Summary

Selectivity and Discord addresses the fundamental question of whether there are grounds for belief in experimental results. Specifically, Allan Franklin is concerned with two problems in the use of experimental results in science: selectivity of data or analysis procedures and the resolution of discordant results.

By means of detailed case studies of episodes from the history of modern physics, Franklin shows how these problems can be--and are--solved in the normal practice of science and, therefore, that experimental results may be legitimately used as a basis for scientific knowledge.

Choice Review

Franklin (physics, Univ. of Colorado) presents a scholarly account of two key issues in scientific methodology, aptly described in chapter 1, "Selectivity and the Production of Experimental Results," and chapter 2, "Resolution of Discordant Results." In-depth discussion is provided for eight scientific investigations in high-energy physics and two in gravity. Franklin's rationalist approach is laid out as a counter to constructivist views about the history of science. In this "conjectural realism ... we have good reasons to believe in facts, and in the entities involved in our theories, always remembering, of course, that science is fallible." Franklin argues that reasoned arguments, incorporating methodological and epistemological considerations, can eventually resolve the discord between experimental results. Selection criteria or "cuts," applied either to data or the analysis procedures, can create artifacts in experimental results, but does this ultimately undercut the possibility of progress in science? Through an extensive set of examples and techniques such as robustness, replication, and blind analysis, Franklin skillfully shows how "Cuts may be ubiquitous, but they need not be fatal." In his effort to root out problems of constructivism, Franklin tends to neglect the social context for the knowledge production activity. ^BSumming Up: Recommended. Upper-division undergraduates through professionals. T. Eastman formerly, University of Maryland