The current paper investigates the structural validity of the Children's Concentration and Empathy Scale (CCES), which was designed based on Montessori's developmental theory to assess, using teacher ratings, the coherence of attention-related characteristics (concentration, empathy, and normalization) in children from 1.5 to 12 years old. The dimensionality analysis was carried out using exploratory graph analysis (EGA), and the stability of the items and factors were checked using a bootstrap version of EGA, and the results contrasted to exploratory factor analysis. The results point to a four-factor structure (emotion regulation, fantasy, task engagement, and empathy) after eliminating items with low replicability across bootstrapped samples. Beyond pointing to ways to improve the CCES, our paper presents a number of data analytical strategies that can be useful for studies investigating the structural validity of measurement instruments and demonstrates how EGA can effectively be used in the scale construction and validation process. Our manuscript and results are fully reproducible and are available on the Open Science Framework.