WASHINGTON — Washington. -- Is there a growing epidemic of crime? Of starving American children? Of dead babies? Of ill-educated teen-agers? Or is it a growing epidemic of media mindlessness and maliciousness?
Data-doctoring has long been an American public-policy disaster. It comes with the noble territory of a free press and free politics. But is statistics-spinning getting worse? Can something done?
In recent weeks, we heard:
* That the violent crime rate is going up.
* That ''a startling number of American children [are] in danger of starving'' (said Dan Rather, in a CBS lead story).
* That schools are a disaster.
* That the infant mortality rate -- remarkably! -- has improved.
These distorted stories will have a public-policy impact, as politicians echo the hype to get funds, make laws, and sometimes -- yes! -- get personal publicity.
Mickey Kaus of The New Republic called the hungry-children story ''crap'' that ''oozes phoniness.'' The ''data'' came from the Food Research and Action Center; the project was funded by Kraft General Foods, which Mr. Kaus notes is ''a major corporate beneficiary of federal food subsidies.'' The report says that if, even once in a year, you rely ''on a limited number of foods,'' because you are short of money, then your children are ''at risk'' of ''hunger.'' Pass the pasta.
Infant mortality is down. Surprise! Much touted as a disaster, it has declined steadily, by half, for blacks and whites, since 1970.
If I hear that the drop-out rate is ''soaring'' one more time, I will vaporize. It's at an all-time low.
The Washington Post headline about the 1990 FBI crime data read ''Violent Crimes Up 10 percent Last Year.'' On ABC's ''Good Morning America,'' one alleged expert said it means ''the battle is being lost,'' and the second agreed.
Missing was the fact that the Bureau of Justice Statistics had recently issued 1990 data that showed violent crime was not up, but down! The decrease was tiny (0.4 percent) but the rate held constant for five years, at levels 14 percent lower than a decade earlier.
FBI data deal only with ''crimes reported to the police.'' Bureau of Justice Statistics conducts a ''victimization'' survey that asks about experiences (sample size: 97,000). Which is better? BJS is, says Christopher Jencks, writing in The American Prospect. (Police are likelier to report crimes to the FBI these days.) I think he is right. In any event, it is irresponsible to report one set of data without the other.
Everyone now has an interest in pumping up the data. Conservatives want to show moral decay and get more money for cops. The liberal Sen. Joseph Biden used the FBI report to demand ''tough measures to ban killer assault guns and combat the epidemic of violence against women.'' (BJS data show rape down by 18 percent.)
Journalists, who should sort out facts from hype, say, ''don't shoot the messenger that brings the bad news.'' Hah! Why do messengers who have 1,000 messages decide which 11 to deliver, all with bad news? Might it be because ratings and
circulation are sinking and hype (allegedly) boosts both?
Perhaps liberals are beginning to act. The New Republic is liberal-eclectic and the Kaus piece contained a dozen other examples of data-distortion, including the liberal myth that real income has been going down. The liberal American Prospect bills Mr. Jencks' piece as the first in a ''series of articles scrutinizing widely believed statistical facts -- and artifacts.'' Godspeed.
But what about workaday journalists? If there's a news desk, why not a data desk? Why don't journalists criticize themselves? If network X ran a story that network Y made a fool of itself, it could shape up both of them. Are they chicken?
Ben Wattenberg, of the American Enterprise Institute, is author of ''The First Universal Nation.''