Data Analysis Using Regression and Multilevel/Hierarchical Models (Analytical Methods for Social Research) (英語) ハードカバー – 2006/12/25
Kindle 端末は必要ありません。無料 Kindle アプリのいずれかをダウンロードすると、スマートフォン、タブレットPCで Kindle 本をお読みいただけます。
Data Analysis Using Regression and Multilevel/Hierarchical Models, first published in 2007, is a comprehensive manual for the applied researcher who wants to perform data analysis using linear and nonlinear regression and multilevel models. The book introduces a wide variety of models, whilst at the same time instructing the reader in how to fit these models using available software packages. The book illustrates the concepts by working through scores of real data examples that have arisen from the authors' own applied research, with programming codes provided for each one. Topics covered include causal inference, including regression, poststratification, matching, regression discontinuity, and instrumental variables, as well as multilevel logistic regression and missing-data imputation. Practical tips regarding building, fitting, and understanding are provided throughout.
'Data Analysis Using Regression and Multilevel/Hierarchical Models' … careful yet mathematically accessible style is generously illustrated with examples and graphical displays, making it ideal for either classroom use or self-study. It appears destined to adorn the shelves of a great many applied statisticians and social scientists for years to come.' Brad Carlin, University of Minnesota
'Gelman and Hill have written what may be the first truly modern book on modeling. Containing practical as well as methodological insights into both Bayesian and traditional approaches, Data Analysis Using Regression and Multilevel/Hierarchical Models provides useful guidance into the process of building and evaluating models. For the social scientist and other applied statisticians interested in linear and logistic regression, causal inference, and hierarchical models, it should prove invaluable either as a classroom text or as an addition to the research bookshelf.' Richard De Veaux, Williams College
'The theme of Gelman and Hill's engaging and nontechnical introduction to statistical modeling is 'Be flexible'. Using a broad array of examples written in R and WinBugs, the authors illustrate the many ways in which readers can build more flexibility into their predictive and causal models. This hands-on textbook is sure to become a popular choice in applied regression courses.' Donald Green, Yale University
'Simply put, Data Analysis Using Regression and Multilevel/Hierarchical Models is the best place to learn how to do serious empirical research. Gelman and Hill have written a much needed book that is sophisticated about research design without being technical. Data Analysis Using Regression and Multilevel/Hierarchical Models is destined to be a classic!' Alex Tabarrok, George Mason University
'… a detailed, carefully written exposition of the modelling challenge, using numerous convincing examples, and always paying careful attention to the practical aspects of modelling. I recommend it very warmly.' Journal of Applied Statistics
'Data Analysis Using Regression and Multilevel/Hierarchical Models is the book I wish I had in graduate school. … The text is an obvious candidate for use in courses or course modules on multilevel modeling, especially in Part 2. Beyond that, where should it be used? Instructors of first-year graduate methods courses should consider complementing their texts with material from Part 1. Many use Kennedy's A Guide to Econometrics (2003) to provide an alternative take in the essentials. Data Analysis is better suited for taking on this role. Students will find its coverage less redundant of what they get from standard texts, and the use of non-economics based examples should also help sell quantitative research to skeptical incomers into the profession.' The Political Methodologist
For me the best way to work with this book was to use the data and examples from the Github page and run each example and replicate the analysis and plots. The exercises do suffer from poor data availability, and unfortunately the best bet is to use the data from the website and try them out. Nonetheless, if one is willing to take hard knocks and keep up the effort to work through each of the examples in the book, the rewards are worth it.
Since the first reading, I've been proselytizing this book to all my colleagues and our interns. Consequently, most now own it and use it frequently. Because I've loaned it out so much, I bought a second copy.
Why do I like it? Simply, I am a quantitatively capable person without much patience. I can't sit down for hours with Greene and feel like I've learned much. I need examples, good writing, arguments, and lots of practice data/code. This book has those things.
Who is this book not for? I'd not recommend it to people who get a lot out of mathematical statistics, engineers, and the like. If you enjoy formalism, you'll get frustrated at the authors' desire for practicality.
The book covers a lot of material, and so nothing is delved into very deeply. There is more than one section you will need to look elsewhere in order to supplement the one or two sentences dedicated to a topic.
You should be experienced with using R in order to get much out of the examples. For example, they use the arm library a lot, but never show in the code samples to load the library. If you've never used R (I have) then you might be lost. The most frustrating thing I'm finding is the data that accompanies the book. It is not documented and not formed to match the examples in the book. For example, the height and earnings data uses 1 and 2 classifications for sex, but I couldn't find it documented which one was for male and female. I little trial and error and I eventually figured it out. Little things like that don't make it easy to work with. The data comes in numerous formats too (Stata, DAT, asc) instead of a generic csv file.
If the ratio was the other way around it would be worth 5 stars, but right now it's an exercise in frustration.
They tackle a complex topic from many different angles. They present enough code and theory to get people up and running with the techniques, assuming some prior familiarity with likelihood based inference and R. Otherwise you might need to dig through some of the references to understand everything. Regardless, this book is a valuable reference to keep in your library.
They use matrix notation sparingly and this helps the reader focus on the important concepts of multilevel modeling. I am not even remotely a statistician so my attention would have been lost if I had to sort through a bunch of matrix transpositions and inversions in addition to all of the multilevel notation.
The authors provide many useful references that help reinforce difficult ideas/concepts and that elaborate on topics that are not explored in depth.
I had no prior experience using WinBUGS and the authors provided enough information for me to successfully execute some models that integrate R and WinBUGS. That is no small feat and the authors should be commended because somehow I understood what was going on.
The organization of the book seems scattered and could be a little more consistent. On pp 245-246, the authors go on a diatribe about "fixed" and "random" effects terminology, claim that much of the literature that applies these terms does so inconsistently, disown these terms by saying they will avoid using them entirely, and then continue using these terms throughout the book.
The website needs some work. You need to already know how to use R to open different types of files (and maybe some basics of variable assignment)in order to reproduce all of their examples. This book will not hold your hand through the steps like many R books.