Academic journal article Demographic Research

"I Didn't Write the Questions!" Negotiating Telephone-Survey Questions on Birth Timing

Academic journal article Demographic Research

"I Didn't Write the Questions!" Negotiating Telephone-Survey Questions on Birth Timing

Article excerpt


This paper examines interviewer-respondent interaction in the collection of demographic data where interviewer and respondents speak the same first language. Conversation analysis (CA), or the analysis of talk in interaction, makes transparent the interaction between an interviewer and 25 respondents on a question about pregnancy and birth timing in an Australian telephone survey, Negotiating the Life Course. The analysis focuses on the troubles that occur and the work that interviewers do to fit respondents' answers to the survey researcher's categories. Interviewers are shown to act as mediators in difficult interaction, with responses often distorted by question format, the imperative of achieving an allowed response, and the need to keep the respondent in the survey. The analysis suggests that conversational resources could be used constructively to ensure a better fit between questions and responses.

1. Introduction

Where most households have telephones, population data are increasingly gathered through telephone surveys. In general, survey researchers work on the assumption that standardized surveys with fully scripted identical questions, delivered to each respondent exactly as worded by well-trained interviewers and probed in a neutral way, will on the whole obtain reliable and valid responses. At the same time survey methodologists acknowledge that full standardization is tricky and some suggest a more flexible approach to standardization (Maynard and Schaeffer 2002; Schober and Conrad 1997; Suchman and Jordan 1990).

Demographers, heavily reliant on survey data, have also been part of the 'larger movement' (Massey 1987:1516) of those advocating a more flexible approach. In the face of dissatisfaction with data quality and criticism of the efforts of the World Fertility Survey (WFS), and later the Demographic and Health Survey, to collect comparable data in a variety of settings (Caldwell 1985; Davis 1987; Stone and Campbell 1984; Ware 1977), some researchers incorporated a variety of adaptations and alternatives in their international fieldwork. These new approaches combined elements of both quantitative and qualitative research strategies (Axinn, Fricke, and Thornton 1991; Caldwell, Caldwell, and Caldwell 1987; Dyson and Moore 1983; Massey 1987). Demographers also incorporated the use of tape-recordings and transcripts (Adeokun 1981, Davis 1973, Gibril 1979, Pool and Pool 1971, Quandt 1973, all cited in Thompson, Nawab Ali, and Casterline 1982:8; Weinreb 2006). Yet, despite the importance of surveys in the collection of demographic data, investigations of the role of interaction in both quantitative and qualitative demographic data collection have been rare.

In the 1975-76 Bangladesh Fertility Survey report, Thompson et al. (1982:7) observed:

The interchange between the interviewer and the respondent is the keystone of the whole enterprise, yet evidence about this interchange is entirely indirect. Tape-recordings of the interviews offer a more direct view into the interviewing itself? with tape-recordings an opportunity is presented to consider in unusual detail and depth many aspects of interviewing.


The great value of the transcripts is the vivid picture they provide of the dynamics of the interview, including subtleties of social interaction not easily observed by any other means (Thompson et al. 1982:11) .

The focus on 'foreign' survey populations reflected the concern of researchers that translation of the survey instrument from English or French into the languages of respondents, an integral part of large-scale survey enterprises such as the WFS, might affect the reliability and validity of the data (Thompson et al. 1982:8). This concern does not seem to have been explicitly or convincingly expressed in relation to surveys where interviewers and respondents spoke the same language until Suchman and Jordan's (1990) analysis of five videotaped interviews and transcripts from the US General Social Survey and the National Health Interview Survey. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.