Altman / Gill / McDonald Numerical Issues in Statistical Computing for the Social Scientist

1. Auflage 2004
ISBN: 978-0-471-47574-3
Verlag: John Wiley & Sons
Format: PDF
Kopierschutz: Adobe DRM (»Systemvoraussetzungen)

E-Book, Englisch, 324 Seiten, E-Book

Reihe: Wiley Series in Probability and Statistics

At last--a social scientist's guide through the pitfalls ofmodern statistical computing
Addressing the current deficiency in the literature onstatistical methods as they apply to the social and behavioralsciences, Numerical Issues in Statistical Computing for the SocialScientist seeks to provide readers with a unique practicalguidebook to the numerical methods underlying computerizedstatistical calculations specific to these fields. The authorsdemonstrate that knowledge of these numerical methods and how theyare used in statistical packages is essential for making accurateinferences. With the aid of key contributors from both the socialand behavioral sciences, the authors have assembled a rich set ofinterrelated chapters designed to guide empirical social scientiststhrough the potential minefield of modern statisticalcomputing.
Uniquely accessible and abounding in modern-day tools, tricks,and advice, the text successfully bridges the gap between thecurrent level of social science methodology and the moresophisticated technical coverage usually associated with thestatistical field.
Highlights include:
* A focus on problems occurring in maximum likelihoodestimation
* Integrated examples of statistical computing (using softwarepackages such as the SAS, Gauss, Splus, R, Stata, LIMDEP, SPSS,WinBUGS, and MATLAB¯®)
* A guide to choosing accurate statistical packages
* Discussions of a multitude of computationally intensivestatistical approaches such as ecological inference, Markov chainMonte Carlo, and spatial regression analysis
* Emphasis on specific numerical problems, statisticalprocedures, and their applications in the field
* Replications and re-analysis of published social scienceresearch, using innovative numerical methods
* Key numerical estimation issues along with the means ofavoiding common pitfalls
* A related Web site includes test data for use in demonstratingnumerical problems, code for applying the original methodsdescribed in the book, and an online bibliography of Web resourcesfor the statistical computation
Designed as an independent research tool, a professionalreference, or a classroom supplement, the book presents awell-thought-out treatment of a complex and multifaceted field.

Autoren/Hrsg.

Altman, Micah

Gill, Jeff

McDonald, Michael P.

Inhaltsverzeichnis

Preface.
1. Introduction: Consequences of Numerical Inaccuracy.
2. Sources of Inaccuracy in Statistical Computation.
3. Evaluating Statistical Software.
4. Robust Inference.
5. Numerical Issues in Markov Chain Monte Carlo Estimation.
6. Numerical Issues Involved in Hessian Matrices (Jeff Gill& Gary King).
7. Numerical Behavior of King's EI Method.
8. Some Details of Nonlinear Estimation (B. D. McCullough).
9. Spatial Regression Models (James P. LeSage).
10. Convergence Problems in Logistic Regression (PaulAllison).
11. Recommendations for Replication and AccurateAnalysis.
Bibliography.
Author Index.
Subject Index.

Über Autor(innen)

MICAH ALTMAN is Associate Director of the Harvard-MIT DataCenter in Cambridge, Massachusetts.
JEFF GILL is Associate Professor of Political Science atthe University of California, Davis.
MICHAEL P. McDONALD is Assistant Professor of Governmentand Politics at George Mason University in Fairfax, Virginia.