This program calculates one or more matrices of pearson correlations or double-scaled Euclidean distance coefficients from a Statistica raw data file. However, what makes it unique is that it optionally computes these coefficients in a Jaccard-coefficient style; that is, it excludes pairs of observations from an agreement calculation where there are values of 0 for each variable (in addition to exclusions based upon one or the other, or both variables possessing missing values. The program is primarily designed for data which consists of counts, frequencies, or percentages, and where 0 indicates no occurrence or presence of something. Thus, a joint zero (a pair of observations equal to 0.0 for each variable) within these data is indicative of a joint “absence”, as might be the case when scanning transects within Marine Biology for the occurrence of fish species (Wellenreuther, M., Barrett, P.T. and Clements, K. (in Press) Ecological Diversification in Habitat Use in Subtidal Triplefin Fishes (Tripterygiidae). Marine Ecology Progress Series), and finding no fish present in say two kinds of habitat. Because this kind of analysis invariably takes place using categorized variables, the program allows for up to two categorical variables to be used such that the automatic production of agreement matrices can be constructed for each combination of categories for one or two variables. This is pretty handy – as STATISTICA seems unable to do this when using two “By Group” variables. The matrices produced and saved automatically by the program can be input directly via Statistica into say multidimensional scaling or a principal components analysis. A 15-page Microsoft Word screenshot step-by-step helpfile is included with the program, along with a Statistica v. 7.1 spreadsheet example/test file, and all three are available in the zip archive here.

A 7-minute Flash audio-visual walk-through file is available