Methods for Sensitivity Analysis When Nonresponse is Nonignorable

Access this Article

Other Title: Methods for Sensitivity Analysis When Nonresponseis Nonignorable

博士論文

資料形態 : テキストデータプレーンテキスト

コレクション : 国立国会図書館デジタルコレクション > デジタル化資料 > 博士論文

総研大甲第421号

Contents
PREFACE
ACKNOWLEDGEMENT
1 INTRODUCTION AND LITERATURE REVIEW
1.1 BACKGROUND
1.2 MISSING DATA IN SAMPLE SURVEY
1.3 MISSING DATA MECHANISM AND ANALYSIS TECHNIQUES
1.4 SELECTION BIAS AND LOCAL SELECTIVITY
1.5 MISSING DATA PATTERN
1.6 TEST TO DETERMINE THE MISSING DATA MECHANISM
1.7 INFORMATIVE CENSORING
1.8 SELECTION MODEL AND PATTERN-MIXTURE MODEL
1.9 BIVARIATE DISTRIBUTION
1.10 SEMIPARAMETRIC APPROACH
2 SENSITIVITY ANALYSIS OF THE SURVIVAL FUNCTION WITH INFORMATIVE DROP-OUTS FOR WEIBULL DISTRIBUTION
2.1 INTRODUCTION
2.2 MODEL
2.3 SURVIVAL FUNCTION
2.4 EXAMPLE
2.5 CONCLUSIONS
2.6 APPENDIX A
2.7 APPENDIX B
2.8 APPENDIX C
2.9 APPENDIX D
2.10 APPENDIX E
3 REGRESSION MODELS WITH NONIGNORABLE NONRESPONSE
3.1 INTRODUCTION
3.2 NONIGNORABLE MODEL
3.3 PROPERTIES
3.4 ESTIMATION
3.5 EXAMPLE
3.6 CONCLUSIONS
4 UNIFICATION OF SELECTION MODEL AND PATTERN-MIXTURE MODEL WHEN THE MISSING MECHANISM IS NONIGNORABLE
4.1 INTRODUCTION
4.2 COPAS AND EGUCHI'S MODEL
4.3 UNIFICATION
4.4 CONCLUSIONS
BIBLIOGRAPHY