Figures & data
Figure 1. Screenshots of the transcoding game (left), addition game (middle) and mole game (right).
Figure 2. Problem difficulty and child ability estimates by grade. The horizontal axis displays child ability and problem difficulty: both on the same scale. The further towards the right, the more able the child and the more difficult the problem.
Figure 3. The distribution of the proportion of inversion errors in different grades.
Figure 4. Results of the mediation analyses by grade: the relation between visuospatial working memory and addition ability, mediated by transcoding ability (upper figure) and inversion accuracy (lower figure). The sizes of the coefficients are presented adjacent to each path of the analysis. The middle graph displays the proportion of the total effect that is explained by the indirect path: .
Table 1. Results of t-tests investigating the difference in math ability between the poorest 10% at inversion and all other children.
Figure 5. Observed (line) and predicted (dotted) problem difficulties of all two-digit numbers.
Table 2. Results of multiple regression analyses with difficulty of items as dependent variable.
Figure 6. Illustration of the high speed high stake principle. With a time limit of 20 seconds, a correct answer after 5 s yields a score of 0.75; an incorrect answer after 12 s yields a score of −0.40.
