HOW MANY HOMELESS ARE THERE?

THE EXTENT OF HOMELESSNESS IN AMERICA HAS BEEN A continuing source of dispute between homeless advocates and government officials, with each side accusing the other of using figures to its own advantage.

Until May of 1984, homeless activists had estimated the number of homeless to be at least 2 million. Then the Department of Housing and Urban Development released a study that put the number at closer to 350,000. Outraged leaders of the homeless movement said the federal government was trying to minimize the number to justify not spending money on the problem. The report caused such an uproar that a House committee held hearings, and witnesses testified that HUD's numbers were distorted because census takers had applied results obtained from smaller geographic areas, like cities, to larger areas, like entire metropolitan areas.

Mitch Snyder's Community for Creative Non-Violence responded with a study of its own. Sampling a limited number of shelter providers in large cities, it extrapolated the estimates obtained there (about 1 percent of local populations) to the entire U.S. population. The result was a tally of 3.5 million.

Last summer a Harvard economist, Richard Freeman, came up with new figures based on interviews with 700 homeless people in New York City. He found that for every person sleeping in a shelter on an average night, another 2.3 were living on the streets and in parks. Multiplying that figure by the 76,500 known (and presumably occupied) shelter beds around the country, he concluded that the total number of homeless Americans was 350,000.

The discrepancies reflect the fact that counting the homeless is a demographer's nightmare. Many homeless men and women, including illegal aliens and the mentally ill, avoid shelters, which is where the homeless are counted. Another problem is that those who do use shelters do so inconsistently. They "migrate," according to Leona Bachrach, research director of psychiatry at the Maryland Psychiatric Research Center, in ways that make definitive counts difficult.

First, homeless people migrate daily and seasonally within localities, going from uptown to downtown soup kitchens, from shelter to park bench to alley. They also travel regional and cross-country routes, a phenomenon noted especially by shelter workers in Virginia (along the north-south route), and in Oklahoma City and Phoenix (along the east-west axis). And finally, many homeless people move in and out of homelessness. A relative may take a homeless family member in for a while. A homeless woman and her children, forced onto the street because of domestic violence, may return home only to be driven out again in several months. Little systematic effort has been made to distinguish between the chronically homeless and the episodically homeless, according to Bachrach.

The result, she says, is that substantially different results will be obtained depending on how counts are taken, whether they are the record of a particular day (the method currently used to count the homeless), or of a more representative time period, like a whole year.

In a report last spring on the growth in homelessness, New York City's Partnership for the Homeless acknowledged that "it has become axiomatic that the nature of homelessness prevents any scientifically reliable estimate of the total number of homeless in the nation." At best, the report concluded, what is possible is to track the growth in levels of demand for food and shelter.

Articles appear as they were originally printed in The Washington Post and may not include subsequent corrections.