data informs beliefs, beliefs do not inform data
I picked up a Bayesian Statistics textbook. It's common in everyday life to collect data to support our existing beliefs. Rather, the Bayesian approach starts with a prior belief, then assessing the data, and then most importantly, updating those beliefs based on how well your hypothesis explains the data. Evaluating competing hypotheses can help you update beliefs based on those hypotheses with greater probabilities. A bit untechnical for a daily-learning but hey ho, it be that way sometimes.