News for hackers. Stuff that matters.
Re:Facebook founder finds a way to profit from its members' ... by kragen on Tuesday February 03, 2009 @12:18AM (#463555)
Relevant is this snippet from http://norvig.com/experiment-design.html:

In my own work, we do a dozen or more experiments every day, and our typical value is p=0.0000001% -- one in a billion or so. Sometimes we get a p of one in a trillion. And even then we agonize over whether we should believe the results of the experiment.

Don't you wish you could do a dozen or more experiments on human subjects each day, with a p of one in a billion?

There are no comments on this comment yet.