Methodological Studies on Data Quality and Data Collection

MPRC faculty are actively involved in a series of studies to improve the quality of data and the quality of the documentation of data. MPRC faculty associate Frauke Kreuter (JPSM) has investigated how interviewers can affect the quality of reports. She has shown that some groups are more susceptible to interviewer effects than others. For example, due to cognitive impairments and higher dependency on the interviewer, elderly respondents tend to be more sensitive to interviewer influences. With computer-assisted interviewing systems, more and more paradata are being collected during the survey processes. Kreuter is investigating how this auxiliary information on the nature and context of respondents’ interactions with interviewers can be used to better understand survey responses. She has been testing hypotheses about the mechanisms for interviewer effects in real survey settings, constructing proxy indicators to be measured in routine practice, and demonstrating the beneficial statistical use of these proxy data for substantive analysis. She has an R03 pending to support this work.

Katharine Abraham (JPSM) and Suzanne Bianchi (Sociology) along with graduate student Aaron Maitland have examined nonresponse in the ATUS, an important issue given response rates below 60 percent. In a paper forthcoming in Public Opinion Quarterly, they show that noncontact is far more common than outright refusals to participate in the survey. Patterns suggest that it is not the busiest people who are missed in the ATUS but rather those who are loosely connected to households or to the community. Because the ATUS is embedded in the CPS, more is known about nonrespondents than in the typical survey context. This analysis contributes to the ongoing debate in the survey research literature about the relationship between response rates and nonresponse bias; low response rates do not always introduce bias just as high response rates do not always ensure against bias.