I recently finished a book on Bad Science. It is a really interesting and excellent book and I highly recommend it. It’s written by Ben Goldacre, a charming humourish chap from the UK. Dr Goldacre writes extensively about how Big Pharma and other shady characters use poor statistical methods to show that their drugs and treatments carry a certain efficacy. Not all the perpetrators are shady. Some are merely ignorant or sloppy, or simply use poor methods to draw statistical significance.
Dr Goldacre is quite thorough and the book is packed with references. He makes a few claims which he can’t back up, much the same as the people he is railing against, but they are mostly harmless and very very infrequent. Anyway, it is definitely worth picking up, especially if you are interested in understanding how certain drugs make it to market and others don’t, or why that conspiracy theory you think exists (Goldacre makes a good example of the MMR vaccine) is really just fear mongering by some people with nothing better to do or ulterior motives.
If you want to read more about shady, or optimistically misinformed, people who tell people to stop taking ARVs to deal with HIV and to replace it with vitamins, start with this book.
One of the 3 most interesting things that I learned from the book is the existence of Medline – a collection of published medical studies compiled by the US National Library of Medicine. You can find the link here.
The book has opened my eyes to the errors that I perform in trying to parse data from noise in engineering work. I will be digging into the end notes more extensively to see what else I can learn. A great read. Pick it up.
You can also read more about Dr Goldacre and his work from his blog, www.badscience.net