ABSTRACT
In order to explain the Bayesian philosophy, it would be important to first recap the philosophy of the more common traditional frequentist approach (Bland 1998). Consider the case of a trial that was conducted with the frequentist approach. When analyzed, the relative risk of having a bad event is estimated between the two comparators, and a confidence interval for that estimate is generated. In the frequentist philosophy, the true relative risk is always an unknown quantity (a parameter), and all we can do is to come up with an estimate of the parameter (a statistic). Technically speaking, in the frequentist world, it is incorrect to state the parameter value, and instead, the correct erudition is to state ‘the estimate for the parameter value is . . . ’ In the case of a single trial, we did not include all patients who had the same exact indication and who were measured the same exact outcome, in the same exact setting. Because trials are only a sample of patients, we can never be 100% sure of what the true value is. Even for estimates that we think we are 100% certain, you can also come up with a set of possible reasons why the estimate is not the absolute truth.
