Bayesian methods for statistical analyses require a different interpretation of probability than traditional “frequentist” methods. The use of Bayesian methods is increasingly common and its flexibility has facilitated a wide range of scientific advances, especially in medicine.
The first example is of an adaptive design examining the dose effect of inoculating healthy volunteers with Haemophilus influenza and estimating the probability that nasal colonization will occur. Different sequential strategies for design will be examined and compared. The second example will be of subgroup analyses in clinical trials where quantification of the evidence is more useful than traditional significance testing. These two examples are, respectively, joint work with Yu-Hui Huang Chang and Emine Bayman.
More information on Kathryn Chaloner may be found at: http://cph.uiowa.edu/faculty-staff/faculty/directory/faculty-detail.asp?emailAddressfirstname.lastname@example.org