Abstract
“All Empirical data are probable only.”
A group of observations taken to establish the value of an empirical constant must be considered as a sample of an infinitely large population, or universe, of possible readings. By the application of appropriate statistical methods to the sample one may arrive at certain conclusions regarding the entire universe. The purpose of this paper is to discuss some basic concepts of statistical thinking, and to offer a graphical solution to the problem of estimating universe values from the sample, values.
The data are prepared and plotted on special graph paper against a probability scale. From the plot are taken the most probable value of the constant sought; and the standard deviation. This latter is the slope of the plotted line, and is the key to further prediction from the data such as the confidence limits of the estimated average; a numerical expression of the precision of the test method used; and a means of comparing this method with another as to accuracy and precision. It also furnishes the answer to the question: "In the future how many readings should I take in order to have a high assurance (say of the order of 90 to 95 percent) that my estimated average will be within plus or minus 1 percent its true value."
