Thursday, June 25, 2009

Integrity & Responsibility In Scientific Research: What Are The Basic Rules?

Are phrases such as “a code for integrity and responsibility in research” and “the basic rules of good scientific practice” mere platitudes? Even at the start of this millenium, scientific organizations were trying hard to correct the situation[1]. Fabricating data (fantastical and low-level) is definitely a heinous scientific sin.

To prevent fraud, preservation of data and records of the investigations are especially relevant for large-scale (experimental or computational) projects since the verification of results require resources and time beyond the scope of the majority. Such projects are properly done when the supervisors and subordinates (usually students and post-docs present for a short term) are equally responsible and if both parties have hands-on experience with the work. It might help if each publication has a footnote stating the role of each author in the publication. This is still anathema to most researchers.

Unfortunately, fraud is not the only sin committed in both experimental (presumably more relevant) and theoretical work. It should be noted that there is ambiguity in the rules of the game and that it is a tough task to decide when or whether the rules are broken. Consider papers in which a theory is “cooked-up” (recipe including simple analyses, a set of assumptions and convenient free parameters) to “fit” experimental results. The difficulty to disprove some of these “theories” could be comparable to the difficulty to explain the experiment. Doubt and criticism remain within an old-boys network and does not become knowledge in the public domain. It is not uncommon that even if a work is erroneous by way of calculation and concept, the paper still gets cited by those who are unaware of the deficiencies and take the peer-reviewed results for granted. By the time the errors are known in public, the publication would have ceased to matter even though it proved to be useful for promotions and new positions. It is usually a lucky day for Science when these errors are mentioned as inadvertent errors in footnotes. Next, consider papers submitted for peer-review. Even if errors are suspected (and probably because brilliant results are as rare as geniuses), a paper could be accepted due to complicity in the network or with the unwritten comment “it does not matter since it is going into that journal”. Meanwhile, the reviewer could start simulations/experiments of their own on the same topic during the review process. Are these low-level sins?

Scientists are definitely as human as everyone else. The job of a scientist is just like any other job in a close community with ethically correct individuals being the minority and rules broken by the majority – if not closely watched.

[1]For example, please refer to “Rules for Integrity” (in MaxPlanckResearch 2/2001, p. 90).

No comments:

Post a Comment