Maximum likelihood analysis of partially-classified contingency tables via the EM algorithm
Methods for the analysis of fully-classified contingency tables are very well known, especially among social science practitioners. However, certain process of collecting data may yield some observations which do not clearly fall under any of the underlying categories and hence result in partially-...
Saved in:
| Main Author: | |
|---|---|
| Format: | text |
| Language: | English |
| Published: |
Animo Repository
1994
|
| Subjects: | |
| Online Access: | https://animorepository.dlsu.edu.ph/etd_bachelors/16163 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | De La Salle University |
| Language: | English |
| Summary: | Methods for the analysis of fully-classified contingency tables are very well known, especially among social science practitioners. However, certain process of collecting data may yield some observations which do not clearly fall under any of the underlying categories and hence result in partially-classified tables. Such situations arise frequently in practice, but ironically, the techniques to handle them are not so well known.This thesis presents an exposition on the methods to handle such data. In particular, this thesis discusses the analysis of partially-classified contingency tables when the process that leads to the nonresponse is ignorable. The discussion is mainly based on Fuch (1982) and Little and Rubin (1987). |
|---|