Terms That Refer to Homeless Populations: National Surveys of Several Stakeholder Groups

Over the past three decades, attention has increased about the importance of language when referring to marginalized and vulnerable groups in society, concurrent with movements toward patient-centered care and respect for personal preferences, as well as sociocultural forces against stereotyping and stigma (1). For example, in contemporary psychiatry, referring to someone as a “person with schizophrenia” is now more common and accepted than the noun “schizophrenic.” The former term uses “person-first” language and a description of a person’s state or condition instead of referring to the person by their state or condition. However, this terminology has not been universally accepted across health conditions (2).

In the field of homelessness services and research, a variety of different terms have been used to refer to people experiencing different types of homelessness (e.g., unsheltered vs. sheltered homelessness), who are at risk for homelessness (e.g., unstably housed vs. housing insecure), or who were homeless (e.g., formerly homeless vs. currently experiencing homelessness). No consensus exists on these terms (3, 4). This variation in language and lack of consensus may be important because advocates for person-first language have stated that “respectful language about homelessness can lead to more positive attitudes and actions” (5).

This study addressed the knowledge gap on the topic of language usage for persons with histories of homelessness by using data from national surveys of five stakeholder groups (adults in the general population, adults with lived experiences of homelessness, providers of homelessness services, researchers of homeless populations, and policy makers for homeless populations).

Methods

Data were from two national surveys. The first consisted of a national survey of 6,607 middle- and low-income U.S. adults, conducted in May–June 2020 (6). Eligibility criteria were adults ages ≥22 years, living in the United States, and reporting an annual personal gross income of ≤$75,000. Participants were recruited through Amazon Mechanical Turk (MTurk). To ensure data quality and following best practices (7), we included in the sample only participants who completed ≥50 approved Human Intelligence Tasks (HITs) and who had an HIT approval rating ≥50%. Poststratification weights were applied so that the sample represented the population with respect to age, sex, race, ethnicity, and geographic region. All study procedures were approved by the institutional review board at the University of Texas Health Science Center at Houston.

On the basis of a survey item that asked about homelessness history (“Have you ever been homeless, i.e., don’t have a stable night-time residence, such as staying on streets, in shelters, cars, etc.?”), the sample was divided into participants with lived experiences of homelessness (N=1,346, weighted 20%) and participants with no lived experiences of homelessness (N=5,261, weighted 80%).

Participants were asked: “If you saw somebody sleeping on the streets, what term would you use to describe him/her?” Participants were asked separately about their use of terms “on social media” and “with friends and family.” Response options included “unsheltered individual,” “homeless person,” “person experiencing homelessness,” and “street person.” These terms were based on an environmental scan of the most common terms found in publications and reports on homelessness.

A second national survey was conducted from July to August 2021, through the REDCap survey platform, and distributed to research affiliates of the U.S. Department of Veterans Affairs (VA) National Center on Homelessness Among Veterans, frontline VA homelessness service providers, and members of the Practice-Based Research Network and Clinicians’ Network of the National Healthcare for the Homeless Council. The survey asked participants about their roles (provider, researcher, or policy maker) and geographic location, followed by a series of questions asking about their preferences when writing and speaking about persons living in places not meant for human habitation, persons staying in a shelter, persons at risk for homelessness, and persons who have resolved their homelessness but are receiving homelessness services.

In total, 354 individuals participated in the second survey, and some participants reported holding more than one role as a provider, researcher, or policy maker. Among participants who reported being providers (N=319), 77% (N=247) worked for the VA, 30% (N=96) for a hospital or clinic that serves homeless populations, and 2% (N=6) for a university; some providers worked in multiple settings. Most providers were located in the Southeast (38%, N=121), Midwest (22%, N=70), and West (20%, N=64), with fewer being located in the Southwest (13%, N=42) and Northeast (7%, N=22) regions. Among participants who reported being researchers (N=40), 83% (N=33) worked for the VA, 48% (N=19) for a university, and 23% (N=9) for a hospital or clinic that serves homeless populations; some researchers worked in multiple settings. Most researchers were located in the Northeast (43%, N=17) and the West (25%, N=10), with fewer being located in the Southwest (15%, N=6), Southeast (10%, N=4), and Midwest (8%, N=3) regions. Among participants who reported being policy makers (N=53), 70% (N=37) worked for the VA, 43% (N=23) for a hospital or clinic that serves homeless populations, and 4% (N=2) for a university; some policy makers worked in multiple settings. Policy makers were distributed across the Southeast (25%, N=13), Northeast (21%, N=11), West (21%, N=11), Midwest (17%, N=9), and Southwest regions (17%, N=9).

To analyze the data from the first survey, we divided the participant sample into those who did and did not report lived experiences of homelessness, and we conducted chi-square tests to compare the use of terms in the two groups. Supplementary analyses were conducted to compare use of these terms on social media versus with family and friends among participants with or without a history of homelessness by using McNemar’s chi-square tests.

For the second survey, the participant sample was divided into nonexclusive categories of provider, researcher, and policy maker because some participants held multiple professional roles. Frequency analyses were conducted to examine the most common terms used by participants in each professional category. Finally, participants who reported using “other” terms were asked to specify their terms, which were categorized and summarized with descriptive statistics.

Results

In the survey of the middle- and low-income U.S. adults, most reported that they used the term “homeless person” on social media (70%) and with family or friends (73%) (Table 1). The second most common term was “person experiencing homelessness” on social media (18%) and with family or friends (16%). Participants with no lived experiences of homelessness were more likely to use the term “homeless person” on social media and with family or friends than participants with lived experiences of homelessness; participants with lived experiences of homelessness were more likely to use the terms “person experiencing homelessness” and “unsheltered individual” on both social media and with family or friends compared with participants with no lived experiences of homelessness.

Among participants with lived experiences of homelessness, the term “person experiencing homelessness” was statistically significantly more likely to be used on social media than with family or friends (χ²=5.96, N=1,346, df=1, p<0.01), whereas “homeless person” was significantly more likely to be used with family and friends (χ²=18.57, N=1,346, df=1, p<0.001). Among participants with no lived experiences of homelessness, the terms “unsheltered individual” (χ²=15.46, N=5,261, df=1, p<0.001) and “person experiencing homelessness” (χ²=26.36, N=5,261, df=1, p<0.001) were significantly more likely to be used on social media than with family and friends. In contrast, use of “homeless person” was significantly more likely to be used with family and friends than on social media (χ²=65.88, N=5,261, df=1, p<0.001).

In the second survey, focusing on researchers, providers, and policy makers, the most frequently used term (≥39%) across groups in speaking and writing with respect to persons living in places not meant for habitation was “person experiencing homelessness” (see Table 1 in the online supplement to this report). Providers, researchers, and policy makers also consistently used “person in shelter” and “person at risk of homelessness” as the most frequently used terms in both writing and speaking (30%–40% [N=16–120] and 35%–56% [N=14–180], respectively; ranges of percentages and Ns are across the three groups surveyed [see the online supplement]) when referring to persons staying in a shelter and persons at risk for homelessness, respectively. With respect to persons who have resolved their homelessness but are receiving homelessness services, responses diverged among providers, researchers, and policy makers. Among providers, the most frequently used term was “person who was formerly homeless” in writing (33%, N=106) and “formerly homeless person” in speaking (34%, N=109); researchers most frequently used the term “person who was formerly homeless” in both writing and speaking (45% [N=18] and 43% [N=17], respectively). Among policy makers, the most frequently used term was “formerly homeless person” in both writing and speaking (43% [N=23] and 47% [N=25], respectively).

Table 2 in the online supplement catalogs “other” terms that providers, researchers, and policy makers reported using. Frequently reported terms were “person is homeless/experiencing homelessness,” “imminent homelessness/risk of homelessness,” or simply the homeless person’s name. A few participants elaborated that their use of terms was based on the context and purpose of communication: “depends, can’t select a single term for all situations.” Some participants reported using a combination of terms; for example, one researcher wrote, “I generally say, ‘Person experiencing homelessness,’ regardless, and then qualify as sheltered or unsheltered,” and a provider reported using “a homeless person living in [insert housing circumstance].”

Discussion

Among both adults in the general population and adults with lived experiences of homelessness, the most commonly used term was “homeless person,” whereas a convenience sample of providers of homelessness services, researchers, and policy makers most commonly used “person experiencing homelessness.” This finding suggests a difference in language use between the lay public (i.e., the general population and adults with experiences of homelessness) and professionals working in the field (i.e., health care and social services, research, and policy institutions). The implication of this usage difference is that professionals may benefit from seeking input from other stakeholders on preferred language use.

Professional stakeholders also differed in the terms for persons who are no longer homeless but are receiving homelessness services. Providers most frequently used the terms “person who was formerly homeless” and “formerly homeless person,” whereas researchers most often used the term “person who was formerly homeless,” and policy makers most often used the term “formerly homeless person.” These stakeholder groups reached greater consensus when referring to persons at risk for homelessness and persons staying in a shelter as “person at risk of homelessness” and “person in shelter,” respectively. These findings are relevant to psychiatric services because they suggest that a diversity of terms are used to refer to people with lived experience of homelessness. The results highlight the importance of obtaining input about appropriate usage of descriptive terms from the general public and clients who are homeless to ensure that communication from service providers reaches those clients.

Professional stakeholders’ use of terms showed almost no difference between writing and speaking, with a few exceptions. Providers were more likely to use the term “homeless person” when speaking and more often used the term “person experiencing homelessness” in writing. When we asked adults with or without lived experiences of homelessness about terms used on social media versus in private communication with family and friends, we found that person-first language (i.e., person experiencing homelessness) was more likely to be used among both adults with and adults without lived experiences of homelessness on social media than in private with family and friends. These findings highlight the different contexts in which stakeholders use these terms and suggest that they use different terms in public and private communications.

This study was meant to be descriptive and not proscriptive, so we did not judge which terms were appropriate for which settings. Use of popular terms is part of a broader social negotiation of terms, and our study did not address the ethical and political implications of language. We did not ask participants why they used specific terms and in what contexts and whether they were aware of alternative terms; future qualitative research may be valuable in this respect. Notably, we found that whereas both adults with and adults without lived experiences of homelessness were more likely to use the term “homeless person,” those with lived experiences were more likely to use “person experiencing homelessness” than adults with no lived experiences of homelessness. We therefore conclude that usage of terms may vary by social background, experience, and individual preference. It is likely good general practice when interacting with a homeless person to ask what terms they prefer to use. Political, ideological, or social environment factors may also influence what terms people use. Nonetheless, our data showed that when speaking or writing about homeless persons in general, most stakeholders agreed that some version of “homeless persons” or “persons experiencing homelessness” is commonly and widely used, which furthers knowledge on the topic of usage of these terms.

These findings should be interpreted in the context of several limitations. Results of the surveys may not have been generalizable to all members of these populations, and many stakeholder participants were affiliated with the VA, a sample that may not be representative of these stakeholder groups in the general population. We did not sample adults who were actively homeless at the time of the survey, so future surveys that include them are needed. These limitations were counterbalanced by the strengths of the study, which included surveys of five U.S. stakeholder groups and examination of a variety of terms related to homelessness and in different contexts. This study represents one of the first contributing to the literature of language usage that describes homeless populations.

Conclusions

We found similarities and differences in descriptive terms that refer to adults who are homeless or at risk of homelessness as used by the general adult population, adults with lived experiences of homelessness, and a convenience sample of professionals providing homelessness services, as well as researchers and policy makers. No definite rule exists about the best terms to use, and language tends to evolve over time on the basis of consensus, although there should be room for individual variability as well. Person-first language was used by professionals more often than by the public and by those with experiences of homelessness themselves. These findings are important to consider when trying to engage and communicate with different groups. Further research is needed to examine how these terms may or may not affect attitudes, knowledge, and actions around homeless populations (5, 8, 9). As societies continue to move toward respecting and empowering individuals to be involved in their care and recovery, a greater understanding of the terminology used for and by these individuals may help guide us in reaching these goals.