Representation of vulnerability and the elderly. A Total Survey Error perspective on the VLV survey

TitleRepresentation of vulnerability and the elderly. A Total Survey Error perspective on the VLV survey
Publication TypeBook Chapter
Year of Publication2016
AuthorsOris, M, Guichard, E, Nicolet, M, Gabriel, R, Tholomier, A, Monnot, C, Fagot, D, Joye, D
EditorOris, M, Roberts, C, Joye, D, Ernst Stähli, M
Book TitleSurveying human vulnerabilities across the life course
Series TitleLife Course Research and Social Policies
Place PublishedNew York
ISBN Number978-3-319-24157-9
Keywordselderly, methodology, non-response, participation rate, survey, Switzerland

This chapter focuses on the “Vivre-Leben-Vivere” (VLV) research and explains how the team in charge has dealt with the issue of representing old people in Switzerland and the vulnerable persons within this population. For this purpose, we draw inspiration from the “total survey error” perspective and discuss the procedures that were used to collect the information, but also the quality of the information itself. We present some of VLV’s methodological choices, such as the contact procedures that were designed to “capture” the most vulnerable individuals, especially in a context where refusal rates for participation in surveys are increasing. We also consider the missing values in responses to some “complex” topics. Results show that participation rates are highly related to the effort that was made in order to recruit the individuals. This is critical for the VLV project, where people can show different degrees of vulnerability related to age, health, and/or poverty. In this sense, efforts such as repeated calls, home visits, and the proxy procedure were successful both in a general perspective and more specifically for a fair inclusion of vulnerable persons in the survey. At the same time, the non-response rates show to what extent it is difficult to explain the personal decision not to respond to a survey or to a question. We can find some traces of “interviewer effects” in the analysis of item non-responses, which we have done by using a logistic regression for one of the psychological tests included in VLV. Some aspects relate to interviewer and respondent characteristics, but others relate to the interview context, all highlighting the complex social interactions at stake.