Web Survey Bibliography
Under the JRA2 ESSi programme, four approaches on detection and adjustment of non-response bias are studied using different types of information such as population statistics, reluctant respondents, observable data and non-response survey. Taking the perspective of previous approaches, our focus of this paper is to explore suitable information/variable for nonresponse bias detection and adjustment. Based on the conceptual framework of survey cooperation from Groves and Couper (1998), this paper firstly identifies the type of information recorded in the paradata, such as dwelling and neighborhood as well as contact procedure variables in relation to survey cooperation, which is suitable for non-response detection. Different types of respondents are compared based on these paradata variables. Then responses of reluctant respondents are calibrated on the basis of paradata variables among non-respondents following previous approaches (propensity scores and stratification). Responses between weighted and unweighted responses among cooperative respondents are thereafter compared. Because reluctant respondents are taken as a proxy for nonrespondents, proportion of reluctant respondents in relation to total number of respondents in cross-national context is also being discussed.
Conference Homepage (abstract)