Abstract
When data are used which contain random errors, there is always some question as to whether a supposed anomaly may not be real but due only to errors in the data. This paper describes a method of testing this point, which is worked out in detail assuming the Gaussian error law for the case of independent values. Possible applications of the method to the problems of contouring data, and evaluating and ranking anomalies are suggested. Examples of the use of the method on gravity and soil gas analysis data are also given.

This publication has 0 references indexed in Scilit: