《Use of paradata derived from vocal》.pdf

Use of paradata derived from vocal properties of interviewers and respondents in researching survey participation Frauke Kreuter June 26, 2014 Acknowledgement Support: Charles Cannell Fund; NSF Grants No. SES-0819734 and No. SES-0819725 Co-Investigators: José Benkí, Jessica Broome, Frederick G. Conrad, Robert M. Groves Statistical Consulting: Susanne Hüttinger, Alexandra Birg, Helmut Küchenhoff Coding, Transcribing, Acoustic Analysis, Sampling: P. Batra, R. Benner, K. Franckowiak, H. Gu, B. Jarvi, E. Kordupel, P. Kotvis, a. Lincoln, L. Linstrom, M. Littlefield, D. Lopez, C. McClain, J. Matuzak, C. McCarty, P. Maher, G. Moss, K. Mull, D. Nielsen, D. Perkins, F. Pacheco, D. Popielarz, C. Sheffler, A. Tatro, D. Vanette, and D. Vollans Kreuter (JPSM IAB/LMU) Voices RSS 2 / 21 The Puzzle in CATI Interviewing CATI introductions are usually fairly scripted. Very little information is conveyed in the first few seconds or minutes of a CATI call. We see variation across interviewers in their success in gaining cooperation over the phone (both within a survey and across surveys). Hypothesis: Differences in interviewers verbal attributes play an important part in outcomes (Oksenberg, Coleman Cannell, 1986; Oksenberg Cannell, 1988). Kreuter (JPSM IAB/LMU) Voices RSS 3 / 21 Past Findings of Speech Patterns on Participation Positive effects on sample unit participation: Higher pitch for interviewers’ voices (Sharf Lehman, 1984; Groves, O’Hare, Gould-Smith, Benki, Maher 2008). Lower pitch for male interviewers (Benkí, Broome, Conrad, Groves, and Kreuter 2011). Less scripted, more extemporaneous deliveries of surve