The Bayesian information criterion (BIC) is one of the most popular criteria for model selection in finite mixture models. However, it implausibly penalizes the complexity of each component using the whole sample size and completely ignores the clustered structure inherent in the data, resulting...