DefinePK

Abstract

Sample Size Calculation and Sampling Techniques:
Calculating and justifying sample size for a study can be an intimidating task for new researchers. Nonetheless, it is one of the most important aspects of any study. In this article, we aim to provide basic and fundamental information about calculating sample size which should be helpful for most research studies.
In research studies, a sample has to be obtained if it is not possible to include all the subjects in the study population. For example, if you want to study the prevalence of smoking in the first year medical students of your medical college, you could just ask all of the students in that class to fill the questionnaire, without doing any sampling. However, if you want to know the prevalence of smoking in all the medical students of Pakistan, then it will be very difficult for you to have each and every medical student to fill your questionnaire. In that case you will need to do sampling.
For sampling to work correctly, it is critical that the sample is free from bias. This means that everybody in the study population should have a more or less equal chance of being included in the sample. If sample is collected in such a way that some members of the study population have less chance of being included than others, then the resulting sample is a biased sample, and the results obtained from such a sample will not be valid or generalizable to the entire study population.
While calculating sample size, the first thing to note is that methods to calculate sample size vary depending on the study design. For example, calculating sample size for a survey has a different methodology than to calculate sample size for a case control study or a clinical trial. However, the fundamental principles remain the same.
There are many statistical softwares and online websites which can help you in calculating sample size for your studies (Figure). However, you need to have following four parameters in hand before you can use those resources:
(1) The effect size
(2) The population standard deviation (for continuous data);
(3) The desired power of the experiment to detect the postulated effect;
(4) The significance level.
The effect size reflects the expected difference in the two study groups of your study. For example, if you know that the current cure rate for disease X with drug A is 20%, and you think that the new drug B will have cure rate of 40%, then your expected effect size for drug B is double that of drug A. The larger your expected effect size, the smaller will be your required sample size. However, if you overestimate your effect size, you will be at risk of having false negative results. While you cannot know the effect size for sure before doing the study, you must have some rationale for choosing a particular estimated effect size. This is done based on previously published studies, or your own preliminary data.
If you are doing a survey, then the effect size means the estimated response rate to your study question in the population. For example, if you are surveying the