Since only one test showed a statistically signicant improvement, performance was considered overall to be stable, and, to give the best estimate of each subject’s pre-training status, baseline and pre-test scores on each variable were pooled into a variable called pre-training.