How should police respond to homelessness? Results from a survey experiment in Portland, Oregon

ABSTRACT

Recent years have seen increases in citizen complaints and legislation about homelessness. Police are often tasked with responding to these complaints and violations. This paper asks: What do people want the police to do when they encounter visible homelessness, and how are these preferences related to characteristics of observers and of individuals who are homeless? It presents findings from a survey experiment delivered to residents of Portland, Oregon, USA. Respondents were given a series of vignettes involving a hypothetical homeless man whose race (Black or White) and background characteristics (substance abuse, mental illness, combat veteran, or control) were randomly assigned. Respondents were then asked to endorse an aggressive (‘arrest’), therapeutic (‘help’), or hands-off (‘ignore’) response by police. Results reveal support for a therapeutic response to visible homelessness, though this was mediated somewhat by the race of the homeless person. The findings contribute to research on public perceptions of police actions.

KEYWORDS:

• Policing
• homelessness
• public opinion
• race

Acknowledgments

Thanks to Adam Slez for suggestions on data analysis and Dina Milovanovic for research assistance. Data collection was funded by a grant from the OSU Research Office. All errors are our own.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. In 2018, the 9^th Circuit Court of Appeals ruled in Martin v. City of Boise (920 F.3d 584 (9^th Cir. 2019)) that cities cannot enforce public camping bans if they do not provide adequate shelter space for unhoused individuals. In 2019, the US Supreme Court declined to hear an appeal from the city of Boise, Idaho, and thereby set precedent for the nine states in the Circuit.

2. Empty cans and bottles are redeemable for $0.10 each in Oregon.

3. Portland has also been the site of sustained protest surrounding policing. Following the murder of George Floyd in May of 2020, the city saw daily protests for more than a year. Our survey was administered before these events.

4. This number represents a 22% increase from the 2017 count. However, some of this increase may be due to additional resources and new sampling approaches implemented in 2019 (Regional Research Institute, Portland State University, 2019).

5. The data for constructing quotas is from the 1-year ACS 2017 (https://censusreporter.org/profiles/16000US4159000-portland-or/). While our sample contains adults aged 18 and above, the ACS race figure is based on the entire population (including those under 18), while the education figure is based on residents 25 or older. Therefore, these quotas are approximations of the adult population in Portland.

6. Qualtrics screened out respondents that completed the survey too quickly or predictably.

7. The three questions are: ‘Where you live, is homelessness a problem?’; ‘Where you live, do homeless people make public spaces less pleasant?’; ‘Where you live, are there too many homeless people in public spaces?’ These three questions have a Cronbach's alpha value of 0.72.

8. We acknowledge some ambiguity in the meaning of the hands-off (‘ignore’) police response. Support for this response may capture either a desire for an intervention by someone other than a police officer (e.g. social worker) or a desire for a homeless person to be free from intervention by any kind of authority figure or state actor. We thank an anonymous reviewer for this insight.

Additional information

Funding

This work was supported by the Oregon State University Research Office.