The Nature of the Refutation Text Effect: An Investigation of Attention Allocation

Undergraduates reading a refutation or expository text on seasonal change (exact n not obtained)

US undergraduates; topic the cause of the seasons - the canonical astronomy misconception (that the earth is closer to the sun in summer).

Reading-time study with conceptions measured at pretest, immediate post-test AND delayed post-test - one of the minority of refutation-text studies that bothered with a delay, which is why it is here despite being undergraduate-only and grade D. Sentence-level reading times were recorded to test the standard mechanistic story (that refutation text works because the refuted misconception captures attention and is read more slowly). Read at abstract depth; the delayed-post-test numbers were not obtainable from the sources we could reach.

The refutation-text group held fewer misconceptions at post-test. But the reading-time result went the OPPOSITE way to the standard mechanistic account: readers spent LESS time on the refutation paragraph than on the corresponding expository paragraph, which the authors interpret as attention being allocated differently rather than more - i.e. the refutation cue makes the conflicting idea easier to process, not harder. This matters for the archive because "increased attention to the refuted misconception" is the mechanism most often cited to justify refutation text, and here it is contradicted by the direct measure. Zengilowski et al. 2021 make exactly this point in their critique - that attentional measures in this literature are interpreted post hoc and inconsistently. The delayed post-test was administered but we could not extract its result; the wider review literature (Tippett 2010) reports that across such studies more readers revert to the misconception at delay than at immediate test.

Not applicable.

Not directly replicated. Later eye-tracking work on refutation text has produced mixed and paradigm-dependent attentional results, which is the substance of the Zengilowski et al. 2021 complaint.