Weighting adjustment techniques, adopted in the estimation phase to compensate nonresponse effect, are based on the use of auxiliary information known both for respondent and not respondent units, usually without distinguishing among different types of nonresponse. This paper proposes to treat separately the components of nonresponse, i.e. noncontact and refusal, with the aim of reducing nonresponse bias. Our weight adjustment method takes into account the sequential nature of the response process. Nested models are used to construct nonresponse adjustment weights in a two-stage response process, i.e. contact and participation conditional on contact. The assumption underlying this approach is that, conditional on auxiliary information, the different types of response are linearly independent.
Treatment of total nonresponse via sequential weight adjustment in the Italian disability survey
COCCHI, DANIELA;
2012
Abstract
Weighting adjustment techniques, adopted in the estimation phase to compensate nonresponse effect, are based on the use of auxiliary information known both for respondent and not respondent units, usually without distinguishing among different types of nonresponse. This paper proposes to treat separately the components of nonresponse, i.e. noncontact and refusal, with the aim of reducing nonresponse bias. Our weight adjustment method takes into account the sequential nature of the response process. Nested models are used to construct nonresponse adjustment weights in a two-stage response process, i.e. contact and participation conditional on contact. The assumption underlying this approach is that, conditional on auxiliary information, the different types of response are linearly independent.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
