Sequential nonparametric procedures for testing the equality of two lifetime distributions based on progressively censored data

• Published in: Sequential analysis Vol. 35; no. 2; pp. 158 - 174
• Main Authors: Zhu, Xiaojun, Ng, Hon Keung Tony, Tang, Man Lai
• Format: Journal Article
• Language: English
• Published: Philadelphia Taylor & Francis 02.04.2016; Taylor & Francis Ltd
• Subjects: Censorship; Computer simulation; Equality; Lehmann alternative; life testing; Monte Carlo methods; Monte Carlo simulation; Monte Carlo simulations; Performance evaluation; precedence statistics; Recommendations; Samples; Statistical analysis; Statistical methods; Statistical tests; Statistics; two-sample problem
• ISSN: 0747-4946; 1532-4176
• DOI: 10.1080/07474946.2016.1165525

Many nonparametric tests have been formulated to test the equality of two lifetime distributions (e.g., Wilcoxon rank-sum test and sign test), and some of these nonparametric tests can be applied to censored samples (e.g., precedence-type tests and log-rank test). In this article, we study several sequential nonparametric test procedures based on the precedence-type statistics when the lifetime data are observed sequentially and one of the two samples is progressively Type-II censored. Exact power function under Lehmann alternatives is derived. Monte Carlo simulation is used to evaluate the performances of these sequential test procedures. Conclusions and recommendations are provided based on these results.