Studying Statistics with DeGroot & Schervish 3e: Reviews and Solutions

[elementbrdr]

I'm independently studying stat. The book I'm using is DeGroot & Schervish 3e. I think it's pretty decent for the most part. However, there is no solutions manual available that covers all the problems in the book. Further, as I was searching on amazon for the solution manual, I found many reviews of D&S 3e indicating that it is not particularly good as an introductory text. Given that I can't even buy a solution manual to D&S to aid my study, I'm wondering whether I should just switch over to another book and accompanying solution manual that is more appropriate for a beginner. Thoughts and suggestions are appreciated.

Thank you.

 

[Stephen Tashi]

I should just switch over to another book and accompanying solution manual that is more appropriate for a beginner. .

I don't know the answer to that and I don't know about the book. But let me suggest you explain your goals for self study and your background.

Are you studying for an exam or in preparation for taking a statistics course? Or do you want to apply statistics to a particular field of research?

Why do you feel you need a complete solutions manual? Some people are found of saying that you learn the material by working problems. In my opinion, this is not a precise statement. Many people lack the skill (or perhaps the habit and will power) for understanding definitions, theorems and general explanations. They only understand mathematics through examples. For such people, working problems is essential. This is particularly true of people trying to learn a subject when they don't have a good background in the "prerequisites" - like a statistics student who doesn't fully grasp calculus or even proability theory and elementary algebra. Such a student is trying to learn several different subjects at one time.

Some people are able to understand math from reading mathematical statements and thinking about them in general terms. They must work problems in order to gain speed at working problems since there is a difference in understanding and drill. You can understanding something and still "have to think" in order to work a problem. If you drill at certain kinds of problems, you perform them faster - almost without thinking!

 

[elementbrdr]

Stephen,

Thanks for your response. I am studying stat because I'm interested in it. In my previous experience learning math, I have found that working problems is very helpful in learning how to apply concepts. D&S is very abstract in its discussion of concepts, which is not a problem for me, as I understand the concepts that the book discusses. However, I believe that actually applying those concepts through concrete examples is the only way to develop problem-solving skills. So I think a supplemental book with exercises and solutions would be helpful.

 

[elementbrdr]

I should also mention that I do understand algebra, calc (1-3ish) and probability theory. I also have a limited knowledge of stat (distributions, sampling distributions, mean, expectation etc.). The next topics I intend to study are special distributions (binomial, poisson etc.) and regression.

 

[Stephen Tashi]

From reading reviews of DeGroot and Schervish, the book does give some space to Bayesian methods. In my opinion, students can come through the typical introduction to statistics course that uses only non-Bayesian methods, learn how to work the problems and end up having a complete misconception about what that sort of statistics does (e.g. thinking that "confidence intervals" act as "credible intervals" or that hypothesis testing quantifies the probability that then null hypothesis is true or false). So I suggest that whether you switch books or not, that you continue to use materials that treat Bayesian methods.