Usually your sample size would be a good indicator of the accuracy of your poll/test/experiment.

Of course, like everything it could be biased by the used methodology.