The benefits of a metacognitive lesson on elementary school children’s mathematics understanding

• In Event: Poster Session 02
  In Poster Session: PS 02 Topic - Cognitive Processes

Abstract

Metacognition is the ability to think about our own thinking and reliably predicts academic achievement (e.g., Ohtani & Hisasaka, 2018). Previous studies show that metacognitive training embedded in a topic can improve metacognitive skills in that target topic (Mevarech & Kramarski, 1997). However, most work has been done with older students and questions remain regarding the generalizability of metacognitive training across topics. In the current study, we experimentally tested the benefits of a metacognitive lesson on elementary school children’s understanding across three mathematics topics.

We worked with 138 children from 8 first- and second-grade classrooms. All students completed a pretest, lesson, and posttest. The pretest and posttest were administered in a whole-class setting and included 18 items across three topics: arithmetic (e.g., 7 – 4), place value (e.g., which is closer to 51: 54 or 49), and equivalence (e.g., 4+2=3+). Children solved each item and rated their certainty on a four-point scale (Figure 1). Metacognition scores were calculated based on the match between children’s accuracy and their certainty rating (Fyfe & Nelson, 2019).
For the lesson, children within each classroom were randomly assigned to one of two conditions: control or metacognitive. Children in both conditions were taught how to solve equivalence problems by a trained researcher in a classroom setting. Children in the metacognitive condition were also taught to use reflective questions as they solved problems: 1) what information is given (comprehension), 2) what steps do I need to take to get the right answer (strategy), and 3) how can I check that my answer is right (reflection).

Both lessons were equally effective at supporting children's ability to solve equivalence problems. Equivalence accuracy scores were higher at posttest than at pretest (68% vs. 55%), t(137) = 4.91, p < .001. Further, after controlling for pretest scores, posttest accuracy scores were similar for children in the metacognitive condition (M = 70%, SE = 3%) and children in the control condition (M = 66%, SE = 3%), F(1, 132) = 1.09, p = .29, ηp2 = .01.
To examine condition differences in metacognition, we conducted three ANOVAs – one for each topic (see Figure 2). In each model, we included condition as the between-subjects factor and posttest metacognition scores in a given topic as the dependent variable. We included pretest accuracy and pretest metacognition as covariates. There was a significant effect of condition for children’s metacognition on equivalence items, F(1, 129) = 4.09, p = .04, ηp2 = .03, but not for metacognition on arithmetic items, F(1, 129) = 0.9 , p > .05, ηp2 = .007, or place-value items, F(1, 129) = 3.9 3, p > .05, ηp2 = .03.

Children who received a metacognitive lesson had better metacognition and similar performance compared to children who received a control lesson. Consistent with prior research, these results suggest that brief metacognitive training can improve metacognition. We extended this work to elementary school children and demonstrated that these benefits appear to be topic-specific – the training only improved children’s metacognition on equivalence items.

Authors

• Collin A. Byers, Indiana University - Bloomington
  Presenting Author

• Emily Fyfe, Indiana University - Bloomington
  Non-Presenting Author