Comparison of Estimator Parameter in Different Methods with Missing Outcome under Logistic Regression / 邏輯斯迴歸模型存在反應變數缺失下各種估計法及有效性比較—敏感性問題之探討

碩士 / 逢甲大學 / 統計與精算所 / 99 / Randomized response technique (RRT) is usually invited to help researchers to acquire more reliable data in a survey research when sensitive questions are involved to which respondents tend to give dishonest responses or resistance.
Based on Greenberg et al.(1969) technique of unrelated-question and the outcome data are missing at random (MAR), use some randomized response models, Full Data Estimator (FD), Complete-Case Estimator (CC), Validation Likelihood Estimator (VL),Joint Conditional Likelihood Estimator (JCL), Weighted Estimator (WE), and Joint Weighted Estimator (JWE),
are adopted to implement estimation of logistic regression of data. In the process of analysis, statistic simulations have been proposed to make a comparison of reliability among varieties of estimators adpoted in this study. Finally, the above six estimators are altogether nvited to complete the goal of this research -- the analysis of the proportion of illegal piracy of TV channels.

Identiferoai:union.ndltd.org:TW/099FCU05336017
Date January 2011
CreatorsYi-chin Lin, 林怡親
ContributorsShen-ming Li, 李燊銘
Source SetsNational Digital Library of Theses and Dissertations in Taiwan
Languagezh-TW
Detected LanguageEnglish
Type學位論文 ; thesis
Format51

Page generated in 0.0129 seconds