Abstract:
|
I have spent over 35 years educating under-graduates and graduate students and have consulted for the pharmaceutical industry for over 26 years, most notably for pharmacokineticists and pharmacometricians. Non-statistician scientists normally take 1 to 2 basic statistics courses as part of their training without developing a deep understanding of the mathematical foundation of statistical modeling, inference and hypothesis testing. Pharmacokineticists and pharmacometricians are part of this group of non-statistician scientists who normally use quite complicated statistical methodology including linear and nonlinear mixed effects models, to fit their data. They are well-trained in how to fit these complicated models using the software specifically designed for their purpose and are skilled in interpreting the biological meaning of the modeling results. However, when unexpected complexities arise while fitting and understanding the statistical nuances of their models, diving into the statistical literature is beyond their training. At the same time, statistical consultants rarely have training in this particular area of statistics. We will discuss some of these problems in this talk.
|