[sumo] Question somewhat related to death of rikishi
joepetrow at gmail.com
Fri Jun 29 23:05:02 EDT 2007
On 6/30/07, Doreen Simmons <jz8d-smmn at asahi-net.or.jp> wrote:
> It's a question I often get asked myself. The first thing is that there
> are never enough rikishi to be statistically significant. Statistics
> only work with millions, not a few hundred at any one time.
Um...better not tell that to the market research industry. Or
doctors. They seem to think that a sampling of few hundred at any one
time is enough to give them information to reach conclusions, provided
the data falls within a certain standard margin of error.
Of course, data from 100% of the known universe is ideal, but human
science has evolved to the point where people can draw reasonable
conclusions based on a limited quantity of data.
More information about the Sumo