Bring Clipboard, Flashlight

THE Census Bureau deserves a commendation and a catcall as it gears up to count the homeless for the first time in 1990. The commendation recognizes the bureau's intention to improve its historical undercount of the poor and minorities. The catcall concerns the bureau's data-collection techniques, which are fraught with peril.

For example, the Census Bureau reasons:

``Sleeping persons will not be awakened to answer questions. Rather, enumerators will estimate as best they can the person's age, sex, and race. If a sleeping person is covered up so that characteristics cannot be determined, the person will be counted and characteristics will be assigned later by a computerized algorithm.''

Just who are these enumerators whose tricky and potentially treacherous job it will be to check sleeping or sleepy-eyed homeless people between the hours of 6 p.m. and 6:30 a.m. in March 1990? The Census Bureau will hire about 9,000 census-takers - some of them homeless. Others will be gang leaders or ex-gang leaders who will help spread the word that the census is important. Clergy also will be enlisted to promote the census in their sermons.

The census-takers, though, have a risky job ahead as they are issued a flashlight and basic safety tips, but no weapons. One wonders if the Census Bureau will provide health insurance to these part-time head-counters. Police cannot accompany the census-takers, who will be instructed to leave if trouble brews.

Trouble could come as census-takers move through alleys, check park benches, and survey other places where the homeless congregate. From 6 p.m. until midnight, people in shelters and hotels or motels that cost under $12 a night will be counted. Those on the street will be counted between 2 a.m. and 4 a.m. Those in boarded-up buildings will be counted between 4 a.m. and 6:30 a.m. Dumpsters will not be checked, thank goodness.

A knock on someone's door in broad daylight or a stranger roaming around a neighborhood often raises alarm today. When the same thing happens after dark, you have to wonder whether it makes any sense to be a census-taker under these circumstances.

Questionable, too, is the bureau's plan to have census-takers go into areas at night where gang warfare is commonplace and where a male wearing the wrong color clothing could be a target.

The Census Bureau has trouble contacting the 18 percent of Americans who don't mail back their census questionnaires or have no permanent address. In 1990, as much as 22 percent of the population may opt out because of growing concern over government intrusion.

Counting these ``small cities'' of people within larger cities is a problem, but to put census-takers in jeopardy of losing life or limb for a few dollars exacts too high a price for figures that can be cranked out by computer.