The latest annual report on King County’s homeless population from All Home King County found an overall decrease in the number of unsheltered people experiencing homelessness in Seattle, from 4,488 last year to 3,558 in 2019—a reduction Mayor Jenny Durkan touted in a letter announcing the expansion of the Navigation Team as “the first decrease since 2012″ and evidence that ” our shared work to address our crisis of affordability and homelessness is having an impact.” Over the same time period, the number of people experiencing homelessness who were in some form of shelter or transitional housing increased from 4,000 to 4,239.

However, those numbers conceal a few important details: First, that the number of unsheltered people living in tent encampments actually went up in this year’s count, from 1,034 to 1,162. Second, this year—at the request of Mayor Durkan’s Human Services Department—All Home redefined “shelter” to include five “tiny house village” encampments that were previously categorized as encampments. This added 180 people to the “sheltered” count, which accounts for all but 59 individuals added to the “sheltered” category. (A sixth village, at Northlake, was excluded “until it is up to ADA code,” according to the board minutes.) Including the tiny houses—communities where people live in wooden structures the size of a small garden shed—in the “encampment” count would have raised that number to 1,342. The board vote on the redefinition was split 10-4.

In a letter to the All Home board in March, Seattle Human Services Department Director Jason Johnson requested that the tiny houses be moved to the “shelter” category, arguing that they meet “the most relevant” criteria set by the US Department of Housing and Urban Development for “shelter”—namely, that the structures are ADA accessible, that there is security on site, that the site has hygiene facilities, that the structures are ventilated, and that they include sanitary food preparation areas. In the letter, Johnson also notes that the five tiny house villages have case management and offer extended hours or 24/7 access.

“If basic shelters, which only allow people to come in overnight and sleep on floor with no services and amenities are classified as shelter, then permitted villages that meet the HUD requirements of shelter, and have amenities, services and outcomes that far exceed that of basic shelter, should also be classified as such,” Johnson wrote.

Alison Eisinger, head of the Seattle/King County Coalition on Homelessness, which was responsible for what was then called the One-Night Count until All Home took over in 2017, called the reclassification of tiny house villages as shelter “Orwellian” and out of keeping with decades of established practice.

HUD’s minimum criteria for emergency shelter (Appendix A) also include additional requirements, such as smoke detectors in each unit, structural standards, compliance with fair housing rules, heating and cooling, and other requirements that Johnson did not mention in his letter.

The report also found a reduction in the number of veterans, young people, and chronically homeless people living outdoors. Of those three categories, the decrease in veteran and youth homelessness is a clear result of new investments in shelter and housing targeted at those specific populations. The apparent decline in the number of people experiencing chronic homelessness, however, could be a result of the methodology used to come up with that number, which is an extrapolation based on in-person interviews with chronically homeless individuals—defined as individuals who have experienced homelessness for a year or more or on four separate periods during a three-year span, and who also have a disabling condition that prevents them from working or going to school.

Extrapolating these numbers to Seattle (based on the percentage of the population , this finding would suggest that the number of chronically homeless unsheltered people—increased from just over 1,200 in 2017 to nearly 1,800 in 2018, then decreased to just over 600 people between 2018 and 2019. Since chronically homeless people are, by definition, people who are homeless year after year, and since there has not been any massive investment in new permanent supportive housing for hundreds of chronically homeless people in Seattle, the obvious conclusion is that these numbers are not an accurate guide to the actual number of unsheltered chronically homeless people in Seattle from year to year. A similar fluctuation can be seen in the number of unsheltered people with mental illness and substance use disorders—a pattern that probably reflects the challenges with the methodology All Home’s researchers use, rather than any wild fluctuation in the number of people living on the streets with mental illness and addiction from year to year.

Daniel Malone, the director of the Downtown Emergency Service Center, says the surveys that serve as the basis for the counts of unsheltered people in various sub-populations may be to blame. “They survey people, then extrapolate out to the total number of people who are unsheltered, so if one year if you happen to interview a bunch of people who meet the criteria for chronic homelessness, and the next year you interview a bunch of people who don’t, then you’re going to end up multiplying a factor and applying it to the total number of unsheltered people,” Malone says. “I think you naturally have to be much less confident in that kind of demographic extrapolation.”

Kira Zylstra, All Home’s acting director, acknowledges that “there is fluctuation with all of these numbers” based on survey data, “particularly with more refined slices of the data. … For chronically homeless and people with disabilities and other characteristics and needs, it’s dependent on a representative survey, which has even further limitations, as well as reported data” obtained through other sources.

To put a finer point on it, information obtained on sheltered people through the county’s Homeless Information Management System (HMIS) is generally pretty good, because it’s based on tracking individual people from year to year—a fact that’s reflected in the number of chronically homeless people in shelter, which has fluctuated only slightly between 2017 and now.

Information on chronically homeless people living on the streets is much less reliable for a number of reasons , including the fact that interview subjects are located by formerly homeless people themselves, who may gravitate to people and places they already know; the fact that people with major disabilities may face extra challenges that make them less likely to participate in lengthy, in-person interviews with researcher; and the fact that the survey results are extrapolated to apply to much larger populations, despite the fact that in the case of unsheltered people in particular, the survey itself may be unrepresentative.

This year,  the data on all chronically homeless individuals in King County is extrapolated using surveys with about 180 people, some of whom did not respond to all questions. Anything unrepresentative about this population will be multiplied and magnified when the researchers extrapolate from that small sample to the entire homeless population in King County and Seattle. For example, the researchers reached conclusions about the chronically homeless population by figuring out what percentage of survey respondents fit into certain categories—sheltered vs. unsheltered, individual vs. families, etc.—and multiplying that percentage by the total number of people in the general street count in those categories.

Malone, whose organization works primarily with chronically homeless people, says he hopes the extrapolated surveys of unsheltered people won’t be used to dictate policy or funding decisions or to fuel self-congratulatory press releases. He maintains that the best use of the count is as a general comparison of homelessness from year to year—by that standard, he says, the real story is that the unsheltered homeless population has declined as the number of shelter beds in Seattle has increased.