Cascade analysis for breakoffs in a computer-assisted telephone interview

Abstract

The purpose of the article is to design a simple and reliable method of a posteriori analysis of the questionnaire allowing identifying the problems and testing the overall design of the questionnaire. The study is based on the data array containing 4000 completed interviews and 1571 break-offs carried out in August 2013; the expanded questionnaire of the telephone survey involved questions concerning social and economic status and financial income of the households. The object of the study was break-offs. Three types of refusals were singled out: (1) retrospective refusal, when respondent is searching for a certain reason; (2) current refusal, i.e. actualization of the inadmissibility of the question; (3) prospective refusal based on the expectation from the questionnaire. Quantitative and qualitative analyses of the break-offs data were conducted. Quantitative assessments of the contextual integrity of the questionnaire were presented; tensions caused by communicative failures were spotted. Transcripts describing failures were analyzed in details. A typology of the reasons for the communication breakdown was designed. The survey revealed that the breakpoints in the questionnaire are not heterogeneous; there are upsurges and communicative tensions which are connected with the contextual troubles rather than the wordings. Questions with breakpoints border on semantic blocks. Interview breaks are dependent on the communicative strategies of politeness. The authors` findings suggest that it is needed to forecast and control the refusals; the tool might be the analysis of the break-offs.