Conclusions: Good Servants but Bad Masters

Abstract

Researchers who ignore sampling error’s impact while drawing conclusions from sample based research are like pilots who ignore gravity’s impact while flying an aircraft. P values help us determine the influence of sampling error on the conclusion we draw from research programs, by encapsulating the sampling errors contained in the research’s data. However, that is all that they do. P values have risen to an unjustified role of preeminence in medical research. Bradford Hill (speaking of the chi square test) described them as good servants, but bad masters. That is a lesson in health care research that we must still learn. By ceding our best judgment to their rule, must is lost.