Collect the desired information from the entire population if the population is small. If the population in question is the students in a high school classroom, the data can easily be taken from the whole group. If the population is the entire human population of the earth, this is not possible. The exact point at which sampling the entire population becomes unfeasible depends upon the cost of sampling and the resources available to the researcher.
Calculate the desired sample size using the formula n = N/(1+ N(e)²), where “n” is the sample size, “N” is the population size and “e” is the desired margin of error, when the population size is less than 100,000. For example, if data was needed from a population of 3,500 college students with a margin of error of 5 percent, and polling the entire group was out of the question, the minimum number needed for the sample would be determined by n= 3500/(1+ 3500(.05)²), or 359. This formula is for dichotomous data, in which the attribute either is or is not present.
Calculate the desired sample size using the formula n = (Z²pq)/(e)², when the population is more than 100,000. In this formula, “p” is the proportion of the population having an attribute, “q” is 1-p, “e” is the desired margin of error and “Z” is the number of standard deviations needed to contain the desired confidence level for the study. For a confidence level of 95 percent, Z would be 1.96, since 95 percent of a normal population are contained within plus or minus 1.96 standard deviations of the mean. This formula is also for dichotomous data.
Calculate the sample size for continuous data using the formula n= (Z²σ²)/e², where “σ²” is the variance of the attribute within the population, and “e” is the desired margin of error. This formula works well when a good estimate of the population variance is known, but that is not always the case.