Introduction to Random Survey Sampling
It may be impossible to interview everybody within the group you'd like to survey. For example, assume that you're surveying the people who live in your city. Unless the city in which you live is very small, your staff won't be able to reach everyone. Likewise, assume that you have been given the task to survey the employees at your company. If your company is large and employs thousands of people across several locations, it may be unrealistic to collect data on each person. In both cases, random sampling may be required. Below, we'll describe the basics of random survey sampling and how to judge the accuracy of your data.
Selecting A Representative Sample
The pivotal question regarding random sampling for surveying large groups is "how many randomly chosen people do I need to survey in order to draw reliable conclusions about the group as a whole?" The answer to that question depends upon two factors. First, you'll need to determine how many people comprise the entire group. Second, you'll need to decide how accurate you'd like your data to be. In short, the accuracy of your data will be directly correlated with the relative size of the group that you survey. That is, the greater percentage of the original group that you interview, the more confidence you should have in your results.
Deriving Conclusions
In order to interpret the data that you collect from your population (that is, your randomly chosen sample), you should have an appreciation for the breadth of error in your results and the confidence level you accord the data. You're likely familiar with the concept of error in surveying. Consider political polls. You may remember hearing that the results of a poll were accurate within a certain percentage of error above and below the result (i.e. "plus or minus 4%").
You may be less familiar with the concept of confidence level in survey results. It reflects the likelihood of similar results occurring for an action that is performed 100 times. In the context of surveys, it reflects your confidence that your survey will yield similar data if it is given 100 times. It works hand in hand with your error percentage. For example, you may have a "95% confidence level, given an error of plus or minus 4%." Both factors are a byproduct of the size of your survey population in comparison to the original group.
What Is The Best Sample Size?
The "best" sample size is a subjective variable that is completely up to you. That is, the size of your population will depend upon how large an error rate you can tolerate and the level of confidence you would like to have in your survey results. For example, you may be able to tolerate a 5% error along with a 90% confidence level for one survey while requiring a 3% error with a 95% confidence level for another survey.
Executing Your Survey
Ideally, you'll be able to achieve a 100% confidence level with a 0% error rate in your data whenever you interview a random sampling of a larger group. Unfortunately, it's impossible. You'll need to tolerate an element of error in order to execute your survey with a reasonable amount of effort and within a reasonable time frame. The key is to determine in advance how accurate you want your results to be given your limitations. Surveying a randomly chosen population from a larger group is not an exact science. However, as long as you're willing to assume the potential for variance in your results, the data can still be valuable.
SurveyGizmo is a leading provider of online survey software, check them out on the web for more great ways to use surveys to enhance your business.
Article Source: ArticlesBase.com - Introduction to Random Survey Sampling