Survey Sampling Basics: Analytical Plan, Population Variability & Confidence Intervals

Last time, we discussed some questions we should ask ourselves when we need to determine how many people we should approach to participate in our survey. Let's continue the conversation now with some more factors to consider when determining sample size:

Deciding on sample size is one of the steps we have to take in the early planning stages of any survey. In order to determine a survey sample, we need to consider these factors:

Our Analysis Plan for Data:
Although it may seem odd to plan out how you're going to analyze data you haven't even collected yet, that's exactly what needs to happen! The planned analysis for the results should be considered before you ever ask anyone to participate in your survey. Why? Well for one reason, there are certain analyses that can be done that require a certain number of responses per question item. Additionally, if you'd like to see if there are statistically significant differences between groups (e.g. males and females), the sample size needs to be adjusted based on that planned analysis.

Differences in Your Population: This is the amount of variability in your target population. If there is a lot of variability in the issue you'd like to research (for example: gun control in the United States), a large sample should probably be sought. A behavior or opinion in your population that occurs either very infrequently or very frequently will have less variability than if it were to occur 1/2 the time, so a smaller sample size can be allowed.

Level of Confidence: 
This is the level of risk we are willing to tolerate, usually expressed as a percentage (e.g. 95% confidence level or interval). The more confident you want to be that your results are falling within a certain range, the larger the interval that will contain the true value of the estimate is needed, which leads to lower levels of precision. More on Type I and Type II errors.

Next time we will continue the discussion of sample size for your survey efforts-there's more you need to know before you begin your survey research!
blog comments powered by Disqus
Crimes in Design Webinar
Subscribe to our Monthly Newsletter