![[logo] US EPA](https://webarchive.library.unt.edu/eot2008/20090511160326im_/http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Mean Similarity Analysis
MEANSIM6 contains software for Mean Similarity Analysis, a method of assessing the strength of a classification of many objects (sites) into a relatively small number of groups. Classification strength is measured by the extent to which sites within the same groups are more similar to each other, on average, than they are to sites in different groups.
Users supply a matrix of pairwise similarities (or dissimilarities) for all possible pairs of sites to a program (RNDTST6) that calculates mean between and within-group similarities. The program also performs a permutation test of whether within-group similarities are larger than would be expected by chance.
MEANSIM6 also contains SAS code for computing common similarity measures from tables of multiple attributes (for example species abundances) at each site. In addition, a program is supplied for finalizing a mean similarity analysis from the ouput of the MRPP programs contained in BLOSSOM and PC-ORD.
The programs run on any DOS or WINDOWS-based machine.
MEANSIM6 closely follows the language and notation of the article: Van Sickle, J., 1997, Using Mean Similarity Dendrograms to Evaluate Classifications, Journal of Agricultural, Biological and Environmental Statistics 2, 370-388.
Example applications, along with a more precise definition of "classification strength", can be found in Van Sickle, J. and R.M. Hughes, 2000. Classification strengths of ecoregions, catchments, and geographic clusters for aquatic vertebrates in Oregon. Journal of the North American Benthological Society 19:370-384.
Additional examples are given in other articles from this same journal issue.
Both articles may be downloaded below.
Download:
-- MEANSIM6.ZIP. UnZips into several files. Documentation is on MEANSIM6.TXT.
Read:
-- Van Sickle, J., 1997, Using Mean Similarity Dendrograms to Evaluate Classifications (PDF)
-- Documentation for MEANSIM6.
To obtain further information contact John Van Sickle (vansickle.john@epa.gov)