Null Effects of Perceptual Disfluency on Learning Outcomes in a Text-Based Educational Context: a Meta-analysis

摘要

Recent research in a text-based educational context has demonstrated a seemingly paradoxical disfluency effect in reading, namely that learning with hard-to-read (disfluent) materials helps learners recall more details than learning with easy-to-read (fluent) materials. Many follow-up studies using a variety of participants, learning materials, and experimental designs have been conducted to verify the effects of disfluency manipulation on recall, transfer, judgments of learning, and learning time. However, a number of them have failed to replicate this effect and the mixed findings bring into question the generality of the disfluency effect with respect to learning. In this meta-analysis, we tested the overall effect of perceptual disfluency on learning with texts, as well as moderators of this effect, based on 25 empirical articles involving 3135 participants. Results showed that overall, there was no effect of perceptual disfluency on recall (d = - 0.01) or transfer (d = 0.03), but perceptual disfluency did reduce participants' judgments of learning (d = - 0.43) and increase learning time (d = 0.52). Tests of moderation focused on the most commonly studied dependent measure, namely recall. There was no evidence that characteristics of the participants, learning material, or experimental design moderated the effect of perceptual disfluency on recall. In general, though perceptual disfluency can be used as an effective metacognitive cue to reduce judgments of learning and increase learning time, there is not enough evidence to show that it either stimulates analytic processing or increases extraneous cognitive load.