Autonomy, Explanation and Evidence in Belief Revision

<ns7:p>Across three studies (n = 155) adolescent learners freely investigated a multivariable database to ascertain the extent to which their initial beliefs were affirmed or warranted revision. Our purpose was (a) to examine the strategies used (Studies 1-3), (b) to identify the respective roles of explanation and evidence in the process (Studies 1-3), and (c) investigate the role of autonomy (Studies 2 and 3). Participants were randomly assigned to one of three conditions. In Condition 1, explanations were solicited prior to and following data investigation. In Condition 2, explanations were solicited only following data investigation, and in Condition 3, no explanations were solicited. Participants completed an autonomy assessment pre- and post-random assignment to three conditions manipulating explanations. Consistent were the ways in which explanation influenced beliefs and benefitted belief revision. Only 8% in sample 1 and 10% in sample 2 applied the investigatory strategy of controlled comparison (COV) that is a staple of the middle-school science curriculum. Instead, they mostly used a less reliable strategy of examining the data broadly to identify favorable outcomes and then implicating antecedents often co-occurring with them. Accurate belief revision nonetheless occurred, slightly more successfully among the young adolescent sample, who made more frequent reference to the data in justifying their conclusions. Explanation facilitated belief revision in both conditions (1 and 2) in which they were solicited. We additionally find increased levels of reported autonomy from learners by condition, respectively. Individual differences in autonomy supporting belief revision is also found. Overall, explaining facilitated belief revision, all the while, underscoring the need to consider the impact of autonomy for adolescent learners confronted with disconfirming evidence prompting belief revision.</ns7:p>