There are many difference between Bayesian and Frequentist inference, for example:

- From Bayesian viewpoint, the parameters are treated as variables.

- It is possible to incorporate prior information in the analysis, which is updated by the information obtained in the experiment.

- It is possible to value the credibility of the null hypothesis, this one is a great advantage, does not force us to take a dichotomic decision, as in Frequentist approach.

- The Bayesian Inference calculate the probability P(Ho given data), which is what really is interested, and not the probability P(Data given Ho) as in Frequentist approach (p value).

These are only some adventages. It is true also that the Bayesian analysis has certain disadvantages, such as the difficulty of the calculation, but the development of new software like the WinBUGS facilitate the calculation of the posterior distribution.

I think that the Bayesian approach is an excellent alternative and I'm doing my thesis in Bayesian Analysis of RxS Contingency Tables. From my viewpoint both approaches (Bayesian and Frequentist) are complement each other in many situations. I like to use both.

Well...Sorry for my precarious english. Greetings from Peru.