What Makes a Good Experiment? ─ Reasons & Roles in Science

What makes a good experiment? Although experimental evidence plays an essential role in science, as Franklin argues, there is no algorithm or simple set of criteria for ranking or evaluating good experiments, and therefore no definitive answer to the question. Experiments can, in fact, be good in any number of ways: conceptually good, methodologically good, technically good, and pedagogically important. And perfection is not a requirement: even experiments with incorrect results can be good, though they must, he argues, be methodologically good, providing good reasons for belief in their results. Franklin revisits the same important question he posed in his 1981 article in the British Journal for the Philosophy of Science, when it was generally believed that the only significant role of experiment in science was to test theories. But experiments can actually play a lot of different roles in science, as he explains—they can, for example, investigate a subject for which a theory does not exist, help to articulate an existing theory, call for a new theory, or correct incorrect or misinterpreted results. This book provides details of good experiments, with examples from physics and biology, illustrating the various ways they can be good and the different roles they can play.

Allan Franklin is professor of physics at the University of Colorado. He has twice been chair of the Forum on the History of Physics of the American Physical Society and has served two terms on the Executive Council of the Philosophy of Science Association. He is the author of eleven books, including Are There Really Neutrinos? An Evidential History;Selectivity and Discord: Two Problems of Experiment; NoEasy Answers: Science and the Pursuit of Knowledge;Shifting Standards: Experiments in Particle Physics in the Twentieth Century, and coauthor ofEnding the Mendel-Fisher Controversy.