StatLib is sponsored by Carnegie-Mellon University in Pittsburgh, PA. The StatLib website is designed for distributing statistical software, datasets, and information by electronic mail, FTP and WWW. According to the information at the beginning of the site, StatLib started out as an e-mail service. They go on to state that some of the organization still reflects that heritage, although I am not in a position to offer such opinion. It contains a wealth of archived information such as data sets/examples, statistical software packages, meeting information, and an assortment of specialized analysis packets, codes, and algorithms.

The general nature of the site indicates to me that it once served as a clearinghouse of information. By initial design, the site brought together an array of information focused toward statistic information. Most of the information currently of value to me would be the teaching of statistical applications. The data sets and examples provide a large assortment of cases for demonstration and application. For example, the Data and Story Library provides a variety of examples in a wide selection of categories. The link to Disease Data provides a small set of files for a few specific aliments. I am unable, at this time, to estimate the value of the various statistical packages available. I did have time to thoroughly learn the process of download and using any of the share-ware provided. Similarly, I don’t believe I currently have the expertise to send StatLib my contributions.

My more critical remarks are aimed toward how relevant and current many of the sites were. Some information was as old as 1993 while some was as new and up to date as last month. I would suggest the StatLib site be purged of all out of date materials, notices of meetings, and resource links. I would also indicate that, when compared to other sites, the StatLib site is not nearly as user-friendly, graphically exciting, attracting or interest holding. Approximately 49 different sites can be accessed through the initial home page for this site. The collection seems too eclectic to me.

In the future, I believe I would bookmark a few of the selected sites linked to StatLib rather than to the entire www.stat.cmu.edu site. I don’t mind having many usable sites marked compared to a single site with limited application.

Gary S. Straquadine