Locate the SISA web page. Find the link for "Correlations" on the left-hand side, under "Statistics," and click on it. This will take you to the correlations page.
Input the correlation from the first sample into the space labeled "r1," and the correlation from the second sample into the space labeled "r2." For example, if you were studying whether gender had an effect upon the degree to which weight affects the age of onset of diabetes, you could input the correlation between weight and age for a group of male subjects in "r1," and the correlation for a group of female subjects in "r2." Input the first sample size into the space labeled "r3-N1," and the second sample size into the space labeled "N2."
Click on the word "Calculate." The results will be displayed to the right as a series of confidence intervals for the differences between the correlations, with levels of statistical significance of .80, .90, .95 and .99.
Locate the SISA web page and click on the "Correlations" link.
Input three correlations from the sample into the spaces "r1," "r2" and "r3." These three correlations must be inter-linked, involving only three variables. For example, the first correlation may involve the variables age and weight, the second correlation may involve the variables weight and cholesterol and the third may involve cholesterol and age. This might be an appropriate calculation if you were looking at such questions as whether age or weight had a stronger correlation with cholesterol levels.
Input the sample size in the space labelled "N2," then click "Calculate." The results will be displayed as differences between the pairs of correlations, and the t-values of the pairs. The t-values would show the strengths of the different correlations.