Probability and Statistics (Classic Version) (Pearson Modern Classics for Advanced Statistics Series)

It has been some time since I read this in its entirety; however, I find myself referring to it so often I felt it was appropriate to write a review. My perspective is that of an investment professional who uses large sets of data to make investment decisions for my investment firm. I have a bachelors degree in Economics and no graduate degrees.Despite my frequent use of statistics and probability, I felt that I was missing a lot of the underlying theory necessary to implement more sophisticated techniques--especially Bayesian ones. So I decided to start back at the basics and refresh my knowledge from my college math classes. I begin reading both this book and "Introduction to Mathematical Statistics" by Hogg, McKean, and Craig (HMC). After sampling the first few chapters and skimming through the rest, I decided to focus exclusively on DeGroot and Schervish's (D&S) book for two reasons:1) I liked D&S's padegogical approach of also beginning with a motivating example providing some context for *how* some subsequent concept might be useful. They then proceeded with the theory and more detailed examples. I didn't gleam the same intuitive understanding of topics from HMC as I did from D&S.2) I liked the integrated and thorough approach D&S took to both frequentist ("classical") and Bayesian statistics. DeGroot (especially--Google him) and Schervish are both heavyweights in Bayesian statistical theory and they don't pull any punches in outlining the deficiencies and inconsistencies of many frequentist approaches. They often do this through end of section vignettes that illustrate these issues with real examples. You will still definitely learn frequentist theory (in fact it always precedes the Bayesian discussions); however, if you to build a solid base in Bayesian theory, this is a good start. I have found no deficiencies in my foundation as I have progressed to actually implementing Bayesian techniques. In fairness, HMC do incorporate Bayesian theory as well, but I felt that based on my limited skimming of their book it was more like separate sections added it to be comprehensive and not something that was truly integrated throughout the book.I do have one critique of the book and that is the lack of computer code for examples. I understand D&S probably wanted the book to be "code agnostic" and this isn't a huge deal for most of the book, but it would be awesome if in the next edition the remaining author Schervish (DeGroot sadly passed away) could integrate R code (given it's free and by the graphic designs it looks like the author is already using it) directly into the book. It's already a mammoth book, so that might be asking too much, but the last several sections dealing with simulation techniques (Markov chain Monte Carlo, Gibbs sampling, etc.) are difficult to understand without computer code to replicate the results, which were the topics I was most interested in. I would recommend "Doing Bayesian Data Analysis" by Kruschke or "Bayesian Data Analysis" by Gelman et al. (I'd go for the third edition as it has STAN code) for actually understanding how to implement the simulation techniques required of modern day Bayesian techniques.

In the math world there are two general audiences:1. Those that want an informal though rigorous example driven approach.2. Those that like formal succinct lemma->proof approach.This book attempts to do both. In doing so it has produced a monster of a volume that can be off-putting to both camps. I am in the former camp and at first I was really disappointed with this book. The author is formal for the most part and does not try to introduce intuition into many places where it would be warranted, especially in the early "counting" and set chapters. This book is dry and uninspiring.What saves this book are the excellent examples and exercises that the author has accumulated over the years. In a sense your intuition is developed while doing the exercises and looking at the examples.I have found that in conjunction with Frederick Solomon's book "Probability and Stochastic Processes", which is more informal, one can learn basic probability theory quite well.

DeGroot & Schervish are accessible to students after three semesters of calculus and one semester of linear algebra. There will now be more proof than in earlier courses, yet the proofs here are delivered relatively informally. I like that in some of the early proofs, notably the multiplication rule for conditional probabilities and the Gambler's Ruin, the authors explicitly write down tautological equations into which substitutions will be made.Episcopal with clear, crisp writing and abundant, well-stated exercises help you learn. It is true that there are some asides the beginner can ignore on first reading, for instance that a set union of uncountably many events may not be an event, and a proof that the reals are an uncountable set. Joint distributions and Markov chains appear early, as soon as needed theorems have been stated.A few irritants: Failure to completely exploit the endpaper space, the traditional haunt for hints and reference formulas. The index gets short shrift, especially regarding ideas that occur only in the exercises. To keep 900 pp. from outweighing the student, very thin paper was used, so page-finding can be a bit awkward.More serious is the author's (probable) attitude that because they are mathematicians, they need pay little attention to the real world when picking numbers for examples and exercises. A disease screening is never described by its manufacturers as "90% reliable." This example also had the type I and II error rates both at 10%, which can misleadingly suggest these rates are always equal. No office complex draws 150 million kWh per day--enough to power a big city--while using only 4,000 gallons of water daily, an amount 6 homes could easily dispose of.Despite these issues, DeGroot & Schervish performed swimmingly in a textbook world that's become all about money and pushing out editions as fast as presses will run. Their text will outlast the changes of instructional fad to become part of my library afterward.

I really like the book. The authors go to great lengths to explain several details that are simply brushed aside in other books at the same level. A plus is the font size used in the book. It makes it a real pleasure to read the book for several hours. With this book, you'l learn the basics, and much more. There's a seamless transition between frequentist and bayesian perspectives in many chapters.Not 5 stars since for this International Edition they took out the last chapter on simulation, which on this day and age, is almost a minimum requirement.

Very comprehensive. A little formal in places, and an annoying lack of diagrams that make the exposition slightly laboured, but definitely recommended.

If you want to have a clear overview for all what is necessary to really understand statistics this is the book you are looking for.It doesn't get very deep in Markov Chains, stochastic process, inference and regression but you can learn everything to later read a specific book about the topic. It is a great book. The best

I am about to return to university to complete a masters in economics. It has been 4 years since completing my undergraduate degree (also in economics) and so I purchased this book in order to brush up on my statistics. I found this book to unnecessarily complicate the most simple of concepts. Whilst comprehensive in its approach, this book ultimately fails to deliver intuitive explanations of statistical principles.That said, this book does provide a large resource of practice exercises (for which answers for the odd-numbered questions are available). These are very useful.I have ended up using alternative resources to learn the theory before completing the exercises to confirm my understanding.

o livro é uma referência indispensável para o conteúdo programático da disciplina de Estatística em várias áreas do conhecimento (Engenharias, Ciências da Saúde e Biológica, Ciências Exatas e de Computação, Economia e Administração).