Hi Universe of 2110 documents have to be sampled. Using the sample calculator, with 95% confidence level and 4% Margin Error the sample size arrived is 467. However, when I selected randomly 10% sample (211 documents), the sample mean and sample median is not too much deviating from the population mean and population median. The SE Mean is also very low (0.00787). Sample Mean - 2:08 (h:mm) - 2:07 is population mean. Sample Median - 1"04 (h:mm) - 1:04 is population median CI for Mean = 1:45 (h:mm) and 2:30 (h:mm). CI for Median = 0:47 (h:mm) and 1:16 (h:mm) My question is, considering this example, if 10% sample can have good representation of the population then how significant is the 467 sample size which came out of the sample size calculator of 95% CL and 4% CI? I am not sure if this question is limited to the example I have, but wanted to understand the concept. Please help. Thanks Madhan