An article yesterday on a national household survey by the National Institute on Drug Abuse quoted the agency's chief statistician, Joseph Gfroerer, as saying figures showing declines in weekly cocaine use were not statistically significant. One reference to Gfroerer's comments should have made clear that he was referring only to the reported decline in weekly cocaine users, not the survey's reported decline in overall cocaine use. (Published 12/21/90)

The largest-ever federal survey of drug use has found that 72 percent fewer Americans regularly use cocaine than used it five years ago, a finding hailed yesterday as "wonderful" news by President Bush.

But the National Institute on Drug Abuse (NIDA) survey of 9,259 people touched off a political debate with congressional Democrats contending the Bush administration was using "misleading" numbers to give the impression it is winning the drug war. Critics said the survey woefully undercounted hard-core crack abusers in inner cities, and NIDA's chief statistician conceded that the reported declines among cocaine users -- described by administration officials as the most astonishing finding -- were not statistically significant.

The survey found sharp declines in illicit and legal drug use among virtually every age group and category. It estimated that 27 million Americans tried some illegal drug in the past year, a 25 percent decline from a 1985 survey. In the past two years alone, it found that the number of monthly cocaine users dropped 45 percent, to 1.6 million, and the number of monthly marijuana users declined 12 percent, to 10.2 million.

The survey, conducted earlier this year, included a special sampling in the Washington metropolitan area that found slightly higher rates of PCP and crack use here than in other metropolitan areas.

But the survey's most unexpected and controversial finding was that the number of hard-core crack and cocaine abusers, defined as persons using the drugs once a week or more, had also declined, from an estimated 862,000 in 1988 to 662,000 this year. While federal anti-drug officials have been reporting declines in "casual" cocaine use among the middle class for some time, the new survey results, combined with a 33 percent drop in cocaine-related hospital emergencies over the past year, were described by the officials as the strongest evidence that the inner-city crack epidemic of the mid-1980s has abated.

Secretary of Health and Human Services Louis W. Sullivan called the findings a "startling turn" and said they "document millions of individual decisions to reject the peer pressure and false claims made by the drug culture." Bush, who appeared with Sullivan and acting national drug policy director John Walters to announce the results, said it "suggests that our hard work is paying off."

But some independent critics yesterday called the figure of 662,000 hard-core cocaine users virtually meaningless because the homeless, those in jail and others living in institutions are not counted in the survey, and 18 percent of all those approached refused to participate. Officials acknowledged that that group is most likely to contain illegal drug users.

The estimates of weekly cocaine and crack users "don't make sense," said Eric Wish, director of the Center for Substance Abuse Research at the University of Maryland. "The survey depends on a person willingly reporting his or her drug use to the government. . . . We all know the stigma that is currently attached to drug use. . . . We don't know how much of this decline is merely the result of people unwilling" to say they used drugs.

In addition, Joseph Gfroerer, chief statistician for the survey, said the estimate of weekly cocaine users was extrapolated from only 63 responses out of the 9,259 people questioned, a response rate he said was so low it was not statistically "significant" enough to reflect any trend. "It's a good chance it's just random error in the sample," he said.

Calling the numbers "wildly off the mark," Senate Judiciary Committee Chairman Joseph R. Biden Jr. (D-Del.) released his own report estimating the number of hard-core cocaine and crack abusers at 2.4 million, more than three times the household survey estimate and a slight increase from the 2.2 million Biden's panel estimated last May.

"These figures show that America's hard-core drug problem is getting worse, not better," Biden said. The staff report added: "The administration's undercount . . . misleads the American public . . . {and} lulls the country into a false sense of security about how far we have come" in stamping out drug abuse.

Biden's estimate was extrapolated from urinalysis drug-testing figures, which continue to show between 50 and 70 percent of those arrested in major cities testing positive for cocaine, as well as admissions to drug treatment centers.

Walters said that critics of the absolute numbers miss the larger point that the trends are consistently improving and public attitudes changing far more rapidly than anybody anticipated. In the 1985 survey, the number of monthly cocaine users was estimated at 5.8 million. By 1988, that had dropped to 2.9 million and to 1.6 million in this year's survey.

"What has happened is a faster and more extensive realization of the dangers and undesirability of drug use," Walters said. "There's a hidden story here: the mobilization of the American people against drugs. . . . There's no other way to explain it."

But the survey also showed the declines to be part of a larger societal turn against substance abuse of any kind. The number of cigarette smokers has dropped from 60.3 million in 1985 to 53.6 million this year, the survey found, while the number of alcohol users declined from 113.1 million to 103 million during the same period.

The survey was conducted for NIDA by the Research Triangle Institute and included personal visits in which each respondent was asked to take about an hour to answer a detailed questionnaire on personal drug habits. The sample size was larger than any previous household survey and contained an additional sampling of minorities and young people in an effort to get more reliable estimates of drug use in those groups.

In the special survey of 1,931 people in the Washington area, drug use rates were not notably different from the rest of the nation except that a slightly higher rate -- 4.7 percent compared with 3.4 percent nationally -- have used PCP at least once, according to Edgar Adams, director of NIDA's division of epidemiology.

The rate of individuals who acknowledged using crack cocaine once in their lives in the Washington area was 2.9 percent, compared with 1.9 percent in other metropolitan areas, but Adams said this difference was not statistically significant.

Cynthia Harris, special assistant to Mayor Marion Barry for drug policy, said yesterday she agreed with the findings and noted that a range of indicators, including drug arrests, hospital emergencies and urinalysis drug-testing of arrestees, show declines in drug use over the past year.

Harris also said there had been a sharp drop in the number of open air drug markets in the city, from about 90 two years ago to less than half that today. Noting the city's record homicide rate this year, she said: "I think we may have erred early on by thinking that if we solve the drug crisis we would solve the homicides. . . . When we're talking about homicides, we're talking about violence, conflict resolution, firearm use and other social issues that far, far exceed drug use and abuse."