A Comparison of ABS Mail and RDD Surveys for Measuring Consumer Attitudes

The increasing cost and decreasing coverage of Random Digit Dialing (RDD) landline telephone surveys motivated The Surveys of Consumer Attitudes (SCA) at the University of Michigan to conduct monthly experimental mail survey studies using address-based sampling (ABS). The primary objectives of the experimental studies were to evaluate the feasibility of transitioning the data collection operations from telephone to mail and to investigate differential survey errors between the two modes. Overall mail survey response rates were comparable to the RDD landline survey. Coverage improved using ABS, with more than 20% of the mail responses from non-landline telephone households not covered by the RDD landline telephone surveys. Mail survey respondents from households without landline telephones were more likely to be younger, have lower income, be renters and live in one-person households. There were no apparent measurement or reporting differences between the telephone and mail self-administered modes of data collection. Furthermore, inclusion of non-landline telephone households did not result in any substantial demographic or economic attitude differences between the two approaches.