ON THE RAINY evening of Jan. 25, 250 municipal workers and volunteers were deployed to our city's streets and alleys to achieve a tally of San Francisco's homeless residents. On Feb. 14, after an interval longer than was required to tabulate more than eight million Iraqi votes, Mayor Gavin Newsom revealed to a singularly uncritical press that homelessness citywide decreased "precipitously" since the last such count, in the fall of 2002. According to the new survey, overall homelessness in San Francisco declined from 8,640 in 2002 to 5,642 this year, a 35 percent reduction, while the number of people counted on the streets plummeted 41 percent, from 4,535, to 2,655.

This seemingly unprecedented decline was proudly attributed to the success of Newsom's Care Not Cash program in drastically diminishing the number of homeless single adults seeking welfare assistance, as well as the mayoral administration's sweeping shift to a "housing first" homelessness policy model.

Assuming the numbers are valid, this represents a unique achievement. Every locality in the United States seeking federal homelessness-assistance dollars must now regularly conduct homeless surveys and censuses, and recent reports reveal the majority of cities and counties that had conducted prior counts have experienced marked increases of homelessness.

Unfortunately, numbers derived from snapshot-type surveys like San Francisco's are useful only if: (a) the methodology employed is consistent from sample to sample, and (b) the results are put in the proper context.

The Coalition on Homelessness has found fundamental methodological flaws, frequently of the obvious variety, in every past mayoral administration's attempts to attach a number to San Francisco's homeless population. The current administration's recent count raises similar criticisms.

Volunteers were instructed to not interact with those counted, so determinations about who was homeless are, at best, subjective. Rain also played a major factor, forcing homeless bodies deep into whatever cover they could find.

A coalition intern, keen to perform homeless census activities in her Richmond District neighborhood, was told there would be no canvassing of that area at all, owing to the low number recorded there in 2002. Fortunately, she took it upon herself to cover the district on her own, adding data that would have otherwise been absent from this year's census total.

Then there's the issue of the city's parks: forbidden zones for surveyors. We imagine these represent potentially fertile ground for finding homeless people, since the number of citations for camping in them (Park Code 3.12) has nearly tripled.

But when queried by coalition staff, Department of Human Services representatives asserted that Recreation and Park Department employees were to count the evening's campers as they exited the parks the next morning, leaving us questioning how they might marshal enough personnel to monitor each park's perimeters.

And let's put this in a more realistic time frame. Last December, Boston conducted a homelessness survey revealing a modest 7 percent decline in homeless people from the previous year's survey, but this decrease hardly represented an opportunity for even guarded optimism. That's because Boston's total of 5,819 homeless people still represented a nearly 10 percent increase from 5,299 tallied a decade earlier.

It's also troubling that these numbers haven't yet been analyzed in the context of other homelessness indicators. If there's really been a 35 percent reduction in the overall number of homeless people in the city, shouldn't there also be a decrease in requests for homeless services? And this is where the danger lies: declines in homeless numbers leave San Franciscans with the lasting impression that the city's rapidly diminishing funds can now be directed to other needs.

If Newsom wants to point to numbers as a measure of short-term success, the 690 homeless people his policies have placed in stable housing, with a 95 percent housing-retention rate, is a number of which he can be rightly proud.

But throwing around sensational and uncorroborated statistics is a disservice to both homeless people and a public hungry for tangible improvement.

Chance Martin is editor of the Street Sheet.