Blogs (3) >>
Sat 23 Mar 2024 11:10 - 11:35 at Meeting Rooms B113-114 - BPC Chair(s): Michael Miljanovic

Assessments that can measure student understanding of concepts in a reliable and valid way are incredibly valuable in research. Unfortunately, assessments can be a source of bias, differentially impacting students along various demographic lines. Differential Item Functioning (DIF) is a method to explore assessment bias. However, DIF is primarily limited to a single binary demographic variable (e.g. white and non-white; male and female). In this paper, we describe a novel expansion of DIF methods to explore intersections of student identities. We demonstrate the use of classic DIF on a data set of 255 complete responses to a CS1 assessment using binary race and gender variables in our analyses. Then, we present the importance of intersectional DIF by running a similar analysis on intersectional data. Using these methods, we identify problematic items on the assessment that bias against certain groups of test-takers. Our work contributes an innovative method to help interpret assessment results and inform changes to assessments.

Sat 23 Mar

Displayed time zone: Pacific Time (US & Canada) change

10:45 - 12:00
BPCPapers at Meeting Rooms B113-114
Chair(s): Michael Miljanovic Ontario Tech University
10:45
25m
Talk
Fostering Race-Conscious Literacies in Computer Science Teacher EducationMSI
Papers
Sukanya Kannan Moudgalya University of Tennessee, Knoxville
DOI
11:10
25m
Talk
Intersectional Biases Within an Introductory Computing AssessmentMSI
Papers
Miranda Parker San Diego State University, He Ren University of Washington, Min Li University of Washington, Chun Wang University of Washington
DOI
11:35
25m
Talk
U.S. Latines in Computing: A Literature ReviewMSI
Papers
Ismael Villegas Molina University of California, San Diego, Audria Montalvo University of California, San Diego, Adalbert Gerald Soosai Raj University of California, San Diego
DOI