emerged as an important influence on several parts of the process of generating and evaluating evidence. In this section we look more closely at the specific ways that prior knowledge may shape part of the process. Prior knowledge includes conceptual knowledge, that is, knowledge of the natural world and specifically of the domain under investigation, as well as prior knowledge and beliefs about the purpose of an investigation and the goals of science more generally. This latter kind of prior knowledge is touched on here and discussed in greater detail in the next chapter.

Beliefs About Causal Mechanism and Plausibility

In response to research on evaluation of covariation evidence that used knowledge-lean tasks or even required participants to suppress prior knowledge, Koslowski (1996) argued that it is legitimate and even helpful to consider prior knowledge when gathering and evaluating evidence. The world is full of correlations, and consideration of plausibility, causal mechanism, and alternative causes can help to determine which correlations between events should be taken seriously and which should be viewed as spurious. For example, the identification of the E. coli bacterium allows a causal relationship between hamburger consumption and certain types of illness or mortality. Because of the absence of a causal mechanism, one does not consider seriously the correlation between ice cream consumption and violent crime rate as causal, but one looks for other covarying quantities (such as high temperatures) that may be causal for both behaviors and thus explain the correlation.

Koslowski (1996) presented a series of experiments that demonstrate the interdependence of theory and evidence in legitimate scientific reasoning (see Box 5-2 for an example). In most of these studies, all participants (sixth graders, ninth graders, and adults) did take mechanism into consideration when evaluating evidence in relation to a hypothesis about a causal relationship. Even sixth graders considered more than patterns of covariation when making causal judgments (Koslowksi and Okagaki, 1986; Koslowski et al., 1989). In fact, as discussed in the previous chapter, results of studies by Koslowski (1996) and others (Ahn et al., 1995) indicate that children and adults have naïve theories about the world that incorporate information about both covariation and causal mechanism.

The plausibility of a mechanism also plays a role in reasoning about cause. In some situations, scientific progress occurs by taking seemingly implausible correlations seriously (Wolpert, 1993). Similarly, Koslowski argued that if people rely on covariation and mechanism information in an interdependent and judicious manner, then they should pay attention to implausible correlations (i.e., those with no apparent mechanism) when the implausible correlation occurs repeatedly. For example, discovering the cause of Kawasaki’s syndrome depended on taking seriously the implausible cor-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement