b4Math: a digital assessment tool for early symbolic numerical skills in kindergarten

Early symbolic numerical skills constitute a foundational component of later mathematical learning, yet rigorously validated assessment tools for preschool-aged children remain scarce in Latin American contexts. The present study aimed to develop and psychometrically validate b4Math, a digital screening tool designed to assess early symbolic numerical skills in children attending the final year of preschool (age 5). b4Math evaluates five domain-specific numerical processes: cardinality, counting sequence, number identification, symbolic number comparison, and ordinality, aligned with curricular objectives for early mathematics education. Data were drawn from two cohorts in Chile (total N = 357), collected across two waves (2023–2024). Construct validity was examined using confirmatory factor analysis (CFA). Four dichotomous-response tasks (cardinality, identification, comparison, and ordinality) were modeled using a correlated four-factor CFA with the WLSMV estimator, showing excellent fit and clear multidimensional structure. The counting-sequence task, yielding continuous responses, was analyzed separately using a single-factor CFA with robust maximum likelihood estimation (MLR), demonstrating high factor loadings and internal consistency. Measurement invariance across gender and temporal stability (test–retest) were supported. Internal consistency indices were satisfactory across subscales. External validity was established through associations between b4Math factor scores and mathematical achievement as measured by the Applied Problems subtest of the Woodcock–Muñoz Battery. Overall, results support the reliability and validity of b4Math as a culturally pertinent, digitally delivered assessment of early symbolic numerical skills. b4Math offers educators and researchers in Latin America an efficient tool for screening and monitoring foundational numerical competencies critical for later mathematics learning.