To determine the sample size, three pieces of information are needed: (1) the degree of confidence necessary to estimate the true value, (2) the precision of the estimate, and (3) the amount of true variability present in the data. In practice, the researcher and client discuss and agree on the desired level of confidence and the precision of the estimate. Further, since the true variability is unlikely to be known, it is typically estimated based on judgment or through a pilot study. An emerging requirement regarding sample sizes with secondary data is to select a random sample from large datasets to evaluate the effect of the larger sample sizes on statistical significance.