Conference rusure::math

    I'd also be curious if there's anything substantial out there which
    is not indigestibly abstract (and not a childish cookbook.)

    Two that I have are are Morris DeGroot's book "Intro to Probablility and
    Statistics" (or something like that) and Cramer's old book "Mathematical
    Statistics".  That's probably still in print, from Princeton U. Press.

    I also have a little book on analysis of variance by Mann called "Design of
    Experiments" - got used.  "Well worth a dollar..." as they say.

    I think DeGroot's book is fairly good as an introduction.

    - Jim

    Many -- what does your friend want to learn about statistics?  Theory
    or practice and at what level, or just a general introductory overview?
    Traditional or Bayesian?  If practice, design of experiments,
    hypothesis testing, parameter estimation (including regression, factor
    analysis, etc) or exploratory data analysis?  Slanted towards
    engineering, hard sciences, social sciences, medicine, economics,
    or biology?
    
    Need some background to make a recomenation.
    
    					Topher

    Re .2:
    
    I think it is more for reference than learning.  It should cover
    general theory, not application to any particular field.  Something
    touching on introductory at its low end is probably appropriate. 
    
    
    				-- edp 

    Cramer's book is excellent, but requires a fair amount of mathematical
    sophistication (he uses theory of measures).  Another excellent
    set of 2 books is the books by Feller, called essentially Theory
    of Probability (the exact title escapes me), but this also requires
    quite some mathematical sophistication (in particular volume 2,
    which deals with continuous probability distributions).

    Cramer's book is somewhat concise mathematically, but is still
    relatively down to earth - the measure theory in there is good because
    there's motivation for it.  I like to see clear derivations of
    things, and not just some magic formula out of nowhere.

    Feller's books are great - lots of examples and detail!  But they're
    not really statistics books.

    Re .2 - if you could mention some good examples you've seen in
    various areas that would be interesting.  Lots of the statistics
    books out there seem pretty terrible...  some of it's like punk
    mathematics.

    - Jim

    Despite my miserable performance in note 1058 I'll recommend the book
    "Introduction to Mathematical Statistics" by Hogg� & Craig.  (Wiley(?))
    It's an undergraduate-level introduction and requires undergraduate
    calculus.  They deal in Moment-generating functions, not Characteristic
    functions (which kind of implies the undergraduate level).
    
    Anyhow, one of the nice things about the book is that you can read a
    chapter, work all the problems and then go on to the next.  It's a good
    blend of theorem/problem.  Up until this real hard one late in the book...
    
    Another interesting possibility is maybe picking up one or more of the
    Schaum's Outline Series.  Some of those are really good if that's your
    style of working.
    
      John
    
    --------------------
    �Pronounced "Hoague"

    My favorite is _Statistics: Probability, Inference, and Decision_
    by R. L. Winkler and W. L. Hays, Holt Rinehart and Winston, NY,
    1975.
    
    It's a big thick book which starts slow and easy, with a chapter
    each on sets and probability theory, then three chapters on probability
    distributions, then chapters on estimation, hypothesis testing,
    Bayesian inference and decision theory, concluding with three chapters
    on correlation, analysis of variance and non-parametric statistics.
    
    The audience seems to be people who need to make decisions based
    on statistical data, and who have access to programs for carrying
    out the calculations.  So there is a lot of discussion of the
    underlying assumptions of this or that approach, where there may
    be practical difficulties, and how the results may be used to make
    decisions.
    
    It is fairly light on the math, uses only some simple calculus,
    and does not include a lot of fascinating but less useful
    generalizations and mathematical superstructure.
    
    There are heaps of examples, many quite practical, in the text and
    in the problems.
    
    I prefer it to Feller because the authors don't waste my time attacking
    anything beyond a narrow frequentist interpretation of probability.

    Re: .3 I think .. Since you are looking for fundamentals and rudiments
    without a lot of heavy mathematical treatment, try 
    
    STATISTICAL ANALYSIS FOR DECISION MAKING by Morris Hamburg.
                                      
    It is quite good for introductory level statistical concepts.

    Try Probability,Random Variables,and Stochastic processes
    Athonasios Papoulis