Item Response Theory

Instructor: Myeongsun Yoon
Assistant Professor, Texas A&M University
email: myoon@tamu.edu
Phone: (979) 862-3515 (Yoon)
Course Time: May 21-23  [5/21 (1:30 to 4:30 pm); 5/22-5/23 (9:00 to 12:00 pm and 1:30 to 4:30 pm)]
Course Location:    Harrington Education Center Tower (HECC), Room 503

Roger Millsap distinguished himself as our 1 TRADITION at the TAMU Summer Statistics Workshop.  Over the past six years he made space in his private and professional life to be part of our group, providing cutting edge information about detection of bias in psychological measurement and models in Item Response Theory.  On Wednesday, May 7, Roger had a brain hemorrhage and on May 8 he passed away.  He was a gifted teacher.  More importantly, he was a terrific person, mentor, colleague, and someone who really cared about his family.  His former student and TAMU faculty member, Myeongsun Yoon, will teach the IRT Workshop.

This course will provide an overview of item response theory (IRT) and its application in psychological measurement.  We will begin with concepts and assumptions common to nearly all IRT models, such as the item response function, local independence and dimensionality.  We will then move to IRT models for binary test items (e.g., items scored pass/fail), covering model specification, estimation and fit evaluation.  Next, we will discuss some IRT models for polytomous response formats (e.g., Likert items), again focusing on specification, estimation and fit evaluation.  Major applications of IRT in adaptive testing, test construction and item bias analyses will be described.  For software, we will alternate between the IRTPRO, BILOG, and Mplus software programs.  The course will be built around IRTPRO, but Mplus will be used for dimensionality work with some mention of BILOG.

References:

Embretson and Reise text
Embretson, S. E. and Reise, S. P. (2000).  Item response theory for psychologists.  Mahwah, New Jersey:  Erlbaum.
deAyala text
de Ayala, R. J. (2008).  The theory and practice of item response theory . New York, NY:  Guilford.
Millsap text
Millsap, R. E. (2011).  Statistical approaches to measurement invariance.  New York, NY:  Routledge.

Software:

Two software packages will be used in this workshop, IRTPRO and Mplus 7.2.  Demonstrations will be based on the student versions of the software.  The student versions can be downloaded for free by clicking the links below and following the instructions.  In addition, some class demonstrations may reference the software programs BIOLOG and PARSCALE.  Trial versions of these programs are also available from the SSI website that hosts IRTPRO.