1117 - Ricky Tang

SOME PRACTICAL ISSUES IN STUDY DESIGNS AND DATA ANALYSIS

 

Many practical projects collected data without a careful study design. This, together with possible inappropriate statistical approaches used in the data analysis, likely results in questionable study conclusions. The report ``Development of Site-Specific BAFs for Se in the Elk Valley" by Golder Associates (2009) is an example and partly motivated this thesis project. We analyze the data with alternative statistical models to address the concerns raised by C. Schwarz in his review on the fishery study. Further, aiming at providing a useful guideline on data collection, we conduct a simulation study with various settings to explore inefficiency in data analysis caused by imbalance of the data.

Keywords: Linear Mixed Effects Model; Simulation; Unbalanced Data; Univariate and Multivariate Response.