Data Science in Educational Testing with Dr Yunxiao Chen
Facebook Event Link: https://www.facebook.com/events/802122663869060/
Thanks to : LSE Department of Statistics' Dr Yunxiao Chen
"This summer, GCSE and A-level exams were cancelled. The government created a rather controversial algorithm to award grades to students as around 40% of results were lower than the grades predicted by their teachers. In this speaker event, Dr Chen (Assistant Professor, LSE) will introduce statistical methods for grade prediction based on the educational testing system in the United States.* The talk will be divided into 3 sections: how to input missing data, what statistical framework to use and how to ensure fairness."
Recap
-
Item Response Theory is a statistical (latent factor) model that allows us to predict a students performance on unseen questions, accounting for the students' skill and difficulty of the question.
-
Differential Item Function accounts for differences in groups (ethnicity, gender), which informs us if the questions are fair. Under DIF, we would assume that the null hypothesis is that there are no differences between the groups in performance for the question.
Further Learning
MY455 Multivariate Analysis and Measurement: https://www.lse.ac.uk/resources/calendar2020-2021/courseGuides/MY/2020_MY455.htm
ST405 Multivariate Methods: https://www.lse.ac.uk/resources/calendar/courseGuides/ST/2019_ST405.htm
Lord, F.M. (1980). Applications of item response theory to practical testing problems. Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
30/10/2020