There is now a long "dishonor roll" in science. Contemporary physicists are notably absent. Why? Is misconduct less common in physics than in other sciences? Or are there skeletons waiting in the closet?

One way to make a start at answering such questions is to ask physicists. I asked R, P, B, and S (as I shall call them). What follows is some of what I was told.

R (whose motto is "Do it right") suggested that in high energy physics, teamwork makes the difference. Very few experiments are done only by one or a couple of persons these days. There would have to be a huge conspiracy to pull off real fraud. People would not keep their mouths shut. The temptation would be greater if you were a solitary person, with no one looking over your shoulder. True, physics is very competitive, but physicists check each other. "People are interested in turning out the correct results. If it is not correct, it is garbage."

According to P (a physics department chairman), experiments are getting increasingly automated. It is not uncommon to have data taken totally by computer. That, he thinks, is part of the difference between physics and fields such as biology. But there are other differences: "As I understand it, in physics the questions are: Under what controlled conditions are you taking the data? How accurately do you make the measurements? What are the sources of random error? What is the signal to noise ratio? and so on. In biology, they don't seem to worry about things like that at all . . . The important thing is whether you have an effect or not. If you see it, it tends to knock you over the head."

Another thing differentiating physics from other fields, according to P, is that physics is seeking fundamental knowledge about simple objects. Physicists try to analyze the simplest of all possible situations, using the simplest of materials. Then they go on to study complexities. This means that the experiments are easily repeatable: "When one lab made high temperature superconductors and described the procedure, hundreds of people could do the same thing and did it overnight. The minute these people published, the materials were accessible to everybody. So-if they hadn't been absolutely careful in their description-credit, Nobel prizes, fame, fortune, and funding would all have gone out the window. They wanted to be very careful before they announced the result."

P also suggested that reproducibility is easy because physics is a much more unified discipline than chemistry or biology. In physics, everybody has the same fundamental training.

How then do you learn "good science"? According to P, "We try to teach our students to estimate things within an order of magnitude so that they have a feeling for when things are going astray. Then, when they see something unusual, it has to be something that doesn't go away, something that is perfect; no matter how you look at it and try to eliminate it as an extraneous effect, it stays. It is not sufficient just to see something unusual, it happens all the time. . .You'd better do everything, because you'd better be right. My thesis advisor once told me: 'Be sure you are right whenever you publish something, because other people will soon forget, but you will always remember."'

R tries to teach his students by example. "They grow up in a certain environment, like my children." Asked whether he could formulate any guidelines for good physics, he simply said: Do it right!

The biophysicist B has a darker view of the situation: "There is fraud in physics, but the physicists respond to it in a more clubby manner." B offers as an example Joseph Weber's experiment with gravitational waves (19613-1975). Weber had a gravitational detector at Princeton and was doing time series analysis. He claimed that he was seeing pulses in his computer data. An independent experiment was made at Argonne, where they didn't find any pulses. Weber had the computer tapes sent to him and claimed he found the same pulses at the expected places in the Argonne tapes. The story I've heard is that-unknown to Weber-the clocks at Princeton and Argonne were set at, respectively, Eastern standard and Greenwich mean time... So he proved himself to be a fraud."

"Fraud?" I asked, "You mean self deception?"

B replied: "That is the way the physics community would choose to interpret it. Whereas the biological community would say: the guy is a fraud! the physicists say: over enthusiastic interpretation."

P also mentioned Weber's gravitational waves, adding that Weber himself has never changed his mind. R also knew the story. Physicists seem to have only a few good stories.

From B's point of view, non reproducibility is a fact of life in science, but it is not a problem as long as you have followed good laboratory practice: "There comes a point in science where there is no percentage in trying to correct something. All you do is try to measure it to the best of your ability and don't worry about why it differs from somebody else's. This is one of the mistakes that Feder and Stewart make. They pick on these little details. They are interesting, but they are not going to lead to anything useful. Whether it is 2.9 or 3.1, does this have any bearing on the principal points of the science? If [not], you should make sure that things are calibrated correctly, that proper care is being taken, and that things are reproducible in your own hands. What is central in science is that you use good laboratory practice..., and that you find an observation that is self-consistent in your hands. Then you proceed. It is usually no benefit to go and pursue why [someone else] found a near but noticeably different number than yourself."

What about cases of fudging data? I brought up Millikan s oil-drop experiments (1910-13) and the fact that Millikan stated in a publication that his result for the charge of the electron was based on the average of all the oil drops over a period of time when he had in fact omitted some bad readings. R said that of course it is a lie to say something like that, adding that it also happens nowadays: "'This is the average over the entire period.' That is a lie, but people do make statements like that. 1 don't think it is a terrible crime."

P's reaction to Millikan's claim that he had included all the oil drops was similar: "That is a misleading, possibly even a false statement, but I wouldn't say it is fraudulent. Things can go wrong with experiments, and sometimes you know some readings are not good but you don't know why. That was probably the case with Millikan. Subsequently, his experiment has been repeated and automated at Argonne by Ray Hagstrom. Hagstrom didn't find a single droplet that was mysterious, not a single deviation from the unit charge."

Physicists, it seems, are not so interested in how you came by your result, if later experiments confirm it. This is important for whistleblowing. P said that if you are going to call a scientist a cheat and liar, you challenge his most basic reason for existence. P's experience is that you'd better be right, and that in two ways. You have to show not only that the conclusion is unwarranted based on the data, but also that the conclusion is wrong: "Because if you simply accuse people of throwing away bad data, or of improving the statistics a little bit, or not taking into account systematic errors, etc. and the results are ultimately published as a number, if that number holds up in the future, then no matter how the person came up with the conclusion, he isn't going to look that bad .. .. On the other hand, if it is a straightforward experiment and someone in the future gets it to disagree by a substantial amount, then he will look bad."

But the skeptical physicist S thinks that physicists are unduly pleased with the present situation. He says he knows of thirty cases where teams have been wrong. The problem is that everybody trusts one another, while no one has access to the raw data. At different levels of analysis, important information tends to get lost. So, for S, physicists may well be doing it right, and still be wrong.