Innovations in Computer-Assisted Data Collection
|Convenor||Ms Beth-ellen Pennell (Institute for Social Research, University of Michigan )|
|Coordinator 1||Ms Patty Maher (Institute for Social Research, University of Michigan )|
|Coordinator 2||Ms Gina Cheung (Institute for Social Research, University of Michigan )|
More and more survey researcher organizations utilize more than one mode of data collection (i.e., mixed-mode surveys/multi-mode surveys) in survey projects in order to capitalize on greater potential response rates and/or improved coverage of the population of interest. However, designing and managing the collection of mixed-mode surveys for appropriate quality dataset is challenging for all survey organizations. This presentation will discuss the technical and methodological challenges and opportunities in mixed-mode data collection.
Open-ended questions provide potentially richer data than closed-format responses. In CAPI surveys, interviewers have typically typed responses, leading to variation in the quality of typing and coding. Audio-recording verbatims offers potential advantages in data quality, however consent to be audio-recorded is unlikely to be random. In this paper we weigh the costs and benefits of audio-recording OEQs, in a nationally representative face-to-face survey. We conclude that while recording reduces interviewer variation and improves data quality, consent is non-random and both consent and this question type favour the more knowledgeable.
This presentation focuses on using keystroke data, household, and geographic characteristics available to monitor interviewer’s behavior in two national household surveys using CAPI in the Kingdom of Saudi Arabia and Ghana. We propose a list of quality control metrics to improve data collection efficiency and data quality of interviewer administered surveys. We demonstrate how making quality control metrics available is extremely important when conducting international work that is centrally coordinated. These tools in conjunction with the analysis of rich paradata allow for real-time monitoring and control even in widely dispersed and geographically distant field operations.
This presentation will describe innovative uses of rich paradata (process data) to monitor production and quality indicators in household surveys in international settings. Data collection technologies that have been widely used in the developed world for decades are increasingly being adapted and used in innovative ways in new contexts. With these innovations come new challenges, however. The presentation will discuss advantages, challenges and lessons learned across a diverse set of projects as well as make recommendations for new developments in this space. The presentation will provide examples from projects in China, Ghana, India, Nepal, and the Kingdom of Saudi Arabia.