Bias in a least square method of analysing decay data

Abstract

The very common method of analysing decay curves by fitting a stright line to the logarithm of the decay curve using the reciprocals of the number of counts per channel as weights, has been found to introduce an error of the order of 0.5% in the lifetime. The dependence of this error on the number of counts accumulated, the number of half-lives analysed and the background can be described by an empirical formula. For accurate results it is better to use an iterative procedure to fit an exponential to the counts observed, using the numbers produced by the fit as weights instead of the actual counts. © 1970.