Aron Roberts
Campus researchers and others using statistics programs should be aware that the integrity of their analyses could be vulnerable to Y2K problems. There are two primary areas of concern:
To resolve these problems, you may need to install your vendors' no-cost Y2K updates or, in some cases, upgrade to newer versions of these applications.
An example of such a century-ambiguous date is "02/18/00". Unless you specify otherwise, some statistics programs will assume that all such two-digit-year dates fall into the 1900s, so this date might be interpreted as February 18, 1900, rather than February 18, 2000, as you may have intended. (Other programs may make different default assumptions when handling such dates.) If your program's default assumptions aren't consistent with your actual data, some dates in your observations might be interpreted as falling into an incorrect century.
Some resources to help you identify and resolve these potential Y2K problems are listed in the document Campus Researchers: Specific Y2K Computer Concerns (http://y2k.berkeley.edu/computers/fixpcs/issues/researchers.html) on the UC Berkeley Year 2000 Information website. As an example, white papers from UCLA's Office of Academic Computing and from statistics program vendors such as the SAS Institute, Inc. and SPSS, Inc. suggest techniques that you can use to safely handle data containing two-digit-year dates.
On a related note, an overview which may help you identify and resolve other potential Y2K risks to your research data and analyses can be found in the article Is your research data at risk from Y2K? in the Summer 1999 issue of Berkeley Computing and Communications.
[ Next Article | Contents | Search BC&C | BC&C Main Menu | IST | UC Berkeley ]
Berkeley Computing & Communications,
Volume 9, Number 5 (November-December 1999)
Copyright 1999, The Regents of the University of California