Contextual Analysis (Quantitative Applications in the Social Sciences) [Paperback]

Gudmund R. Iversen (Author)

Book Description

ISBN-10: 0803942729 | ISBN-13: 978-0803942721 | Publication Date: August 14, 1991 | Edition: 1

The effective use of contextual analysis - the study of the role of group context on actions and attitudes of individuals - often involves overcoming a number of issues such as group boundaries, the mobility of individuals within a group, overlapping groups, missing individual data and the choice of statistical models. This straightforward volume offers researchers a guide for selecting the best model to use, and discusses such topics as contextual analysis with absolute effects, with relative effects and the choice between regression coefficients as fixed parameters or as random variables.

Product Details

• Paperback: 84 pages
• Publisher: Sage Publications, Inc; 1 edition (August 14, 1991)
• Language: English
• ISBN-10: 0803942729
• ISBN-13: 978-0803942721
• Product Dimensions: 8.4 x 5.4 x 0.2 inches
• Shipping Weight: 3.7 ounces (View shipping rates and policies)
• Average Customer Review: 4.0 out of 5 stars  See all reviews (1 customer review)
• Amazon Best Sellers Rank: #2,360,419 in Books (See Top 100 in Books)

More About the Author

Discover books, learn about writers, read author blogs, and more.

Customer Reviews

Most Helpful Customer Reviews

4.0 out of 5 stars Contextual effects, July 1, 2009

By 

not a natural "Bob Bickel" (huntington, west virginia United States) - See all my reviews
(VINE VOICE)   

Amazon Verified Purchase

This review is from: Contextual Analysis (Quantitative Applications in the Social Sciences) (Paperback)

This is a useful, well-written book by one of the the first authors to explain procedures for including contextual effects in multiple regression models (see Boyd and Iversen, 1979). Iversen departs from what has become convention in textbooks discussing contextual effects in multilevel models by using only OLS estimators, rather than maximum likelihood or residual maximum likelihood estimators. Beyond that, he does not provide a method for adjusting standard errors for non-trivial intra-class correlations, though this is not difficult to do. Furthermore, differences in the number of degrees of freedom applicable for individual effects and contextual effects are not discussed.

Whatever its limitations, I was able to co-author three modest empirical pieces which were published in creditable journals by using Iversen's account of contextual effects. In part, this may be due to the fact that, in spite of problems posed by non-trivial intra-class correlations and differences in degrees of freedom from level to level, procedures such as those proffered by Iversen usually yield regression coefficients very similar to those produced using multilevel modeling or generalized estimating equations.

Through no fault of Iversen's, I found the transition from his book to conventional multilevel texts to be difficult. As best I can determine, this is due to the fact that Iversen explicitly acknowledges and never loses sight of the fact that his work is an extension of regression analysis, while multilevel texts generally fail in this regard.

All tolled, I think that Iversen deserves a good deal more credit than he has gotten for progress in using individual and contextual explanatory factors in the same regression model. Barr and Dreeben's widely read and influential book How Schools Work (1983), as an example, could not have been written without using contextual effects after the fashion of Iversen.