Schwarz's Bayesian information criterion (BIC) is a description of a model fit statistic that is most appropriate when which conditions apply?

Schwarz's Bayesian information criterion (BIC) is a description of a model fit statistic that is most appropriate when which conditions apply?

• A. Small sample sizes and many parameters.
• B. Large sample sizes and many parameters.
• C. Small sample sizes and few parameters.
• D. Large sample sizes and few parameters.

Answer: (B)

BIC balances how well a model fits the data with how complex it is, using a penalty that grows with both the number of parameters and the sample size. The penalty term is proportional to the number of parameters times log(n). As the dataset gets larger, this penalty becomes stronger, so extra parameters must substantially improve the fit to be worth including. This makes BIC especially reliable when you have a large amount of data and a relatively complex model to choose from—it helps prevent overfitting and, under suitable conditions, tends to pick the true model as the sample size grows. In small samples, that growing penalty can be overly harsh, which is why BIC is best described as most appropriate for large sample sizes with many parameters.

BIC balances how well a model fits the data with how complex it is, using a penalty that grows with both the number of parameters and the sample size. The penalty term is proportional to the number of parameters times log(n). As the dataset gets larger, this penalty becomes stronger, so extra parameters must substantially improve the fit to be worth including. This makes BIC especially reliable when you have a large amount of data and a relatively complex model to choose from—it helps prevent overfitting and, under suitable conditions, tends to pick the true model as the sample size grows. In small samples, that growing penalty can be overly harsh, which is why BIC is best described as most appropriate for large sample sizes with many parameters.