2003 Q2 Survey Response Rates and Confidence Levels
Invariably, when conducting surveys, there is often confusion regarding acceptable response rates. Response rates can vary considerably depending upon a number of factors. These factors include the methodology used, i.e. mail, on-line or telephone, the number of questions asked, incentives used and how “user friendly” the questionnaire appears to the respondents. Although it is common to judge the success of a survey based on the percent of responses that are returned, this could be very misleading. Rather, the survey success should be predicated on achieving a high level of confidence in the results. The industry standard is generally a 95% level of confidence with a + 5% margin of error.
For example, if our survey population is 100, then we need to achieve a return of 80 (80% response rate) in order to have a 95% level of confidence in the results. Likewise, if our survey population is 500, we need to achieve a return of 218 (43.6% response rate) in order to have a 95% level of confidence in the results. For survey populations of 1,000 and 2,000, returns of 278 (27.8%) and 323 (16.2%) are needed respectively. You will notice that as the size of the population base increases the response rate percentage goes down significantly.
Often times, the results of surveys can be considered invalid if there are an insufficient number of responses. The importance of achieving a 95% level of confidence in the results is vitally important if you are basing management decisions on the information, establishing a baseline for comparing results from different periods or basing employee compensation on the results.