Frequentist vs Bayesian
The main difference between frequentist approach and bayesian approach in Statistics is that in the frequentist approach, the value we want to estimate is assumed to be a constant across the population.
The bayesian approach, on the other hand, assumes that the estimate follows a probability distribution, called prior.
E.g. Let’s say we’re interested in a binary action, like conversion on an upsell screen. We want to make a change to the upsell screen and first want to test how effective it is. In this case, the unknown statistic of interest is the probability of conversion — the chance that a user landing on this screen would convert.
In frequentist approach, the statistics is assumed to be a single value, a ground truth. In bayesian approach, the probability of conversion is a distribution.