Search

This article presents information related to a theorem related to statistical interference developed by a minister, Thomas Bayes. Bayes theorem is an alternative to conventional statistical testing. Two types of distribution are involved in the theorem. The one containing background knowledge at disposal before examining a new data set is called a prior distribution, which is the hypothetical probabilities of particular states of the world as it is thus far revealed. The posterior distribution is a revised estimate of knowledge about the state of nature after new data have been collected and incorporated into prior knowledge. In the interactionist perspective, social interaction is seen as a process of continually interpreting meanings from others, and defining meanings to others. The Bayesian approach is likewise a continual process of refining distributions of expected outcomes in order to predict certain occurrences. The Bayesian approach seems especially suited for longitudinal, panel and experimental designs in which data are collected at two or more periods. When used in conjunction with causal regression models in such designs, emphasis could be placed on probabilities for change rather than amounts of change in succeeding variables.