Abstract
The application of the method of maximum likelihood (ML) to the determination of the slope of the number-flux-density distribution of radio sources is discussed. It is shown that the ML estimate of the slope, which uses the maximum available information, may be obtained by a simple calculation from the individual flux densities. The efficiency and bias of other methods are compared with the ML method by means of numerical calculations on simulated data. The error of using integral counts is stressed and an example is given from the literature to demonstrate that incorrect conclusions may be drawn through the use of statistically invalid procedures. The need to test the adequacy of the power-law model is stressed.
Author
Crawford, David F.; Jauncey, David L.; Murdoch, Hugh S.
Journal
Astrophysical Journal
Paper Publication Date
November 1970
Paper Type
Astrostatistics