contribution. Minor though they may appear, the cumulative effect of these departures from honest practice is corrosive for the spirit in which science is practiced. Referees can (and do) identify some errors of this kind, but there is also a general need for ordinary news reporting in these and related fields. Only by appreciating how widespread are some of these practices will members of the research profession come to dangers in the continuation of these practices, both for those who engage in them individually and for the research profession as a whole. But, in general (in my opinion), the scientific press has a responsibility not merely to draw research people's attention to of conduct but also to lead ethical issues

that should

departures from the norms discussion of the novel concern the research

profession. The past two decades seem to have bristled with a host of issues with which the subcommittee will be familiar, as follows:

Should research in genetic engineering be

regulated in the interests of public safety?

How should laboratory animals be used and cared

for?

What conflicts of interest arise when researchers seek, or are encouraged (by governments or their universities) to exploit their discoveries commercially, and how should such conflicts be resolved?

What duties of open disclosure do individual researchers owe their colleagues, even those who are competitors? (The convention that a researcher should be ready to share samples of his research materials or data arising from his research with legitimate enquirers seems to have been seriously eroded in the past few years.)

Do present methods of financial support for research conflict with the traditional procedures by which the integrity of research has been assured? (There is some reason to believe that the US system, by enhancing the reputation and power of successful principal investigators within their institutions, may have diminished the degree to which people discuss research in progress with others than their immediate colleagues.)

My general impression that these and other cognate issues are not sufficiently discussed in the scientific press. The result is that many researchers are less aware they should be of the need to be alert to the pitfalls that abound in the present competitive climate of research. That, I believe, is perhaps the most important challenge for the scientific press.

########

Mr. JOHNSTON. All right, thank you very much, Mr. Maddox. We will address some questions after the other two panelists have spoken.

Dr. Koshland.

TESTIMONY OF DR. DANIEL E. KOSHLAND, JR., EDITOR,

“SCIENCE,” WASHINGTON, DC

Dr. KOSHLAND. I first of all would like to echo my colleagues and say I thank you very much for this opportunity. I think an evenhanded examination of this very important issue can not only benefit both the scientific community and the public, which has so much to depend on in that category.

As I listened to this discussion today, I think one of the things that has intrigued me is that we have got to be very careful of distinguishing between wanting a perfect world and talking about legislation which is to deal with the worst imperfections of the system.

I think the analogy I might make between the scientific world life and to say 81 Congress is that you would clearly have laws and rules about bribery or conflict of interest. I think if you have laws about whether people should be totally meaningful whenever they get up and address the House or totally agreeing with a consensus, most of you would feel that is legislation you don't want.

So I think the situation we face is that we have on one extreme of 82 what we call misconduct, something like fraud, which we are all uniformly against and is extraordinarily detrimental to science, and then you go between sloppiness, honest error,83 poor scholarship, and you are dealing with matters which may best be dealt with within the scientific community, and the journals, I think, as John Maddox just pointed out, can be very helpful in the sloppiness, poor data, scholarship. We are not perfect, but peer review does help a lot with that.

In the case of fraud I think we are occasionally helpful, in the sense that, ironically, when a manuscript comes in that is too perfect it arouses suspicion, and frequently then a referee does pick up something or an illogical 84 or a figure that happens to have appeared once before and a reviewer picks it up.

But fraud usually is sufficiently clever that the way it is picked up by journals is by some other scientist duplicating the work and then calling 85 into question-most often, rather interestingly enough, most scientists will then call the original authors and most authors really don't want to be wrong. They hate to retract, but they even hate worse to have somebody else point they are wrong, and they usually then duplicate the experiments themselves and we then-we had a case in Science a few years ago of a retraction in exactly that manner. The laboratory that originally did the work retracted the work when others pointed out that it was in question.

81 Dr. Koshland amended his statement to read "world and, say, Congress. . . ."

82 Dr. Koshland deletes the word "of."

83 Dr. Koshland adds the word "and."

84 Dr. Koshland adds the word “argument."

85 Dr. Koshland adds the word "it."

So I think one of the real duties of the scientific journals is to publish retractions, corrections, criticisms when they do exist, and I think each journal has a real obligation to publish-to correct the record, certainly with its own journal and probably will criticize other journals if they do not respond appropriately.

The question was brought up in regard-by Barbara Mishkin, which I agree with-fraud is a very serious matter to charge anybody with. It is something that occurs, I think, very infrequently. I have got some numbers that I didn't know how 86 to quantitate this and lots of people have been discussing it. The Library of Medicine lists 3,300 journals, and that is far from 87 I think the rough estimate is there are 20,000 scientific journals, but this is the top journals 88 that are recorded by them. These publish about 300,000 articles a year in science, and we are talking a lot about fraud, and I think there are, as I saw, about a dozen well-known cases of fraud. There are probably more, as Dr. Sprague and others have mentioned. I don't think that is all of them, but 89 that is a very small number of the total amount of articles that are published, and of course some of those people published a number of articles. So I thoroughly agree, when you deal with a high consequence but low incidence type of problem, the way to do it is punish the people who do it very severely, and I think

Mr. JOHNSTON. Doctor, let me interrupt you, though.

Dr. KOSHLAND. Yes.

Mr. JOHNSTON. A low incidence of discovery, could that be a low incidence of policing for fraud?

Dr. KOSHLAND. Let me

Mr. JOHNSTON. 12 out of 300,000?

Dr. KOSHLAND. Yes. I think that is a very good question. Let me answer it this way.

I think practically no important-I should say major-fraud goes undiscovered. If the fusion-low cold fusion debate 90 that has just gone on I think is illustrative. When that discovery was made, laboratories all over the world jumped to it. If you tomorrow had somebody who fraudulently claimed they had a cure for AIDS, I think it wouldn't last very long, and I think people who do things either by honest error, which I think was possibly true in the cold fusion case, or deliberate fraud, know that the lifetime of an important 91 finding is going to be very short. If you now go down that scale to less and less important findings, maybe even to trivial findings, I think those could last in the literature a long time without being discovered.

So I think when-I would say all the major fraud is fairly minor 92 and gets discovered very rapidly. The minor fraud-I

86 Dr. Koshland amends his statement to read "I have got some numbers that may help. I didn't know how

87 Dr. Koshland amends his statement to read "Far from the total . . ."

88 Dr. Koshland amends his statement to read "but these are the top journals

89 Dr. Koshland amends his statement to read "as I said, about a dozen well-known cases of fraud in the last 10 years. There are probably a few more, as Dr. Sprague and others have mentioned. I don't think that is all of them, but the total

90 Dr. Koshland amends his statement to read ". goes undiscovered. The cold fusion debate

91 Dr. Koshland adds the word "fraudulent."

92 Dr. Koshland deletes the word "minor" and inserts the word “infrequent..."

can't tell you how many of those there are, but I don't think those who say it is very prevalent can say so either.

So I think the incidence 93 on the progress of science of fraud is relatively small. That is what I am saying. So I think the way you do that is sort of like murder-and I don't want to analogize those two at all-but on the other hand, you have a very severe punishment-and what I was leading to-then if that is true-what Barbara Mishkin mentioned is true, the accusation of that is very seri

ous.

As far as I know, in almost every case of fraud the person accused has ended a career in science, and I think that if Mishkin's comment is correct that there should be something in the machinery, that that person, if they may sort of plea bargain, to say "keep this all quiet and I will resign quietly," some further university or agency should be able to know about that.

But I sympathize with universities because as an editor the question comes up of fraud and people say, "well, have a great deal of courage, stand up for your rights." Our lawyers estimate a minimum cost of going to litigation of $500,000. Probably a million 94 is a better estimate. So that is even if you are right and defend yourself. So any journal editor, any university department is going to have to think about that, and I think that should be considered.

But I think there is no question that when a duly constituted committee makes a charge it seems bad to keep that quiet. So I would go along with that kind of consideration.957

On the other hand, I do think, as I said, the incidence of that is low. I think the other thing is that the parts of science that we would like to improve-and I am not going to be here to defend 96 the scientific community-that we are perfect, but I think the procedures that have been started, that Dr. Friedman and others talked about before, are being taken much more seriously than they were in the past and that an attempt should be made to give those a chance possibly before we deal with further legislation even in the area of immunity, and so forth, because that can be abused. But I do think if you want-if the Congress is going to ask for certain things, like immediate publication, and pass this information on to other people, there has to be some protection against these heavy financial burdens which a journal might undergo. At least, I know that has affected us.

So I think I guess my final conclusion would be that it is very important to look at these. We are under a spotlight which everyone recognizes, and journals have a heavy responsibility which we are willing to assume. We can't take an impossible responsibility, and we do have the other responsibility that most scientists are honest. They work very hard.

Competition was mentioned here. My graduate students have a slogan that if you are only willing to work 40 hours a week you shouldn't be in graduate school, and so the competition is working

93 Dr. Koshland amends his statement to read "So I think the influence on the progress of science.

94 Dr. Koshland added the word "dollars."

95 Dr. Koshland deletes the word "consideration" and inserts the word "immunity." 96 Dr. Koshland amends his statement to read "and I am not here to defend ...

23-941 0-90-9