Unreliable diagnoses (e.g., based on inconsistent criteria, subjective) may be inaccurate and even inequitable. This study uses an event history approach with yearly child- and school-level data from 378,919 children in a large urban school district in the southwestern United States between 2006–2007 and 2011–2012 to investigate contextual reliability in the designation of cognitive health conditions (e.g., autism, learning disabilities). This study’s findings suggest the likelihood of designation is higher in schools with more resources (higher teacher-to-student ratio, student population with more resources at home, charter school or magnet program), controlling on student-level differences. Cross-level interactions suggest children’s likelihood of designation also may be higher if they are distinctive relative to other students in their school, sometimes even in terms of nonclinical qualities (race, English Learner status).
A discrete time event-history approach to informative drop-out in mixed latent Markov models with covariates
