What follows is an account of how I gradually case to realize that a published scientific paper contained serious inaccuracies and what happened when I attempted to get a correction published.
I was awarded an N.I.H. training stipend and started work as a postdoctoral fellow in the laboratory of Dr. Thereza Imanishi-Kari at the Massachusetts Institute or Technology.
July and August 1985:
Some of my experiments yielded results that contradicted Dr. Imanishi-Kari's claims. I had many discussions with her about the discrepancies. I brought her my data. She told me her results from memory. I reported to Dr. Wortis, my former advisor and a close associate of Dr. Imanishi-Kari's, that I considered many of the procedures on the laboratory unsound and improper.
[I continued my search for the reasons that our results disagreed until March 1986. The matter was resolved only when, as described in detail below, in April and May 1986, Dr. Imanishi-Kari told me and others that my results were actually the sane as hers and that the data as published were misrepresented.]
November 1985 - March 1986: I received a pre-print of the Weaver et al. paper. Dr. Imanishi-Kari was the senior author and much of the work had been done by her and Dr. Reis in our Laboratory. This was the first time I had seen these data in written form and there were indeed a number of disagreements between my results and those in the manuscript. Dr. Imanishi-Kari encouraged me to rely on the previous results instead of my own. However, the experiments included controls for my own project and so I could not do this.
Despite many attempts, I was unable to replicate the results of an important experiment Dr. Reis and I had done in September. When subsequent experiments continued to contradict my original results, Dr. Imanishi-Kari decided that she herself would perform these experiments. She told me to confine my laboratory duties to mouse husbandry. In March, I discovered that Dr. Reis had made an error which invalidated our initial results and explained the failure to replicate the finding. Dr. Imanishi-Kari, however, still insisted that she would eventually repeat my initial results and she continued to do the experiments.
I then I informed Drs. Henry Wortis, Herman Eisen and Mary Rowe of these facts. (Dr. Eisen was the director of the N.I.H. training grant from which I received my stipend and as such I assumed he had supervisory responsibility towards me. Dr. Rowe is an assistant to the president at M.I.T.) Drs. Eisen and Wortis expressed an unwillingness to become involved in the "dispute". Dr. Rowe stated that, in her opinion, by so pressuring a subordinate to generate the expected results, Dr. Imanishi-Kari was actually guilty of harassment and/or scientific misconduct. I discussed my dilemma with a number of my peers but there was no clear consensus about what course of action I should follow.
Dr. Imanishi-Kari told me to prepare for publication a study I had done that was closely related to that described in the Weaver et al. paper. She gave me instructions to omit certain facts about my results. This would have resulted in a misleading paper. I pointed this out a number of times to her but she persisted in her instructions. I eventually resolved to simply stop arguing and just not do what she told me. I decide to leave Dr. Imanishi-Kari's laboratory and seek employment elsewhere. I gave my notice but agreed to continue taking care of the mice through May.
Dr. Imanishi-Kari showed me the results of her own first cell transfer experiment. By manipulating the data she forced the results to conform to her hypothesis. She pointed out that, if my data were manipulated in the same way, they also would support her hypothesis. She said that I was too "nit-picky" and that "better reagents" would confirm her results.
Dr. Imanishi-Kari admitted to me that one of the graphs in the Weaver et. al paper was misrepresented and that the true results agreed with mine. This startling disclosure, coupled with my increasing knowledge of the available laboratory reagents and procedures, led me to doubt that the experiments could have been performed as the described in the Weaver et al. manuscript. On two occasions I informed Dr. Weaver that I was concerned about the results reported in his paper. On both occasions, however, his schedule did not permit a thorough scientific discussion. The Weaver et al. study appeared in print.
On a chance meeting, Dr. Huber, a professor at Tufts University, told me that Dr. Imanishi-Kari had characterized her disputes with me as a ''personality conflict." I said that our differences were of a scientific nature and that I considered many of the conclusions to be unsound. Dr. Huber, a friend of both Dr. Imanishi-Kari and mine, agreed to review the data to see if she could help me and asked me come and see her in two weeks when she would be less busy.
May 7th: I had occasion to check the pedigree of a certain mouse. While examining the breeding records Dr. Reis had left in my care, I recognized that some of the entries were not breeding records, but the results of some critical experiments for the Weaver et al. paper. In order to study these records more carefully, I xeroxed the notes and examined then during my lunch break. I became convinced that the several of the major assertion of the paper were actually contradicted by the experimental results. Moreover, certain crucial facts, much like the facts she had instructed me to omit from my manuscript, were also omitted in the Weaver et al. study. I then approached Dr. Imanishi-Kari and, without mentioning the notes I had xeroxed, asked if I could examine the records for a related set of experiments that she said she had performed herself. (I had asked her for this data previously also). She searched the laboratory for some time. Not only did she not find the records I requested, she was unable to find any of her records concerning the claims in the Weaver et al. paper.
May 8th: I again asked Dr. Imanishi-Kari for the notes. Again she stated that she had no idea where they were.
May 9th: I presented my reasons for doubting the published claims to Dr. Huber and gave her a copy of the notes. Before I could give her a complete analysis, however, she interrupted stating that she did not want this "on her shoulders", that she was ill-equipped to deal with such a serious matter and that she must consult a more expert scientist for advice. By this time we were both very upset and I was unable to decide what to do. Dr. Huber told me that there was no way that I could be expected to deal with a problem of this magnitude alone and that I did not deserve to have it on my shoulders either. She telephoned Dr. Woodland and reported to me that Dr. Woodland had advised us to bring the matter to Dr. Wortis and let him decide what to do. I replied that, because of the close personal and professional associations of everybody involved, and the apparent seriousness of the discrepancies, I thought that it would be wise to seek the advice of someone with no apparent conflict of interest. While she did not rule this out, she was adamant that, at least as a first step, we should let Dr. Wortis know since he was in the process of applying for a grant as a co-investigator with Dr. Imanishi-Kari on this work.
I presented the data to Dr. Wortis. He said that he would confront Dr. Imanishi-Kari and, if my objections were verified he would "rehabilitate her." (I should mention that Dr. Wortis has expressed anger that I quoted him to the chairman of his department on this point. Although I could not understand his reason for saying I had done something improper, I believe in fairness he, himself, should best explain why he thought I had). I protested that this course of action placed me in an ethically compromised position and that I felt bound to report to a less involved person. He said he understood my point and he believed that, if I must report to anyone else, it should be to Dr. Baltimore. Since Dr. Baltimore was an author on the study in question and had no supervisory responsibility for either Dr. Imanishi-Kari or me, this did not make any sense to me. I consulted a colleague who insisted that I show the notes to Dr. Flax, the chairman at Tufts. Dr Flax informed me that this was an M.I.T. matter and should rightly be investigated there. He added, however, that he would allow Dr. Wortis to continue his own investigation.
May 12th: I met with Drs. Wortes and Huber for over two hours to further explain the scientific issues.
May 13th: Dr. Huber informed me that Dr. Wortis had contacted Dr. Imanishi-Kari and asked for a conference to review data. In a conversation with me, Dr. Wortis assured me that he would give Dr. Imanishi-Kari the opportunity to voluntarily produce her records. He stated that only it she was not forth-coming with her notes would he let her know of my involvement.
May 14th: Dr. Wortis informed me that his meeting with Dr. Imanishi-Kari is scheduled for the next morning. I told him I would stay at home.
May l5th: Within half an hour after the meeting had been scheduled to begin, Dr. Imanishi-Kari called me. She accused me of vindictive motive and threatened to sue me. I attempted to calm her by pointing out that the matter was in the hands of her good friend and strong supporter; if there was a legitimate explanation for the discrepant data she could present it. She ordered me to come to the laboratory and resume my mouse husbandry responsibilities. I stated that I wanted a mediator present while she and I discussed my remaining responsibilities. She rejected this proposal and told me not to return to the laboratory. I related this conversation to Drs. Flax, Wortis and Huber. Dr. Huber informed me that Dr. Imanishi-Kari was "not taking this well" and a meeting had been scheduled for the following evening: she and Drs. Wortis and Woodland would then review the data.
May 17th: Dr. Huber reported to me that the above mentioned meeting had indeed occurred on the 16th. She said that Dr. Imanishi-Kari was unable to produce most of the relevant records. She described some records that they did examine and agreed with me that these tended to support my objections and actually confirmed one of my most important assertions. Dr. Huber stated that Dr. Imanishi-Kari was getting the rest of the data together and that they would have another meeting when she was ready.
May 20th and 21st: I related these events to Dr. Mary Rowe at M.I.T. She informed me that this was an M.I.T. matter and should not have been taken to Tufts. She pressed me to "bring charges" through the official channels. I refuse saying that, if my suspicions proved correct, I would insist only on a retraction of the paper. I stated that I had no interest in causing Dr. Imanishi-Kari unnecessary pain and that formal charges would likely adversely affect both of us, would definitely poison the academic environment and abolish any chance of a collegial resolution of the scientific issues.
I called Dr. Huber and protested that the "investigation" was being mishandled. I said that their failure to examine the records as agreed was a glaring error and they should disqualify themselves. She agreed to push for another meeting and I said I wanted to be present. She said she would discuss this with Dr. Wortis.
In a subsequent conversation, Dr. Wortis told me that my protests about the way he was handling the case called into question my motives. Again, I was unable to understand his reason for saying this. Despite our differences however, we agreed on the following: l.) I would be present at the next meeting. 2.) If my concerns were substantiated he would make sure the paper was retracted and I would not bring misconduct charges. 3.) If the records could be explained in a manner consistent with the published claims, I would apologize to all concerned and the matter would be dropped.
May 23rd: I received a call from Dr. Wortis in the afternoon stating that he had scheduled a meeting for that evening. He said that only himself, Drs. Huber and Imanishi-Kari and myself would be present. Dr. Woodland was not able to be there. I told Dr. Wortis that I thought we should have at least one other person present. He rejected the proposal. Further, he stated that he would limit his questions to those concerning table 2 and 3 of the published work. I called Dr. Flax and informed him of the meeting, the restrictions and my misgivings. I asked him if he thought I should attend the meeting under these circumstances. He said that he had no objection. I attended the meeting and examined all of the records presented by Dr. Imanishi-Kari in support of the published claims. My assertions that the data did not support the published claims were completely confirmed. I left the meeting under the impression that the paper would be retracted.
May 24th: In a telephone conversation, Dr Huber informed me that, although she and Dr. Worits agreed with me scientifically, they felt that fraud was not the cause of the discrepancies. They had decided that a retraction "would be devastating" and was not necessary. They were dropping the matter entirely and were strongly advising me to do likewise. I disagreed saying that other laboratories were relying on the data. She said the decision was for Dr. Wortis to make. I said that I needed to talk to him to make sure he had fully understood the issues. She said she would try to arrange this.
May 29th: Dr. Wortis telephoned me, informed me that he did indeed understand and accept the validity of my objections. He reviewed many of them and acknowledged their importance but he declared the subject closed. In response to inquiries by Dr. Rowe, I admitted that my course of action had not achieved the desired result. She cited the damage being done to my reputation by rapidly traveling rumors that I had falsely accused my supervisor, insisted that I speak to Dean Brown, called his office and arranged for me to speak with him that afternoon. Dean Brown told me that I should either bring charges of fraud or forget the matter entirely. I told him that of those two options I would choose the latter.
May 30th: Dr. Eisen called me and said that he wanted to meet with me regarding this matter and we met that afternoon. I showed him the xeroxed records and explainned my concerns. He told me to put my objections on writing This I did in the form of a memorandum to him dated June 6, 1986.
On receipt of my memorandum Or. Eisen called me and asked how one could explain how these misrepresentations could have been published without invoking fraud as a cause. 1 told him my interest and responsibility were with the accuracy of the science and that judgement of these conduct issues was not my concern. The next week I saw him in person and he informed me that he had given my memorandum to Drs. Imanishi-Kari, Weaver and Baltimore. He said that Dr. Imanishi-Kari had then left town on business and that a meeting was scheduled on her return. I requested that I be allowed to bring someone to the meeting. He rejected this request calling it "intrusive on the science." I understood from Dr. Rowe that she also asked Dr. Eisen if someone could be present to represent my interest and he rejected this suggestion.
June 16th or 23rd: I attended a meeting with Drs. Eisen, Baltimore, Weaver and Imanishi-Kari. Dr. Baltimore opened the meeting by asking Dr. Eisen to report on his conversations with Dr. Wortis at Tufts. Dr. Eisen said that Dr. Wortis had not presumed to judge Dr. Imanishi-Kari's science but that he was satisfied that she had not committed fraud. I said that I wished to approached this as a scientific issue.
We then went through the points in my memorandum. (As I have described above, I wrote this only after I had examined the all data presented by Dr. Imanishi-Kari in support of the published claims). Drs. Baltimore and Eisen spent a lot of the time examining graphs Dr. Imanishi-Kari said she had generated in Germany many years previously and which did not form any of the basis for the study in question. Some records of Dr. Weaver's were examined and they again proved that my concerns were well founded. None of my objections were answered and I made this point clear a number of times. Dr. Imanishi-Kari told Dr. Baltimore that she had not brought the critical records to the meeting because she had not thought she would need them and because they were written in Portuguese. Dr. Baltimore assured her that he did not need to see any other data.
With regard to the scientific issues raised in my memorandum, Dr. Baltimore then stated that "this sort of thing is not at all unusual" and that he would deal with it in "the usual manner". Further, on examining the records that formed the basis for table 2 he said "you can't tell anything at all, one way or the other, from this". Dr. Imanishi-Kari confirmed without dispute that she had misrepresented the data in figure 1. Dr. Baltimore asked her why she had done this and she said she thought they must have "got that result once". He said that was not acceptable and he would talk to her in private. He outlined a series of new experiments she could do that could, if successful, support her hypothesis. He said he thought that should satisfy me. He strongly advised me to drop the matter "for my own good" and told me that a lot of his time went into calming people like me down. He appeared to make this remark in the context of what junior people often do when asked to review a manuscript. He said that in these cases, however, these people had not examined the laboratory records. He appeared to cite this fact as an argument against my position and I could not understand his reason. I stated I had legitimate access to the records in question; Dr. Imanishi-Kari concurred. He said that my only recourse, as he saw it, was to write to the journal. He added that if I did this he would personally wrote a rebuttal "similar to the recent one in Nature". I stated that I considered my responsibilities discharged and intended to drop the matter. Dr. Baltimore said "Oh well, then there is no problem." He stated that he had reason to believe that part of the study was true and that therefore no correction was warranted. He asked Dr. Eisen to write up a memorandum stating that the matter had been "aired."
Following the meeting I protested to Dr. Eisen that, despite the acknowledged misrepresentation, he did not examine the records and was not advocating that the paper be retracted or corrected. I pointed out the damage his actions had caused to my reputation. He replied that my continuing pursuance of the matter "would seem to indicate vindictiveness." Eventually, however, he did agree that he would never write a report stating that "everything was all right." I protested the handling of the case to Dr. Rowe. She said that Dr. Eisen had not reported any misrepresentation to her and that the matter was now "in the hands of God." I believed I had exhausted all my options and I dropped the matter completely.
In July or August. Walter Stewart and Ned Feder contacted me and told me they had been aware of this case since before my meeting with Dr. Eisen. I had not known this, although I was well aware that a large number of people knew of my challenge to the accuracy of the paper and that I had been rebuked by two universities. Mr. Stewart and Dr. Feder told me that they wanted to examine the data themselves and persisted over a period of months in presenting me wlth their arguments in support of my co-operation. These I found very persuasive, but I resisted at first for a number of reasons, one of which was my fear that my husband's career would suffer. Another reason was that I still thought there was a possibility that the authors would retract or correct the paper. When this did not happen, despite my fears for my husband's career, I eventually sent them the records I had xeroxed.
In October 1986 I was told by Dr. Huber that Dr. Imanishi-Kari had since repeated the experiments described in the Weaver et al. study and that this time the results supported the previously published claims. In January 1987, Mr. Stewart informed me of the existence of a report by Dr. Eisen and I requested a copy from him on March 17, 1987. On July 20, after I agreed to his conditions of confidentiality, Dr. Eisen's secretary finnally sent me a copy at a memo dated December 30 1986. (I noted that Dr. Eisen's memorandum was written more than six months after he received my memo and also after Mr. Stewart and Dr. Feder had made Inquiries to the authors). I have written twice to Dr. Eisen asking him to correct his memo. I have recelved no reply.
In fall 1987, I saw reference to a report written by Dr. Wortis in some documents sent to me by Mr. Stewart. I wrote to Dean Banks requesting a copy. Dean Banks replied that the report was "independent of the informal discussions" in which I participated. He added that "....distribtion of the report to anyone who was not a party to the report or who is without formal standing in this matter would be inappropriate." I have since received a copy of the report, but only through the efforts of this sub-committee.