March 2007
Current
Research Program
of
Don A. Dillman
Washington State University
My current research emphasizes the development of effective survey methods for use in a mixed-mode data collection environment. My goal is to produce knowledge helpful for providing estimates of opinions, behavior and attributes in policy-relevant surveys, which are consistent and precise regardless of data collection mode.
Traditionally, telephone, Internet and mail survey methods were used individually for the conduct of surveys. They remain effective when used alone, for certain surveys in certain situations.
However, the shift of telephone ownership away from household phones towards individual cell phones, as well as a significant decline in response rates, is causing survey designers to rethink the ability of telephone surveys to provide high quality data. The culture of how the telephone gets used in the U.S. has changed dramatically, bringing into question our ability to use it as a sole data collection method.
In addition, lack of Internet access and skills to use it among a significant portion of the general public limits the replacement of telephone surveys by surveys conducted solely over the Internet. Use of the Internet is further limited by barriers to sampling Internet addresses and contacting individuals for whom there is no prior relationship. In addition it cannot be assumed that all members of most households have the ability to provide answers to survey questions via the Internet, or that those who do are comfortable doing so.
Although response rates to mail surveys have not undergone the decline associated with telephone interviewing, they have been limited in their use because of the lack of sample frames for conducting general public survey. That may be changing because of the availability of the U.S. Postal Service Delivery Sequence File, but its effectiveness remains to be proven.
These issues, as described in the Appendix of my book, Mail and Internet Surveys: 2nd edition, 2007 Update, and in other writings have contributed to the development of increased interest in the use of mixed-mode surveys. This use of mixed-mode surveys has resulted in the need to have a greater understanding of how and why different survey modes produce somewhat different answers to the same survey questions.
My current research program involves the use of cognitive laboratory evaluations of questions and questionnaires as well as formal experimentation, and is addressing a variety of issues:
| Under what conditions does asking identical questions across survey modes, e.g. telephone and web surveys, produce the same vs. different answers? |
| How does the structure of survey questions, e.g. converting forced-choice to check-all formats interfere with obtaining equivalent answers across survey modes? |
| How is the quality of data obtained from open-ended questions affected by survey mode, and can changing the way the way these questions are asked improve the quality of answers? |
| How can sample frames for the conduct of general public surveys by mail and the web be improved? |
| How do alternative visual formats for web surveys influence response rates and the survey answering process? |
Primary support for this research program is provided by the Department of Community and Rural Sociology and the Social and Economic Sciences Research Center at Washington State University. Research assistance is provided by SESRC staff and graduate students in the Department of Sociology. Additional support in recent years has been provided by the National Science Foundation Division of Science Resource Statistics, The United States Department of Agriculture National Agricultural Statistics, and the Gallup Organization.
Results from this research program are communicated regularly to survey designers through short courses and applied frequently in local, state, national and international surveys, and in particular those conducted by public agencies.