LE3 .A278 2016

2016

Peng, Jianan
Hooper, Jeff

Acadia University

Master of Science

Masters

Mathematics and Statistics

Mathematics & Statistics

Risk assessment studies, in which human, animal or ecological data are used to set safe low dose levels of a toxic agent, are challenging as study information is limited to high dose levels of the agent. In the literature simultaneous inference for low-dose risk estimation with quantal data for multistage models and Abbott-adjusted models have been proposed but there is no simultaneous inference for log-probit models. In the first part of this thesis, we consider both hyperbolic bands and three-segment bands for log-probit models which are widely used in risk assessment. The proposed method is illustrated with real data applications. The second part of this thesis is about the parametric bootstrap (PB) approach which has been proposed by Krishnamoorthy et al. (2007). In certain types of applications, the prior knowledge imposes a prespecified order upon the population means. The PB approach is extended to construct one-sided simultaneous confidence intervals for pairwise comparisons in which order restrictions are imposed on the normal means when the variances are unequal.

The author retains copyright in this thesis. Any substantial copying or any other actions that exceed fair dealing or other exceptions in the Copyright Act require the permission of the author.

https://scholar.acadiau.ca/islandora/object/theses:1393