Some aspects of the efficiency of statistical methods of particle size analysis

Abstract

In statistical methods of particle size analysis by microscopy the analyst seeks to isolate a parameter of the particle profile which can be measured with a minimum of decision-taking and then relates mathematically this prime parameter to other parameters which have physical significance for the system being analyzed. The reason why these methods are called statistical techniques is that usually they, rely on some statistical averaging of the prime parameter in order to obtain some quantity which can be related to the physical parameter of interest. In this study the efficiency of current statistical methods is explored. It is shown that considerable semantic confusion concerning “random chord” populations exists in the literature and standard definitions are proposed. The use of Martin's and Feret's diameter in size analysis is explored. Techniques for improving the efficiency of statistical methods using antithetic variable principles and averaging techniques in which errors cancel instead of accumulating are suggested.