|
Activity Number:
|
79
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 4, 2003 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Government Statistics
|
| Abstract - #300283 |
|
Title:
|
Applications of Item Response Theory for Questionnaire Evaluation
|
|
Author(s):
|
Bryce B. Reeve*+ and Louise Masse
|
|
Companies:
|
National Cancer Institute and National Cancer Institute
|
|
Address:
|
Outcomes Research Branch, Bethesda, MD, 20892-0001,
|
|
Keywords:
|
item response theory ; questionnaire evaluation ; psychometrics ; confirmatory factor analysis ; classical test theory
|
|
Abstract:
|
Using Classical Test Theory (CTT) to evaluate the psychometric properties of questionnaires provides a limited understanding of the behavior underlying item responses and scale performance. Item Response Theory (IRT) provides a more powerful framework for assessing a questionnaire's psychometric properties. IRT modeling offers an instrument developer information about how each item functions within a scale, and how the scale functions over the continuum of the construct measured by the questionnaire. This knowledge integrates quantitative and qualitative analysis to help the developer select the appropriate items for a questionnaire based on measurement properties and item content. This paper will illustrate the added information that IRT provides over CTT in evaluating a questionnaire. As a didactic example, responses to the SF-36 Mental Health Summary Scale used in a breast cancer survivor study will be evaluated for its psychometric properties. First, the psychometric properties of the scale will be evaluated with CTT techniques followed by a more powerful psychometric evaluation using confirmatory factor analysis and IRT.
|
