Files in this item



application/pdf9624289.pdf (6MB)Restricted to U of Illinois
(no description provided)PDF


Title:Sequential analysis of duration data with application to "reemployment bonus" experiments
Author(s):Bilias, Yannis
Doctoral Committee Chair(s):Koenker, Roger W.
Department / Program:Economics
Degree Granting Institution:University of Illinois at Urbana-Champaign
Economics, Labor
Economics, Theory
Abstract:Suitable methodology and an asymptotic theory for the sequential analysis of time-to-event (duration) data is developed and its application in "Reemployment Bonus" experiments is studied.
First, a proportional hazards regression model with possibly censored observations, time-dependent covariates and staggered entry is assumed. Sequential analysis of the partial likelihood score process and other related statistics cannot rely on martingale theory. Instead a modern empirical process theory is used to deal with the underlying two-dimensional character of the problem. A very general treatment of the large sample theory with minimum technicalities is made developed.
Next, sequential testing procedures for the analysis of duration data are employed to reconsider the Pennsylvania "Reemployment Bonus" Demonstration. The efficiency and flexibility gained by the monitoring of large-scale experiments is illustrated. Boundaries that preserve the overall type-I error are constructed based on the asymptotic theory developed previously. Several treatments show their efficacy earlier than the termination day dictated by the fixed-sample design. Stopping the study at that time would have achieved important pecuniary savings.
Last, techniques developed earlier are applied to test statistics arising from the analysis of right-truncated data in which both the duration and calendar time are naturally involved. Asymptotic theory for the one-sample and the regression problems is established.
Issue Date:1995
Rights Information:Copyright 1995 Bilias, Yannis
Date Available in IDEALS:2011-05-07
Identifier in Online Catalog:AAI9624289
OCLC Identifier:(UMI)AAI9624289

This item appears in the following Collection(s)

Item Statistics