Cognitive Skill Assessments for Foundational Literacy and Numeracy in the Native Language

Foundational learning assessments in Indian government schools often report whether children can answer curriculum-linked items, but provide limited diagnostic insight into why children struggle, especially in multi-level classrooms. This pilot will test the feasibility, fidelity, usability, and measurement properties of CLASS (Cognitive Learning Ability Screening Score), a modular, audio-supported assessment that generates immediate composite and domain-level profiles (literacy, numeracy, working memory) in Hindi. CLASS has been iteratively developed and tested with 750 children; this pilot focuses on school implementation and teacher-facing use. In 15 Uttar Pradesh government primary schools within one district of Varanasi (across 2–3 blocks), teachers will administer CLASS to 900 students in Grades 2–5 in one-student-one-tablet sessions scheduled to minimise disruption. We will track completion, time per module/child, connectivity interruptions, protocol fidelity, teacher burden, and classroom disruption. We will assess teachers' ability to interpret reports through comprehension tasks, vignettes, and interviews, and refine dashboards and operating procedures for scalable delivery. To anchor measurement, we will administer an independent ASER-like custom short module and examine convergent patterns with CLASS domains and teacher perceptions. The pilot will estimate outcome variances and intra-cluster correlations, and produce an RCT-ready package to evaluate CLASS-informed teacher support and remediation.