Index Last Week Contact DQ News Latest Issue
DARPA Research Contract Awarded to SIMS
According to a press release, the University of California, Berkeley School of Information Management and Systems has been awarded a $954,180 contract by the Defense Advanced Research Project Agency to make searching for documents and data easier and more cost-effective. The project is called "Search Support for Unfamiliar Metadata Vocabularies."
The project is designed to provide a cost-effective remedy for Internet searches when the searcher is unfamiliar with the classification, categorizing, and indexing schemes ("metadata vocabularies") being searched. The rapid rise in the number in the number of databases that can be accessed over the Internet means that searches will increasingly require the use of indexes and classification schemes that are unfamiliar to the person searching. To provide a cost-effective remedy, researchers will develop entry vocabulary modules that accept topical statements in the searcher's terms and respond with a ranked list of terms in the system's vocabulary. See the project website (http://www.sims.berkeley.edu/research/metadata) for more information.
NIST Releases Statistical Reference Datasets Website
The National Institiute of Standards and Technology (NIST) has released a new website that contains reference datasets for objectively evaluating the computational accuracy of statistical software. NIST states that both statistical software users and developers can use these datasets - the Statistical Reference Datasets (StRD) - to ensure and improve software accuracy.
The StRD web pages contain datasets with NIST-certified computational results in the areas of analysis of variance, linear regression, non-linear regression, and univariate summary statistics. According to NIST, the collection includes both generated and "real world" datasets, of varying levels of difficulty. Generated datasets are designed to challenge specific computations. Empirical datasets include "benign" and "challenging" datasets from the NIST laboratories and other sources. NIST certification methods are also described in the website. The website is located at: (www.nist.gov./itl/div898/strd/).
Big Social Changes Revive the
False God of Numbers
An article in the August 17th issue of The New York Times explores the recurrent attempts of politicians and others to draw categorical conclusions from scanty data. According to the article, there is intense pressure on politicians to make sweeping statements about the "success" or "failure" of programs like welfare reform, although there is little or no information available to show whether such programs are working or not.
Moreover, whatever data is available can often be interpreted in two or more ways. For example, soaring divorce rates are cited to "prove" a breakdown in the family and worsening conditions for children. But women and their children may not be better off enduring abusive relationships. Similarly, the growing numbers of women in the American workforce may mark the American family's efforts to stay afloat financially, or indicate the professional liberation of women.
Another problem is that amassing enough high quality data to meaningfully evalute social trends may take years - even decades. As the recent United Parcel strike dramatized, more employees are working part-time or on a contract basis than ever before. Are they doing it because they cannot find full-time employment? Or are they trading full-time work and benefits for more freedom and family time? Do both situations occur simultaneously? The article concludes that there will always be some things that cannot be measured in a rapidly changing marketplace. Written by Times staff writer John Broder, the article appears in the "Week in Review" Section.