Conferences

ESRA2009:

Conference main page

Overview of sessions

Time table

Warsaw 2009: Presentations and short courses

The use of behaviour coding for the convergence evaluation of proxies and direct informants responses in a disability survey.

Session: Interviewer and Respondent Behaviour in Survey Interviews

Authors:

Isabel Benítez; University of Granada, Spain
Jose Luis Padilla; University of Granada, Spain
Yfke P. Ongena; University of Twente, Netherlands

Abstract:

The use of proxy informants in household surveys has increased the interest in analysing the agreement of their answers with the direct informants’ answers. Behaviour coding can provide useful additional information in the analysis of the proxy informant’s behaviour when answering to survey questions.

The aim of the present study was to evaluate the agreement between direct informants and proxies with additional information based on behaviour coding. The evaluation of the convergence between both types of informants was realized during the pretest of a questionnaire used in a disability survey. This questionnaire includes questions related to different limitations about daily life activities. 16 direct informants and 16 proxies answered to 11 “target” questions about the same content for both types of informants. The agreement level was evaluated between the information provided for both types of informants. The influence of informants’ characteristics and questions characteristics on the convergence levels was also evaluated. The analyses were carried out using the program Sequence Viewer (Wil Dijkstra, version 4.4.a. 2008).

The results showed a 52% of agreement in the answers of the informants, being these answers different in 19% of the cases. The remaining 30 % concern other situations in which the answers of one of the informants or of both were inadequate (as detected by means of behaviour coding) and therefore the comparison between them was not possible. Differences were also found in the percentage of agreement between the different couples and between the questions of the questionnaire.

A possible explanation might be the difference in the level of salience for type of informant at the moment of answering, or the quantity of available information, which can facilitate or inhibit the work of translating their response in an alternative. Future research will be focused on the study of other factors which can influence the convergence between proxies and direct informants, as well as on the definition of the “ideal profile of the Proxy “.