Chapter 7: Evidence Evaluation and Scientific Progress
[Image removed for copyright reasons] (Wikisource contributor note)
Scientists and philosophers of science share a concern for evidence evaluation and scientific progress. Their goals, however, are quite different. The philosophers find the process of science intrinsically interesting. Most are not trying to ‘straighten out’ the scientists and tell them how science should be done. Some of their conclusions do have possible implications for future scientific methods, but scientists seldom listen. Perhaps scientists’ reactions are somewhat analogous to those of creative writers toward literary critics and academic literary analysts: often the doer is unappreciative of the outside reviewer.
Each scientist unconsciously selects criteria for evaluating hypotheses. Yet clearly it would be both confusing and professionally hazardous to adopt substantially different criteria than those used by one’s peers. Judgment, not irrefutable evidence, is a foundation of science. Judgments that observations confirm or refute hypotheses are based on personal values: accuracy, simplicity, consistency, scope, progressiveness, utility, and expediency.
Different types of laws, or hypotheses, require different evaluation criteria [Carnap, 1966]. A universal law such as ‘all ravens are black’ is best tested by seeking a single exception. In contrast, a statistical law such as ‘almost all ravens are black’ or ‘99% of ravens are black’ requires a statistical test that compares observed frequencies to hypothesized frequencies. Theoretical and empirical hypotheses call for contrasting evaluation techniques and standards. For example, a theoretical-physics hypothesis may concern properties that are not directly measurable and that must be inferred indirectly, and it may be judged more on simplicity and scope than on accuracy of fit to observations.
This chapter considers all of these aspects of evidence evaluation.
Critical thinking skills and mistakes begun in childhood survive the transition to adult. Naive conceptions do not simply disappear when a more mature thinking skill is developed; they must be consciously recognized as wrong and deliberately replaced. For example, children learn about causality first by treating all predecessors as causal (“if I wear a raincoat and avoid getting wet, I won’t get a cold”). Only later and rather haphazardly is this superstitious approach supplanted by the skill of isolation of variables.
