Q: I am looking at class survey results and am wondering if there is a benchmark for the percent of surveys that we need back in order for the results to be reliable.
A: If you have 400 responses, your data will have a margin of error of 5%.
Let's say, however, that you have only 200 responses. Your margin of error will be higher.
To use this chart, you need to compare two responses at a time to determine the degree of accuracy. If 80% of respondents answer one way, and 20% choose another response, then the margin of error will be 6%. Therefore, at a 6% margin of error, the likelihood of accuracy is between 74% and 26%. The 80% of responses will be 74% accurate, in other words.
However, if you have 200 responses and 60% choose one response and 40% choose another, your margin of error goes up to 7% and the difference in response is much closer to 50-50. (53% to 47%).
If you have only 50 responses, but the response ratio is 75% for one response and 25% for another response, your margin of error is 12%. Here is an example: Let's say you ask people whether they prefer courses that meet on Saturday morning or Wednesday evening. Fifty people respond to your survey. 75% say Saturday Morning and 25% say Wednesday evening. However, this is not a large sample, so there is more chance for error. Although 75% of your respondents say they prefer Saturday, you can not confidently project that this applies to your entire market, because you have a 12% possibility for error. You can, however, feel confident that the majority prefer Saturday morning. With a 12% margin of error the percentage preferring Saturday is 63%. Likewise, as many as 37% of respondents prefer Wednesdays, based on the margin of error.
I know this is may seem complicated, so if you have questions, let me know and I will try to make it more clear. The nice thing about the chart is that you don't have to know statistics to figure the reliability of your survey responses. Just use the chart, and you will have a very good idea of where you stand. Thanks for asking!
Questions
Benchmark for Reliable results
May 23, 2018 12:36:03 AM
by
Jordan Kivley
|