When ABC sent undercover reporters to infiltrate Food Lion stores in the Carolinas in 1992, it set off a rousing debate over the ethics of investigative journalism. Now it looks like the debate missed the main point.

With lipstick-size video cameras concealed in their hair and using faked resumes to get Food Lion jobs, the investigators captured sounds and images of what appeared to be stomach-turning food-handling practices by the authentic employees.

For journalism professors, the case was a windfall. I have shown that Prime Time Live segment to every media ethics class that I have taught since 1993. The debate -- in my class, in textbooks and in other commentaries on media morality since then -- was over the appropriateness of undercover reporting. Is it moral to conceal one's identity in order to nail the bad guy?

The discussion typically takes the form of a cost-benefit analysis, what philosophers call "utilitarianism." Did the good in exposing the risk to public health outweigh the harm to those who were unknowingly providing the evidence?

Viewed that way, it's pretty easy to justify intrusive reporting methods. Even the objection that the information could have been collected in other ways, such as through the testimony of former employees, is arguable. The hidden cameras produce entertainment value, and TV needs to be entertaining to get our attention. While witnesses might lie, the camera surely cannot.

Well, maybe it can. An article in the current issue of The IRE Journal, published by the Missouri-based organization of elite investigative reporters and editors, invites us to consider the possibility.

Written by Sandra Davidson of the University of Missouri School of Journalism, the article points to a different ethical issue: lying through selective editing. After viewing some of the footage ABC didn't use -- among the 45 hours' worth obtained by Food Lion through discovery procedures -- Davidson concluded that these images provide some support for the grocery chain's accusations of wrongdoing by ABC. And they do so, she says, "with a power that is difficult for the printed word to match."

ABC News spokeswoman Eileen Murphy has promised a response to Davidson and her cited Food Lion materials. I have seen Food Lion's selection of footage, and its effect is indeed powerful.

For example, the original broadcast showed a worker's feet sliding on the meat department's floor while a voice-over blamed a greasy surface. A scene from the outtake offers an alternate explanation: The employee was washing equipment, and the source of the slipperiness was spilled soapy water.

A conversation that in the broadcast seemed to record a Food Lion employee complaining about having to prepare outdated chicken for sale reveals something else when editing cuts are restored. She had asked her supervisor for permission to discard the spoiled chicken. "He told me to dump it, go back and get fresh chicken . . ." and, she adds, she learned that no prior approval for getting rid of spoiled food was needed.

Not all of the broadcast's accusations were refuted in this way, but the ability of picture selection to support a preconception was clearly demonstrated -- whether the selection is ABC's in the original broadcast or Food Lion's in its editing of outtakes.

The main problem with undercover investigations is not the invasion of privacy or faking credentials. The real problem is that undercover observers find it very difficult, if not impossible, to go to work with open minds. Their projects involve so much of their company's money, not to mention their own sweat, that almost anything would be better than returning empty-handed.

Such a strong bias could find incriminating evidence anywhere, even in newsrooms. Conversations there tend to be energetic and irreverent. I know that any 45 hours of secret videotaping could provide enough images and sound bites for splicing to ruin the reputation of the most respected among us. Think about your own workplace, and you will get the idea.

Carl Sagan, in his plea for scientific thinking, The Demon-Haunted World, argued that all undercover observation has a built-in bias for perceiving evil, especially if the observers are paid. The witch hysteria in medieval Europe was one of his examples.

But he also cited a case of paid undercover postal inspectors in Cleveland who fabricated criminal evidence against 32 innocent colleagues. That was in 1994.

Several things could have been done to make the Food Lion investigation more scientific. One would have been to send crews to the stores of different companies without telling them which one was suspected of unsanitary practices. That would at least have indicated whether Food Lion was abnormal by industry standards.

Another would have been to train the reporters to be passive and neutral observers, as eager to record the good as the bad. In the ABC outtakes, you can hear someone vocalizing disappointment with soft expletives when an anticipated wrongdoing or damaging statement fails to materialize. At another point, someone murmurs, ". . . get these guys."

Equating journalism with science might sound pretentious, but journalism should not be shy about adopting the well-tested rules of scientific method that are designed to counter our human tendencies to fool ourselves with prejudice and wishful thinking. There is a small but growing band of reporters using some of the tools of science, including statistical sampling and analysis, who buy into that model.

But they are running against the tide of the moment, trying to push journalism toward science while the rest of the trade is pushing it toward art and entertainment. Rather than argue about hidden cameras, let's debate that one.