Using Item Response Theory and Post Test Interviews to Improve Assessment in Large General Chemistry Classes

Abstract

In the University of Georgia's Freshman Chemistry program we have assessed students for seven years using computerized tests. All of the students' responses are stored in our server's database. In 2005, we began analyzing student performance on our assessments using Item Response Theory (IRT). From our IRT analysis we have learned which chemical topics are the most difficult for our students. Based upon that knowledge we have modified our instruction slightly and detected an increase in student performance data. Concurrently, we have performed post test interviews of approximately 80 students. In these interviews we asked students with varying IRT ability levels to show us how they approached certain problems. From the interviews, we have gleaned stumbling blocks that impede performance, particularly for lower ability students. We will present IRT and interview data as well as student performance data in our talk.

Location

Room 2903

Share

COinS
 
Nov 1st, 3:00 PM Nov 1st, 3:45 PM

Using Item Response Theory and Post Test Interviews to Improve Assessment in Large General Chemistry Classes

Room 2903

In the University of Georgia's Freshman Chemistry program we have assessed students for seven years using computerized tests. All of the students' responses are stored in our server's database. In 2005, we began analyzing student performance on our assessments using Item Response Theory (IRT). From our IRT analysis we have learned which chemical topics are the most difficult for our students. Based upon that knowledge we have modified our instruction slightly and detected an increase in student performance data. Concurrently, we have performed post test interviews of approximately 80 students. In these interviews we asked students with varying IRT ability levels to show us how they approached certain problems. From the interviews, we have gleaned stumbling blocks that impede performance, particularly for lower ability students. We will present IRT and interview data as well as student performance data in our talk.