25 of 29 people found the following review helpful
5.0 out of 5 stars
As yet, an unparalleled reference on the science of epidemiology, April 8 2008
By GKatz51 "NewYorker" - Published on Amazon.com
This review is from: Modern Epidemiology (Hardcover)
Masterly. Rothman and coworkers' 3rd edition of the classic `Modern Epidemiology' is a landmark step in presenting with cohesiveness and clarity traditional + cutting-edge concepts and methods for observational research (and by extension for experimental research on human subjects as well, since problems like unblinding or loss-to-follow-up may reintroduce bias in clinical trials after randomization, with a corresponding need for epidemiological interpretative and analytical tools).
This highly regarded volume has no equal, as it has been not only an authoritative source of information but `the' reference on epi methods for almost a quarter century. For those looking for an introductory level textbook, Rothman's one by Oxford University Press is highly recommended, since the comprehensive `Modern Epidemiology' requires some previous exposure to the concepts and biostatistical methods presented.
The 3rd edition is an encyclopedic effort, brings methodological coherence to a whole new level, is highly readable, and confirms itself as the standard reference on epidemiological and clinical research for many more years to come. An outstanding scholarly achievement.
Definitely a must-have for anyone who needs to learn and apply basic/advanced epidemiological methods rigorously in clinical as well as general population settings.
5 of 5 people found the following review helpful
5.0 out of 5 stars
excellent and very up-to-date, Jun 17 2010
By Cyrus Samii - Published on Amazon.com
This review is from: Modern Epidemiology (Hardcover)
I am a social scientist, not an epidemiologist, and I found this book to exceptionally good. It is the most current, complete, and clear presentation of methods for causal inference for observational (i.e. non-experimental) studies that I have seen. The things that really set this book apart for me include:
1. It synthesizes contributions by Pearl and Rubin on the foundations of causal inference, and contributes its own perspective via the sufficient cause model. This is truly cutting edge, not to mention impeccably coherent.
2. The first third of the book is on study design, including measurement, sampling, and defining effects. This is just fantastic. Many methods textbooks jump right into approaches to analyzing data with little time taken to discuss how to make the data in the first place. This book provides a major corrective to that tendency.
3. In data analysis, a lot of attention is given to sparse data problems, which again is just great. So many textbooks overlook this problem, which is a huge omission.
4. The data analysis section includes discussion of up-and-coming data mining and non-parametric methods (e.g. BART, boosted regression, etc.) to characterize response surfaces in the service of causal inference. That's amazingly cutting edge for a textbook.
5. The meta-analysis section emphasizes simplicity and provides a very nice list of common errors that should be avoided.
6. The references are to state of the art literature not only in epidemiology, but also in econometrics, education research, and statistics. It's great to see such cross-fertilization across disciplines, and it shows how these various disciplines are converging, it seems, on common analytical tools for causal inference in observational studies.
There are lots of nice examples throughout the book too. For other social scientists out there, I highly recommend this as a primer on state of the art methods for carrying out observational studies.
12 of 15 people found the following review helpful
4.0 out of 5 stars
Disappointed with paper/printing quality, Mar 5 2009
By E. Grant - Published on Amazon.com
This review is from: Modern Epidemiology (Hardcover)
I borrowed the 2nd edition from a colleague and was so impressed with the book that I ordered it immediately. This book has no equal in terms of content (the word 'encyclopedia' is aptly used in a previous review). The statistical explanations are what really set this book apart from other epidemiological textbooks. The price was surprisingly low compared to other textbooks on the market.
Unfortunately, after receiving the book, I see how they managed to keep the price down. The type is small, the margins are small and the paper is thin. Also, to me (highly subjective), the paper has an unpleasant desiccating feel. Finally, I like to mark up my books with highlighting and notes in the margins -- but this just makes a mess out of this book. Highlighting bleeds through and pencil notations are visible from the opposite side of the page. The 2nd edition was more robust.
Publishers can't win. Most books are too expensive. This (outstanding) book comes in at a great price -- but the compromises are painfully evident.