ConQuest - Generalized Item Response Modeling
Margaret L. Wu, Raymond J. Adams, Mark R. Wilson and Sam Haldane
ConQuest Version 2.0 is a computer program for fitting item response (Rasch) and latent regression models. It provides a comprehensive and flexible range of item response models (IRM) to analysts, allowing them to examine the properties of performance assessments, traditional assessments and rating scales. ConQuest 2.0 also offers the wider measurement and research community the most up-to-date psychometric methods of multifaceted item response models, multidimensional item response models, latent regression models and drawing plausible values.
ConQuest Version 2.0 Enhancements
ConQuest Version 2.0 has incorporated many enhancements to the 1998 release. These include:
plots of item characteristic curves user-defined fit statistics estimation of population characteristics such as percentages above a cut-point on a scale, more user-friendly interface
Models ConQuest Can Fit
Rasch’s Simple Logistic Model Rating Scale Model Partial Credit Model Ordered Partition Model Linear Logistic Test Model Multifaceted Models Generalized Unidimensional Models Multidimensional Item Response Models Latent Regression Models
How Does ConQuest Fit These Models?
ConQuest produces marginal maximum likelihood estimates for the parameters of the models. The estimation algorithms it uses are adaptations of the quadrature method described by Bock and Allen (1981) and the Monte Carlo method of Volodin and Adams (1995). The fit of the models is ascertained by generalizations of the Wright and Masters (1982) residual-based methods developed by Wu (1997). The manual contains a summary of these procedures.
Some Applications of ConQuest
Performing item analysis Examining differential item functioning Exploring rater effects Estimating latent correlations and testing dimensionality Drawing plausible values ConQuest is available with both a graphical user interface (GUI) and a simple command line, or console, interface. The console version of the program is available for all of the ConQuest platforms except Windows 3.1x. The GUI version is available for all Windows platforms.
Program Limits
Item response categories: 65 Item codes: 200 Number of generalised items for each case: 1000 Response model parameters: 3000 Latent dimensions: 15 Elements in each facet: 1000 Characters in a line of data and length of filename path: 3072 Lines specified in a formal statement: 20 Variables in a model statement: 1000 The number of item steps is not limited except by system memory.
Comparing ConQuest and Quest
Quest fits Rasch’s simple logistic model (for sets of dichotomously scored items), Andrich’s Rating Scale Model (for analyzing Likert scales) and Master’s Partial Credit Model (for analyzing polytomous items or mixtures of dichotomous and polytomous items). Quest also implements Mantel-Haenszel approaches to assess differential item functioning. ConQuest can fit a wider range of item response models than Quest. ConQuest can also fit both unidimensional and multidimensional latent regression models. Quest uses joint maximum likelihood to estimate parameters, whereas ConQuest uses marginal maximum likelihood. Quest is available for Macintosh OS, DOS, Windows 95 and Windows NT, but it does not have a GUI interface. ConQuest is available for Windows and Windows NT with a GUI interface and a console interface.
Books on Rasch Models and Related Methods
Applying The Rasch Model Multivariate and Mixture Distribution Rasch Models: Extensions and Applications Polytomous Item Response Theory Models Rasch Models for Measurement Rasch Models: Foundations, Recent Developments, and Applications
1.ConQuest 試用版立即下載
政府、教育、單位需索取報價單?
|