When the The Miami Herald won this year's Pulitzer Prize for investigative reporting that found voting fraud so massive that the city's mayoral election was overturned, 23 reporters, editors, and researchers shared in the applause and champagne.
The 24th contributor could not raise a glass or hear a toast, but the winning series, "Dirty Votes: The Race for Miami Mayor," might not have been possible without the SAS suite of so-called data-mining software, says the Herald's research/technology editor, Dan Keating.
"As journalists, we get information all the time in some humble-jumble format, and we have to recreate and analyze that information," Keating says. "Basically what SAS allows you to do through its programming language and its power and flexibility is to take statistics from many databases and many fields and put them together in a meaningful way. When you combine that with primary, shoe-leather reporting, you can do these investigations."
The Herald's Pulitzer is raising the profile of the journalistic potential of SAS software that until now has been far more at home in newspaper business offices than newsrooms. A small coterie of investigative journalists is urging more editorial staffs to harness the ability of SAS to bring meaning to vast amounts of data.
"I'm now sort of a big proselytizer for SAS," says Steve Doig, a former Miami Herald journalist who began using SAS 10 years ago when the software resided on the paper's big IBM mainframe. "In fact, I've got a thing I make clear upfront, that I'm not a consultant for SAS or getting anything from them."
SAS, says Doig, now the Knight Chair in Journalism at Arizona State University, has "the matching tools, the logic tools, statistical tools all the kinds of thing you need to get around big amounts of information."
Another fan is The Charlotte (N.C.) Observer's database editor, Ted Mellnik. "I use it every day for just about everything," Mellnik says. "When you've got data and you want to explore, analyze and report on it, SAS with its power and flexibility is the best statistical tool that I have ever seen."
The integrated software suite marketed by Cary, N.C.-based SAS Institute Inc. is sometimes generically known as data-mining software. The idea is to go beyond the row-and-column displays and reports offered by such spreadsheet programs as Microsoft Excel, and permit users to more easily manipulate and make sense of the seas of information stored in big databases. Another data-mining software maker, Chicago-based SPSS Inc., has made somewhat smaller inroads in the newspaper industry with its less expensive desktop packages.
SAS runs on all popular platforms from mainframes to personal computers, including operating systems such as Macintosh, OS/2, Windows 95, and Windows NT. More than two-thirds of its installed base of 33,000 customer sites in 115 countries serve business users, with universities accounting for another 18%.
The user base with the greatest potential to drive the growth of SAS in newsrooms, however, is the 16% of user sites at government agencies. "Certainly, the government has really bought into SAS," says Neill Borowski, the director of computer-assisted reporting/analysis for The Philadelphia Inquirer. Though Borowski himself uses SPSS software more heavily, he recently convinced his paper to increase its SAS licenses in preparation for analyzing data to come from the 2000 U.S. census.
Similarly, The Charlotte Observer just ordered a site license and expects to have two or three more journalists using the software by the time of the census.
"SAS really is pretty much the native language of the census program," says the Miami Herald's Keating. "With just a few lines of code, you can get information you couldn't get with any other software."
So far, the biggest selling points for SAS in the newspaper industry are the ways individual papers have used it in dramatic investigative projects. …