Relations among Three Aspects of First-Year College Students’ Epistemologies of Science

At the start of their first semester, 35 college freshmen were given an interview probing (a) their differentiation of scientists’ ideas from evidence, and hypotheses from theories; (b) their understanding of the inherent uncertainty of scientific knowledge; and (c) their reasoning about scientific controversies. The most common responses were in terms of an epistemology in which scientists’ ideas and evidence are differentiated, and theories are understood as tested hypotheses (Level 2 in our system based on Carey, Evans, Honda, Jay, & Unger, 1989), although students varied in how consistently they differentiated theories and evidence across all questions. Responses in which theories are understood as broader explanatory frameworks guiding hypothesis testing (Level 3) were virtually nonexistent, but some students gave responses that showed awareness of processes of interpreting and reinterpreting patterns of results (Level 2.5). Responses across the three parts of the interview were significantly related. Consistently differentiating scientists’ ideas from evidence was strongly related to appreciating the inherent uncertainty of scientific knowledge and with having a deeper understanding of the reasons for scientific controversies and how to resolve them. 2006 Wiley Periodicals, Inc. J Res Sci Teach Reform efforts in science education at the K–12 and college levels have stressed the importance of developing students’ epistemological understandings of science (American Association for the Advancement of Science, 1993; National Research Council, 1996; National Science Foundation, 1996). At the same time, there is concern that most college students and preservice science teachers have at best a very naı̈ve understanding of knowledge construction in science as following an atheoretical ‘‘scientific method’’ (Abd-El-Khalick & Lederman, 2000; Lederman, 1992; Windschitl, 2004). Developing more sophisticated epistemological understandings may be important for diverse reasons: to help create better science students, science teachers, scientists, and evenmore informed citizens capable of reasoning critically about science Contract grant sponsor: NSF; Contract grant number: REC-99 80519; Contract grant sponsor: Howard Hughes Medical Institute; Contract grant number: HHMI 71100–503803. Correspondence to: C.L. Smith; E-mail: [email protected] DOI 10.1002/tea.20113 Published online in Wiley InterScience (www.interscience.wiley.com). 2006 Wiley Periodicals, Inc. issues. More specifically, if one’s epistemological understandings in a given field function as a kind of metacognitive control structure (Hofer, 2004), then they would guide one’s goals, reasoning, and sense making in situations in which they are activated. For example, they could guide how one structures firsthand inquiry (Sandoval, 2003; Windschitl, 2004), reasons about scientific controversies (Bell & Linn, 2002; Driver, Leach, Millar, & Scott, 1996; Sadler & Zeidler, 2004), conducts searches for reliable information on science topics (Hofer, 2004), approaches learning difficult science content (Hammer, 1994; Songer & Linn, 1991), and designs classroom inquiry experiences for students (Chinn & Malhotra, 2002). Although there is broad agreement that developing students’ epistemological understandings may be important as an educational outcome, there is less consensus about how these understandings should be conceptualized and assessed. One issue is the extent to which epistemological views are ‘‘tacit’’ versus ‘‘articulated.’’ Other issues concern what ‘‘grain size’’ it is useful to use in describing these understandings, theways that context affects them, the extent to which even the novice’s thinking (as well as the expert’s) is coherent and organized, and the extent to which epistemological understandings are domain general versus domain specific. At present, researchers have conceptualized student epistemological understandings as a domain-general developmental structure (King&Kitchener, 1994), as domain-general beliefs (Schommer, 1990), as domain-specific theories (Hofer, 2004; Hofer&Pintrich, 1997), and as highly situation-specific resources (Hammer & Elby, 2002). We believe that debate about these issues can be productive at this early point in charting out this complex terrain, especially if those with competing views take care to articulate their positions clearly, generate testable predictions from their views, and identify the important phenomena about student thinking and reasoning that any theory of epistemological reasoning needs to account for. Further, it might even be the case that each tradition is trying to understand an important ‘‘part of the elephant’’ and that we may ultimately have more comprehensive theories that integrate diverse views. The initial studies of epistemological development among college students concerned students’ personal epistemologies rather than their epistemologies of science and were conceptualized as a domain-general developmental structure (Baxter Magolda, 1992; King & Kitchener, 1994, Perry, 1970). Work in this area has been influenced by William Perry’s (1970) pioneering longitudinal studies of Harvard University students, with subsequent researchers attempting to further develop and refine his basic coding scheme. These schemes consider the extent of student recognition of the inherent uncertainty of knowledge, the role of viewpoint and knower in the knowing process, and the need to make commitments in the face of uncertainty and to identify some standards of argument that transcend individual perspectives. Researchers have identified a progression of student views of increasing complexity from (a) the naı̈ve view that knowledge is unproblematic, known by authorities with a high degree of certainty to (b) transitional radical relativist views in which all knowledge is uncertain and everyone uses evidence to support his or her own opinion to (c) more sophisticated views of knowledge as uncertain, affected by the interpretative perspectives of the knower, but in which knowledge claims can nonetheless be evaluated by standards that take account of and transcend individual frameworks. According to the longitudinal studies mentioned earlier, college freshmen typically still believe in the certainty of knowledge, although some are beginning to embrace radical relativism (Baxter Magolda, 1992; King & Kitchener, 1994; Perry, 1970). These studies paint a picture of limited progress in epistemological development during college, in which sophisticated epistemological positions still elude most college graduates (Baxter Magolda, 1992; Hofer & Pintrich, 1997; King & Kitchener, 1994). One limitation of this approach is that it is silent with respect to issues of important domainspecific differences in epistemological views or how particular kinds of educational experiences 2 SMITH AND WENK