Abstract
This paper discusses the testing of SN ratios for homogeneity based on experimental data to compare the performance of signal-response systems. When the (passively acquired) data in these systems is independent or dependent, this problem is known by testing the homogeneity of the SN ratios in the error distributions as normal distributions. In this paper we use the SN ratio (proposed as a dimensionless evaluation measure for positive-valued data) to derive the distribution of test statistics based on a certain hypothesis.
