge it on appeal. Even though the instruction may misstate the applicable law, that misstatement binds the parties because no one complained when the instruction was given. See, e.g., Benigni v. City of Hemet, 879 F.2d 473, 477 (9th Cir.1989) (plaintiff prevailed on first amendment claims that were possibly invalid under Supreme Court precedent, but defendant’s failure to object precluded review on appeal). This rule makes a great deal of sense. Justice is not served by going through all the bother and expense of a jury trial only to have one of the parties point out long afterward that there was some glitch in the way the issues were framed for the jury so that the process must start from scratch.

Here, the jury was instructed that Parker acted with just cause if company officials reasonably believed plaintiffs were smoking marijuana on the rig. The district judge charged the jury as follows: “You are instructed that just cause is generally a fair and honest cause or reason, regulated by good faith on the part of the employer in light of all the facts and circumstances.” CR 94a, Instruction No. 7 (emphasis added). Plaintiffs didn’t challenge the propriety of this instruction. Just before reading the charge, the district judge gave the parties an opportunity to object. The company objected to three of the instructions, including the one defining just cause; plaintiffs, on the other hand, stated they were “satisfied with the instructions” and had “no objections.” RT 6:2.

Nor can there be any doubt that plaintiffs understood that the instruction focused on what the company subjectively believed, not on what plaintiffs actually did. 3 The record is replete with acknowledgments by plaintiffs’ counsel that jury instruction number seven limited the jury to considering whether the company acted in good faith. Plaintiffs’ strategy was to cast doubt on Parker’s motivation by getting the jury to draw inferences of bad faith from certain “objective evidence”— i.e., (1) that plaintiffs never actually smoked marijuana while on Rig 191; (2) that they were treated unfairly in comparison to other employees suspected of using drugs; and (3) that they were threatened with “blackballing” if they contested their termination. As plaintiffs’ counsel explained it,

RT 7:11-12. 4

On appeal, plaintiffs adhere to this reading of the jury instruction and make no

complaint as to its accuracy:

Brief for Appellees at 19-20 (emphasis added). In fact, plaintiffs have never wavered from this position.

In reviewing a jury verdict, it is our job to determine whether there was sufficient evidence to support the jury’s answer to the question posed, not to some other question. After all, we review the record of the trial actually conducted below, not that of some hypothetical trial that might have been conducted. The majority does not dispute that Parker had a good-faith belief that the plaintiffs used drugs on the job; 5 *209 instead, the majority changes the relevant inquiry:

Majority at 194.

This is where the majority loses me. That there was insufficient evidence to cast doubt on the company’s good faith should be the end of the matter; we are obliged to overturn the verdict. 6 We may not inquire whether plaintiffs might have prevailed if just cause had been defined differently; 7 like the parties, we are stuck with the definition set out by the district court in instruction number seven. Rule 51 would clearly preclude us from considering a last-minute challenge to this instruction by plaintiffs; the same rule also precludes the majority from disregarding the instruction on its own initiative.

II. Alaska Law

While I find the majority’s disregard of a long-standing procedural rule troubling, my most serious disagreements with the majority are on issues of substantive law.

I find especially troubling the following statement in the majority opinion: “Alaska’s common law is clear in wrongful termination cases; the jury is entitled to decide whether the alleged conduct that led to the termination actually took place unless the facts are so one-sided the issue can be decided as a matter of law.” Majority at 194. I understand why the majority would like Alaska law to be “clear” on this issue: A plain error exception to Rule 51, if one existed in this circuit, 8 could only help *210 plaintiffs if the district court’s just cause instruction so grossly misstated the law as to amount to a miscarriage of justice. I do not understand, though, how the majority could conclude that Alaska law is “clear” based on the scant caselaw it cites in its opinion. Obviously, neither the district judge nor plaintiffs’ counsel, all based in Alaska, shared the majority’s understanding of state law. Nor do I.

The majority cites only two cases supporting its assertion that Alaska law is clear as to what constitutes just cause. First, it cites a decision of the Alaska Supreme Court, Rutledge v. Alyeska Pipeline Serv. Co., 727 P.2d 1050 (Alaska 1986). Rutledge stands for the unremarkable proposition that, where fighting on company property is a terminable offense and the employee admits he was involved in a fight, the trial court can properly direct a verdict for the employer. Id. at 1052, 1056 . Because Rutledge was not a case where the employee denied having committed the terminable offense, the court did not discuss the issue concerning us — whether, in a disputed case, the jury is to determine if the employee actually committed the offense, or merely if the company reasonably believed he did.

All the same, the majority finds in Rutledge a single stray remark that, if read a certain way, might conceivably support the definition of just cause the majority believes is correct. See Majority at 194. To the extent it chooses to rely on stray remarks, though, the majority could as easily have quoted another passage; only two sentences earlier, Rutledge contains the following statement: “All of the testimony and correspondence regarding Rutledge’s termination indicates that Alyeska discharged him because it determined that Rutledge was the instigator in the altercation.” Id. at 1056 (emphasis added). While Rutledge admitted he was involved in the fight, he vehemently denied having started it; thus, this passage strongly supports the notion that just cause means only that the employer believe the discharged employee committed the bad act. That’s the danger of placing too much reliance on stray remarks: both sides can find them just as easily. The better approach, it seems to me, is to admit that Rutledge doesn’t address the scope of the jury’s inquiry in disputed cases. No Alaska case does.

The majority cites a second case in support of its position, Toussaint v. Blue Cross & Blue Shield, 408 Mich. 579 , 292 N.W.2d 880 (1980). Unlike Rutledge, Tous-saint does share the majority’s view of just cause; it does nothing, however, to advance the majority’s argument as to the clarity of Alaska law. First and most fundamentally, Toussaint is not an Alaska case; it’s from Michigan. It has been cited by Alaska courts a few times, but never for the proposition for which the majority relies on it, nor anything remotely close. Toussaint is known primarily for its holding that, under Michigan law, the terms of an employee handbook can convert an at-will employment relationship into one requiring just cause. While the case discusses peripherally the role of the jury in determining whether there was just cause, this is not the heart of the case, nor is this the portion that is normally cited. No Alaska court has ever embraced the Michigan court’s decision that the factual basis for a discharge decision is to be reviewed by the jury.

In the ten years since it was handed down, Toussaint has been cited by Alaska courts on only three occasions. In Eales v. Tanana Valley Medical-Surgical Group, Inc., 663 P.2d 958 (Alaska 1983), the court cited Toussaint as support for the proposition that promises not to fire an employee so long as he properly performs his duties can modify an employment contract of indefinite duration. Id. at 959-60. Then, in Rutledge v. Alyeska Pipeline Serv. Co., 727 P.2d 1050 (Alaska 1986), the court cited Toussaint for two propositions: that an employee manual can modify an employment contract, id. at 1056, and that prior disciplinary actions might be relevant to *211 establishing a breach of the covenant of good faith and fair dealing. Id. 9 Most recently, in Jones v. Central Peninsula General Hosp., 779 P.2d 783 (Alaska 1989), the court cited Toussaint for the proposition that an employee manual can modify the employment contract, id. at 785-88, and that it is a jury question whether plaintiff was fired for the stated reason or, instead, some other motive. Id. at 789. 10 To say that the Alaska courts, having cited Tous-saint for specific, unrelated propositions, are bound to accept as law every jot and tittle in the case is to announce a rule of stare decisis that would surprise most lawyers. The majority’s reliance on Toussaint as an expression of Alaska law is simply unfounded.

To make the factual basis for all personnel decisions reviewable by a jury is an immensely troubling prospect. Indeed, even the Toussaint court equivocated on the issue. On the one hand, it did not want to give employers carte blanche to terminate employees on a whim; on the other, it recognized the danger in robbing management personnel of the flexibility needed in running a business. 292 N.W.2d at 896-97 . While the court ultimately came down on the side of jury review, it tacked on a limitation: “While the promise to terminate employment only for cause includes the right to have the employer’s decision reviewed, it does not include a right to be discharged only with the concurrence of the communal judgment of the jury.” Id. 292 N.W.2d at 896 . Since then, courts applying Michigan law have limited Tous-saint by circumscribing the jury’s role. See, e.g., Boynton v. TRW, Inc., 858 F.2d 1178, 1182 (6th Cir.1988) (decision to discharge as part of economically motivated reduction in work force not reviewable by jury for “just cause”); Joumas v. Maryland Casualty Co., 698 F.Supp. 675, 678-79 (E.D.Mich.1988) (same); Parker v. Diamond Crystal Salt Co., 683 F.Supp. 168, 172-73 (W.D.Mich.1988) (same); Fischhaber v. General Motors Corp., 174 Mich.App. 450 , 436 N.W.2d 386, 389 (1988) (rejecting plaintiff’s claim that jury must decide whether he was constructively discharged without just cause by means of demotion), app. denied, 432 Mich. 902 (1989).

Given that Alaska courts have never addressed this precise issue in a published opinion, there is no principled reason for imposing the rule of Toussaint on this case. It is clear from the jury instructions that the trial did not proceed under this theory. Even if our circuit had a plain error rule that would permit us to rewrite instructions that grossly misstate the relevant law, there is simply no indication that the definition of just cause read to the jury misstated Alaska law.

If, on the other hand, we adopt the Tous-saint rule, state and federal courts will soon be in the business of reviewing employment discharge decisions on a wholesale basis. Given the number of jobs where, arguably, the employer has adopted a just cause requirement, the courts will become Merit Systems Protection Boards for all private employment relationships. I think we ought not impose that burden on Alaska courts — and Alaska businesses— lightly, particularly in the absence of definitive guidance from the state supreme *212 court. 11

III. Public Policy

The majority acknowledges that Parker’s drug ban finds strong support in public policy, see Majority at 193, 194-95, but fails to adequately consider the impact of its decision on that policy. In this, I believe, the majority falls short. In imposing on Alaska courts a new standard for what constitutes just cause, this court ought not shut its eyes to the factual context in which the case arises. We ignore the decision’s effect on safety and related public policy issues at our peril.

A. In upholding the constitutionality of post-accident drug testing of railroad employees, the Supreme Court recently point *213 ed to the frightening statistics linking substance abuse to work-related accidents. Skinner v. Railway Labor Execs. Ass’n, 489 U.S. 602 , 109 S.Ct. 1402 , 1407-08 & n. 1, 103 L.Ed.2d 639 (1989). 12 Oil rig accidents are no less serious or frequent than train wrecks. Working on a rig has been called “one of the most arduous and risky jobs in the world” and “a daily gamble with death.” Associated Press (AP), March 28, 1980. And with good reason: There have been at least 141 major oil rig accidents since the mid-1950s. United Press Int’l (UPI), Oct. 30, 1982. In 1983, a rig capsized in the China Sea, killing 81 workers. AP, May 21, 1985. In 1980, the Alexander Kielland capsized off the coast of Norway, taking 123 lives. AP, Feb. 16, 1982. And in 1979, a rig blow-out in Mexico’s Gulf of Campeche created an oil spill that lasted 10 months and released 140 million gallons of oil, more than 12 times the amount released by the Exxon Valdez. N.Y. Times, June 17, 1980, at B7, col. 6. 13 These major catastrophes are in addition to the countless smaller accidents which, according to a 1981 International Labour Organization report, claim 250 to 500 lives annually. Reuters North European Service, Dec. 3, 1981. 14

Drug or alcohol abuse by oil rig workers magnifies tremendously the job’s inherent dangers. As the Supreme Court of Alaska noted recently, “marijuana can impair a person’s ability to function normally.... [And] work on an oil rig can be very dangerous.” Luedtke v. Nabors Alaska Drilling, Inc., 768 P.2d 1123, 1136 (Alaska 1989) (citing cases involving deaths and debilitating injuries resulting from oil rig accidents). But this matter-of-fact statement does not adequately describe the dire consequences that can flow from even a momentary lapse in concentration. For example, on February 6, 1982, a worker on the rig Ocean Ranger was not paying attention and pushed the wrong button, inadvertently letting water into the ballast tank that helped keep the rig afloat. The Ocean Ranger listed five degrees as a result, triggering alarm bells. It was a minor emergency but not considered all that serious at the time: just a mistake. Ten days later, the Ocean Ranger capsized and sank without a trace; all 84 men aboard were lost. Reuters, Feb. 17, 1982. The rig sank because water splashed on the ballast control console from broken portlights, causing an electrical malfunction. National Trans. Safety Bd., Marine Accident Report: Capsizing and Sinking of the U.S. Mobile Offshore Drilling Unit Ocean Ranger on February 15, 1982, No. NTSB-MAR-83-2, at 75. While it could not be conclusively established what caused the portlights to break, id. at 76, there was suspicion that the “wrong button” incident 10 days earlier was responsible. See AP, Mar. 22, 1982. It is fair to say, then, that oil rig employees “discharge duties fraught with such risks of injury to others that even a momentary lapse of attention can have disastrous consequences.” Skinner, 109 S.Ct. at 1419 .

In the face of these weighty considerations, the United States Supreme Court and the Supreme Court of Alaska alike have approved the use of highly intrusive measures for assuring that drugs and alcohol are not used by people responsible for the lives and safety of themselves and others. In two cases decided on the same day, Skinner v. Railway Labor Execs. Ass’n, 489 U.S. 602 , 109 S.Ct. 1402 , 103 L.Ed.2d 639 (1989), and National Treasury Employees Union v. Von Raab, 489 U.S. 656 , *214 109 S.Ct. 1384 , 103 L.Ed.2d 685 (1989), the United States Supreme Court upheld drug and alcohol testing of private railroad employees and U.S. Customs Service employees.

Particularly relevant to our case, the Supreme Court of Alaska recently approved random drug testing of employees who operate oil rigs. Luedtke v. Nabors Alaska Drilling, Inc., 768 P.2d 1123 (Alaska 1989). As had the United States Supreme Court, Alaska’s highest court acknowledged that such testing amounted to a serious intrusion into the privacy of the affected employees, as to most of whom, presumably, there would not be the slightest suspicion. Nevertheless, the court concluded that safety concerns justified the intrusion. In fact, Luedtke held that oil rig operators could inquire even into their employees’ off-the-job activities: “Where the public policy supporting the [employees’] privacy in off-duty activities conflicts with the public policy supporting the protection of the health and safety of other workers, and even [these employees] themselves, the health and safety concerns are paramount.” 768 P.2d at 1136 .

Straining the law to affirm the substantial award to the employees in this case cannot be squared with the rulings of our Supreme Court and, more importantly, with those of the supreme court of the state whose law we apply here. At a time when concerns about safety and drug-related accidents are being used to justify wholesale abrogation of the privacy rights of employees about whom there is absolutely no suspicion, the majority sustains a substantial award of damages to employees who were identified as drug users by no fewer than three of their co-workers, including one who risked termination himself by admitting he had joined plaintiffs in using drugs on the job. See n. 2 supra.

While we are properly reluctant to overturn jury verdicts, a meaningful degree of judicial oversight is required where important public safety concerns are implicated. Even in the context of labor arbitration awards, where judicial review is normally precluded, a narrow exception has been carved out for awards that run afoul of “well defined and dominant” public policies. United Paperworkers Int’l Union v. Misco, Inc., 484 U.S. 29, 43 , 108 S.Ct. 364, 374 , 98 L.Ed.2d 286 (1987) (internal quotations omitted). Accordingly, courts have struck down awards reinstating a pilot who flew a commercial passenger plane while drunk, Delta Air Lines v. Air Line Pilots Ass’n, 861 F.2d 665, 666-68 (11th Cir.1988); a truckdriver whose drinking on duty and speeding may have caused his eighteen-wheel rig to overturn, Amalgamated Meat Cutters v. Great Western Food Co., 712 F.2d 122, 123-24 (5th Cir.1983); and a power company employee responsible for checking various meters and gauges to ensure that high pressure equipment did not overheat, who turned out to be a chronic drug user, Georgia Power Co. v. IBEW, Local 84, 707 F.Supp. 531, 533-34 (N.D.Ga. 1989), aff'd, 896 F.2d 507 (11th Cir.1990). Similar considerations of public safety should cause us to reverse a jury verdict where, as here, it is clear that management acted entirely in good faith and with no motive other than to assure safety in the workplace. 15

Neither Rutledge nor Toussaint , the two cases on which the majority relies for its new standard, nor any other case of which I am aware, deals with a situation where public safety was so directly implicated. The plaintiff in Rutledge was a night shift employee; he got into a scuffle with a co-worker about whether or not to leave a window open so that one of the men could smoke. 727 P.2d at 1052 . Neither man was seriously hurt and no one else was put in danger by their activities. The plaintiff *215 in Toussaint was a paper-pusher; he administered Blue Cross’s fleet of company cars. He was fired because people complained, inter alia, that his odometer readings were inaccurate. 292 N.W.2d at 897 n. 39, 902-03. Hardly weighty stuff. No one would have been endangered had Blue Cross engaged in more extensive investigation of these complaints before letting the axe fall.

Plaintiffs in our case, by contrast, were engaged in work so dangerous that a single slip could easily kill a co-worker or unleash an environmental catastrophe. In spite of this, Parker Drilling obtained no fewer than three eyewitness reports that plaintiffs were using drugs on the job before firing them. To wait any longer or look any closer would have been reckless; the dangers being what they were, the company had no responsible choice but to act decisively. By affirming the jury verdict against Parker, we are saying that management erred grievously by failing to send the employees back onto the oil rig after receiving three eyewitness reports that they were observed regularly abusing drugs on the job, and that the company must now pay hundreds of thousands of dollars for its mistake.

This strikes me as a result so preposterous it would be laughable if it were not so scary. Is this the type of decision we want to take out of the hands of management and give to a jury? Is it fair (or safe) to put company officials to a choice between risking an environmental catastrophe and a crushing jury verdict? It seems to me that the most we can reasonably ask of managers under these difficult circumstances is that they act responsibly and in good faith. Whatever may be the rule when dealing with employees who are not in safety-sensitive positions, it seems to me that substantial deference is owed to managers who discharge employees out of a legitimate, good-faith belief that they are acting to preserve life, limb, property and the environment. 16

The record here discloses in chilling detail that drug abuse on oil rigs was not an imaginary problem for Parker Drilling. In 1982, company officials investigated rumors of drug abuse on some of its platforms in Cook Inlet. Because no eyewitness would come forward to finger the drug users, the company had to resort to a surprise drug raid which included a drug-sniffing dog and unannounced searches of employees’ personal quarters. The searches, conducted in a single afternoon, yielded a small mountain of drugs; at least 10 drug stashes — ranging from marijuana and amphetamines to cocaine and LSD— were discovered. Yet Parker could do little to rid itself of the offending employees. While 14 employees were implicated, only two were fired; two others were issued warnings. The reason was simple: Because employees on the platform shared living quarters, it was virtually impossible to identify the owner of a particular drug stash. 17

If we are going to read the rule of Tous-saint into Alaska law, therefore, we should temper it by including a public policy exception that takes into account the practical realities of running a dangerous business. 18 *216 The fact is, misconduct is especially difficult to prove in such cases. Drug abuse is illegal, as well as contrary to company policy, so employees who use drugs will tend to deny accusations leveled against them. Additionally, the tremendous hazards posed to co-workers by drug use on an oil rig mean that employees using drugs will tend to do so surreptitiously, attempting to conceal their dangerous habit from those they work with. We should not require company officials to catch employees with a lit joint hanging from their lips before they can dismiss them for misconduct. See Note, Employer Opportunism and the Need for a Just Cause Standard, supra p. 212 n. 11, at 525-29.

Failure to recognize and allow for these difficult problems of proof will bring us again and again to a situation where an employer has very compelling reasons for believing employees are engaged in on-the-job drug abuse, yet the employees manage to prevail before a jury on the basis of paper-thin evidence that the employer’s suspicions were unfounded. Here, for example, plaintiffs relied entirely on their own self-serving denials, coupled with the testimony of three rig employees who never personally observed them smoking pot on the job. See Majority at 198-99. Is management required to have a signed confession, or an eyewitness report from every co-worker before it can feel confident in discharging drug abusing employees? To ask that management wait until it has more proof than Parker had against Cantu, Howard and Sanders before pulling an employee off an oil rig — or a nuclear power plant, for that matter — is to beckon disaster.

B. But the cause of public safety is not the only one that takes a beating as a result of today’s ruling; so does that of employee privacy. We are presented here with that admirable but all too rare a creature: an employer who took seriously its responsibility to assure safety in the workplace, but who approached the task in a responsible and non-hysterical fashion, cognizant of the need to be fair to its employees and respectful of their privacy. While Parker Drilling banned all drugs and alcohol from its rigs, it did not attempt to control its employees’ activities when they were on leave, as it could well have done under Alaska law. 19

*217 Even as to on-duty employees, Parker did not conduct wholesale drug testing or routine patrols with drug-sniffing dogs; rather, it acted primarily in response to reported violations. Employees found with drugs in their rooms, even in their dresser drawers, were not automatically fired; the company recognized that the drugs could have been stashed there by a roommate. When the company learned that plaintiffs were using drugs, the decision whether to take disciplinary action was not left to their immediate supervisors who might have been swayed by personal animus; the task was assumed by top management personnel, who everyone agrees were strangers to the accused employees prior to the incident. These high-level managers had received two eyewitness reports incriminating plaintiffs, yet, in an abundance of caution and fairness, they sought further corroboration. Only after obtaining a third eyewitness verification did management discharge plaintiffs.

If one were writing a manual describing how an ideal employer ought to behave under these difficult circumstances, one could hardly do better than to list the steps taken by Parker Drilling in this case. Employers are now free under Alaska and federal law to adopt far more oppressive measures to discourage substance abuse by employees who operate oil rigs and other dangerous instrumentalities. One nevertheless hopes that employers will not take full advantage of that freedom, but will approach the task of assuring safety in a more restrained and responsible fashion, much as Parker Drilling did here.

Our hopes may be in vain. Decisions like the majority’s have already encouraged employers to abandon less intrusive means of ensuring a drug-free workplace. Supervisors are becoming increasingly unwilling to risk mistakenly singling out workers for “for-eause” drug-testing. See Motorola Aims High, So Motorolans Won’t Be Getting High, Wall St. J., June 26, 1990, at A17, col. 3, 4. So what do employers do? They institute mandatory drug testing for all employees — even in the absence of safety concerns — seriously invading the privacy of all their employees. Id. For the right to have a jury second-guess management’s termination decisions, my colleagues trade away not only the workers’ safety but their privacy and dignity as well. This, to me, is too high a price. 20

C. If today’s morality tale teaches anything, it is the wisdom of the aphorism: nice guys finish last. For its trouble, Parker Drilling is rewarded with a bill for $360,107, years of litigation and a truckload of attorney time sheets. The moral of this story will not be lost on other, similarly situated, employers.

Ideas have consequences and the ideas embodied in judicial opinions have very direct and immediate consequences. I suspect two things will happen. First, employers will be discouraged from discharging employees suspected of drug use. Absent concrete evidence — a positive urinalysis test or a signed confession — employers may well feel constrained to send employees observed using drugs back to work. The clear lesson of this case is that, unless the employee is caught red-handed by someone with authority to fire him, he can always manufacture a triable issue of fact by finding a few co-workers who never saw him using drugs, inventing some threat, or whatnot. And if a jury buys into the story, the courts will cheerfully uphold the award, no matter how little sense it makes in light of the record as a whole. It is difficult to say how many drug abusers will be permitted to remain on the job by litigation-timid *218 managers, but there will surely be some. 21 Consider, for example, how gun-shy Parker Drilling officials might be if tomorrow they were to hear a reliable report that a particular employee on one of their oil platforms in Cook Inlet had been observed smoking a joint. Knowing that drug use on an oil platform at sea creates substantial risks to public safety and the environment, do we really want to encourage Parker and similarly situated employers to send suspected drug abusers back to work until they get caught lighting up?

The second and more direct effect of today’s decision is to tell employers they had better adopt draconian measures to screen employees for drug abuse. What Alaska employer, on reading this decision, would continue to ban drugs only in the workplace, as Parker did? Having received the green light from the Supreme Court of Alaska to extend this ban to employees between shifts and on leave, and having seen the trouble Parker got into by having failed to do so, see n. 19 supra, an employer would be foolish to impose a drug and alcohol ban less broad than the law allows. And, what employer would be dumb enough to rely on mere eyewitness testimony in any future discharges? Even if an appellate court eventually vindicates the employer, the litigation expense makes the process uneconomical. Instead, companies involved in safety-sensitive activities— those that operate planes, trains, boats, manufacturing plants and thousands of other hazardous activities — will have to seriously consider instituting procedures that routinely and indiscriminately trample the privacy of their employees. Invasive and unpleasant though they may be, these methods are effectively jury-proof. The judgment we affirm today will thus be paid for dearly by thousands of employees engaged in high risk enterprises, whose employers will steer clear of the mistakes made by Parker Drilling. Urinalysis machines, drug-smffmg dogs, surprise locker inspections and whole body searches will become the way of life in the American workplace. Is this really the world we want to live in?

Conclusion

The jury system is a wonderful thing. Born of the English common law, it stands as a tribute to democratic government and a bulwark against tyranny. But juries can function effectively only under close judicial supervision. Jurors are people and people make mistakes. Total deference to jury verdicts can occasionally be as oppressive as the denial of the right to jury trial. The jury in this case reached a verdict that has no support in the record; rather than overturn the verdict, though, the majority chooses to sidestep the law of the case and to allow plaintiffs to prevail on the strength of an argument they never raised.

The court has a responsibility to overturn a verdict not supported by the record. That responsibility is particularly grave in a case such as this where timely and fundamental issues of public policy are at stake: on the one hand, the need for safety; on the other, the responsibility to avoid gratuitous invasions of personal privacy. Because I fear my colleagues today have failed in this responsibility by upholding a verdict that is clearly erroneous and by ignoring the likely consequences of this action, I must respectfully part company with them.

. Parker is not concerned with drug use by employees while they are off the rig; only drug use on the rigs, or immediately prior to transportation thereto, is prohibited.

. The concurrence suggests that Chapman never directly confirmed that he had seen the plaintiffs use drugs and that Chapman moved out of the room he shared with Howard and Cantu because it was “too noisy” and "there was too much traffic.” Concurrence at 201 n. 11. However, the quote reproduced in text shows that Chapman moved out because he feared that his former roommates' on-the-job drug use would get them into trouble. Later, Chapman signed an affidavit to the same effect: “In [the] course of my duties I roomed and worked with Rhodney Cantu, George Howard, and Bowen Sanders. On more than one occasion I saw all three of these people smoking marijuana while on the job and/or while at camp. I specifically warned each of them that they should knock it off and what the consequences would be. They chose not to listen.” RT 3:27-28. Chapman so testified at plaintiffs’ unemployment compensation hearing. At that time he also confessed having joined them in smoking marijuana. In so doing, he risked termination and did, in fact, suffer a severe reprimand.

. The majority suggests at one point that the jury was instructed to determine whether plaintiffs had actually used drugs on the rig. See Majority at 193. My colleagues are mistaken. See pp. 207-09 & n. 4 infra.

. Plaintiffs made the same argument when they opposed the company’s motion for judgment notwithstanding the verdict or, in the alternative, for a new trial on the ground that instruction number seven placed too heavy a burden on the company. Plaintiffs pointed out in their papers that “the complained of instruction did exactly what defendant contends it should have — it put the focus of the inquiry on the employer’s subjective motivation and good faith, rather than on the objective evidence supporting the conclusion that the employees committed the offense.” CR 108, at 22 (emphasis added). At oral argument on the motion for JNOV, plaintiffs’ counsel reiterated that the focus of the instruction was "that the jury find that there was a good faith belief on the part of the defendants that what they were doing was correct,” and that there was "ample evidence” to show "there was a lack of good faith.” RT 7:12-13.

. The concurrence's attempt to portray Parker’s belief as erroneous, see concurrence at 200 n. 11, is beside the point. It does not dispute the key fact of this case — that at the time Parker acted it had reason to believe and did believe that the plaintiffs were using drugs on the job. The concurrence instead relies on information that Parker did not have when it had to decide whether to send the men back on the rig. Because the information was not available to Parker at the time it was required to make the decision, it is irrelevant to whether Parker acted in good faith.

Contrary to the concurrence’s assertion, no rational jury could have found that Parker lacked good faith. Although the way the majority has resolved the case makes the existence of good faith irrelevant, the evidence of lack of good faith was in fact insufficient. The key element missing from plaintiffs’ case is the slightest hint that Parker acted out of an improper motive; indeed, it would have been exceedingly difficult to find one as Mosely, McCar-rell and Upchurch, the managers who made the decision to discharge, did not even know the plaintiffs. The failure to allege an improper motive alone should have kept this case from the jury. See p. 211 n. 10 infra.

Plaintiffs utterly failed to prove the elements of their disparate treatment claim, that Parker denied them the benefit of procedures it made available to other employees by agreement, policy or practice. See Jury Instruction No. 9 (disparate treatment claim requires showing of practice or procedure denied to plaintiff but given to others); see also Arco Alaska v. Akers, 753 P.2d 1150, 1155 (Alaska 1988). Contrary to plaintiffs’ contentions, there was no policy of searching employees’ rooms or conducting urinalyses before discharging them for substance abuse: In 1981 or 1982, a high ranking employee was terminated when a single eyewitness reported having seen him drink alcohol on the rig. RT 4:81, 5:41. Another employee was fired on the spot when he climbed off a helicopter onto an oil platform and his supervisor suspected that he had been drinking. In neither case was a urinalysis or a body or room search done. RT 4:81, 5:41. On two occasions employees were disciplined or dismissed for drinking; there was no indication of any search or urinalysis. Plaintiffs' Exhibit 16 at 9; id. at 15. The only room searches in the record were part of a massive drug raid conducted on July 13, 1982, which yielded a large quantity of drugs. See p. 215 infra. Although the raid implicated fourteen employees, only two were discharged. The reason? In general, identifying the owner of the stash was impossible because the employees shared rooms and there were no eyewitnesses. RT 2:225-36. Of the eight non-raid related disciplinary actions in the record, at least two and perhaps as many as four were based on the statement of a single eyewitness. Here, Parker had two eyewitnesses. Parker nonetheless gave the plaintiffs the benefit of the doubt and waited until it got confirmation from a third eyewitness. If anything, plaintiffs were treated better than other employees accused of substance abuse.

The implausible threat to blackball the plaintiffs if they contested their discharge — which of *209 course management vehemently denied, RT 3:90-92, 3:178-82, 5:70 — also fails to raise an inference of bad faith discharge. Parker actually preserved plaintiffs’ ability to find future employment by listing the cause for discharge as “violation of company policy” rather than branding plaintiffs as drug users. Plaintiffs' Exhibits Q, AC, AK; see RT 2:215-16, 3:96-97, 3:191-92, 4:91. And plaintiffs did in fact find comparable employment. RT 1:96, 2:41, 2:43-44, 2:152-53. In any event, a threat to blackball employees, even if shameful or actionable in its own right, is irrelevant to whether the decision to discharge was wrongful. Threatening employees after dismissing them to persuade them not to contest their discharge does not taint an otherwise proper termination. Cf. Funk v. Sperry Corp., 842 F.2d 1129, 1132-33 (9th Cir.1988) (where “layoff was otherwise fully consistent" with company policy, claim of post-discharge scheme to conceal employer’s failure to conduct contractually required review of the layoff was insufficient to establish a "plot to disguise the true motive for ... termination”).

.The jury found in plaintiffs’ favor on two theories, absence of just cause and breach of the implied covenant of good faith and fair dealing. The majority’s implicit acknowledgment that the record evinced no lack of good faith by Parker renders the verdict clearly erroneous on both counts. As explained above, the instruction on just cause demanded a verdict for Parker if the jury determined that company officials reasonably believed plaintiffs had smoked marijuana on the rig. The instruction on the good faith issue stated the same proposition even more clearly: “If you find that Parker honestly believed that termination was for a legitimate business reason, you must find for Parker on this issue, even if you also find that Parker was mistaken in that belief.” CR 94a, Instruction No. 9.

. Nor can the majority plausibly argue that the two definitions of just cause somehow merge— for example, that the jury, in concluding that plaintiffs did not in fact use drugs, determined a fortiori that Parker failed to act in good faith. The jury was told quite explicitly that Parker could have acted in good faith even if it was wrong about plaintiffs’ drug use. See n. 6 supra. Thus, whether plaintiffs had or had not smoked marijuana was not directly relevant to the jury’s deliberations; all that mattered was whether company officials reasonably believed they had. Plaintiffs have consistently conceded that this was the issue. See pp. 207-09 & n. 4 supra.

. In holding that the jury was properly entitled to answer a different question than that posed in the district court’s instructions, the majority appears to invoke sub silentio a plain error exception to Rule 51. Our circuit is unique, though, in holding that there is no plain error exception to Rule 51. See, e.g., Brett v. Hotel Employees and Bartenders Union, Local 879, 828 F.2d 1409 , 1414 n. 7 (9th Cir.1987); Bock v. United States, 375 F.2d 479, 480 (9th Cir.1967); *210 Bertrand v. Southern Pacific Co., 282 F.2d 569, 572 (9th Cir.1960). The leading treatise on federal procedure explains the wisdom of our practice. 9 C. Wright & A. Miller, Federal Practice and Procedure § 2558, at 674-75 (1971).

. Even the majority concedes that, although Rutledge cited Toussaint , it did so “on other grounds." Majority at 194.

. The employer’s motive for discharging the employee is the central jury question in virtually every wrongful termination case. Indeed, every Alaska wrongful termination case I’ve come across has involved a claim that the discharge stemmed from an improper motive, not that the employee did not actually commit the suspected offense. See, e.g., Arco Alaska, Inc. v. Akers, 753 P.2d 1150, 1154-55 (Alaska 1988) (alleging personal animosity); Klondike Indus. v. Gibson, 741 P.2d 1161, 1164, 1167-68 (Alaska 1987) (alleging desire to avoid paying contractual bonus); Rutledge, 727 P.2d at 1056 (alleging reverse race discrimination); S.B. Mitford v. de Losala, 666 P.2d 1000, 1007 (Alaska 1983) (alleging desire to prevent sharing future profits).

Plaintiffs here failed to allege any ulterior motive. Their case rested entirely on “objective evidence” from which they hoped the jury would infer some kind of wrongful motive. See pp. 207-09 & n. 4 supra. As I read the law, their failure to suggest even the slightest ulterior motive should have precluded them from reaching the jury at all.

. In a spirited display of hyperbolic exuberance, my concurring colleague suggests this approach would lead to "a significant reversal of our national labor policy.” Concurrence at 196. These concerns are wide of the mark. We are not dealing with rights explicitly spelled out in a collective bargaining agreement which carries with it mutually agreed-upon procedures for resolving labor disputes. Rather, we deal here with judicially created rights and remedies which by their nature require judicial enforcement. There is a significant difference between the two.

Collective bargaining agreements, negotiated by union representatives for the benefit of all employees, provide for swift and inexpensive resolution of disputes such as these by way of grievance and arbitration. Moreover, consistent with their duty of fair representation, union officials must promote the interests of all employees, including those whose life and limb might be jeopardized by the misconduct of their colleagues. Judicial proceedings, by contrast, are expensive and protracted, with litigation costs amounting to 80% of the average total recovery in settled cases and 130% in cases that go to trial. See Note, Employer Opportunism and the Need for a Just Cause Standard, 103 Harv.L.Rev. 510, 525-28 & nn. 70-71 (1989). They require the courts to second-guess management decisions made years earlier; they largely ignore the legitimate interests of others, including those of fellow workers who may be affected by the decision.

The well-defined body of labor law to which my colleague refers would not be affected in the least by a contrary decision today. To say it is very bad policy to have twelve randomly selected men and women second-guess good-faith, honest judgments made by management personnel acting out of genuine concern for worker safety, is not to trumpet a "counter-revolution against worker’s rights." Concurrence at 8443. I certainly don’t suggest that courts should refuse to apply the terms of a written contract, such as a collective bargaining agreement, which calls for review of an employer’s decision to remove employees for cause. Parties should be free to bargain about the terms of employment and courts should enforce those contracts, regardless of their perception of what is good policy. See, e.g., United Food & Comm’l Workers Union v. Lucky Stores, Inc., 806 F.2d 1385, 1386 (9th Cir.1986) (“Wise or not, a deal is a deal.’’) However, when the contract is unwritten, and often unspoken, the court must fill in the terms. In so doing, it is entirely appropriate to consider the public policy implications of particular rules of decision, including the effect upon safety in the workplace. See Pugh v. See's Candies, Inc., 116 Cal.App.3d 311, 330 , 171 Cal.Rptr. 917, 928 (1981) (just cause as an implied contractual term differs greatly from the standard applicable to contracts for a specified term of employment; where the employee occupies a sensitive position the employer must be allowed to exercise her subjective judgment). There is nothing at all revolutionary about this approach. See Rulon-Miller v. IBM, 162 Cal.App.3d 241, 253 , 208 Cal.Rptr. 524 (1984) (”[A]n employer who acts in good faith on an honest but mistaken belief that discharge of an employee is required by legitimate business interests has not committed a wrongful discharge....”); Khanna v. Microdata Corp., 170 Cal.App.3d 250, 263 , 215 Cal.Rptr. 860, 868 (1985) (suggesting that termination due to “dissatisfaction with services” is the same as “good faith”). No less an advocate of employee rights than Justice Joseph Grodin adopted this position: Just cause, he wrote, connotes “a fair and honest cause or reason, regulated by good faith on the part of the party exercising the power.... [Where] the employee occupies a sensitive.. .position, the employer must of necessity be allowed substantial scope for the exercise of subjective judgment.” Pugh, 116 Cal.App.3d at 330 , 171 Cal.Rptr. at 928 (internal quotes and citations omitted; emphasis added). This is precisely the view of the law adopted by both parties and the district court below. See pp. 207-209 supra.

My distinguished colleague’s concern for the rights of working men and women is well known and entirely justified. We do not disagree as to ends, only as to means. In my view, the interest of everyone involved would have been better served had this dispute arisen in the context of a collective bargaining agreement, which would have provided an effective mechanism for dealing with the issues presented. I fear, however, that decisions such as these ill-serve the cause of voluntary unionization, shifting to the courts an increasing number of labor disputes that have traditionally been handled by union-sponsored grievance procedures. While it is difficult to judge such matters, I suspect that a significant cause of the recent trend away from union membership may be the availability of judicial remedies that give employees the same — and sometimes superior — rights as those available under a collective bargaining agreement.

. Forty-five train accidents in eight years have been directly attributed to employees under the influence of drugs or alcohol. Skinner, 109 S.Ct. at 1408 . No less alarming are the statistics involving ships and planes. Drug or alcohol abuse has been linked to no fewer than 50 commercial shipping accidents in a five-year period. L.A. Daily Journal, May 17, 1989, at 6, col. 3. And, in Florida alone, more than 6400 pilots failed to disclose drug and alcohol-related arrests when applying for their licenses. Id.

. Also, at least 70 crewmen were killed when a Chinese rig toppled into the Bo Hai Gulf in 1979. UPI, Feb. 16, 1982. In this country, few have forgotten the disastrous 1969 blow-out of a rig off the California coast, releasing 2 million gallons of oil into the Santa Barbara channel. N.Y. Times, Oct. 27, 1985, at L26, col. 1; AP, Apr. 23, 1977.

.I am aware that the oil rig here was on Alaska’s North Slope, not at sea. But the standards we set today will apply to all oil rigs, whether on land or sea, and to all other dangerous instrumentalities. Parker Drilling, for example, has crews working oil platforms in Cook Inlet.

. The concurrence does much breast-beating about how these plaintiffs are innocent, going so far as to suggest that, were we to uphold their dismissal, we would next be bound to let the government imprison people at will. Concurrence at 202 & n. 13. The fact is, we don’t know whether plaintiffs are guilty. Contrary to my colleague’s assertion, id. at 202, the jury was not instructed to base its verdict on whether plaintiffs were guilty of misconduct but on whether the defendant acted in good faith in dismissing them. The whole case was tried on the theory that plaintiffs were treated disparately from other employees and both sides admitted that the case did not turn on whether the plaintiffs in fact smoked marijuana but on Parker's good faith. See pp. 207-09 & n. 4 supra.

. The good faith standard is not, as the concurrence would have it, “an almost meaningless concept.” Concurrence at 196. It protects the employee from being dismissed based on personal animosity, ill will and any other ulterior motive; and it requires the employer to conduct a reasonable investigation of the alleged safety violation.

. The undisputed testimony of John Haynes, who conducted the raid, explains the difficulties of disciplining employees whom eyewitnesses have not identified as drug users:

RT 2:227-28. The proof problems were so severe that one employee who was initially discharged was able to convince the company to reduce his punishment to a warning, even though four drug stashes were discovered in his room — in a plastic bag in his closet, in a film canister, in a chewing tobacco container and in bottles underneath his bed. RT 2:233-35.

.Raising the specter of judicial activism — a concept he no doubt finds anathema — my concurring colleague suggests that a narrow public policy exception for situations involving severe *216 risk to life, property or the environment would amount to "judicial legislation.” Concurrence at 8452. What he fails to appreciate is that this entire area of Alaskan law is judge-made; case-law, not statutes, set forth all the governing principles. To the extent the courts have established the applicable rules, it is up to the courts to also announce the limiting principles. It is a well-established tradition of our jurisprudence that, in the absence of statute or custom, we must decide as if we were legislators. B. Cardozo, The Judicial Process 140 (1921); see also O.W. Holmes, The Common Law 35-36 (1881). In so doing however, we should not let sound judgment yield to "spasmodic sentiments.” Cardozo, supra, at 141. Call it common law or common sense, if there is a judicially created rule that allows juries to second guess all employee terminations, there ought to be a judicially created exception for situations where the employer moves quickly and in good faith to ensure the safety of its employees.

Workers may not all agree with the concurrence that “job security is the most fundamental employment right," concurrence at 197; some may consider the right to a safe work environment at least as important. Where there is a collective bargaining agreement, we need not worry about the employees' preferences: Their needs and those of management will be reflected in the negotiated agreement; we can trust the parties to know what is in their interest and how to achieve it. However, when we are called upon to fill in the terms of a contract never reduced to writing, we must account for the interests of all those who would have input into and be affected by it. See n. 11 supra. Much as my concurring colleague would have it be otherwise, the more difficult we make it for employers to get rid of employees whom they reasonably and honestly believe are using drugs on the job, the more likely it is that employees who in fact use drugs will be sent back to safety-sensitive positions. The inattentiveness of a drug-impaired worker could convert Judge Reinhardt's rhetorical "off with the worker's head,” concurrence at 202, into gruesome reality-

. Parker certainly paid the price for its respect of personal privacy. As explained above, see p. 206 supra, plaintiffs were on leave when the company first learned of the drug allegations. Because company policy did not prohibit drug use while on leave, drug testing the employees as they were returning to the rig would have been pointless. Parker thus faced a hard choice between sending the men back to the rig and risking a disaster, and discharging them and risking a jury verdict. Clearly, Parker would have been better off prohibiting all drug use, on and off the job, and testing plaintiffs as soon as suspicion arose. For failing to adopt this draconian solution, Parker is now left holding the dirty end of the stick.

. Judge Trott suggests that random drug testing might be less intrusive than investigating individuals about whom there is suspicion. Majority at 195 n. 3. I can only wonder what kind of nation we would live in if the Framers had concluded that random invasions of our privacy were less troublesome than those based on probable cause or reasonable suspicion. Judge Reinhardt’s apparent agreement with me on this point, see concurrence at 196 n. 1, is cold comfort. Railing about the need for "a greater sense of commitment to the fundamental constitutional principles,” concurrence at 196 n. 1, does little to help the cause of worker privacy which, as a consequence of today’s decision, suffers another body blow.

. Of course, even those employers who are reluctant to fire employees who endanger themselves and others may find themselves in court. By not discharging an employee he reasonably believes is using drugs, the employer may be opening himself to criminal sanctions, see People v. Hegedus, 432 Mich. 598 , 443 N.W.2d 127 (1989); People v. Chicago Magnet Wire Corp., 126 Ill.2d 356 , 128 Ill.Dec. 517 , 534 N.E.2d 962 (1989), cert. denied sub nom. Asta v. Illinois, U.S.-, 110 S.Ct. 52 , 107 L.Ed.2d 21 (1989), not to mention wrongful death lawsuits, if that employee makes a mistake that gets someone killed.