Below are FAX-BACK comments we received for topics raised in the November-December and January-February issues.
On investigating scientific-misconduct allegations
I do not think that NIH should waste time and effort in attempting to set up internal mechanisms for the examination of alleged misconduct. [see Science Ethics Forum, November-December 1994 issue] My reason for this view is that institutions are inherently incapable of investigating themselves honestly. The historical record shows that intra-institutional pressures, extraneous to the scientific-misconduct issue, have invariably produced whitewash and/or cover-up. For example, a faculty member charged with investigating alleged misconduct at his own institution may be conflicted between fact finding and the pressure to be a "team player." Therefore, he or she might not find misconduct that could possibly result in the loss of prestige and (more importantly) funds to the parent institution.
The Mikulas Popovic - Robert Gallo and the Thereza Imanishi-Kari - David Baltimore cases are the two most well-known but are not atypical examples of how the current system has "worked" in government and academic environments. The fact that these cases are still ongoing after approximately 10 years is a gross injustice to the accused, to the whistle-blowers, to all scientists, and to the public at large. Clearly, we need a new system.
The establishment of an organization that is devoted entirely to the fact-finding process and is wholly independent of government and of any particular university has been suggested. Universities and government agencies would agree in advance to accept the results of the objective investigations of this body. The existence of such an organization would eliminate intra-institutional conflict of interest and would also separate fact-finding from disciplinary actions, the latter being left to the relevant institutions. This organization should be directed by an outstanding individual of unimpeachable integrity and staffed primarily by people with scientific backgrounds and capable of critical analysis.
I call upon Dr. [Harold] Varmus to take a leadership role and convene a meeting for refining the many details involved in establishing such an organization. Such details would include, among others, procedural safeguards, mechanisms for obtaining documents, policies relating to confidentiality, and means of funding support for such an enterprise. Once it is accepted that some amount of scientific misconduct exists and that the current procedures have not worked satisfactorily, it should be possible to develop techniques to solve this problem.
-- Philip D. Ross, NIDDK
On NIH's new affirmative action plan
The new affirmative action plan appears to be going in the right direction.[see Insights from OEO's New Leader, January-February 1995 issue]. The concepts -- a) making the leadership at the top responsible but allowing them to be flexible to handle the problems inherent in small group statistics, b) comparing NIH representation to that in truly comparable nationwide groups, and c) focusing on education to reduce the real bottleneck in the process of creating minority professionals -- all make sense. Most important is a spirit of cooperation to improve race relations and to increase the interest and ability of minorities in science, as opposed to an adversarial attitude.
We do note an egregious systematic error in the table on page 17. Only the URI [underrepresentation index] results for Hispanics and Native Americans are correct. If the computation is done as the asterik-marked footnote within the table dictates, grossly different results would be obtained. It appears that any result over 100% has been limited to 100% without mention. ... If one tries to excuse this on the grounds that "we forgot to mention that we limited the results because a URI over 100% is meaningless," with which we disagree since it is very informative of reverse discrimination, then how do you explain that the figure for men, which computes to about 69% on one table and 68% on the other, is also given as 100% ? One senior researcher in our lab has noted that if he submitted such data to a journal, he might never be allowed to publish again. ...
We wonder if you would be willing to publish a similar article, but corrected as above, in The NIH Record, with its very different readership, or do you have two standards?
-- Anonymous
The Underrepresentation Index figure was based on tables provided by the Office of Employment Opportunity from its draft of the new NIH Affirmative Action plan. Because the plan's emphasis is on recruiting members of underrepresented groups, not removing members of overrepresented groups, placing an upper limit of 100% on the underrepresentation index (URI) seems appropriate. That aside, you are indeed correct that there is an error in the URI calculations for men. During the editing process, 100% was mistakenly inserted under the "men" category. As you note, the correct URI for men should be about 69% for the National Research Council data comparison and 68% for the 1990 Census data comparison. We apologize for the error.
As for publishing a similar article in The NIH Record, you may be interested to know that The NIH Catalyst, published by the Office of Intramural Research by and for NIH scientists, is editorially independent from The NIH Record, which is published by the Office of Communications for the entire NIH community.