Dr Grover Powers's reputation as a clinician, teacher, and sympathetic doctor is worldwide. I do not need to waste words in telling you how appreciative I am of the honor of being invited to stand here.
A more accurate title for this talk might have been simply "Statistics," but that would have led members of any potential audience to find more urgent duties elsewhere, especially those who, like me, know that I am not a statistician. I do, however, respect both faith and doubt, the latter sometimes being elevated to hypothesis, and concerning which we can gather evidence and which our statistician friends can quantify, if we abide by certain ground rules in the accumulation of information. Uncertainty can be quantified but never eliminated. It can be given a comforting value, such as P < .01. My talk is intended to show how necessary it is for us to give the statistician data of high quality, and why we often fail to do so.
The ancient method of making progress has been by observing events and comparing the results with previous experience–in medicine, with "my clinical experience." In Latin, the technique is described by the phrase "post hoc, ergo propter hoc." In the stone age, it led to the pressure flaking of razor-sharp knives from flint and obsidian. Later, our ancestors discovered agriculture, cheese-making, brewing, and the fermenting of grapes; the smelting of tin, copper, bronze, and iron; carbon steel; the recognition of all poisonous plants and animals; and tea, coffee, tobacco, hashish, opium, cocaine, quinine, oil of wintergreen, digitalis, mercury, cascara, curare, castor oil, and the Model T Ford.