An Analysis Of The Factors Affecting No Response To Surface

Response rate is an important indicator to measure the quality of data. The value ofresponse rate can influence the bias which is generated by statistical inference. Thetopic of this article which is based on the analysis of paradata of “Chines GeneralSocial Survey (2008)” is the factor to unit nonresponse. By this study, it is found thatthere are area effect, interviewer effect and survey-execute effect. In different area, inorder to gain higher response rate, survey-execute should adjust measures to localconditions. Actually, different characteristic of interviewers also affect response rate. Itis better to choose the interviewers with higher-education,20-30aged students. At thesame time, the workload of interviewers should not be oversize or undersize. Theworkload of each interviewer in one survey should not more than50. Based on thesethree levels variables, this study establish logistic nesting regression model, andanalyze the explanatory power of each level. At last, this article gives some advice todoing survey.