All Times ET
Program is Subject to Change
Transitioning to a Mixed-Mode Data Collection: Questionnaire Considerations and Data Quality Impacts (308207)Danielle Mayclin, U.S. Energy Information Administration
*Joelle Davis Michaels, U.S. Energy Information Administration
Keywords: mixed-mode, data quality, questionnaire design
The Commercial Buildings Energy Consumption Survey (CBECS) is a periodic survey that produces the only independent, statistically representative source of national-level data on the characteristics and energy use of commercial buildings. It has been conducted since 1979 and has evolved over the years from an interviewer-administered paper-and-pencil questionnaire to CAPI/CATI, and most recently to a mixed-mode design including both CAPI/CATI as well as a self-administered web version. To accommodate the new web mode, the 2018 CBECS was designed using Blaise 5, which allows a single database that can be used for both the CAPI/CATI and the web mode.
Adding a self-administered web mode after more than thirty years using an in-person questionnaire involved many design considerations. Some of the elements that required changes were: allowing for don’t know or refused responses, eliminating most edit checks within the questionnaire, providing if volunteered categories for rare responses, rostering numeric responses, and collecting other specify responses. This paper will first describe the variety of questionnaire considerations encountered, as well as the design solutions implemented in the 2018 CBECS. Then, the response data from both modes will be used to examine the effects of these design changes on data quality. This will include looking at item nonresponse rates, interview question timings, edit failure rates, and other variable attributes.