Judging from the media coverage of a recent taxpayer-funded survey of "customer satisfaction" with the federal government, Washington bureaucrats now give your local Sears repairman a run for his money when it comes to prompt, friendly service. "Customers Satisfied With Government," an AP headline read. "Uncle Sam Gets Thumbs Up," The Washington Post declared. In reality, the survey is little more than a $1.8 million public-relations stunt.

Conducted by the University of Michigan's Business School, the first-ever government-wide survey used a computer model known as the American Customer Satisfaction Index (ACSI) to rate 30 federal agencies. On a scale of 1 to 100, the government received a 68.6, compared to a private-sector rating of 72. This prompted Vice President Al Gore, the self-proclaimed "re-inventor" of government, to announce that "several agencies are delivering services equal to or better than private sector organizations."

The vice president wasn't alone in touting the survey. A spokesman for Michigan's Business School said it "allows for the first time … a look at what is the satisfaction with the government versus satisfaction with the private sector." Philip Diehl, director of the U.S. Mint and chairman of the Committee on Excellence for the Customer Satisfaction Initiative, said the survey is a "scientific, independent source of information for government leaders and change advocates" to measure government performance. The scores reportedly will be used to justify future budget requests.

Left unexplained is the fact that each federal agency was able to choose the sampled beneficiaries of its programs and which aspects of those programs would be examined. In fact, 18 agencies provided the University of Michigan with lists that included the names of 3,000 individuals they wanted to have questioned -- hardly the random samples typically encountered in such surveys. This is like having parents poll their children after all the Christmas presents have been opened.

For example, the Environmental Protection Agency's survey sample was reference librarians who happen to use the EPA web site. Its rating: 69. The Department of Education decided to query people who ordered its publications in the last year, and wound up with an 80. And no, NASA didn't survey members of the public who watched millions of their hard-earned dollars disappear along with the Mars Polar Lander. Instead, the agency went with a group of educators who attended a week-long NASA program. Its rating: 80.

Not surprisingly, agencies that provide money or services polled better than ones that regulate or enforce laws. Of the five agencies surveyed that provide direct services or funding to the public, only one agency (the Veterans Administration) included individuals whose application for services was denied. Its rating: 61. As the survey wryly noted, "inclusion of those denied benefits in the segment probably lowers the score." The question is, why weren't other agencies bold enough to survey their "dissatisfied customers"?

The dual results for the Internal Revenue Service suggest an answer. Unlike the other 29 agencies, the IRS had two groups polled: taxpayers who filed their returns electronically, and all tax filers. The first group, which received its tax refunds in record time, gave the agency a 74. The second group, which Commissioner Charles O. Rossotti admits has to "struggle with getting tax returns prepared correctly," assigned the IRS a 51.

As the IRS results show, the problem goes beyond methodology. The survey didn't ask the right questions. No doubt the American public is thrilled to learn that reference librarians are happy with the EPA, but what does this say about "customer satisfaction" with the onerous regulations the agency enforces annually at great cost to taxpayers? Yes, the folks who receive federal education mailings may consider themselves happy "customers," but this hardly translates into public satisfaction with the federal role in education.

A more useful survey -- requiring a wide, random sample from among the general public -- might have asked participants if they were even aware of the many programs overseen by each agency. It could have asked if these agencies are spending public money wisely, or if the services provided could be better handled by state or local government (or the private sector). Instead, Americans get a meaningless poll that elevates perception over performance and amounts to a "drive-by" wasting of taxpayer money.

"So many government agencies have no idea how their performance ranks with any outside measurement," Diehl told The Wall Street Journal. Unfortunately, they still don't. Most of their true "customers" -- the ones who foot the bill -- have yet to be asked.

Virginia L. Thomas is a senior fellow at The Heritage Foundation (www.heritage.org), a Washington-based public policy research institute.