The accuracy of AIC and BIC is evaluated under simulated multiple regression conditions, varying number of total and valid predictors, R2, and n. AIC and BIC were increasingly accurate as n increased and as total predictors decreased. Interactions of the ratio of valid/total predictors affected accuracy.
Gagné, Phill and Dayton, C. Mitchell
"Best Regression Model Using Information Criteria,"
Journal of Modern Applied Statistical Methods:
2, Article 57.
Available at: http://digitalcommons.wayne.edu/jmasm/vol1/iss2/57