ANALYZING ENVIRONMENTAL BIOASSAYS WITH SPATIAL ODDS, RISK, AND SURVIVAL PROBABILITY RATIO REGRESSIONS

• 成果类型: Article
• 署名作者: Mondal, Debashis; Chang, Xiaohui
• 署名单位: Washington University (WUSTL); Oregon State University
• 刊物名称: ANNALS OF APPLIED STATISTICS
• ISSN/ISSBN: 1932-6157
• DOI: 10.1214/24-AOAS1923
• 发表日期: 2025
• 页码: 1086-1109
• 关键词: confidence-intervals; attributable risk; inference; toxicity; FIELDS; table

摘要

Environmental bioassays, such as sediment toxicity tests, provide abroad survey of toxicity that is crucial for the conservation and protection of marine and estuarine ecosystems. Using odds, risk, and survival probability ratios, this paper presents a critical evaluation of sediment toxicity tests data collected in the New York-New Jersey harbor area. It further derives spatial regression analysis to combine test results, predict toxicity at unsampled locations, and determine the effects of specific contaminants. The proposed spatial analysis is based on non-Euclidean distances and is applicable to complex sampling domains with nonconvex boundaries. The findings suggest that current practices can be improved with the use of relevant statistical methods with odds, risk, survival probability ratios, and attributable effects.