Weighting Solution: Verifying SEM Trees for National Data

• In Event: AERA Roundtable Session Friday 9:50 am Four Seasons Ballroom 1
  In Roundtable Session: Nuanced Approaches to Informing Policy Implementation, SEM Tree Modelling, and Factors Impacting Opportunities to Learn (Table 3)

Abstract

Structural Equation Model Trees (SEM Trees) combine Structural Equation Modeling and Decision Trees to analyze complex data but struggle to incorporate sampling weights essential for national data. This study addresses this gap by validating the unweighted SEM Tree results using the R package MplusTrees, followed by a weighted analysis in Mplus. We offer a comprehensive validation process and demonstrate its utility using national data from the High School Longitudinal Study of 2009 (HSLS:09). The proposed method enhances the accuracy and generalizability of research with complex sampling surveys, enabling researchers to advance current theories by reexamining distinct patterns within national data using SEM Trees.

Authors

• Chi-Ning Chang, Virginia Commonwealth University

• Xun Liu, Virginia Commonwealth University

• Yuyan Xia, University of Kentucky

• Tzu-Wei Wang, Virginia Commonwealth University