Children's Task-Switching Efficiency: Missing Our Cue?

Abstract

In simple rule-switching tests, 3- and 4-year-olds can follow each of two sorting rules but sometimes make perseverative errors when switching. Older children make few errors but respond slowly when switching. These age-related changes might reflect the maturation of executive functions (e.g., inhibition). However, they might also reflect children's ability to use task cues. Cue-processing difficulties predict switch costs in adult task switching (Logan & Schneider, 2007). It is unknown whether they explain children's task-switching errors or slowing. The current study tested whether inhibition, cue interpretation, or both predict 3- to 6-year-old children's switch-related errors (Experiment 1) and slowing (Experiment 2). Children performed a computerized task-switching test in which most trials were preceded by an audiovisual cue that instructed them to switch rules, or to stay—that is, continue using the current rule. Interspersed control trials used no cue. In Experiment 1, 3- and 4-year-olds made as many errors on cued stay trials as on cued switch trials; however, children were significantly more accurate on uncued stay trials. The presence of cues, not switching demands, predicted errors. Accuracy was predicted by children's speed in a simpler task in which children matched stimuli on only one dimension (shape or color), with no stimulus conflict or rule switches. Additional variance was predicted by an unrelated measure of processing speed. In Experiment 2, switch costs in 4.5- to 6-year-olds were similarly predicted by speed in the simpler unidimensional matching task.

Notes

¹Also, the stimulus word pairs cat and duck, and brown and blue, are similar in word length and phonological complexity.

²Although game asymmetries have been reported for switch-to-animal vs. switch-to-color trials in older children (Ellefson, Shapiro, & Chater, 2006), our design did not allow for in-depth analysis for switch asymmetries because there were only eight switch trials of each type and only correct trial RTs were analyzed. Preliminary within-subjects t tests did not reveal any significant game asymmetries; however, game was not included as a factor in any ANOVA.