International Conference on Total Survey Error

Do Interviewers with High Cooperation Rates Behave Differently? Interviewer Cooperation Rates and Interview Behaviors.

Dr Kristen Olson (University of Nebraska - Lincoln), Dr Jolene Smyth (University of Nebraska - Lincoln), Dr Antje Kirchner (University of Nebraska - Lincoln)

Abstract: Interviewer skills for obtaining cooperation from sampled households require flexibility, tailoring to the respondent, and maintaining interaction. On the other hand, administration of survey questionnaires requires reading questions exactly as written, nondirective probes, and a clear set of regimented behaviors during the interview. That is, interviewers are required to be flexible during recruitment, but standardized during measurement. These skill sets may be at odds. This paper will examine behavioral differences in the survey interview between interviewers who are more successful at gaining cooperation and those who are less successful. We examine whether question misreadings, probing, feedback, disfluencies, and clarifications differ for interviewers with higher versus lower cooperation rates. We use the Work and Leisure Today Survey (n=450, AAPOR RR1=4.7%), including survey data, paradata, and behavior codes. Preliminary analyses indicate that interviewers with higher cooperation rates deviate more from the question wording, introducing (major) changes to the question stem or response options more often than interviewers with lower cooperation rates.

Paper accepted for presentation at the International Conference on Total Survey Error, Baltimore, MD. September 2015

Relationships between data quality and confidentiality

Dr. Jerry Reiter (Duke University)

The agency collects microdata that it wants to disseminate to public. Doing so is challenging, because the data missing and faulty values. data subjects' con dentiality must be protected. The key question is how do we release microdata that  are free of implausible values, have low risks of disclosure, and are useful for a wide range of analyses?

The role of statistical disclosure limitation in total survey error

Dr. Alan Karr (RTI International)

Statistical disclosure limitation (SDL) is an alteration of data implemented in order to protect respondent privacy and dataset confidentiality.  It is the only form of survey error that is introduced deliberately. It is fully controllable at effectively no cost. To date, there has been a disconnect. But, this can change! 

Sep 19, 2015 to Sep 22, 2015
Renaissance Harborplace Hotel
Baltimore, MD
United States