Description: 

The survey interview consists of spoken interaction between an interviewer and a respondent. Whether the interviewer is speaking extemporaneously, typical of the invitation, or adhering to a script, such as during a standardized interview, the quality and nature of the spoken language can affect both participation and responses. Theoretical frameworks for explaining variability in spoken language, developed by researchers in psycholinguistics, sociolinguistics, and speech science, have much to contribute to the investigation of the role of language in surveys. In this seminar we will apply theories and findings about speech production (e.g., prosody, accommodation and convergence) and conversation (e.g., grounding, the cooperative principle) to the study of survey interviews and survey recruitment. In addition, the course will include practical introductions to computational speech analysis and interaction coding using the software packages Praat and Sequence Viewer respectively. In addition to considering phenomena within a particular language, we will examine the implications of using two or more languages for surveys conducted in bilingual or multilingual contexts, as is increasingly common in the U.S. and across the world. Course requirements will include 5 exercises on acoustic analysis (transcription and turn-marking, duration measurement, pitch analysis, vowel formant analysis, and resynthesis), 3 short papers, and one or two class presentations on the weekly readings. Paper 1 will combine the 5 acoustic analysis exercises into a single report. Papers 2 and 3 each will present a research proposal or experimental design to investigate spoken language in survey measurement or recruitment.