By Erica C. Barnett

Tonight, temporary Census workers will fan out across King County, and communities all over the country, and attempt to count everyone who is living unsheltered by doing a “head count” of people observed sleeping in tents, vehicles, and on streets and in green belts statewide. Similar head counts, which are a way to include homeless people in the Census rather than an effort to count the number of people experiencing homelessness, began across the nation starting on Tuesday and will wrap up up tomorrow.

The counts are taking place in combination with separate counts of people who stay in shelters or access other homeless services, such as hot meals—the so-called sheltered homeless. This one-night “count,” which will take between four and six hours will be the only effort to enumerate the number of people living unsheltered in the United States—a number that effects not only political representation but the allocation of federal resources to address issues such as homelessness. Because President Trump shortened the Census timeline by a full month, to September 30, the agency will have no ability to recount or recalibrate if local counts go poorly or result in obvious undercounts of people living outdoors.

The ability of the Census Bureau to do an accurate count hinges on whether they follow best practices for counting people who generally don’t want to be found.

So what will tonight’s count look like? According to Los Angeles Regional Census Center spokesman Donald Bendz (whose division includes Seattle), the Census has trained its workers to interview people they encounter and has equipped them with advance intelligence, collected from local homeless service and outreach providers, about where encampments are located in every community. “We work with the city, the county, the state, and all of the partners who work in providing services to people experiencing homelessness and they provide us a list” of places where people are living unsheltered, Bendz said, “and then we have a list that we use from the 2010 census” that will be updated with new information from local service providers.

In practice, homeless service providers and advocates say that outreach to their organizations has been patchy, confusing, and redundant. Nicole Macri, a state representative and deputy director of the Downtown Emergency Service Center, said that “ten or 15 different people [from the Census] reached out to us” asking similar questions. “I don’t know if it’s that COVID made it feel even more rushed and last-minute”—the Census collections, originally scheduled for April, were moved to September due to the pandemic—but “it just felt very confusing and chaotic.”

According to the National Alliance to End Homelessness, the situation is similar in cities across the nation. “A lot of [service providers] expressed that it was confusing, that they had had difficulty reaching people at the Census to discuss issues and problems,” said NAEH president and CEO Nan Roman. “People have had a hard time understanding what was expected of them. The guidance was all over the place.”

Ground-level Census workers say they, too, are confused about how tonight’s “head count” will work. According to two local Census “enumerators,” the training for the overnight count has been scattershot and incomplete, with two weeks of in-person training replaced, due to COVID, by a single in-person orientation and fewer than two days of online exercises. As of late Monday afternoon, one census workers said he hadn’t gotten any details about where his team will be going, the methodology they’ll use for counting people if they don’t want to be interviewed or are asleep, or what to do if they can’t figure out how many people are sleeping in a location. 

“The fact that I don’t even know where we’re doing [the count] tomorrow is a little unsettling,” one worker, who asked to remain anonymous to protect his temporary position, said. “We’ve gotten no instruction at all [on how to count people who are asleep]. I don’t know how we’re supposed to do that—are we supposed to throw the tent flaps back?”

Another temporary Census employee, who originally volunteered to participate in tonight’s count, which comes with a 10 percent pay bonus, backed out after he decided the process was “a shit show”; for one thing, he said, workers were expected to refer throughout the night, and make notations on, a 300-plus-page printout that they only received in electronic form.

The worker said he was also concerned about people who were newly homeless and might not show up a count that only focused on shelters, soup kitchens, and people living outdoors. “When you’re newly homeless, you don’t end up directly on the street—you couch surf or you jump in your car and travel,” he said. “The newly minted homeless certainly don’t have a location where you can count them.”

During tonight’s count, Census spokesman Bendz said, Census workers are supposed to try to talk to anyone who’s present and awake, but that “if they’re asleep, we won’t attempt to wake that person.” A Census training document obtained by PubliCola says that Census workers should try to talk to people in locations like encampments by going through a “group quarters contact person,” such as the “leader: of the encampment—a directive that suggests that ad hoc encampments are significantly more organized than they typically are in practice.

Alison Eisinger, executive director of the Seattle/King County Coalition on Homelessness, told a group of advocates and service providers last week that the coalition has been telling the Census Bureau for more than a year that a going “to places where people are camping by the thousands and ask[ing] them to complete a census form with a total stranger at night is a very poor process that isn’t going to count people effectively.”

Eisinger made her comments during Zoom meeting organized by the Coalition. Instead of telling member organizations to direct Census staff to specific outdoor locations, Eisinger said SKCCH is urging groups to proactively encourage clients to fill out their census forms whenever they come in contact with unsheltered people.

Unlike the annual Point In Time count of the region’s homeless population, the Census won’t be counting tents and cars from a distance and using a standard multiplier to estimate how many people are inside. Instead, Bendz said, they will be going right up to tents and vehicles and attempting to count people individually. “If everyone is asleep in a car, then we will count what we see in the car,” Bendz said. “If it’s a tent and the tent ‘windows,’ for lack of a better word, are open and we can see inside the tent, then we will count the people we see inside the tent.” 

The Census Bureau’s practices differ from the methods used during the Point In Time count in other ways as well. Every year, in the run-up to that count, volunteers spend weeks scouting sites during daylight hours to find encampment locations that might be overlooked at night. On the night of the count, more volunteers, mostly recruited from the ranks of community organizations and groups that work with the homeless population, spread out across a grid carefully designed to avoid double counting. Teams typically include at least one person with lived experience of homelessness who is familiar with the area and able to relate comfortably to unsheltered people the groups encounter.

“It just takes tremendous effort to organize an effective, comprehensive count of people who are unsheltered,” DESC’s Macri said. participated in one-night counts for more than a decade, back when the counts were done by the Seattle King County Coalition on Homelessness. “Compared to the 2010 Census, there are a lot more people who are living unsheltered, and of course there’s a much greater proportion of people who are living homeless who are living unsheltered” in 2020, Macri adds.

Bendz said the Census Bureau has done months of outreach to service providers to figure out the best way to count people living unsheltered, but Macri—whose organization provides outreach and is one of the largest shelter providers in the Seattle area—told me that she was “unaware” of tomorrow’s count until I asked her about it. DESC director Daniel Malone told me that he, too, was unaware of any communications with the Census Bureau about encampment locations.

“The fact that DESC is a major homeless service provider, and it’s not clear to me that it’s well-known within the organization that this is happening, is a big red flag,” Macri said. Roman, from the NAEH, said that she heard from one large city that Census officials told the county that they were working closely with the Continuum of Care—the regional planning body that coordinates homeless services for a county or other jurisdiction—”but none of the [CoC] board and none of the staff had ever talked to them.”

Without local expertise or team members who are actually familiar with local homeless populations, advocates say, the Census Bureau’s version of a one-night count will almost certainly undercount people living on the streets. “I’ve argued with the Census [Bureau] that this is a terrible idea and not a good methodology,” said Kirsten Jewell, the Housing and Homelessness Program Coordinator for Kitsap County. “I was very adamant that doing it overnight was not going to be a very effective or accurate way to count. I got to make my case both in writing and over the phone, and ultimately I was told this was a national strategy and they were doing it the same way across the country.”

The uniformity of the Census Bureau’s strategy doesn’t take into account the meaningful differences between urban and rural homelessness, Jewell added. For example, most unsheltered people in her rural county live in hard-to-access forests and greenbelts where tents may be impossible to find at night. “It’s very difficult in a rural environment,  where you’re literally thrashing through the woods to find an encampment,” she said.

Instead of nighttime bushwhacking, rural point-in-time counts typically rely on daytime interviews with unsheltered people, plus trained volunteers who head out equipped with maps and incentives. “For our small county, we sent 165 citizen volunteers out with maps and goodie bags,” Jewell said. “We know how to do this, and we still have a dramatic undercount.”

Getting an accurate count matters because the decennial Census is used to allocate resources that are critical to people experiencing homelessness, both sheltered and unsheltered. Census data, including accurate information on people’s housing status and income, helps government agencies and nonprofit groups assess the scope of homelessness in their communities, apply for grant funding, and receive fair funding allocations from the federal government.

“For people that are housed, there are multiple opportunities to be counted—there are postcards that come to your house, for example,” Jewell said. “But for people who are unsheltered, there’s just one shot.” Jewell doesn’t fault Census workers themselves, who she said are clearly “trying to do a good job”—she faults the federal policy that has resulted in the process that she and other advocates and service providers call inadequate. “It almost feels like they’re writing off people who are homeless,” she said—”like they’re a footnote, or an afterthought.”