Experiments have been considered the gold standard for informing evidence-based decisions. Most consideration on the design of experiments has focused on Type I errors (false positives) but we call attention to the importance of Type II errors (false negatives). Design complexity can lead to Type II errors due to measurement error and heterogeneous treatment effects. Studying clinical trials in drug development, we find evidence that design complexity may have contributed to some drugs failing to move past Phase 2 trials. Simplifying designs, particularly in Phase 2 trials, therefore might reduce development costs and result in more approved drugs. We also discuss the implications for improving efficiency of experimental research in other settings.
