The Statistical Abstract momentarily seems safe, but scores of other printed reports are simply being eliminated. In 1992 Census issued 1,085 reports: last year the number was 635, and the retreat from print has only begun. Gone are, among others: “Earnings by Occupation and Education”; “Poverty Areas in the United States,” and “Language Use in the United States.” This is absurd. We go to great trouble to collect this information, and now Census is suppressing it.
The losers are not just statistics addicts. Our public conversations depend heavily on these dry numbers. They shape our concept of who we are, of how society is performing and of what government should or shouldn’t do. Political speeches routinely spit out statistics, which can be made to tell stories: some true, some not so true. Keeping the conversations honest requires that the basic data be easily accessible to anyone who wants it.
When I say Census is “suppressing.” I don’t mean that it’s deliberately hiding its surveys. As a reporter. I’ve asked Census for information hundreds of times; I can’t recall an instance when answers, when available, weren’t provided quickly. The culture of the place is to release information. By its lights, Census isn’t abandoning print so much as it’s shifting its data to the Information Superhighway. Statistics are being distributed by CD-ROMs and the Internet. Already, Census brags that its Internet site is receiving 50,000 hits a day. Sounds amazing.
It isn’t. Those 50,000 daily hits are a lot less breathtaking than they seem, even if the figure is accurate (and I have my doubts). In May, Interactive Age, a trade publication, surveyed Internet sites. It reported that Pathfinder (the site for Time Warner publications, such as Time and People) had about 686,000 daily hits. Playboy had about 675,000 and Hotwired (the site for Wired magazine) had about 429,000. I mention these popular Web sites because they belong to magazines. As yet, none is forsaking the printed page for the glories of the Internet.
There are good reasons for this. One is that the number of daily hits on a Web site exaggerates how many people use it; the same person may hit the same site repeatedly. Another reason is that the Internet hasn’t yet evolved into an effective platform for advertising. But the main reason is that, for many purposes, the printed page is still superior to the computer screen. You can flip pages faster than you can search computer files. You can read a magazine standing in a subway or lying in a hammock.
Census’s shift from print clearly discriminates against people (including me) who don’t surf the Internet or use CD-ROMs. We remain the vast majority. American Demographics magazine recently reported a number of surveys that tried to measure U.S. Internet use in 1994. The surveys put usage of the World Wide Web between 2 million and 13.5 million people, which is at most about 5 percent. The average income of Internet households was $67,000, which is the richest fifth of Americans. But it’s not just computer clods or the unaffluent who will suffer.
Carl Haub is a demographer at the Population Reference Bureau in Washington. He’s a big user of Census statistics and is comfortable cruising in cyberspace. “It’s going to be a disaster for the average analyst,” he says. Downloading and printing data from the Internet can take hours. Getting a number from a CD-ROM is often a lot harder than getting it from a book. To Haub, Census is transferring a lot of the cost–in time and money–of making statistical information useful to people like him.
Martha Farnsworth Riche, director of the Census Bureau, admits as much. “If someone else can do it, let’s shift it to the outside,” she says. “We’ve had a hiring freeze since at least 1992, and those [printed] reports take an enormous amount of time from professionals.” They need to concentrate on doing surveys of “an economy and population that are changing dramatically. Our statistics have fallen behind.” Only Census can collect much of this data, she says. Let academics and analysts prepare reports.
Up to a point, Riche has my sympathies. The Constitution created the census (Article 1, Section 2), and social and economic surveys are a basic function of modern government. Some congressional proposals to cut the agency’s budget sharply are stupid beyond words. But that said, the new approach is misguided. The danger of overrelying on outsiders to organize and analyze basic data is that statistics may fall hostage to special pleaders or incompetents. Printed Census reports provide an easy way to check self-interested or faulty claims.
Print’s other great virtue is that it guarantees a historic record. Computer technology is changing so rapidly that data committed to one technology may no longer be easily accessible if that technology vanishes. “The CD-ROMs that we’re so excited about today–20 years from now, no one will use them,” says Richard Rockwell, director of the Inter-University Consortium for Political and Social Research. “The book is a highly advanced technology for preserving some kinds of information.” Exactly.
Let’s not become too infatuated too soon with the Information Superhighway. Census should be issuing its data in computer-friendly ways, but not as a substitute for printed reports. A jaunt on the Internet–piloted by my friend Steve–only affirmed my skepticism. Steve typed the Census Web address (http://www.census.gov) and up popped the “home page” designating me as the 567,352d visitor. Unless the count began 10 days earlier (and it didn’t), that was a lot fewer than 50,000 daily hits. I informed a Census official. He was mystified. After checking, he said there were other ways of accessing the Web site that didn’t raise the count. Hmm. Could be. But it also shows how, on the Information Superhighway, we’re still navigating in the dark.
