|
Abstract:
|
Adequately powering a study is a crucial part of experimental design, necessary - but not sufficient - to ensure scientific validity and reproducibility. It is well documented that under-powered studies are prone to both false negatives and false positives, impeding scientific progress, wasting resources, and raising ethical concerns for both human and animal research. Thus, obtaining guidance on best practices regarding sample size and power calculations is of central concern to mental health researchers and to the NIMH. This talk covers a brief outline of recommended best practices from the Executive Committee of the Mental Health Statistics (MHS) Section of the American Statistical Association (ASA) to the NIMH for power calculations. It is intended to provoke discussion about current practice and how we, as statisticians, can help improve which studies are funded by the NIH and other funding agencies, as well as which studies are published in high-impact journals.
|
