EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT COOK ET AL. A META-ANALYSIS OF RESPONSE RATES IN WEB- OR INTERNET-BASED SURVEYS

Response representativeness is more important than response rate in survey research. However, response rate is important if it bears on representativeness. The present meta-analysis explores factors associated with higher response rates in electronic sur-veys reported in both published and unpublished research. The number of contacts, per-sonalized contacts, and precontacts are the factors most associated with higher response rates in the Web studies that are analyzed. During each election cycle, we are bombarded with the results of national polls updating us on the views of millions of potential voters. These surveys are usually based on responses of only an infinitesimal proportion (e.g., 2,000/ 150,000,000 = 0.000013) of the total population. Yet these portrayals are of-ten stunningly accurate. Election polls make clear that the representativeness of our samples is much more important than the response rate we obtain. A sample of fewer than 1 % of the population can be more representative, indeed much more rep-resentative, than a sample of 50 % or 60 % of the population. Thus Krosnick (1999) emphasized in his recent survey of the paper-and-pencil response-rate literature, But it is not necessarily true that representativeness increases monotonically with increasing response rate... recent research has shown that surveys with very low response rates can be more accurate than surveys with much higher response rates. (p. 540) Nevertheless, we remain concerned about our response rates if we are un-certain with regard to sample representativeness. We know that ultimately, i