مدل سازي خطاي پاسخ در ـمارگيري ها به شيوه ي مصاحبه ي حضوري

Response Error Modeling in face-to-face Interview Surveys

Survey implementation is one of the common methods to data collection. Quality of survey results is of prime importance due to their applications in decision making and planning. To determine quality of survey results, it is necessary to estimate both sampling and nonsampling errors. However, estimation of nonsampling error is difficult because of problem of quantifying other parts of nonsampling error that is nonresponse error, Coverage error and data processing error. Response error is a quantifiable part of nonsampling error. Therefore, modeling of the response error has received a particular attention over years. In this paper, we consider some major statistical surveys which are conducted in a face to face interview and sources of error are determined, then a model of response error in face to face surveys has been proposed based on presentation of an experimental design. Furthermore, computation of precision in surveys with two-stage sampling designs is assessed and an estimator of precision is presented in this case. As an application of proposed model, precision of survey estimators of interest is computed based on resulted estimator, for data of household income and expenditure survey for the year 1382.