Knowledge Revision Processes in Refutation Texts

Abstract

In the present set of experiments, we systematically examined the processes that occur while reading texts designed to refute and explain commonsense beliefs that reside in readers' long-term memory. In Experiment 1 (n = 36), providing readers with a refutation-plus-explanation of a commonsense belief was sufficient to significantly reduce disruption during reading caused by the commonsense belief. In Experiment 2 (n = 36), the refutation alone reduced but did not eliminate the disruption during reading caused by the commonsense belief. However, in Experiment 3 (n = 36), the explanation alone was as effective as the refutation-plus-explanation in reducing disruption during reading. Finally, in Experiment 4 (n = 73), the refutation-plus-explanation manipulation not only reduced disruption during reading caused by the commonsense belief, it also produced long-term learning outcomes. Findings are discussed in the context of the Knowledge Revision Components framework.

Notes

This article is based on Panayiota Kendeou's 2013 Tom Trabasso Young Investigator Award talk at the 2013 Society for Text and Discourse Annual Meeting in Valencia, Spain.

¹ Only F₁ tests against an error-term based on participant variability were conducted for the eight-plus and misconception groups across experiments. This is because not all items appeared in each condition an equal number of times for this subset of participants.

² The degrees of freedom for the delayed test participant analysis are reduced because six students did not return for the delayed post-test session.

³ Potential guessing actually underestimates the extent to which participants possess the misconception. Participants guess when they do not know the answer or when they have the misconception. This in turn inflates the correct responses, so the frequency of the misconceptions in the sample is likely higher than what was obtained, not lower.