Check your understanding
Quiz
You have built the Bayesian workflow from one line of Bayes rule up to a full regression: the prior-times-likelihood update, conjugate posteriors and the priors that earn them, sampling a posterior with the Metropolis algorithm, reading the trust dashboard (R-hat, effective sample size, divergences), pooling small groups toward a learned prior, checking a fitted model by making it invent data, ranking rivals on days they never saw with LOO and WAIC, and finally a GLM with priors, a fit, checks, and honest intervals. This quiz checks what stuck. The last two steps are live R you can run.
Question 1 of 10
What the posterior is
Asha starts with a prior for her sign-up rate and collects a day of data. Which statement describes the posterior she should report?
Correct: posterior is proportional to prior times likelihood. The data does not replace the prior, it reweights it, and with enough data the likelihood dominates.
Only in special cases does it look like a simple average, and even then it is a precision-weighted one. In general the posterior is the normalized product of the two curves.