Sampling Techniques, 3rd Edition

If you work in sampling, survey design, even experimental design, this book remains essential. Don't think that because this book is decades old that its material is any less relevant. Pretty much every page contains crucial insights (though the few pages at the beginning on using random number tables is a bit quaint). Cochran's development of randomization-based sampling theory is simply beautiful. The only thing I regret about the book is that its treatment of large sample/asymptotic theory is too light. But then if it had that, there might not be any reason for anyone else to write in statistics anymore, so we can thank Cochran for his generous omission.

I recommend it for advanced students. Not easy to read but a very helpful book. It is more focused on sampling techniques for surveys rather than industrial applications.

Cohen has written the "bible" on sampling.

This used book is in really good condition, so the seller has really good credit. Also the book itself is bible of its area. Generally excellent ~

A good book needs to know it better.

I have to agree that this book in all three editions was a classic text on the theory of finite population sampling and it is so readily referenced in the research literature on survey sampling that it deserves to be called a classic and should be added to Wiley's series of classic books in statistics. It also may be appropriate to call it the bible on survey sampling. But the reader should note that it is not the only good book on survey sampling. Leslie Kish's book which I also reviewed today and compared to Cochran's is a very applied text with many social science applications and may be more accessible to the non-matematically inclined reader. In reviewing Kish's book I make some comments comparing it to this book by Cochran.

Yes this is the classic - first edition 1953. It is full of wisdom that is oft-ignored. For example that alpha and beta (Type 1 and Type 2 error probabilities = chances of false positives and false negatives respectively) should be set according to the real-world costs of making those errors. In other words, set alpha small if the cost of a false positive is high, but set it larger and set beta small if it's the cost of a false negative that's high in the real world. But we still go around mindlessly setting alpha at 0.05 because R.A. Fisher had that column of t-values worked out (by hand in those days). I recently reviewed a paper for an ecological journal which declared that the author had discovered a marvelous new approach to environmental study design. Turned out he had discovered Cochran's 1953 book! My other complaint is that the price is ridiculous, both from the publisher (which happens to be my publisher) and from 3rd party sellers. Classic books shouldn't mindlessly go up in price every year by X percent. Anybody who knows how compound interest works will know that a very small X will result in a very expensive book after 2012-1953=59 years. There should be a cheap re-issue of Cochran's classic. Please!

I searched for Sampling, this book is about surveys, I should have searched for "Statistic Process Control" as that is my reason for sampling. Also, for this book you already need to have a lot of knowledge about statistics and sampling. And it has very few examples. I need examples for better understanding. So for me not a good book.

great

Very useful!