A meta-analysis and systematic review of reactivity to judgements of learning

Abstract

Judgements of learning (JoL) are often used in memory research as a means for assessing an individual's metacognitive beliefs about their learning. JoL have been shown to reliably predict performance as well as learning behaviours and decisions . Participants may, however, modify their behaviour in response to performing JoL. There has, however, been little consensus as to the reliability and direction of the effect. We report on a meta-analyses that assesses the evidence that memory performance is reactive to JoL. The results indicate that overall providing JoL does not have a significant effect on memory performance (g = 0.054, 95% CI -0.027 to 0.135). However, sub-groups analysis showed that this effect depends on the nature of the stimuli to be recalled, with moderate positive reactivity observed for related word pairs (g = 0.323, 95% CI 0.083 to 0.563) and word lists (g = 0.384, 95% CI 0.146 to 0.622) but no reactivity when pairs were unrelated or a mixture of related and unrelated pairs. These results indicate that researchers should be aware that eliciting JoL may well influence participants' underlying encoding processes, especially when using related word pairs or word lists.

A role for metamemory in cognitive offloading

“…First, the act of making offloading decisions during learning in Experiments 1 and 2 might have affected the processing of semantic associations for easy word pairs. Previous research has shown that the involvement of metamemory judgments affects memory process selectively for strongly-related word pairs (Double et al, 2018, Soderstrom et al, 2015). For example, Soderstrom et al (2015) asked participants to learn a list of strongly and weakly related pairs.…”

The Effect of Confidence Rating on a Primary Visual Task

“…The improvement of the Confidence Rating group in accuracy is congruent with previous studies on metacognition. The positive effect of metacognition-measures assessment, termed positive reactivity, was reported to be caused by Judgment of Learning in mathematical problem solving (Desoete and Roeyers, 2006) reading (Pressley, 2002) learning from visual items (Prins et al, 2006) and more (for a review see Double et al, 2018). Lately, Double and Birney (2017) showed positive reactivity to confidence rating in Raven’s Progressive Matrices.…”

“…They found that delivering confidence rating after each experimental trial significantly improved task performance, constructing positive reactivity. However, the underlying mechanisms of the reactivity effect remain unclear, and reactivity effects on early perception are yet to be studied (Double et al, 2018). Metacognitive reactivity effect is congruent with the literature on task formation, according to which targeted variations in task demands result in adoption of dissimilar cognitive strategies, yielding differences in following performance.…”

Reactivity to Measures of Metacognition

“…Here, we present a review that examines the evidence that measures of metacognition 65 typically used in the research literature are reactive. While recent meta-analyses have provided some 66 clarity as to the overall size of reactivity effects (Double, Birney, & Walker, 2018;Fox, Ericsson, & 67 Best, 2011; Rhodes & Tauber, 2011), we aim to provide a more nuanced review focusing on the 68 variables that determine the conditions under which reactivity does and does not arise. To this end, 69…”