People are often troubled when the government and universities solicit confidential or anonymous complaints; the William & Mary Bias Reporting system controversy is the latest example. I think such systems are principle necessary and legitimate, so long as they are properly run. True, there's a risk that they'll be abused, but that's true of most enforcement systems -- the risk alone should not lead us to reject them, though it may lead us to look careful about how they actually operate.

1. Anonymous Speech: To begin with, note the similarity between anonymous speech in public (such as on blogs) and anonymous speech to the government. Requiring that people reveal their identities may deter them from saying things that will open them to retaliation. Allowing anonymous speech will diminish this "chilling effect." It will encourage people to blow the whistle on what they see as misconduct, whether they're revealing (a) government misconduct to the public, (b) nongovernment misconduct (say, by their corporate employers) to the public, (c) government officials' misconduct to the government (for instance, if they are complaining about alleged misbehavior or poor performance by their public university professors), or (d) nongovernment actors' misconduct to the government.

At the same time, anonymity encourages falsehoods as well as accurate whistleblowing. This is why we as readers should be extra cautious about charges levied by anonymous speakers (though we should be cautious about charges levied by named speakers, too). Likewise, government officials should be extra cautious about charges levied by anonymous complainers.

2. Confidential Speech: The William & Mary system seems to be encouraging confidential complaints -- which is to say complaints for which the complainers' identity will be shielded from the target, and even from many of the investigators -- rather than purely anonymous complaints. (Of course, one can always submit even purely anonymous complaints to any agency, and William & Mary seems likely to consider them; but its system, which requires a William & Mary user name and password, is intended to solicit confidential complaints, not anonymous ones.)

The effects, both good and bad, of confidential speech seem likely to be moderate versions of the effects of anonymous speech: The system encourages both accurate and false complaints more than a your-name-will-be-revealed system but less than a purely anonymous system. (The system could discourage false complaints more if William & Mary takes aggressive steps to investigate whether a complaint is not just unproven or mistaken but a deliberate lie, punishes the liars, and publicizes such actions; but I doubt that William & Mary is likely to do this. Unfortunately, such a system would also discourage true complaints to some extent, though if the system is trusted, it will discourage false complaints more than true ones.)

3. Anonymous/Confidential Speech Systems in Practice: Because of this, many institutions routinely rely on anonymous or confidential evaluations. Most universities, for instance, have student evaluations be anonymous. Likewise, most universities assure outside tenure reviewers that their names will be kept confidential from the person being reviewed, and while this assurance isn't ironclad -- there's always the risk of leakage, and of the rare subpoena.

Naturally, the anonymous evaluations are treated with some extra skepticism because of their anonymity. But the sense is that the university will get far fewer accurate criticisms of the professor if the students knew that their names will be revealed; and the risk of some extra inaccurate criticisms is seen as being worth running, especially given that the anonymous evaluations are read skeptically.

A few states, by the way, expressly require by statute that evaluations of government employees not be based on anonymous criticisms, and thus require that students sign the evaluations. I don't know of any empirical comparisons between such systems and the anonymous evaluation systems. Still, my tentative sense is that the anonymous evaluation systems are better, despite the slightly higher risk of unfairness to the person being evaluated. What do you think?

And note that these confidential and anonymous evaluations are directly used in deciding whether certain employees are to be promoted, hired, or even dismissed. My sense is that confidential and anonymous complaints are even more commonly used when it comes to looking into something, and seeking tangible or nonconfidential evidence.

4. Investigating Based on Anonymous/Confidential Speech: Let me elaborate a bit more on this, and bring in point 1 above. Say that someone e-mails you some anonymous allegation about a politician. You'd often be reluctant to believe the allegation, and act (even if it just means publicizing the allegation with your endorsement) without further investigation. But you'd often be quite willing to start that investigation based on the anonymous allegation, since you expect that the anonymous and possibly unreliable tip may lead you to uncover highly credible evidence.

The same is true, I think, with regard to anonymous (and especially nonanonymous but confidential) complaints about an employee or patron of your establishment. If you run a restaurant, and a patron tells you that a waiter has insulted him but insists that his name be kept confidential (maybe he's afraid the waiter will beat him up, or spit in his food if the patron visits again), what would you do? Would you tell the patron that due process prohibits your taking the complaint?

I take it you wouldn't; you probably wouldn't fire the waiter on the spot (unless you've heard a lot of such complaints, and have little cause to think they're part of a general campaign to unfairly malign the waiter), but you might watch the waiter more closely, or ask the waiter's colleagues if they'd heard the waiter insulting this customer or others. The university context is not quite the same as this, for various reasons; but in both situations, the employer may be acting wisely -- and properly -- in paying attention to confidential or anonymous complaints, and even in encouraging them.

The same, of course, is true of law enforcement. The police naturally prefer nonanonymous and nonconfidential complaints. But in many situations, the complainants are naturally worried about retaliation. To minimize the chilling effect on valuable speech (here, complaints to the police about likely criminals), the police should, to the extent possible, allow anonymous and confidential complaints. They generally shouldn't be admissible evidence at trial; but they can be useful tools for tracking down evidence that is admissible.

5. Severity: Finally, some commenters tried to distinguish systems for encouraging anonymous or confidential reports of possible terrorist planning, or even of street crime, on the grounds that those are serious crimes that justify such measures.

But the crimes also lead to severe penalties, and can involve extremely intrusive investigations. So while the upside of allowing anonymous or confidential complaints as to such crimes (the encouragement of accurate complaints) is greater than the upside of allowing such complaints as to "bias incidents" on campus, the downside (the encouragement of false complaints) is greater, too.

6. The Bottom Line: So, as I've said, the William & Mary policy should be faulted to the extent that it seems to allow the punishment of protected speech. Even to the extent that it aims to punish unprotected attacks, vandalism, or threats, it should generally reject the use of anonymous or confidential evidence in the actual quasi-judicial disciplinary proceeding. And campus authorities should be especially skeptical about anonymous or confidential complaints. But such complaints are a legitimate, and often necessary, part of enforcing campus rules -- or for that matter criminal laws -- just as anonymous speech is a legitimate, and often necessary, means of promoting public debate and whistleblowing.