Ian Ayres',wrote
Super Crunchers; for me he wrote two books in one. One book I liked and the
other I didn't. The book I liked is filled with countless stories and
volumes of data about myriad statistical subjects that fascinate me,
just like Freakonomics (Levitt is a friend of Ayres).
Examples of what is included in the book I liked are data on number of
days from human conception to birth and delivery along with the standard
deviation; he gives details on job assistance programs that provided
good job search training (helpful) and on prison terms (length of term
for the same crime has no effect on recidivism or post prison earnings.)
The
book I didn't like speaks with a voice that sounds like his sub-title:
Why Thinking by the Numbers is the New Way to be Smart. It is preachy
and breathless in telling us how extensive is the use of math,
empirical testing and statistics in the modern world.
Cool it Ian. The use of math, testing and statistics may be increasing but it certainly isn't new or dramatic.
I
used linear regressions on credit card customer default data working
with an economist when we were both at Bank of America in 1964. Together
we developed the first credit scoring system.
I tested nearly
every bank product and service by testing with matched samples. I
didn't just test three or four different designs of statement stuffers and bank statements, I frequently tested 25 variables at the same time (using a Latin Square design) which Ayres doesn't mention. This was in 1965.
By
1968 I had used nearly every statistical conception of testing that was
known. I presented my findings to management and to my professional
peers regularly. Nearly everyone understood what I was doing and
appreciated the empirical results.
I repeat, cool it Ian...no need
to be breathless about something that has been coming to the business
and policy world at a steady and methodical pace for half a century.