Applied Regression

This is a great book to acquaint students with the world of linear models. It is perfect to use in combination with other texts, or as a stand-along book in introductory courses. The Lewis-Beck’s have updated the presentation, provided additional examples, and included more discussion of regression diagnostics. I am sure that it will, once again, be a best seller!

This is an excellent update and extension of a wonderfully clear exposition of bivariate and multiple regression analysis for beginning practitioners and students.  I was a fan of the first edition, and I am even more pleased with the revision.  


This is one of the best resources on basic regression techniques available on the market today and it remains my go-to guide for my own research. Applied Regression is the quintessential text for graduate students pursuing degrees in the quantitative social sciences; it has helped train several generations of social science researchers over the course of the last four decades. The second edition will remain instrumental in training social scientists for years to come.

The new edition of Applied Regression maintains the excellence of the original edition while modernizing and extending it.  Its highpoint is how the Lewis-Becks state everything with complete precision.  From the assumptions of OLS to the ways of coping with outliers and to the methods of detecting multicollinearity, the authors tell readers exactly what they need to know to perform regression analysis.

Colin Lewis-Beck is a PhD candidate in Statistics at Iowa State University.  He holds a BA from Middlebury College and a dual MPP/MA in Public Policy and Applied Statistics from the University of Michigan.  While at Michigan, he received an Outstanding Teaching Award from the Department of Statistics.  Also, he has worked as a Teaching Assistant and a Computer Consultant, during multiple summers at the Inter-University Consortium for Political and Social Research (ICPSR) Summer Program, University of Michigan.  His research experiences in statistics are varied, and including serving as a Statistician in the Economic Analysis and Statistics Division of the OECD (Paris), and at STATinMED, a health outcomes research firm in Ann Arbor, MI.  His interests are applied statistics related to social science research, causal inference, and spatial statistics.  Mr. Lewis-Beck has co-authored papers on the quality of life and work productivity, modeling health care costs, and technology use in educational performance.

Michael S. Lewis-Beck is F. Wendell Miller Distinguished Professor of Political Science at the University of Iowa, and holds a Ph.D. from the University of Michigan.  His interests are comparative elections, election forecasting, political economy, and quantitative methodology.  He has been designated the 4^th most cited political scientist since 1940, in the field of methodology. Professor Lewis-Beck has authored or co-authored over 240 articles and books, including Applied Regression: An Introduction, Data Analysis: An Introduction, Economics and Elections: The Major Western Democracies, Forecasting Elections, The American Voter Revisited and French Presidential Elections.  He has served as an Editor of the American Journal of Political Science, the Sage QASS series (the green monographs) in quantitative methods and The Sage Encyclopedia of Social Science Research Methods.  Currently he is Associate Editor of International Journal of Forecasting and Associate Editor of French Politics.  In spring 2012, he held the position of Paul Lazersfeld University Professor at the University of Vienna. During the fall of 2012, he was Visiting Professor at Center for Citizenship and Democracy, University of Leuven (KU Leuven), Belgium.  In spring 2013, Professor Lewis-Beck was Visiting Scholar, Centennial Center, American Political Science Association, Washington, D.C.  During fall 2013, he served as Visiting Professor, Faculty of Law and Political Science, Universidad Autnoma de Madrid, Spain. In spring, 2014, he was Visiting Scholar, Department of Political Science, University of Gteborg, Sweden.  For fall, 2014, he served as a Visiting Professor at LUISS University, Rome.  At present, he is co-authoring a book on how Latin Americans vote.

&...

1. Bivariate Regression: Fitting a Straight Line Exact Versus Inexact Relationships The Least Squares Principle The Data The Scatterplot The Slope The Intercept Prediction Assessing Explanatory Power: The R^2 R^2 Versus r 2. Bivariate Regression: Assumptions and Inferences The Regression Assumptions Confidence Intervals and Significance Tests The One-Sided Test Significance Testing: The Rule of Thumb Reasons Why a Parameter Estimate May Not Be Significant The Prediction Error for y Analysis of Residuals The Effect of School Size on Educational Performance: A Bivariate Regression Example 3. Multiple Regression: The Basics The General Equation Interpreting Intervals and Significance Tests The R^2 Predicting y Dummy Variables The Possibility of Interaction Effects The Four-Variable Model: Overcoming Specification Error 4. Multiple Regression: Special Topics The Multicollinearity Problem High Multicollinearity: An Example The Relative Importance of the Independent Variables Flexing the Regression Model: Nonlinearity Determinants of Presidential Popularity: A Multiple Regression Example Presentation of Regression Results in a Research Paper What Next?