Scientific Integrity: From Theory to Practice

Maintaining scientific integrity is clearly important, and its centrality to scientific practice is reflected in reports from the NAS and in documents like the Singapore Statement.  But definitions of scientific integrity have become increasingly amorphous, generating lists of various responsibilities for scientists, with little coherence or examination of the grounding of such responsibilities.  In short, ideas of integrity have lost their integrity.  This talk will remedy this problem by providing a clear definition, based on what is valuable about science.  With the value of science as the ultimate grounding for scientific integrity, an account of the proper roles for values in science can be given.  It is on maintaining these roles that scientific integrity rests.  The value-free ideal will not do, but constraints on the use of values in scientific reasoning is crucial.  Scientific integrity consists in honoring the value of science and thus the constraints on values in science.  This notion of scientific integrity narrows the responsibilities that rest upon it, thus revealing the other bases for responsible conduct of research, providing a clearer conceptual organization for the terrain.  Under this view, scientific integrity must be maintained in conjunction with the proper handling of the social epistemic practices of science and the proper addressing of the societal responsibilities of scientists for the responsible conduct of research to be achieved.   Understanding integrity also illuminates the assessment of scientific expertise.   We can assess experts with respect to their integrity by examining patterns of argument in their work over time.  Integrity is separate from an assessment of whether we agree with the values the scientific expert holds.    By having a clearer understanding of what scientific integrity is, on what it rests, and how we can assess its presence (or absence), scientific integrity can become a more functional concept in scientific practice.