ABSTRACT
The two-sample spreadsheet also includes a chi-square test, which compares the shapes of the distributions for the two samples. Although the samples may individually exhibit dependence, the assumption is made that the two samples are independent of each other. There are no assumptions regarding the form of the underlying distributions. To examine shape only, the program adjusts the two data sets by subtracting the respective mean value from each observation such that both series then have a common mean value of zero. Then the overall range of data in the two series is divided into the lesser of 9 or n/4 classes and a frequency distribution is produced for each sample. The frequency distributions are compared with a chi-square test.
