[2] Development of post-error slowing and the role of metacognitive control

Background

Over development, children become increasingly able to effectively regulate their thoughts and actions to attain their goals, reflecting greater cognitive control. For example, they become more able to sustain their attention in the classroom. While research has probed the development of cognitive control, there is still much to uncover about the extent to which children and adolescents have insight into their own thoughts and behaviour, i.e. their metacognition. Metacognition can be defined as thinking about thinking and comprises metacognitive knowledge (e.g. knowing that one can learn from mistakes) as well as metacognitive regulation, which itself encompasses metacognitive monitoring (e.g. noticing a task is difficult and we are making mistakes) and metacognitive control (e.g. adjusting our behaviour to do better on the task) (Nelson & Narens, 1990).

Metacognitive monitoring and control enable individuals to monitor and evaluate their performance, and adjust their behaviour if performance is not optimal. An example of engagement of metacognitive monitoring and control is the fact that after making an error on a task, people usually slow down, a phenomenon called post-error slowing (PES). While PES may initially be maladaptive and reflect capture of attention by an unexpected event i.e. the error, it can be followed by an adaptive increase in cognitive control engagement or adaptation of goal-directed behaviour, e.g. a change in strategy, to improve performance (Rabbitt & Rodgers, 1977; Wessel, 2018). For example, a child may use skip counting instead of retrieving a multiplication fact like 2 x 3 = 6. Adaptive PES is thought to reflect metacognitive monitoring —the error has been noticed— and metacognitive control, through a strategic, goal-oriented, increase in cognitive control engagement after an error is made. While most PES studies have used experimental paradigms such as choice-reaction time tasks, go/no-go tasks or Stroop tasks (Balogh & Czobor, 2016), recently PES has also been observed when children performed mathematics tasks, particularly on slow and/or more difficult tasks (de Mooij et al., 2022).

Research Questions / Hypotheses

This REP study was mainly for piloting the tasks that will be modified and used in the main study that broadly aims to answer the following questions - 1. What are the domain-specific and domain-general patterns of developmental differences in PES in childhood? 2. What are the predictors of individual differences in PES in childhood and adulthood?

Participants

7 participants completed all three tasks and 1 participant completed only two tasks. Preliminary results are described for 6 participants for Task 1 and 7 participants for Tasks 2 and 3.

Methods

The study piloted the following tasks, all involving a switching component very two trials between two rules. All stimuli used in the blocks were incongruent and there were 4 possible responses in each trial. 1. Numerical Stroop Task - requiring a choice response indicating the symbolic digit or the number of items as per the cue what number or how many respectively 2. Spatial Compatibility Task - requiring a choice response as per the orientation (upright or flip) of the pictures shown as per its colour or location respectively 3. Arithmetic Switching Task - requiring a choice response as per the arithmetic operation of addition or multiplication, having multiple choice answers as well Response times and accuracy were recorded for all the tasks to assess post-error slowing.

Results

The results for each of the tasks are described below. 1. Numerical Stroop Task - response times were between 722 ms to 1377 ms, accuracy was greater than 91.67%. Out of 6 participants, 3 had sufficient errors preceded and followed by correct trials to calculate post error slowing, ranging between -88 ms to 583 ms. 2. Spatial Compatibility Task - response times were between 437 ms to 1131 ms, accuracy was greater than 80%. Out of 7 participants, 6 had sufficient errors preceded and followed by correct trials to calculate post error slowing, ranging between - 427 ms to 1189 ms. 3. Arithmetic Switching Task - response times were between 939 ms to 1438 ms, accuracy was greater than 87.5%. Out of 7 participants, 4 had sufficient errors preceded and followed by correct trials to calculate post error slowing, ranging between -258 ms to 1073 ms. The plan for the main study analyses is to calculate post error slowing means and variability, and look for developmental and age effects.

Implications

Individual differences in metacognition impact learning and contribute to the educational achievement gap observed in many countries, including Australia. While the results will not have a direct impact on you personally, the outcomes of this project will help further our theoretical knowledge of children’s developing thinking skills and eventually could inform interventions to foster metacognition at school. The study is currently under piloting stage. To achieve these goals, findings related to your child’s data will be disseminated to the scientific community in the form of journal articles and conference presentations. In any publication, the information will be provided in such a way that you cannot be identified.