A government agency funds research on cancer. The agency funds 25 research projects, all of which are testing various drugs to see if they are effective in reducing brain tumors. One of the projects (project number 12) finds that the drug they are studying significantly reduces the size of tumors with a P-value of 0.028. The other 24 projects found no significant effects (P-values all greater than 0.05) of the drugs that they studied. Is it proper to conclude that the drug in project number 12 is effective in reducing the size of brain tumors?

Question

Quizplus · Accepted Answer

The correct interpretation of the P-value in the context of multiple comparisons (25 projects in this case) requires adjusting for the increased risk of a Type I error (false positive). Simply finding a P-value below 0.05 in one out of 25 studies does not robustly confirm effectiveness without considering the multiple testing problem and the potential need for correction methods (e.g., Bonferroni correction).

A Government Agency Funds Research on Cancer