Innovation and integrity in biomedical research.

Science's reputation for purity suffered two major setbacks in the past ten years: scientists' capability to regulate themselves came into serious question; and burgeoning entanglements between universities and industry created additional incentives for misconduct in research. In this article, the author seeks to shift attention from compliance to the definition of the standards themselves and to suggest that there is less agreement about acceptable norms of behavior than is commonly supposed among critics of science. Further, this lack of clarity is in part a consequence of the fragmentation of research communities at the forefront of science. Contrary to popular misconception, there is no abstract, universal "scientific method" that guides practice in all situations. The most promising way to deal with criticisms of scientists' integrity is to recast the problem of scientific integrity as one of prospectively creating acceptable research practices rather than retrospectively finding and applying them. To achieve conditions that foster integrity, however, will require more than teaching research ethics to graduate students or educating senior scientists in better mentoring. Instead, the culture of science will need to confront and, where necessary, dismantle the structural barriers to collegiality in research. As now organized, science is organized as a winner-take-all game, with no glory or comfort for the also-ran. Would a more collaborative science produce as many dazzling results? The burden of proof rests with those who want to change the present system, but with the public image of science hanging in the balance, the time may be ripe for taking up the challenge.