ABSTRACT
This chapter discusses practical application of the Bayesian and frequentist approaches as related but distinct perspectives on the application of probability models and the likelihood concept. The frequentist approach has its conceptual roots in the experimentalist view of scientific investigation with direct application to the repeated sampling related structures of population genetics, agriculture biology and time series applications. The modern Bayesian perspective has roots in the areas of insurance, loss and risk, and decision-making, areas that must incorporate risk with limited experimental or sampling data. The frequentist approach avoids formalizing prior knowledge regarding the unknown parameter or population characteristics of interest. In the Bayesian setting prior belief is updated using model and observed data through a formal prior density and the likelihood function. In both Bayesian approaches and frequentist approaches, the design of the study, experiment or data collection must be of high quality.
