Skip to main content

CSoI Student Interdisciplinary Team Selected to Present at the Modern Modeling Methods Conference

The Center's most recently formed student-led research team has continued to collaborate and make progress on their joint research topic. Their research title, "Using Markov Chain Monte Carlo Methods and Bayesian Estimation to Investigate Cross-National Teacher Preparedness and Professional Practices", has been selected to be presented at the Modern Modeling Methods conference in 2021. Team members and abstract below:

Title: Using Markov Chain Monte Carlo Methods and Bayesian Estimation to Investigate Cross-National Teacher Preparedness and Professional Practices 

Authors and Affiliations: 

  • Genéa Stewart, Ph.D. Student, Department of Educational Psychology, The University of North Texas 
  • Cary Jim, Ph.D. Student, Department of Information Science, The University of North Texas 
  • Jaret Hodges, Ph.D. Department of Educational Psychology, The University of North Texas 
  • Milushka Elbulok-Charcape Ph.D. Student, Department of Educational Psychology, The City University of New York 

Other team members with past contributions include: Tessa Pham, Bryn Mawr College, Dinuka Gallaba, Southern Illinois University, and Debjani Sihi, Oakridge Laboratories

Abstract

Oftentimes, social science and policy researchers wish to examine how various macro-level country factors relate to differences in outcomes. However, many secondary datasets only offer small cluster sample sizes for multi-country data which severely restricts researchers’ ability to estimate multi-level effects on outcomes and draw any robust conclusions. Bryan and Jenkins (2013) recommended the use of Bayesian estimation methods for more reliable estimations when the country level sample size is smaller than traditionally recommended 30. Using cross-national data from the 2018 International Questionnaire of the Teaching and Learning International Survey (TALIS), we will investigate model fitting of variance components for teachers’ preparation, instructional practices, and beliefs across 15 countries. We will demonstrate this method using the Bayesian Regression Models (brms) package in R. This project is supported by Purdue University’s Information Frontiers Learning Program at the Center for the Science of Information, a National Science Foundation Center. 

 

View the Team's Project page