Rastgeleleştirilmiş Yanıt Modellerinde Oran ve Ortalama Tahmin Edicileri
Abstract
Randomized Response Models(RRM) are used to reduce nonrespondents rates andbiased responses arising from surveys on sensitive subjects such as gambling,alcoholism, sexual abuse, drug addiction, abortion, tax evasion, illegal income,mobbing, political view and many others. Starting from the pioneering work of Warner[37], many versions of RRM have been developed that can deal with both proportionand mean estimation.in this study, we introduce various randomized response models, some based on theadditive and multiplicative RRM, and some from optional RRM s. New ratio andregression estimators for both proportion and mean estimation in various randomizedresponse models are proposed using auxiliary variable information. The efficiency ofthe proposed estimators with respect to existing estimators in the literatüre arecompared theoretically and a simulation study has been performed to controltheoretical results.