Quote:
Originally Posted by renormalised
1200 people is statistically insignificant anyway. ...
|
Whether a given number of people is a statistically valid sample size depends on what your are trying to measure.
If 50% of the population have the characteristic for which you are testing, 1200 is probably oversampling. If it was 1% it is not enough.
Left handed people are around 8% of the population.
For left/right comparisons, 1200 is about the right size for hypothesis testing for statistics valid at the 95% confidence level. Every extra conclusion you draw means you need a bigger sample size, so I'd agree he hasn't tested enough people.
Then again, this is soft science. You can't believe anything they say.