Presentation Title

Bayesian Hierarchical Probit Regression Model

Faculty Mentor

Valerie Poynor

Start Date

23-11-2019 10:45 AM

End Date

23-11-2019 11:30 AM

Location

252

Session

poster 4

Type of Presentation

Poster

Subject Area

physical_mathematical_sciences

Abstract

General Education (GE) courses offered at California State University, Fullerton strive to provide students with important critical thinking skills necessary to enter the global workforce. In order to assess the effectiveness of these courses we looked to Math 338 courses which cater to a high volume of STEM and Non-STEM students. Faculty members have their own style, methods and approaches to teaching giving argument for including a "teacher effect" in our model in a hierarchical fashion. It is important to note that our model does not rate or give context to the effect of the teacher but rather account for the many reasons why students may score differently depending on which teacher they have. Each student was given the same problem on an exam and the option to complete a survey of important individual information such as age, family income/household, ethnicity and work. We analyze these data using a Bayesian Hierarchical Probit Regression model that exploits latent variables for computational convenience. Our results show that gender, being a first generation college student, and ethnicity were all significant factors in determining critical thinking score.

This document is currently not available here.

Share

COinS
 
Nov 23rd, 10:45 AM Nov 23rd, 11:30 AM

Bayesian Hierarchical Probit Regression Model

252

General Education (GE) courses offered at California State University, Fullerton strive to provide students with important critical thinking skills necessary to enter the global workforce. In order to assess the effectiveness of these courses we looked to Math 338 courses which cater to a high volume of STEM and Non-STEM students. Faculty members have their own style, methods and approaches to teaching giving argument for including a "teacher effect" in our model in a hierarchical fashion. It is important to note that our model does not rate or give context to the effect of the teacher but rather account for the many reasons why students may score differently depending on which teacher they have. Each student was given the same problem on an exam and the option to complete a survey of important individual information such as age, family income/household, ethnicity and work. We analyze these data using a Bayesian Hierarchical Probit Regression model that exploits latent variables for computational convenience. Our results show that gender, being a first generation college student, and ethnicity were all significant factors in determining critical thinking score.