Surveying the General Public over the Internet Using Address-Based Sampling and Mail Contact Procedures

We report results from two statewide experiments in Washington designed to test potential methods for using postal mail to obtain survey responses over the Internet from address-based samples of general public households. The five methods we test are: 1) sending Web and mail modes of response sequentially; 2) providing a prepaid 5 incentive; 3) offering an instruction card for responding over the Web; 4) sending the follow-up request by Priority Mail; and 5) providing an additional 5 incentive with this follow-up request. Results are evaluated from the standpoint of response rates, demographic representativeness of respondents, and survey costs and data collection times for Web and mail modes. We find that a “Web-plus-mail” design—mailing an initial Web request followed by a mail request—with the prepaid incentive offers much potential for obtaining Internet responses. In addition, the mail follow-up to the initial Web request significantly increases overall response rates and improves respondent representativeness. However, the results also show that a mail-only design consistently obtains higher response rates and a demographically similar sample compared to Web-plus-mail. We also find significant challenges in obtaining a representative sample of respondents with either Web or mail modes and in realizing the potential cost and speed benefits of the Internet in using postal mail contacts.