A LATENT PROCESS MODEL FOR MONITORING PROGRESS TOWARD HARD-TO-MEASURE TARGETS WITH APPLICATIONS TO MENTAL HEALTH AND ONLINE EDUCATIONAL ASSESSMENTS

• 成果类型: Article
• 署名作者: Jeon, Minjeong; Schweinberger, Michael
• 署名单位: University of California System; University of California Los Angeles; Pennsylvania Commonwealth System of Higher Education (PCSHE); Pennsylvania State University; Pennsylvania State University - University Park
• 刊物名称: ANNALS OF APPLIED STATISTICS
• ISSN/ISSBN: 1932-6157
• DOI: 10.1214/24-AOAS1873
• 发表日期: 2024
• 页码: 2123-2146
• 关键词: exponential-family models; space models; growth-model; inference

摘要

The recent shift to remote learning and work has aggravated longstanding problems, such as the problem of monitoring the mental health of individuals and the progress of students toward learning targets. We introduce a novel latent process model with a view to monitoring the progress of individuals toward a hard-to-measure target of interest and measured by a set of variables. The latent process model is based on the idea of embedding both individuals and variables measuring progress toward the target of interest in a shared metric space, interpreted as an interaction map that captures interactions between individuals and variables. The fact that individuals are embedded in the same metric space as the target helps assess the progress of individuals toward the target. We demonstrate, with the help of simulations and applications, that the latent process model enables a novel look at mental health and online educational assessments in disadvantaged subpopulations.