A text that explains the concepts without much math and formulae? Beginners in medical research in clinical trials/epidemiology often need a basic book on medical statistics that is appropriate for self study.
There is, to my knowledge, no really good book on medical statistics at all. Most books provide collections of "recipies" and do not promote understanding or "statistical thinking". All books I know present medical statistics as something objective and fixed and consider statistical analysis as beeing the thing to decide about what conclusions from a study must have to be drawn. Furthermore, such books generally have a strong focus on hypothesis tests and do not even manage to differentiate significance tests and hypothesis tests.
So I would also be happy if someone will suggest a book that aims on statistical thinking and - if it has to deal with tests - will handle them correctly.
A concern I have about the question: why should no math or formulas occur? Why should this be a problem? Or the other way around: If this was such a big problem for a medical student(!), would it be wise to talk about statistic, i.e. data analysis and interpretation at all?! If this is so, it should be questionable if a study of medicine or any other subject based on empirical data is sensible. Get me right - I am not saing that medical students (or students of any empirical science) should be math-experts. It is only about a very basic understanding, not more than should be known from school time. But this much must be requested.
Ok, there seems to be no really good book, but among the not-perfect books, I can recommend the best one: "What is a p-value anyway". It is for (more or less absolute) beginners, handels all the most important aspects of statistics, is written in a funny but yet instructive way. However, it is not particularily focussing on clinical trials and epidemiology.
I have to agree with Jochen. After reading through a dozen or so applied stats books written by non-statisticians, they look good. Unfortunately, there are some mistakes and improper use of formula and methodology. A lot of the books written by statisticians are pretty dry and boring.
When I read a stats book, I look for the topics covered and how they are covered. Most of the formulae in a stats book are already used by your stats program. I look for a book that motivates why you use certain methods and how you use it properly.
The stats book I like a lot is Optimal Design of Experiments: A Case Study Approach by Peter Goos and Brad Jones. They motivate the designs and the analyses through case studies. They leave the technical, theoretical and formulae to the last section of the chapter. If you want to do medical research, outside of clinical trials, it is a good book to read.
Sometimes for quick reference, I read Medical Biostatistics by Abhaya Indrayan. With lots of example from medical background, it helps to understand the application of key statistical concepts to those who work in medical or epidemiology domain.
Another book which is simple and easy to go through is Essential Medical Statistics by Betty Kirkwood.
Best way understand stat is practice (with good basic stat knowledge). And at start with STAT adventure you should remember one importatnt thing - forgot about post hoc hypothesis.
Practical Statistics for Medical Research by Douglass G Altman is very good, reliable and not boring (for a book on statistics). Kirkwood is very good also.
Norman & Steiner´s Biostatistics is very amenable, and sometimes very funny. It gives a good idea of how statistics work.
However, for a total begginer I would recommend The Cartoon Guide to Statistics, by Larry Gonick and Woolcott Smith:
All good suggestions above. I think Bernard Rosner's, Fundamentals of Biostatistics is an excellent and reasonably-priced choice as well.
It is the reference used for the highly useful page for contingency table analysis: http://statpages.org/ctab2x2.html. I find this page to be more accurate than a couple of expensive software packages, and it includes confidence intervals for many quantities that the others don't.
In my own work however, I find it a bit frustrating that there is sort of a different gold standard text or publication for a wide variety of commonly used methods. And equally frustrating that you can spend a lot of time getting a computation exactly right by the literature only to have a nit-picky journal reviewer still disagree with some subtlety (or even just be dead-wrong). And acquiescing rather than arguing is usually the better path to publication.
All that being said, I find that a good software package, its help files, and the function reference therein to be more useful than any single textbook. i.e. Matlab, SAS, PASS,...etc. They will usually have a programmed method and example for something you are interested in, there will be a succinct and effective explanation, and there is a reference for you to use in the paper. PASS even goes a step further in frequently generating the proper statistical jargon for you to use as well.
I would not recommend the WHO book. I had a look into the chapters about data analysis (Chap. 8 ff) and they contain the "usual" misconceptions and the minglemangle of hypothesis- and significance tests.
I have recently published a basic stats book intended for those preparing to enter or just beginning their studies in health sciences. You may review the book electronically by sending a request to the publisher on the webpage linked to this note. This book begins with an overview of the place of statistics within research, which is emphasized throughout the book with scenarios from real research. It has few formulas, promotes conceptual understanding, and is student friendly. I would be happy to answer any questions about the book.