Universal screening measures are an integral component of any tiered system of instructional delivery. Recent studies of screening measures have often excluded examinations of bias in predictive validity. The present study examined a common screening instrument for evidence of bias in predictive validity across the four disaggregation categories of the No Child Left Behind Act. Performance of 3,805 students in Grades 1-3 on the Nonsense Word Fluency and Oral Reading Fluency measures of the Dynamic Indicators of Basic Early Literacy Skills were examined cross-sectionally in relation to a state criterion-referenced test. Bias in predictive validity was found, but varied by grade and by disaggregation category. Implications are discussed. (Contains 7 tables and 4 figures.)