Measurement and Methodological Challenges and Treatments in Large-Scale Database Analyses

Session Type: Paper Session

Abstract

The four papers in this session illustrate some common measurement and methodological challenges in studies using national and international databases and provide some analytical treatment and techniques to alleviate these difficulties. Researchers use these data to measure how test items could be biased due to testing modes (i.e., paper- vs computer-based assessments); explore if within school sample size could be reduced while keeping the precision of estimates; demonstrate how hybrid fixed effects model with lagged endogenous variables could increase the estimation precision of time-varying variables; and examine how contextual and personal characteristics can jointly relate to students’ advanced placement exam performance. The papers use a variety of analytic approaches, including Rasch modeling, 2PLM IRT modeling, and mixed effects models.

Sub Unit

• SIG-Advanced Studies of National Databases

Chair

• Olivia G Carr, Vanderbilt University

Papers

• PISA 2015: Cross-Mode Comparison Between Paper- and Computer-Based Assessments - Nnamdi Chika Ezike, University of Arkansas at Fayetteville; Wen-Juo Lo, University of Arkansas

• Sample Size and Design Considerations for School-Based Assessments - Martin Murphy, Australian Council for Educational Research

• The Impact of Occupational Needs for Education on Education Aspirations: A Hybrid Fixed Effects Model With Lagged Endogenous Variables - Lu Liu, University of La Verne

• Success in Context: How Student and School Characteristics Jointly Impact Advanced Placement Exam Performance - Carina M. McCormick, University of Nebraska - Lincoln

Discussant

• Haiyan Bai, University of Central Florida