1. What is the research design (i.e. between or within subjects, what type of statistical tests were used, what were the levels of each variable)?

Results

As expected, the difficulty of the reading digit task increased with the pace. In the two youngest groups in which the number of errors and omissions was recorded, 47%, 12%, 3.5%, and 1% of the series of digits were incorrectly read at the paces of 2, 1.2, 0.8, and 0.4 digits per second, respectively. Of interest is the fact that71% of these incorrectly read series were followed by failures in recall, indicating that these reading errors were not due to a trade-off between processing and storage, with children favoring the memory task over the processing activity.An analysis of variance (ANOVA) was performed on the mean spans with age (8, 10, 12, and 14 years) and pace (2, 1.2, 0.8, and0.4 digits per second) as between-subjects factors. This analysis revealed a trivial effect of age with mean spans increasing smoothly with age (2.14, 3.05, 3.72, and 4.76 in 8-, 10-, 12-, and14-year-old children),F(3, 240)107.28,p.001,p2.57 (seeFigure 2A). As already observed by Barrouillet et al. (2004) in adults, there was a main effect of pace with faster paces eliciting lower mean working memory spans (4.79, 3.77, 2.93, and 2.18 for0.4, 0.8, 1.2, and 2 digits per second, respectively),F(3, 240)110.92,p.001,p2.58. More interesting is the fact that this effect was significant in each age considered separately, in 8-year-olds,F(3, 60)9.08,p.001; 10-year-olds,F(3, 60)33.93,p.001; 12-year-olds,F(3, 60)26.16,p.001; and 14-year-olds,F(3, 60)51.84,p.001. The linear trends accounted for96%, 99%, 99%, and 100% of the experimental effects in 8-, 10-,12-, and 14-year-old children, respectively, indicating that in each age, the increase in the number-of-digits-to-time ratio resulted in a smooth decrease in span as Barrouillet et al. (2004) observed in adults (see Figure 1). As we predicted, this effect of pace was more pronounced in older children, as testified by the significant inter-action between age and pace,F(9, 240)3.36,p.001,p2.11. Whereas the youngest group exhibited an increase in span of1.54 from the fastest to the slowest pace conditions (from 1.44 to2.98), this increase was 3.60 in the oldest children (from 2.94 to6.54).However, it can also be observed that these increases in span in the extreme groups were approximately in the same ratio (3.60/1.542.33) as their mean spans across paces (4.57/2.142.23). This was reflected by the results of an ANOVA performed on working memory spans after their log transformation (see Figure2B). Contrary to the effects of age,F(3, 204)91.97,p.001, and pace,F(3, 240)97.91,p.001, which were still significant, the interaction no longer reached significance,F(9, 240)1.31,p.20, suggesting that the effect of pace was proportionate to the span levels.