Misinformation : A Negative Consequence of Multiple-choice Testing

The Persisting Benefits of Using Multiple-Choice Tests as Learning Events Jeri L. Little (jerilittle@wustl.edu) Department of Psychology, Washington University in St. Louis Campus Box 1125, One Brookings Drive, St. Louis, MO 63130-4899 Elizabeth Ligon Bjork (elbjork@psych.ucla.edu) Department of Psychology, University of California, Los Angeles 1285 Franz Hall, Box 951563, Los Angeles, CA 90095-1563 research has indicated that multiple-choice testing may have a benefit for later overall performance that is not produced by the taking of prior cued-recall tests: namely, not only improving the retention of the information tested but also that of non-tested, but related, information—an outcome that would seem to be particularly desirable in educational contexts. When, for example, instructors give quizzes or practice tests to be followed later in the course by a more comprehensive exam, they are likely to ask questions about related information instead of (or in addition to) questions about the specific information tested earlier. Thus, it seems critical that the use of prior testing in educational contexts should help to improve the retention of both types of information. With respect to the question of how retrieving some information affects the later retention of non-tested related information, previous research has demonstrated that the effects can be negative—that is, answering a cued-recall question can lead to impaired recall of competitive related information on a later test, a finding referred to as retrieval- induced forgetting (Anderson, Bjork, & Bjork, 1994). Even in cases in which retrieval-induced forgetting does not occur, however, the retention of non-tested related information is rarely facilitated as a consequence of a cued- recall test, especially when such non-tested information has a competitive relationship with the tested information. (For exceptions to this finding when tested and non-tested related questions were created to be facilitative, see Chan, McDermott, & Roediger, 2006, and Chan, 2009.) In contrast to such findings for cued-recall tests, however, Little, Bjork, Bjork, and Angello (in press) recently demonstrated that—in addition to improving retention of previously tested information—taking an initial multiple- choice test, as compared to taking an initial cued-recall test, can improve retention of related information when the answers to the competitive related questions have occurred as incorrect alternatives in the initial multiple-choice test. Although these findings have clear implications for educational practice, particularly how an instructor might construct a practice exam, the procedure used by Little et al. employed a delay that is considered to lack educational realism (e.g., 5 min). Thus, a primary goal of the present research was to assess whether the demonstrated retention benefits (particularly for competitive related information) occurring as a consequence of taking multiple-choice Abstract Taking a test tends to improve the retention of the tested information. Additionally, taking a test often influences the later retention of non-tested information, provided such information is related to the tested information in a specific manner. To illustrate, recent research has demonstrated that multiple-choice tests containing competitive alternatives can improve retention of both tested and non-tested information pertaining to such incorrect alternatives at least over a short delay. The present research investigated whether such improvements in retention would persist with a delay more likely to occur in educational contexts (i.e., 48 hr). Taking an initial multiple-choice test improved retention more than a comparable cued-recall test—for both previously tested and related information—and over both short and long delays. Moreover, misinformation effects seen for the multiple-choice test at the short delay were reduced. These results thus have important implications for the use of multiple-choice tests as learning opportunities. Keywords: Testing; test effects; multiple-choice; retrieval- induced forgetting, retrieval-induced facilitation Introduction Taking a test does more than assess one’s knowledge: It can also improve one’s long-term retention of the tested information. Not all tests, however, are equally beneficial in this manner. For example, more open-ended tests (e.g., cued-recall), in general, have been shown to improve long- term retention more than multiple-choice tests (see meta- analyses by Anderson & Biddle, 1975; Hamaker, 1986). Moreover, some research has shown that taking multiple- choice tests can actually harm later performance on open- ended tests by exposing the test takers to misinformation in the form of incorrect, but plausible, answer choices, with the consequence of such information sometimes being intruded as incorrect responses to later cued-recall questions— findings referred to as misinformation effects (e.g., Roediger & Marsh, 2005). Thus, multiple-choice tests are often accused of not only being less effective for learning than are more open-ended types of tests, but also to risk negative misinformation effects, bringing their use as learning tools in educational contexts into question. Although, as indicated, multiple-choice tests can produce misinformation effects on later open-ended tests, research, in general, has shown that the positive effects of multiple- choice testing (as compared to no testing) outweigh any such negative effects they engender. Furthermore, recent