Estimation of response errors in complex sample surveys

---

Tezin Türü: Doktora

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Fen Edebiyat Fakültesi, İstatistik Bölümü, Türkiye

Tezin Onay Tarihi: 2013

Öğrenci: FİDAN FAHMİ

Eş Danışman: İNCİ BATMAZ, HÜSEYİN ÖZTAŞ AYHAN

Özet:

In this thesis, simple and correlated response errors are investigated. The optimum interviewer allocation settings has been investigated by using different experimental design models (Nested design, Nested and factorial factors design and Split plot design). The proposed designs considers several stages of an interactive process. Several linear and multiplicative models have been used to explain the factors effecting nonsampling errors. This naturally includes response error models. The response error models are generally evaluated as simple response errors, and some other proposals are used to evaluate them as correlated response errors. Correlated response errors, naturally takes the interaction between the interviewer and the respondent into account. In this study, after an exhaustive literature review, the interviewer fieldwork allocation techniques are investigated by the proposed experimental design models to study the sources of error. The pilot survey and main survey results are presented in detail.