My research has focused primarily on improving large scale educational assessments, including issues of reliability, validity, and test security. Today, one main challenge is to develop theories and methods for the wide-ranging implementation of computerized assessment. More recently, I have been concentrating on developing item selection methods for computerized adaptive testing (CAT). Several new strategies have been developed, such as alpha-stratification (i.e., Chang, Qian, & Ying 2001), global information (1996), and constraint-weighted information (Cheng and Chang, in press). I try to tackle research problems that emerge from real world applications. For instance, one problem was to identify a design flaw in the computerized testing system that failed to generate reliable scores for thousands of GRE and GMAT test takers from 2000 to 2002. I and Zhiliang Ying at Columbia University proposed a theory, and presented both analytical and empirical evidence to support our hypothesis (Chang & Ying 2007). This research effort has been recognized by the 2008 Annual Award of National Council on Measurement in Education (NCME).
I am a practitioner turned professor. I spent nine productive years in the testing industry before moving to academia in 2001. From 1992 to 1999, I was a Research Scientist at Educational Testing Service (ETS), Princeton, NJ. From January 1999 to August 2001, I held the position of Senior Psychometrician and Director of Computerized Testing Technological Research at the National Board of Medical Examiners (NBME), Philadelphia, PA.