I'm writing a book about the many ways well-intentioned scientists can (and usually do) screw up statistics:

http://www.refsmmat.com/statistics/

I started when I was an undergraduate physics major. My entire statistics education consisted of thirty minutes being told not to use R^2. I started reading about statistical abuses and realized that I would probably have committed most of them had I published research -- my training was entirely inadequate. Most scientists do only slightly better.

Combine this with a bit of unhealthy obsession and I ended up with 14,000 words of explanation, which I promptly published online.

My current draft is at 28,000 words and climbing, but I'm having more fun writing than figuring out what to do with it. I'm quietly hoping a serious publisher will notice and approach me, saving me the effort of writing a proposal and getting it rejected however many times before someone picks it up.

(happy to share a draft with anyone interested! Email is in my profile)