Investigate prior research in your field to determine the likely degree of variability of the topic that you are studying. Generally speaking, if relatively few or relatively many people in your population display the trait that you are interested in studying, then you can use a lower sample size. For instance, if you are interested in studying the effect of a certain gene phenotype on behavior, and 50 percent of people possess the gene, then you will need to obtain a relatively larger sample size than if 20 percent or 80 percent of the population possessed the gene.
If you are unsure of the degree of variability inherent to the topic that you are studying then you should assume a degree of variability of 50 percent, or 0.5. This is the population distribution that contains the maximum amount of variability and thus provides a conservative sample estimate.
Determine the level of confidence that is required for your study. The confidence level refers to the number of samples out of 100 that will yield accurate results. For instance, if you use a 95 percent confidence interval then the results of your study will be accurate 95 percent of the time and based on change 5 percent of the time. A 95 percent confidence interval is used for most studies in the social sciences, such as psychology, while studies that require a great deal of precision, such as those done with medical populations, frequently used a confidence interval of 99 percent. Consult similar research to see what confidence level they used if you're not sure what level is appropriate in your field.
Specify the level of precision, or sampling error, that is inherent to your study methodology. The level of precision is typical expressed in percentage points, such as plus or minus five percent. Again, if you are unsure of the level of precision required for your study, you should consult similar research studies that have been published in academic journals.
Estimate the total number of people in the population that you are studying. If the population size is large or is not estimable, you can skip this step.
Navigate your web browser to a free online sample size calculator (see Resources). Input the values that you have determined for all variables and click on the "Calculate" button. If you could not estimate the population size, you may leave that field blank. The calculator will perform the necessary calculations and will output the sample size that is required to satisfy your specified parameters.
If you prefer not to use the online calculator, you may consult a statistical table to find the approximate sample size that is necessary to satisfy your parameter values (see Resources).