Question

When does one know that a sample size is adequate? Can one be too small or too big? Can they be biased? What are you looking for when you see a report and must determine if the information presented is good information or just junk?

Answer 1

There are some thumbs rules for deciding whether the selected sample size is adequate for the experiment or not. The thumb rule cannot be used in all conditions. In general practice, if possible, it is assumed that the selected sample should be 10% of population size. This means, if the population size is 1000, then selected sample should be 100 for drawing unbiased results about the population. But there are some limitations for using this rule. If the experiment is sensitive or it needs high cost, then sample size would be smaller.

We know that if the sample size is small, then there is a possibility of getting biased results. For obtaining minimum adequate sample size, there are some formulas based on the margin of error and confidence level. We can use these formulas for computing adequate sample size. In general, most of the times, the sample size should be at least greater than 30.

When we look at the report, we must check the population size, sample size, method of random sampling used, method of data collection, etc. These factors are very important in determination of sample size. If sample size is too small to present good information, then we cannot believe that the conclusions for the report are unbiased.