The Impact of Informational Interventions about Police Alternatives on Police Reliance in the United States
In the United States, police officers are called upon to address both violent and nonviolent situations. Several government and community based organizations exist to provide alternatives to police in situations related to domestic violence and nonviolent situations such as mental health crises. Researchers conducted a randomized evaluation to test the impact of providing information about police alternatives on individuals' reported likelihood of calling the police and a follow-up randomized evaluation to understand key public-safety stakeholders’ receptiveness of police alternatives. The first study found that information about police alternatives had a bipartisan decrease on police reliance in nonviolent scenarios. The second study revealed that police are more responsive to evidence on public interest in police alternatives when it is presented using some framing devices compared to others. Qualitative interviews shed additional light on the mechanisms underlying these effects.
Policy issue
Police officers are often called upon to respond to nonviolent incidents that other professionals may be well suited to handle. Many governmental and nongovernmental resources exist as alternatives to police for nonviolent scenarios. Some examples include hotlines for suicide prevention and mental health emergencies; social services such as housing aid, nutrition programs, and employment support; and other community-based resources offering support for youth, LGBTQ+ individuals, people experiencing homelessness, domestic violence, and substance use disorder.
Proponents of these alternatives believe they may help improve community safety while reducing police violence, which has increased in recent years across the United States.1 In 2023, over 1,200 people were killed by police.2 Over half of these deaths occurred after police responded to suspected nonviolent offenses or in cases where no crime was reported, such as someone experiencing a mental health crisis.3 This violence disproportionately affects people of color—particularly young men of color4—and other groups such as those experiencing poverty5 or homelessness.6
In the wake of high-profile incidents of police violence in the United States, there has been a growing call to explore alternatives to traditional policing. However, the scale of policing has not reduced as a result of recent anti-police violence movements. It is not clear whether this trend is due to a lack of demand for police alternatives or a lack of information about such alternatives among the general US population. It is also unclear if the police force is supportive of shifting towards alternative crisis responders for nonviolent scenarios. The study seeks to answer whether providing information about police alternatives impacts public reliance on police, particularly in nonviolent situations and if police and other stakeholders are open to these alternatives.
Context of the evaluation
To answer these questions, researchers conducted a series of randomized evaluations.
In the first study, researchers used an online survey platform in the United States, Prolific, to recruit participants in July and August 2024. Survey respondents were 60 percent white, less than half identified as men, and the average age was 47. A majority of participants self-identified as Democrats. After randomization, the three groups were balanced on age, race, gender, political affiliation, marital status, and a baseline value on an index for policing demand. The study participants were fairly representative of the United States population when compared to demographics from the American Community Survey and Current Population Survey, but skewed slightly older, more educated, more Democratic, and had a higher proportion of Black people.
In the second study, conducted from February to March 2024, researchers gathered contact information for relevant stakeholders using data from the National Public Safety Information Bureau, public Department of Justice records, and local official contacts from an online resource. They sought out police officers and sheriffs, local officials, DOJ grantees, and other law enforcement. The final sample included 45,163 valid email addresses. The main analysis included the 11,623 recipients who opened the email, across 2,368 US counties.
Details of the intervention
Study One: Survey
Researchers evaluated the impact of providing information about police alternatives on individuals' reported likelihood of calling the police. In August 2024, researchers recruited 2,745 adults in the United States to participate in the evaluation. Participants were randomly assigned to one of three groups.
- Comparison Group: Participants in this group received a placebo video providing information on economic indicators (e.g., full-time and part-time employment rates).
- Government Group: Participants in this group received a video providing information on government resources that can serve as alternatives to police response, including the 988 Suicide Prevention Hotline, the 211 Community Assistance Hotline, and the 311 City Services Hotline.
- Don’t Call the Police (DCTP) Group: Participants in this group received a video highlighting how police are often the default responders to emergencies in the United States, even for calls not involving violent crimes. The video then presented information on nongovernmental emergency response options through dontcallthepolice.com, a database of community resources suitable for attending to nonviolent crises. The video also included information on the 988, 211, and 311 hotlines from the Government video.
Participants were presented with five scenarios and asked to rate their likelihood of calling the police for assistance in each case. The scenarios included both violent incidents (e.g., armed robbery) and nonviolent situations (e.g., a naked man walking down the street). Scenarios were presented to respondents in a randomized order. The evaluation measured participants' self-reported inclination to call the police and their demand for alternative first responders.
Researchers conducted a follow-up survey with study participants a week after the initial survey to assess the persistence of the effects on awareness of the 988 helpline.
Study Two: Stakeholder Outreach
The second part of the evaluation aimed to estimate receptiveness among key stakeholders who shape policies to promote public safety (e.g., police officers, sheriffs). The researchers conducted a field experiment disseminating a summary of a previous version of the survey experiment (the results of which are similar to those in this evaluation) to a sample of these stakeholders.
Counties were randomized into two conditions with different email subject lines:
- 911 Subject line: “New research: bipartisan support for police alternatives - receive information about the findings and 911alternatives.com”
- DCTP Subject line: “New research: bipartisan support for police alternatives - receive information about the findings and dontcallthepolice.com”
The content of the email was the same for both groups; it informed recipients that providing the public with information about resources available on dontcallthepolice.com reduced the likelihood of calling the police for nonviolent encounters. Stakeholders could request more information on the study and provide feedback on the report.
The researchers also conducted qualitative interviews with sixty participants to deepen their understanding of the findings.
Results and policy lessons
The first study found that providing information about police alternatives significantly reduced the likelihood of relying on police in nonviolent scenarios. Participants who received information about the 988 Suicide and Crisis line were more likely to use the hotline in a hypothetical suicidal ideation scenario a week later. In the second study, police officers—but not other stakeholders—engaged with the DCTP email at higher rates than the 911 email.
Study One: Survey
Demand for Police in Violent Scenarios
There was no meaningful impact of the DCTP or Government interventions on police reliance in the armed robbery scenario presented. The comparison group mean for this scenario was 93 percent, suggesting that individuals may already recognize the important role of police intervention in high-stakes, violent situations.
In the other violent scenario presented, where a woman is screaming and crying while a man makes threats, the DCTP group was 5.5 percentage points less likely to say that they would call the police (a 7 percent decrease from a baseline of 80.7). In contrast, the Government group was 1.8 percentage points more likely to say they would call the police in the same scenario (a 2 percent increase from a baseline of 80.7).
Demand for Police in Nonviolent Scenarios
Participants in the DCTP group were 2.6–17 percentage points less likely than the comparison group to call the police in nonviolent situations ranging from a scenario where someone is disruptively asking for money (a 10 percent decrease from a baseline of 24.7) to suicidal ideation (a 33 percent decrease from a baseline of 51.8).
In contrast, participants in the Government group were 3.1 percentage points more likely to say they would call the police in a scenario where someone is disruptively asking for money (a 12 percent increase from a baseline of 24.7). In the other nonviolent scenarios presented, participants in the Government group were just as likely to call the police as those in the comparison group.
Preferences for First Responders
The study also examined participants' preferences for who should respond in each scenario: the police, a social worker, or no one. The DCTP group demonstrated reduced preference for police response in all but the armed robbery scenario. Preference for police as first responders decreased by 4.75–12 percentage points (ranging from 8–39 percent decreases from respective baselines) across the remaining four scenarios.
Researchers found a corresponding increase in preference for a social worker as the first responder. The DCTP group increased their preference for a social worker by 4.71–14.6 percentage points (ranging from 6–67 percent increases from respective baselines) across the remaining four scenarios.
The Government group did not demonstrate different preferences for first responders relative to the comparison group in any scenario.
Follow-up Survey: Impact on Recall of Police Alternatives
One week after the initial intervention, a follow-up survey assessed participants' recall of the 988 Suicide & Crisis Lifeline. The follow-up survey asked participants to choose a hotline to call within 20 seconds of being faced with a hypothetical suicidal ideation scenario. Over 85 percent of the initial study participants engaged in the follow-up survey (2,342 respondents).
Participants in the DCTP group were 8 percentage points more likely to remember the 988 hotline relative to the comparison group who did not receive such information (an 89 percent increase over the comparison group mean of 9 percent). Participants in the Government group, who also received information about the 988 hotline, were 9 percentage points more likely than the comparison group to recall the hotline information (a 100 percent increase over the comparison group mean of 9 percent).
Study Two: Stakeholder Outreach
Non-police recipients did not respond or engage at meaningfully different rates to the two emails while police recipients engaged more with the DCTP email than with the 911 email.
Compared to study participants who are not police and received the 911 subject line, police who received the DCTP subject line were 3 percentage points more likely to request additional information on the study (a 61 percent increase relative to a comparison mean of 5). The authors hypothesize that the differential response by subject line reveals police resistance to specific framings of alternative responders that would symbolically or substantively challenge the role of police in US public safety. Qualitative follow-up interviews revealed support for this theory.
The first study demonstrated that providing information about police alternatives can meaningfully reduce the demand for police in nonviolent situations while maintaining or even increasing the demand for police in violent scenarios across political affiliations. The impact on recall of the 988 hotline further underscores the effectiveness of targeted educational interventions in promoting the use of alternative crisis resources. The second study demonstrated the importance of generating buy-in from key stakeholders, such as police themselves, and found that specific framing devices may be helpful for engaging these audiences.
Ba, Bocar A., Meghna Baskar, Tony Cheng and Rei Mariman. 2024. "Understanding Demand for Police Alternatives." NBER Working Paper #32418, August 2024