Understanding Regression Assumptions (Quantitative Applications in the Social Sciences) [Paperback]

William D. Berry (Author)

Book Description

ISBN-10: 080394263X | ISBN-13: 978-0803942639 | Publication Date: February 25, 1993 | Edition: 1

Through the use of careful explanation and examples, Berry demonstrates how to consider whether the assumptions of multiple regression are actually satisfied in a particular research project. Beginning with a brief review of the regression assumptions as they are typically presented in text books, he moves on to explore in detail the substantive meaning of each assumption; for example, lack of measurement error, absence of specification error, linearity, homoscedasticity, and lack of auto-correlation.

Editorial Reviews

About the Author

Ph.D., University of Minnesota, 1980. Major research and teaching interests include public policy, budgeting, state politics and research methodology. Berry's primary areas of interest are public policy, American state politics, and research methodology. Throughout his career, his work has focused on explaining why governments make the policy choices they do, and improving the methodologies available for studying policy making. Among his current research projects are analyses of the impact of electoral competition on the policy choices made by state legislators, and the effect of state welfare policy on poverty in the United States. Berry's current research on methodology focuses on the development of techniques for estimating econometric models with binary dependent variables, and methods for studying policy diffusion using geographical information systems. Berry has also been working for several years collecting data on outcomes of all state legislative elections in the U.S. since the 1960s; he has begun to use these data to study the behavior of challengers in state legislative elections, and the role of partisan and incumbent protection goals in determining the nature of redistricting in state legislatures.

Product Details

• Paperback: 104 pages
• Publisher: Sage Publications, Inc; 1 edition (February 25, 1993)
• Language: English
• ISBN-10: 080394263X
• ISBN-13: 978-0803942639
• Product Dimensions: 8.7 x 5.4 x 0.2 inches
• Shipping Weight: 4.5 ounces (View shipping rates and policies)
• Average Customer Review: 5.0 out of 5 stars  See all reviews (2 customer reviews)
• Amazon Best Sellers Rank: #154,744 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

5.0 out of 5 stars Its sterngth is in the examples, February 26, 2010

By 

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

This review is from: Understanding Regression Assumptions (Quantitative Applications in the Social Sciences) (Paperback)

I've owned a copy of Berry's Understanding Regression Assumptions for ten years, but I didn't get around to reading it until a few days ago. Better late than never, I suppose, but Berry's text would have been an invaluable adjunct to any of the more complete regression/econometrics texts -- Gujarati, Wooldridge, Wittink -- I've used for teaching multiple regression in years past.

It's easy to list the assumptions, explain the consequences of their violation, and provide corrective procedures to assure that OLS regression provides BLUE estimates of slopes. It's a good deal more difficult, however, to identify and explain concrete circumstances that give rise to violation of assumptions in the first place. Sure, mis-specification is easy, at least in principle: variables that should be in the equation are, those that should not be are not, and funtional forms of all relationships are correct. But providing specific examples that illustrate violation of even this most fundamental of assumptions requires a good deal of work on the part of an instructor who wants to do the job right.

Futhermore, when we get past the conceptually easier issues and have to provide examples of circumstances that generate, say, heteroscedasticity or serial correlation with cross-sectional data, the task of making these ideas concrete becomes much more demanding. Only a practiced hand who has attended to these issues in purposeful fashion will be able to provided illustrations which will attune students to the conditions that are likely to cause essential assumptions to be violated.

Berry's book is replete with examples that make the usual OLS regression assumptions real and easy to remember, rather than leaving them as odd-sounding abstractions that we memorize and take on faith as essential to best use of OLS estimators. I'm retiring at the end of this semester, so maybe better late than never doesn't really apply after all. Still I'm glad I finally got around to reading this fine book.

5.0 out of 5 stars Delivers all that's hoped, June 13, 2008

By 

Sandy Alabast (Arlington, VA) - See all my reviews

This review is from: Understanding Regression Assumptions (Quantitative Applications in the Social Sciences) (Paperback)

While equation heavy at times, after reading this book one gains a great deal of understanding as to the weaknesses/boundary conditions for inference of OLS regression. For those with little understanding of the basic mechanics of regression, this is NOT a good starting place, but those with some working knowledge this book is invaluable.