The dataset contains all raw data and corresponding standard scores used for analyses in the research described by Walda, Weenk, Van Weerdenburg, and Bosman (in preparation). Also, corresponding edited scores of the Attention Concentration Test (ACT) are displayed. The raw scores of the ACT were edited using computer programs designed by the designers of the ACT. The manuscript addresses attention skill, as assessed with the ACT in children with and without dyslexia. For a full description of the reading and spelling remediation program, and all measures see the manuscript of Experiment 1 (Walda, Van Weerdenburg, Van der Ven, & Bosman, 2022). Data of Experiment 1 were previously deposited in DANS-EASY (https://doi.org/10.17026/dans-258-kq8c). An important finding of Experiment 1 is that many children with dyslexia (about 65%) were unable to finish 25 consecutive bars of the ACT at pre-test. Even at later moments of measurement within the nine months of the remediation program, when they had gotten more used to the task, many children failed to complete the ACT without making any error. Thus, regardless of whether they were exposed to the ACT training, children with dyslexia failed to complete the ACT and also hardly mastered the task of the ACT at a later moment. A prudent conclusion from the findings of Experiment 2 is that an adaptive and intensive training of the ACT lead to better results on all ACT measures (i.e., number of bars completed without making errors, working speed, and distraction time). More interestingly, literacy progress was not affected by improvement of attention skill, which raises the question whether this null-finding is specific for children with dyslexia or general for all children who are learning to read. In the present (third) experiment, we focused on a comparison between children with and without dyslexia in terms of attention skill. In the present experiment, three questions were addressed: 1. Do children with dyslexia differ from children without in terms of performance on ...