Product Details

• Amazon Sales Rank: #52951 in Books
• Published on: 2008-08-26
• Original language: English
• Number of items: 1
• Binding: Paperback
• 320 pages

Customer Reviews

Disappointing
There are many books around about how econometrics and statistical analysis are used in the modern world, often in areas you may not expect. This is another, and makes a reasonable go of it, but certainly isnt the best in the genre.

Ayres, in order (presumably) to keep the book simple and easy to understand for non experts, has removed a lot of detail and replaced said it with sillyness. Hence researchers and anyone else who has ever used an excel spreadsheet are "supercrunchers".

Anyone who has ever done any econometrics before will be familiar with the basic concept of a regression - but Ayres builds up this "supercrunching tool" to mythical proportions, suggesting that this strange technique mysteriously pops out answers to the worlds queries, with the help of the expert "supercruncher" at the wheel, of course.

In all a reasonably entertaining book on regressions and statistical analysis, which can certainly be read by non experts. Anyone familiar with a regression beware, as you may find this book comically belittling.

Interesting, fun, but a bit glib
"Super Crunchers" is a breezy and somewhat glib account of how quantitative methods are replacing traditional expert judgement. The author, Ian Ayres, is a law professor at Yale University that often has used quantitative methods in his research, some of which has been conducted in co-operation with Steven D. Levitt of "Freakonomics: A Rogue Economist Explores the Hidden Side of Everything" fame.

In this age of plenty when it comes to computing power and information, traditional quantitative methods for decision-making such as regressions and randomised trials are spreading to new fields. A regression on meteorological data performed by Orley Ashenfelter is a better predictor of the eventual price of Bordeaux vintages that the expert judgement of Robert Parker. In the USA government is using randomised trials to judge the efficacy of government programs. Ayres also mention methods such as neural networks that can be used to find patterns in data.

The advent of "super crunching" has its downsides. It will automate decision-making, reducing the scope for judgement in many white collar jobs. Loan officers at banks these day do little more than feed the computer with information. There is a movement for evidence-based medicine where doctors are aided by computers in the diagnoses of patients. Even jobs such as teaching could turn out to be less interesting in the future. In the USA some schools have adopted Direct Instruction, where the entire lesson is scripted, the script having been developed and tested using quantitative methods.

In the beginning of the book Ayres seems quite unconcerned about privacy, but this is dealt with in a later stage of the book. What he never really handles are the limitations of the methods he touts. Statistics is a hard tool to master, witness the many mistakes in scientific papers. Data-snooping biases is a real danger in data mining, when testing a huge number of hypotheses some are bound to pass conventional tests for statistical significance by pure chance. A hilarious riposte to evidence-based medicine was an article in the BMJ where they found that there were not enough data to support the hypotheses that parachutes are effective in preventing major trauma related to gravitational challenge.

Despite this the book is well worth a read. It is funny and interesting. If you are interested in super crunching methods in a business setting, you might want to read "Competing on Analytics: The New Science of Winning" by Thomas H. Davenport and Jeanne G. Harris.

an extended magazine article
It is fun and well written and I certainly learnt something from it.
However it just doesn't go into enough detail. It is just like an extended article in a magazine.