Statistics book for bioscience / medicine?

July 19, 2008 9:58 AM

Can you recommend a statistics book which is appropriate for the stats required for basic bioscience / clinical research?

So, my background is I'm a (medical) doctor. I've also just finished the lab research for a (cellular biology) PhD (which I'm still doing data analysis for and writing up).

I know a reasonable amount about statistics but need a wide ranging and pretty detailed textbook on statistical methods, potential pitfalls etc. I'd rather have a book that erred on the side of comprehensiveness, but which is written clearly. Including practical examples along with equations is also a bonus.

Most of the statistics I've used during my PhD were actually just basic descriptive statistics, T-tests, Mann-Whitney tests and ANOVAs. So there's a huge world of stats out there I don't really get.

Any recommendations?

posted by inbetweener to Science & Nature (7 answers total)

Altman's Practical Statistics for Medical Research may be a good choice. It does focus on fairly basic statistics, but the explication of them is comprehensive and lucid. Like you I'm an MD with research experience; I love this book.
posted by ikkyu2 at 10:53 AM on July 19, 2008

Zar and Biometry are both well regarded and pretty comprehensive.
posted by epugachev at 10:55 AM on July 19, 2008

Seconding Biometry.
posted by Blazecock Pileon at 11:15 AM on July 19, 2008

The book from my biostats course in college (which I took 2 years ago) is called Intuitive Biostatistics. It may be too basic for you but it does at least cover most, if not all, of what you asked for. I especially liked the examples, which tend to be clinically based.
posted by dondiego87 at 11:43 AM on July 19, 2008

I thought Zar was okay, and I still use it (it is the only undergraduate textbook I still have), although there has to be something better.
posted by grouse at 3:07 PM on July 19, 2008

I was going to suggest Intuitive Biostatistics but dondiego87 beat me to it. The guy who wrote it also designed a software package (Graphpad Prism) that is used in many biology labs. It would certainly be suitable for the kind of stats that would apply to most experiments in cell biology, but not if you're trying to do heavy-duty bioinformatics. It definitely scores high on readability and clarity.
posted by sevenyearlurk at 5:49 PM on July 19, 2008

Try this book: Practical statistics for field biologists. I found it covered the basics in a very easy to understand way. (I'm in ecology, btw)
posted by dhruva at 8:46 PM on July 20, 2008