Failure to account for prior information
NHST treats each scientific result as if it were the first ever to evaluate a hypothesis. This requires one to ignore all prior (quantitative) information. It provides no quantitative solution by which to incorporate any type of earlier evidence, whereas Bayesian methods can easily do so. Use of this prior information can help dispel the false appearance of unexpectedly heterogeneous research results.
Pingback: Confidence Intervals: The basics | communitymedicine4all