Conferences

One argument in favour of mixing modes of data collection in surveys is that using a combination of modes may help to reduce the likelihood of non-response bias by encouraging a potentially greater number and wider range of respondents to participate. Indeed, alternatives to the principal survey mode are increasingly being considered as a means to compensate for the fact that certain subgroups may be underrepresented in samples, either as a result of coverage bias, or due to variations in contactability and response propensity associated with that method. This paper compares patterns of nonresponse on two implementations of the European Social Survey: 1) the standard face-to-face data collection; and 2) a version of the study conducted by telephone as part of a wider programme of research into the possibility of mixing modes. Using contact data from both surveys, we examine cooperation among different subgroups in each of the two modes and draw conclusions about the potential benefits of combining face-to-face and telephone interviewing as a way to reduce non-response error in survey data.