I believe David B is correct in saying the sample size would be critical but with a large enough sample it would be reliable.
With a small sample carefuly selected you could prove most anything you wanted to.
The intent of the questioner and other replies including the 'best answer' are so much garbage! They cooked the books on this one.