The sample size calculator on the Project Dashboard can be used to identify the ideal number of respondents you should have for your survey and/or the number of people to whom you should distribute the survey, given your expected response rate. This tool is designed for use with larger populations from which you want to draw a sample; if you are dealing with a small population (e.g., 50 people who attended a workshop), it is suggested that you send to everyone in the population rather than trying to extract a sample.
There are several fields in the sample size calculator that will help you identify the appropriate sample size needed for your survey:
- Confidence Level - This measures how accurate or reliable you want your data to be. The percent indicates how confident you want to be that your results are correct. If you aren’t sure what level to use, we recommend using 95%, which means that you can say with 95% certainty that your results are correct.
- Confidence Interval - This asks “How much error are you willing to tolerate?” Think of this in terms of the polls you see in the news where they say the results are accurate within ±3%. For the confidence interval, they are saying that if you gave this particular survey over and over again to different groups of people, the results would always fall within that percent (plus or minus) of the mean you obtained in your survey. If you aren’t sure what percent to use, we recommend using anywhere between 3-6%.
- Population Size - This is the total number of people that you could send the survey out to. This is the total number you hope to pull your sample from. For example, your population may be all residential students on your campus. If you have 5000 students living on campus, you would enter 5000 as the population size.
- Estimated Response Rate - This is the average number of people that typically complete your surveys. You may know this if you have calculated it for past assessments you have completed, but if not, use the national average of 20%.
- Sample Size Needed - This is a number that is calculated based on your population size and the levels of error and confidence you set. This is the recommended number of people for whom you should have data once your survey has been completed.
- Send Out To - This is a number calculated by taking the sample size and multiplying it by the estimated response rate. Typically you need to invite more people to participate so you can get a completion rate that meets your sample size needed figure. For example, if you needed 100 people to complete your survey (100 is your sample size needed) and you typically have a 20% response rate, your “send out to” would be 500 people.