Figure 1.
Example of a training task item with two memory/math cycles.
Figure 2.
Log of math/memory cycles per session as a function of log of cumulative math problems solved in the context of the working-memory training task.
The line represents the best-fitting linear function, which in log/log space is equivalent to the power function with untransformed scores.
Table 1.
Intercorrelations among all pretest measures and Session-1 performance in the training task for the experimental group.
Table 2.
Raw mean scores (SD) for all pretest and posttest measures, separately for the training and the experimental group.
Figure 3.
Gains in raw scores for all transfer measures for the control and the intervention group, estimated through the multi-level model, as well as t-values and corresponding effect sizes (d) of the between-group differences.
Figure 4.
Bars represent empirical effect sizes for each transfer measure and horizontal lines represent the estimate of the expected effect size based on the relation between session-1 training-task performance and pretest transfer-task performance.
Both empirical and estimated effect sizes are shown in reference to the pretest standard deviation of the training group for each measure. The purpose of these effect sizes is to allow a comparison between expected and actual effects; standard effect sizes are shown in Figure 3. In each case, a positive effect size reflects a transfer benefit.
Figure 5.
t-values representing the strength of the relation between individual students' total gains in the training task and pretest-posttest gains in each of the transfer tasks.
In each case, a positive t-values indicates that a larger learning rate is associated with greater transfer benefits.