Factors, preventions and correction methods for non-response in sample surveys

摘要

Missing survey data occur because of unit and item non-response. This is practically independent of the method of data collection. As a result of the bias that non-response sometimes introduces in survey estimates, identifying factors that promote it, and taking measures of prevention and correction methods are clearly necessary. The standard method to compensate for unit non-response is by weighting adjustment, while item non-responses are handled by some form of imputation. This paper reviews factors that give rise to nonresponse and the corresponding methods used for its prevention and control. It also discusses their properties.