Winsteps 項目反應分析理論軟體

WINSTEPS^® & Facets Rasch Software

Dichotomous Analysis - Rasch - 30,000 items by 10,000,000 persons
Rating Scale Analysis & Extended Logistic Model - Andrich - up to 255 categories
Partial Credit Analysis - Masters - up to 255 categories per item
Psychometric and Educational test development
IRT Item Response Theory One Parameter Logistic Model 1PL 1-PL - Lord & Birnbaum
Polytomous - Polychotomous - Analysis - Andersen; Sequential Rasch Models - Glas etc.
Rasch Measurement & Fit Analysis - Wright; Many-facet Rasch Measurement - Linacre
Statistical Adjacent Logit, Transition Odds and Continuation Ratio Models
Item bias, DIF and DPF Analysis, Summary Statistics; Stochastic Guttman Biorder Analysis
Rasch Analysis Training Workshops; and more ...

■ Abstract

Social scientists have great need for the development of valid measures, e.g., of the quantity and quality of health services and of the outcomes of those services. Many researchers are frustrated when existing instruments are not well tailored to the task, since they then cannot expect sensitive, accurate, or valid findings. This workshop presents the theory and practice of classical test theory, the traditional approach. It then provides an overview of "modern" measurement as practiced using item response theory with a focus on Rasch measurement. Rasch analysis provides the social sciences with the kind of measurement that characterizes measurement in the natural sciences. Since item response theory focuses on the items and the persons rather than the test score, the synthesis of quantitative analysis with qualitative issues is experienced in a way that is rare in social science. Ultimately, Rasch measurement can facilitate more efficient, reliable, and valid assessment while improving privacy and convenience to users. The demonstration is useful for anyone who wants to understand the role of modern measurement in research.

■ Attendees will learn:

* Differences between Classical Test Theory and Rasch Measurement
* Why and how Rasch seeks to create linear, interval measures
* The inner workings of the Rasch model
* How to run Winsteps analyses
* Interpretation of Rasch/Winsteps output