Early Identification of Children at Risk for Long-Term Difficulties with Mathematics Learning and Innumeracy in...

Children who start school behind in mathematics are at-risk of staying behind throughout schooling and into adulthood. Unfortunately, there is no consensus on the best criteria for the early identification of these at-risk students. Using a unique 1^st to 10^th grade study of academic development (n = 445), we assessed the sensitivity and specificity of different ability-achievement and achievement-only criteria for identifying risk of mathematics learning difficulties (MLD) and adolescent innumeracy. The outcome criterion was based on the risk of innumeracy in 10^th grade (< 25^th mathematics percentile). The achievement-only cutoffs were mathematics scores below the 10^th to 35^th percentile (at 5 percentile increments, inclusive) in 1^st grade and across 1^st and 2^nd grade, and teacher-reported in-class attentive behavior. The best combination of sensitivity (.589) and specificity (.878) was found for an achievement cutoff below the 25^th percentile in 1^st grade. We then examined growth in mathematics from 1^st to 10^th grade for different at-risk and typically achievement groups using latent score change (LCS) models. The models revealed that at-risk students were behind their peers in 1^st grade, with a widening gap across the school years. However, not all at-risk students were identified indicating that additional factors need to be considered in follow-up studies. Until that time, low-achieving 1^st graders, independent of cognitive ability, attentive behavior, and parental socioeconomic status, are at risk for long-term difficulties in mathematics and innumeracy in adolescence.