Ordinal Data Modeling (Statistics for Social Science and Public Policy)

Rating:
Summary: nice treatment with many examples from education
Review: This book provides both the Bayesian and classical approaches to ordinal data analysis but is unique in emphasizing the Bayesian approach and the latest advances. The authors are academic statisticians and the text is designed for a graduate level course for statistics or social science majors. It includes some very well written introductory material on these two forms fo statistical inference.

The mathematical level is intermediate but is written in a clear way to be accessible to social science students. This is also a good reference book for statisticians especially those involved in educational testing.

Markov chain Monte Carlo methods are provided along with some programmed algorithms for doing Gibbs sampling. A website is available to help the reader get access to datasets and software to implement the procedures.

Although the offer of software is nice, the authors neglect to mention the BUGS software that has been developed in the UK to handle MCMC problems. BUGS or the new window based WinBUGS is easily accessible to the reader and provides a lot of additional modeling aids including diagnostics.

The book covers a lot of interesting and applications oriented topics including logistic regression, ordinal regression, item response models, graded response models and the analysis of ROC curves. Concepts are illustrated and techniques demonstrated through real problems.

Rating:
Summary: nice treatment with many examples from education
Review: This book provides both the Bayesian and classical approaches to ordinal data analysis but is unique in emphasizing the Bayesian approach and the latest advances. The authors are academic statisticians and the text is designed for a graduate level course for statistics or social science majors. It includes some very well written introductory material on these two forms fo statistical inference.

The mathematical level is intermediate but is written in a clear way to be accessible to social science students. This is also a good reference book for statisticians especially those involved in educational testing.

Markov chain Monte Carlo methods are provided along with some programmed algorithms for doing Gibbs sampling. A website is available to help the reader get access to datasets and software to implement the procedures.

Although the offer of software is nice, the authors neglect to mention the BUGS software that has been developed in the UK to handle MCMC problems. BUGS or the new window based WinBUGS is easily accessible to the reader and provides a lot of additional modeling aids including diagnostics.

The book covers a lot of interesting and applications oriented topics including logistic regression, ordinal regression, item response models, graded response models and the analysis of ROC curves. Concepts are illustrated and techniques demonstrated through real problems.