An Introduction to Item Response Theory (IRT), IRT-PRO, and Their Use in Social Work Research

Friday, January 16, 2015: 2:30 PM-4:15 PM
Balconies N, Fourth Floor (New Orleans Marriott)
Cluster: Research Design and Measurement
William R. Nugent, PhD, University of Tennessee, Knoxville
Item Response Theory (IRT) is an extension of classical measurement theory (CMT) derived by adding a set of strong assumption to those of CMT.  Researchers can use IRT to accomplish measurement objectives unattainable with CMT, such as adaptive and tailored measurement procedures.  IRT also provides a powerful approach to differential item functioning (DIF) studies.  DIF concerns the possibility items work differently for different groups of persons, such as males and females.  The purpose of this workshop is to introduce participants to: (1) the fundamentals of IRT; (2) use of the program IRT-PRO for doing IRT analyses; and (3) use of IRT and IRT-PRO for conducing DIF analyses.  Upon completion of the workshop participants will have (a) a basic understanding of IRT; (b) basic skills in using IRT-PRO for conducting IRT analyses of data from measures used in social work research and practice; and (3) sufficient knowledge of IRT and IRT-PRO to conduct DIF analyses for both dichotomously scores items and Likert type items.

Workshop participants will be introduced to basic IRT concepts such as the item response function and item characteristic curve, item and test information curves, and the test characteristic curve using (a) didactic methods, (b) illustrative examples, and (c) demonstrations of use of IRT-PRO for various IRT analyses.  Participants will learn the most important concepts in IRT while simultaneously learning how to use IRT-PRO to conduct analyses that produce results demonstrative of these IRT concepts.  As each fundamental IRT concept is presented and discussed, participants will be taken step-by-step through data analyses, using IRT-PRO, that produce results exemplifying the fundamental concept.  These illustrative results will be used to help participants understand the importance of the IRT concepts for the development and assessment of measures for use in research and practice

Two illustrative analyses will be used.  The first is an IRT analysis of scores on the Geriatric Depression Scale-Short Form (GDS-SF).  This will illustrate an IRT analysis of scores from a measure with dichotomously scored items and use of the two-parameter logistic item response function.  The second is an analysis of scores from the Ultra-Short Suicidal Ideation Scale (USSIS).  This will give participants the opportunity to go step-by-step through an analysis of scores from a Likert type scale using the graded response IRT model.  They will also learn how to obtain graphic representations of item characteristic curves, item and test information curves, and the test characteristic curve; how to obtain results of specific tests of DIF and how to graphically represent these results; and how to interpret, present, and write up these various results.

Participants will leave the workshop with sufficient knowledge to conduct a variety of IRT analyses, using IRT-PRO, in their research and scale development and testing.  They will also have knowledge on how to present and interpret the results of IRT analyses in manuscripts and presentations of their research.  Finally, there will be discussion on how to teach IRT principles to students in classes that the participants teach.

See more of: Workshops