Freakonomics

I'm sure not too many would appreciate the pun that is "freakonomics" - it's about as bad as "blogerithm", none-the-less, this is a book of cute anecdotes, most of which involve stats.

Levitt's philosophy is simple: numbers cannot lie - which runs counter to the lying statistics revelations of 1954. In the spirit of this belief, Levitt's work has been rigorous and seemingly unbiased. In investigating correlations been abortion and crime-rate, I was convinced enough by the controlling variables that were investigated to believe, "yeah, crime-rate dropped in the 90s because aborted children in the 70s would be just reaching their prime criming age!".

I was also really impressed by his algorithm to detect cheating teachers (teachers who doctor some of their students' standardised tests so that they don't look like such bad teachers) and the following statistics used for support. What is interesting is, indeed, that statistics were able to form sole justification for the firing of teachers - against professings of innocence, and traditional innocent before guilty notions. Wow, I say.

There are other nice anecdotes retold in this book, one of which involves a man who brought down an entire chapter of the KKK just by revealing their secrets on a Superman radio programme.

I am always skeptical of this type of book, where the "wow" factor of mathematics and statistics is used to intrigue people, however I think the basis of Levitt's findings have been pretty solid, and he has intelligently answered questions that I'm sure would have been thrown in the too-hard-to-know-for-sure basket. This isn't just basic maths dressed up as magic - in fact, I'm pretty inspired to try and find a really interesting application for my own research. Nice one.