Suppose you measure the sex ratio of a population, and find 1.2 males per female

If the sample size used to measure sex ratio was small, this result could reflect limitations of the experiment. However, if the sample size was large enough to accurately represent the population sex ratio, it is still possible that this is a chance event, rather than an effect of selection. One way to test this would be to measure the population over a long period of time to check for changes in sex ratio. Another would be to remove males from the population, then observe over several generations to see whether the 1.2:1 ratio is reestablished.