Online Program

Return to main conference page
Saturday, February 16
Sat, Feb 16, 8:00 AM - 9:15 AM
St. James Ballroom
Poster Session 3 and Continental Breakfast

Benefits of Using Item Response Theory with Unvalidated Questionnaire Data (303882)

View Presentation View Presentation

Sara Anvari, Baylor College of Medicine 
Carla Davis, Baylor College of Medicine 
Charles Minard, Baylor College of Medicine 
*Kristen Staggers, Baylor College of Medicine 
Daisy Tran, Texas Children's Hospital 

Keywords: Item response theory, IRT, questionnaire, latent variable

Questionnaires that have not been formally validated are frequently used for clinical and educational research to measure a latent variable, those that cannot be measured directly such as knowledge, quality of life, or satisfaction. Item response theory (IRT) is a method for both test scoring (measuring the latent variable) and development of accurate and reliable tests or questionnaires. It is already widely used in education (e.g., SAT and GRE test development) and psychology (e.g., ability assessment). IRT is also becoming more popular in other fields, such as clinical research, quality of life research, and marketing. IRT can be used post-hoc with an un-validated questionnaire to inform future questionnaire development and latent variable scoring.

An un-validated questionnaire was given to school personnel to determine their apprehension related to stock epinephrine. IRT will be used to create item characteristic curves and information functions to see which items can be used to more accurately and reliably measure the latent variable, apprehension. Estimation of apprehension using IRT will be compared to the sum of scores method. Advantages of using IRT will be described.