This article examines the validity of IQ-achievement discrepancy and low-achievement as criteria for the identification of disabilities in Spanish-speaking ELs and the factors that moderate the validity of these approaches as bases for identification. While there has been a long history of examining the validity of different approaches to disability identification in monolinguals, there are no systematic approaches taken for ELs. Data from Grades 1 and 2 of a large longitudinal data set consisting of young Spanish-speaking students attending schools in the U.S. were used to empirically examine criteria for disability identification among language minority children - one of the first large-scale attempts. Findings indicated significant overidentification when the language of assessment was not matched to the language of the instruction, although effects varied predictably over time and by language of instruction. Validation of classifications using measures external to the classification found that low achieving and discrepant children differ from typically developing children, and from one another in predictable ways based on differences in IQ. The study highlights the importance of taking into account the language of instruction and the severity of the cut-off to reduce misidentification of typically developing children. [This paper was published in "New Directions for Child and Adolescent Development" n166 p111-143 2019 (EJ1221827).]