Student Prior Knowledge in Studies of Learning
Research on learning can involve the evaluation of a number of variables observed in a model, and experiments in this vein apply to a variety of situations--studies of classroom instruction or learning from hypertext, for example. A recent study by Shapiro questions two methods that experimenters have historically used to control for one assumed confounding variable, prior knowledge of the learners (usually students). Major points of the study, that encompassed two laboratory experiments, are as follows:
- "One common approach to controlling for prior knowledge is assuming that all subjects are similar enough in background that their prior knowledge can be assumed to be equivalent. This is often the case in experiments on learning that employ subjects who are grade school children or adults enrolled in introductory courses...The practice of assuming equality between subjects' prior knowledge on the basis of a small number of shared factors, such as enrollment in a particular course, major, or year in school, is relatively common..." [pp.167-168] [Note: "Experiment 2" tested this approach.]
- "Yet another common strategy to control for prior knowledge is the use of fictional stimuli...This is sometimes done by taking texts about real events and changing specific names, locations, or dates and removing bits of information to make the content unfamiliar to readers...Alternatively, fully fictional texts are also sometimes used...Although it is possible to control for topic knowledge through fictional information, that is not the case for domain knowledge..." [pp.168-169] [Note: "Experiment 1" tested this approach.]
- [From Experiment 1,] "it can only be concluded that the use of fictional materials to control for prior knowledge, without actually assessing and including prior knowledge in the analyses is methodologically unsound..." [p.174]
- "Experiment 2 showed that topic knowledge is an important contributor to learning even among novices. As such, Experiment 2 casts doubt on the use of domain-novices to remove prior knowledge as a factor in learning outcomes research...Experiment 2 is informative because it indicates that, even among novices, (a) topic knowledge varies, and (b) it influences learning outcomes. Indeed, the variability in subjects' knowledge accounted for as much as 40% of the posttest results..." [p.181]
- "...current theories of text-based learning, [e.g., the construction-integration model, assimilation theory, schema theory, and the MDL]... all predict that prior knowledge will enhance learning and all attempt to explain the process through which that happens. All of these theories propose that meaningful understanding happens when prior knowledge is applied to new information. The results reported here support that view by demonstrating the degree of learners' sensitivity to prior knowledge..." [p.181]
- "Because the effects of prior knowledge are so profound and because the two methods of controlling for those effects were ineffective in the present study, the data underscore the importance of including prior knowledge measures in analyses of learning outcomes..." [p.182]
Amy M. Shapiro (Associate Professor of Psychology, University of Massachusetts, Dartmouth) documents her study in an article ("How Including Prior Knowledge As a Subject Variable May Change Outcomes of Learning Research") in the American Educational Research Journal (Vol.41, No.1, Spring 2004, pp. 159-189). The article includes a concise summary of related research on domain knowledge, topic knowledge, and more than two pages of references.
[Abstract done by Willard Hom, Director, Research & Planning Unit, System Office, California Community Colleges, 2/24/05]
|
