Super-Cocooning Against Property Crime: Do Visual Primes Affect Support and Does Race Matter

Abstract

American citizens are significantly more likely to experience property crime victimization than violent crime victimization. During a staffing crisis, police prioritize limited resources in combating serious crime; however, property crimes remain impactful to the community. Therefore, agencies need to consider innovative ways to control property crime, such as “super-cocooning” strategies that alert residents to recent offenses. These strategies intend to empower the community to implement guardianship and crime prevention measures. For these strategies to be effective, they require public buy-in and support. The present study implements a preregistered information provision survey experiment (N = 2412), similar to the strategy of super-cocooning, to assess whether the public is more likely to support such strategies to combat property crime. Although the sample held overall high support of this strategy, exposure to a super-cocooning door hanger prime produced no significant changes in perceived effectiveness. However, there was observed racial heterogeneity in the treatments: non-White respondents assigned to the treatment relative to White respondents experienced significantly increased support of super-cocooning strategies. Implications for light-footprint crime control strategies, particularly during a staffing crisis, are discussed.

1. Introduction

Americans are more likely to be impacted by serious property crime relative to violent crimes, in which estimates show about 108 per 1000 households experiencing a property crime in 2022 (Thompson and Tapp 2022). While police agencies understandably prioritize preventing, investigating, and apprehending serious crime/offenders due to the severity of their nature, property crimes remain an important concern (Bannister et al. 2025). During the current police staffing crisis (Adams et al. 2023; Mourtgos et al. 2022), many agencies experiencing staffing shortages may struggle to fully staff patrol and investigative units, respond quickly to calls for service, and provide adequate response to reported crime events (e.g., lengthy investigations), among many other issues (Mourtgos et al. 2022, 2024). Police agencies may triage out certain calls (Ratcliffe and Kikuchi 2019), while prioritizing other calls based on agency staffing levels, whereby some reported crimes may not receive the adequate level of attention due to low staffing numbers.

Police agencies can adopt creative strategies, such as “super-cocooning,” which involves police alerting the community of a recent nearby property crime (e.g., residential burglary) and crime prevention tips/materials to empower the community to aid in property crime control. This strategy allows agencies to inform the public of the crime, increasing neighborhood awareness and guardianship, while focusing limited resources elsewhere to respond to other and more serious community issues. Additionally, there is potential that this strategy may serve as a “force multiplier” (Simpson et al. 2020) through increased “eyes on the street” and additional perceived police presence (e.g., distributed agency visuals), serving as a potential cost-effective crime control strategy.

Experimental and field research tends to focus on violent crimes; however, little is known as to whether the public supports property crime prevention strategies, and whether information provision associated with these strategies sways public support. The key mechanism for super-cocooning to be effective is police communication with affected community residents. Therefore, the use of a messaging strategy within an experimental survey mimics that of the actual crime control strategy used by a police department. That is, this study utilizes an original survey with an embedded experiment to assess (1) public support of a “super-cocooning” strategy and (2) whether detailed information about strategy implementation improves public support of this strategy. The public is the key mechanism for super-cocooning strategies to be successful. That is, the strategy relies on the citizens buying in and taking preventative action. Therefore, it is important to assess public support.

1.1. Super-Cocooning Effectiveness

It has been grounded in repeat and near-repeat crime/victimization research (Bowers and Johnson 2004; Johnson 2008; Johnson and Bowers 2010; Pease 1998; Pease and Laycock 1999; Schnell 2024) that after a crime occurrence (particularly burglary), there is an increased risk of victimization at or nearby the recently victimized locations. With this knowledge, various super-cocooning strategies have emerged to control and reduce burglaries and property crime in the recently burglarized area. Johnson et al. (2017), for example, implemented a randomized controlled trial on 46 neighborhoods and found that treated neighborhoods, which received police response, the delivery of resources detailing target hardening strategies, experienced a significant decrease in crime and re-victimization. Recently, Boehme et al. (2025) randomized six neighborhoods to treatment–control, in which the treatment neighborhoods were distributed door hangers at the victimized location and nearby homes. The findings showed that the treatment neighborhoods experienced a significant decrease in total property crimes and larceny-theft.

These strategies are common in the United Kingdom and Australia, where studies provide strong evidence for the impact of various super-cocooning strategies on repeat and near-repeat property crime victimization. Stokes and Clare’s (2019) post hoc analysis of a strategy distributing pamphlets to burglary-victimized and nearby neighbors found a significant reduction in near-repeat burglaries within 200 m of the victimized location and within the first five days after the initial burglary. Other strategies involve police officers visiting and speaking with victimized households and their neighbors, and have shown promise of effectiveness (Pegram et al. 2018; Sherman et al. 2017; Weems 2014). However, studies are limited as to whether the public supports these light-footprint crime prevention strategies targeting property crime.

It is important to note that some scholars caution against overtly identifying a recently burglarized address. Neighboring citizens may stigmatize victims (De Waardt 2016), and the potential offenders may be alerted to vulnerable homes or businesses. Strategies should mitigate these concerns by omitting the victimized address, avoiding language identifying items stolen or value, and distributing notification materials in wider areas rather than only around the victimized location. In addition, researchers should monitor for unintended harms.

1.2. Public Perceptions of Crime Prevention Strategies

Public perceptions are important in the sustainability and success of crime prevention strategies. Models that depend on community cooperation (e.g., neighborhood watch programs, citizen policing, and information approaches like super-cocooning) require both public understanding and support in order to be effective. When the public perceives a strategy as fair, effective, and minimally intrusive, they may be more likely to endorse it and participate during the early implementation phases (Tyler and Fagan 2008). Conversely, prior research demonstrates that when strategies are viewed as punitive, invasive, or misaligned with community needs, the public may not cooperate, or even could resist (Tyler and Huo 2002; Mazerolle et al. 2013). Therefore, understanding how residents perceive various approaches is important in the design and implementation of public safety protocols.

To date, prior research has mostly examined public perceptions in the context of high-contact strategies, like stop-and-frisk, hot spots policing, and aggressive traffic enforcement. The findings often show more skepticism among racial and ethnic minority groups (Boehme et al. 2024; Gau and Brunson 2010; Geller et al. 2014; Tyler et al. 2014). However, less is known about how the public feels about low-contact or light-footprint approaches. That is, it is unknown whether the public views these strategies as effective or only symbolic efforts.

Further, research has not explored whether these perceptions differ across civilian socio-demographic categories. Prior research shows that non-White civilians hold lower support for police (Boehme et al. 2022; Peck 2015). While non-White civilians do not want police to be defunded or abolished (Vaughn et al. 2022), research has shown that certain criminal justice policies (Metcalfe and Pickett 2018; O’Brien 2020), and specifically policing strategies, are less likely to be supported by non-White civilians (Brunson 2007; Hagan et al. 2005; Tyler and Fagan 2008). For example, prior research has found that non-White residents are less supportive of aggressive policies (Hagan et al. 2005; Weitzer and Tuch 2004). Yet, little is known about whether non-White residents are more supportive of less invasive strategies, such as super-cocooning (Tyler and Fagan 2008).

Finally, racially minority residents could respond more favorably to low-contact property crime interventions. Research shows that minority communities experience higher burglary risk and may feel underserved by traditional patrols (Sampson and Lauritsen 1994; Brunson 2007). These factors could increase support for any strategy that can protect them without intrusive contact. First, minority residents may have stronger preferences for visible, but respectful, police presence (Tyler et al. 2015). Second, interventions that shift guardianship responsibilities toward citizens may resonate more with collective efficacy coping norms in minority neighborhoods (Gibson et al. 2002; Reid et al. 1998; Sampson and Lauritsen 1994).

1.3. The Present Study

While some previous studies have evaluated public perceptions of various super-cocooning strategies (Johnson et al. 2017; Stokes and Clare 2019), there remains scarce research on public support of these strategies and whether information provision assigned randomly sways public support and perceptions of these strategies among the public. Therefore, this study presents findings from a survey experiment examining whether information about super-cocooning strategies encourages public support. This study informs research by experimentally assessing the public opinion of property crime interventions. The results offer policy implications for police agencies assessing whether the public generally supports these strategies and if information from police agencies induces support for these strategies. Specific to this strategy, public support and buy-in are vital for super-cocooning to be effective. The following research questions and hypotheses are tested:

RQ¹: What is the level of general support for the presented super-cocooning strategy?

H1. 

Respondents assigned the super-cocooning treatment will have increased perceptions of these strategies relative to respondents in the control condition.

H2. 

Non-White respondents in the treatment group will hold lower support for these strategies relative to White respondents.

2. Materials and Methods

2.1. Survey Distribution

The authors obtained a list of 903,570 emails of South Carolinian heads of households from Mailer’s Haven. Mailers Haven (2023) is a third-party listserv that validates household addresses multiple times per year through multiple data sources, including the U.S. Postal Service, secondary mailing lists, and open-source data. Addresses are excluded from the distribution list when addresses cannot be verified by Mailer’s Haven, when homes are unoccupied and therefore do not have a head of household listed, and when households actively place their address on no-contact lists. Within the distribution list of home addresses is an associated email list of the head of the household, which serves as the sampling frame for the study.

The survey was developed and administered in Qualtrics. The survey data collection began on 6 August 2024, and ended on 17 October 2024. Five periodic email reminders were sent throughout the data collection timeframe. A final working sample of 2412 was obtained, indicating a response rate of <0.1% according to the AAOPOR RR2 calculator (AAPOR 2023). Pilot testing involved the distribution to about 30 test subjects within the author’s networks, and about 20% of email invitations to test subjects “landed” in their main inbox. The other email invitations were either undelivered or landed in a junk or spam folder. Therefore, we believe that our response rate is likely higher for the emails that truly landed in the main email inbox.

Although this response rate may be concerning, studies have shown that a low response rate is not linked to nonresponse bias (Pickett et al. 2018; Pickett 2017). Further, the large sample size obtained in this study (n = 2412), in concert with the experimental design, increases confidence in internal validity, allowing for causal conclusions between treatment and the outcomes. Additionally, a low response alone may not be determinative of bias (Keeter et al. 2017); however, our commercial email frame could lead to underrepresentation of residents with limited internet access. Notwithstanding this discussion, caution should be taken when making claims of generalizability, with greater confidence in the internal validity of this study’s findings.

An important component of experiments is whether observable characteristics are balanced across treatment and control groups.

Table A1. Balance of covariates across conditions.

Variable	Hanger Treatment (N = 1209)	Hanger Control (N = 1209)	Chi2
	Mean/% of Subsample	Mean/% of Subsample	p
Trust and Confidence	3.97	3.99	0.8
Procedural Justice	3.63	3.66	0.3
Willingness To Obey	3.88	3.87	0.9
Political Ideology
Liberal	15.62%	13.24%	0.1
Moderate	36.54%	40.16%	0.1
Conservative	47.84%	46.60%	0.6
Gender
Male	44.89%	45.81%	0.7
Female	55.11%	54.19%	0.7
Race/Ethnicity
Black	12.27%	12.28%	0.9
Other Race	7.31%	8.65%	0.2
White	80.42%	79.07%	0.4
Non-White	19.58%	20.93%	0.4
Education			0.7
No diploma/GED	0.35%	0.69%	-
High school degree	8.28%	7.16%	-
Some college	18.48%	17.69%	-
Two-year degree	14.30%	14.84%	-
Four-year degree	31.82%	32.70%	-
Postgraduate degree	26.77%	26.92%	-
Age			0.3
18–24	0.52%	1.12%	-
25–34	4.96%	6.12%	-
35–44	12.62%	13.71%	-
45–54	20.02%	19.83%	-
55–64	24.11%	24.14%	-
65+	37.77%	35.09%	-
Married	66.90%	65.83%	0.6
Crime Victim	9.57%	9.40%	0.9
Employed	51.35%	53.02%	0.4
Police Contact	31.94%	31.63%	0.9

in Appendix A presents a balance table of observable covariates with associated chi-square statistics to test whether significant differences in covariates exist between the treatment and control groups. As seen in Table A1, there were no significant differences in covariates between the treatment and control groups, offering evidence of successful randomization via Qualtrics and balance of observable characteristics across the two groups. The overall characteristics of our sample, regardless of treatment condition, can be found in

Table A2. Sample characteristic.

Variable	Mean/%	S.D.	Min–Max
Super-Cocooning Outcome	4.04	0.74	1–5
Trust and Confidence	3.96	0.92	1–5
Procedural Justice	3.64	0.99	1–5
Willingness To Obey	3.87	0.74	1–5
Liberal	0.14	0.35	0–1
Moderate	0.38	0.49	0–1
Conservative	0.47	0.50	0–1
Male	0.45	0.50	0–1
Female	0.55	0.50	0–1
Black	0.12	0.33	0–1
Other Race	0.08	0.27	0–1
White	0.80	0.40	0–1
Non-White	0.20	0.40	0–1
Education	4.51	1.29	1–6
No High School Degree	0.52%	-	-
High School Degree	7.72%	-	-
Some College but No Degree (yet)	18.12%	-	-
2-Year College Degree	14.56%	-	-
4-Year College Degree	32.25%	-	-
Postgraduate Degree	26.83%	-	-
Age	4.70	1.28	1–6
18–24	0.82%	-	-
25–34	5.54%	-	-
35–44	13.17%	-	-
45–54	19.92%	-	-
55–64	24.12%	-	-
65+	36.42%	-	-
Married	0.66	0.47	0–1
Crime Victim	0.09	0.29	0–1
Employed	0.52	0.50	0–1
Police Contact	0.32	0.47	0–1

of Appendix A. The average respondent was a White female who was married and employed. The majority of the sample was between the 45–64 age range, with about 24% between ages 55–64 and about 36% of the sample being 65 and older, indicating a relatively older sample. For education, the majority of the sample had either a four-year college degree (~32%) or a postgraduate degree (~27%), indicating a sample with relatively high educational attainment. Most had not been a crime victim in the past year or had contact with the police in the past year, and 47% identified as politically conservative.

2.2. Research Design

We used an information provision experimental survey design, with the respondents randomly assigned to either the treatment or the control condition. Before the experimental portion of the survey, the respondents were asked questions about their perceptions of procedural justice, trust and confidence in police, and willingness to obey the law (described in greater detail below). After answering these questions, the respondents entered the experimental portion of the survey. If assigned the control condition, the respondents were provided baseline information about super-cocooning/door hanger strategies that stated the following: Following a burglary, some police agencies will send police officers to the victimized neighborhood to distribute information (e.g., pamphlets) to alert residents about the recent crime in the neighborhood. If assigned the treatment, the respondents would receive the same baseline statement as above followed by the following statement: For example, your local police officers may hang “door hangers” like the one pictured below on you and your neighbors front door to increase neighborhood awareness of crime, encourage residents to watch over the neighborhood, and bring about increased police visibility in your area:

Below this information, as indicated by the above statement, the respondents were presented with a photo of the front and back of a door hanger used by a South Carolina police agency (see photo of door hanger in Appendix A, Figure A1). If assigned the treatment, the respondents saw the treatment on one page, and then clicked the “next” button to move to the next page, in which they would again see the treatment with the outcome questions resting below. After the experimental portion of the survey, the respondents were asked various socio-demographic questions (detailed below). The survey was piloted by faculty, post-doctoral fellows, and graduate students on two separate occasions before being distributed to the sample.

2.3. Variables

After the respondents received either treatment or control, they were asked about their level of agreement/disagreement (5 = strongly agree, 4 = agree, 3 = neither agree nor disagree, 2 = disagree, and 1 = strongly disagree), which consisted of our dependent variables. These statements were as follows: (1) This is an effective strategy in preventing future neighborhood burglaries. (2) This strategy will increase my trust with police. (3) This strategy will help me feel safer in my neighborhood. (4) This strategy will make the police more visible in the neighborhood. (5) I would support police doing this in my neighborhood. (6) This strategy shows that police are aware of neighborhood crime.1 These items were combined to create a continuous mean scale (alpha = 0.918). The treatment variable was a binary variable coded as one if respondents were assigned the treatment and coded as zero if they were assigned the control condition.

For the controlled models, pre-experimental perceptions of trust and confidence in police (alpha = 0.939), procedural justice (alpha = 0.958), and willingness to obey the law (alpha = 0.779) (see Appendix A 

Table A3. Measures and alpha scores of combined mean scales.

Variables	Measures
Combined Mean Scale of Dependent Variable (alpha = 0.918)	[1] This is an effective strategy in preventing future neighborhood burglaries. [2] This strategy will increase my trust with police. [3] This strategy will help me feel safer in my neighborhood. [4] This strategy will make the police more visible in the neighborhood. [5] I would support police doing this in my neighborhood. [6] This strategy shows that police are aware of neighborhood crime.
Procedural Justice (alpha = 0.958)	[1] Treat everyone equally. [2] Clearly explain the reasons for their actions. [3] Treat people with dignity and respect. [4] Treat people fairly. [5] Respect people’s rights. [6] Listen to suspects before making any decisions about how to handle a case.
Trust and Confidence in Police (alpha = 0.939)	[1] The police protect people’s basic rights. [2] The police are generally honest. [3] Most police officers do their jobs well. [4] The police can be trusted to do what’s right for my neighborhood.
Willingness to Obey (alpha = 0.779)	[1] I always try to follow the law even when I think it is wrong. [2] You should do what the police tell you even if you disagree. [3] You should accept police decisions even if you think they are wrong. [4] People should obey the law even if it goes against what they think is right.

for these measures). Note that these survey items were asked before the respondents entered the experimental portion of the survey. Post-experimental socio-demographics were also included in the controlled models. These variables include several binary variables, such as sex assigned at birth (male = 1, female = 0; female is the reference category); non-White (=1);2 liberal (=1), moderate (=1), or conservative (reference category); married (=1); employed (=1); crime victim in the past year (=1); and police contact in the past year (=1). Age (1 = 18–24, 2 = 25–34, 3 = 35–44, 4 = 45–54, 5 = 55–64, and 6 = 65+) and educational attainment (1 = no high school/GED, 2 = high school degree, 3 = some college, 4 = two-year degree, 5 = four-year degree, and 6 = postgraduate degree) were categorical variables included in the controlled models. An attention check was used in the survey, and sensitivity models were estimated excluding the respondents who answered the attention check incorrectly. The findings did not substantively change from the main analyses presented below.

2.4. Analytic Strategy

The analytic approach proceeds in five steps. First, descriptive statistics assessing the level of agreement with the outcomes across the sample are presented. This approach shows what level of support the public may have for these strategies, regardless of the experimental condition assigned. Second, the results from t-tests comparing mean differences in outcomes across the two groups are discussed. Third, the uncontrolled (unadjusted) OLS regression results on the combined mean scale are presented. Fourth, the controlled (adjusted) OLS models, including the aforementioned covariates, are then presented. Finally, heterogeneity tests interacting the treatment variable with the non-White (and disaggregated Black variable) are presented and discussed.

3. Results

Looking descriptively at the level of agreement with the various dependent variables that consist of the combined mean scale, across the entire sample (regardless of experiment condition assigned), a substantial majority of the sample supported the strategy. Specifically, they agreed or strongly agreed that the strategy would be “effective in preventing future burglaries” (65%), “will increase trust in police” (65%), and will make them “feel safer in my neighborhood” (69%). Additionally, 87% agreed or strongly agreed that this strategy would make “police more visible in the neighborhood,” 88% agreed/strongly agreed that they would “support police doing this in my neighborhood,” and 91% of the sample agreed/strongly agreed that “this strategy shows that police are aware of neighborhood crime.”

Turning to t-tests assessing mean differences between treatment and control respondents, which can be found in

Table A5. T-tests comparing mean differences between the treatment and control conditions.

Dependent Variables	Treatment	Control	p
	Mean (S.D.)	Mean (S.D.)
Combined mean scale (alpha = 0.918)	4.015 (0.74)	4.063 (0.74)	0.113
Effective in preventing future burglaries	3.798 (0.97)	3.792 (1.00)	0.884
Will increase trust in police	3.792 (0.94)	3.855 (0.96)	0.104
It will help me feel safer in my neighborhoods	3.813 (0.99)	3.911 (0.98)	0.015
It will make the police more visible in the neighborhood	4.157 (0.81)	4.174 (0.75)	0.000
I would support police doing this in my neighborhood	4.245 (0.83)	4.271 (0.82)	0.441
This strategy shows that police are aware of neighborhood crime	4.284 (0.73)	4.269 (0.74)	0.625

Notes: Mean = average score by condition; S.D. = standard deviation; p = p-value.

of the Appendix A, the results show non-significant differences in perceptions of the various measures, with two exceptions. The respondents assigned the treatment held significantly (p = 0.015) lower support to the survey item of “it will help me feel safer in my neighborhood” and significantly (p = 0.000) lower support to the survey item “it will make the police more visible in the neighborhood.”

The results from

Table 1. Uncontrolled regression results.

	AME (S.E.)	p	95% C.I.
Treatment	−0.048 (0.030)	0.113	−0.107–0.011
N	2412
R-Squared	0.001
RMSE	0.744

Notes: AME = average marginal effects; S.E. = robust standard errors; p = p-value; C.I. = 95% confidence intervals; RMSE = root mean squared error.

and

Table 2. Controlled regression results.

	AME (S.E)	p	95% C.I.
Treatment	−0.033 (0.027)	0.211	−0.085–0.019
Trust/Confidence	0.180 (0.033)	0.000	0.115–0.245
Legitimacy	0.195 (0.029)	0.000	0.139–0.251
Obey	0.118 (0.023)	0.000	0.073–0.164
Liberal	0.001 (0.044)	0.986	−0.086–0.088
Moderate	0.025 (0.030)	0.399	−0.033–0.083
Non-White	0.082 (0.028)	0.028	0.009–0.156
Male	−0.127 (0.028)	0.000	−0.182–−0.072
Education	−0.026 (0.010)	0.012	−0.046–−0.006
Age	0.001 (0.014)	0.962	−0.026–0.027
Married	−0.004 (0.029)	0.898	−0.061–0.054
Employed	−0.020 (0.032)	0.525	−0.082–0.042
Crime Victim	−0.006 (0.054)	0.916	−0.113–0.101
Police Contact	−0.005 (0.030)	0.866	−0.064–0.054
N	2247
R-Squared	0.288
RMSE	0.625

Notes: AME = average marginal effects; S.E. = robust standard errors; p = p-value; C.I. = 95% confidence intervals; RMSE = root mean squared error.

below present the uncontrolled model and controlled models, respectively. The uncontrolled model revealed that those assigned the treatment, compared to the control, had a non-significant (p = 0.113) decrease (AME= −0.048) in support for super-cocooning strategies. The controlled model also indicated a non-significant (p = 0.211) decrease (AME= −0.033) in support for such strategies. At baseline levels, the respondents holding higher trust and confidence in police, perceptions of police legitimacy, and willingness to obey the law held significantly higher support for such strategies. The non-White (compared to White) respondents also reported higher levels of support for these strategies. However, the male respondents (relative to female respondents) and those with higher educational attainment held lower support for these strategies.

Heterogeneity Tests

As stated earlier, the non-White respondents may have different opinions of police strategies. First, non-White Americans are disproportionately impacted by crime (relative to White Americans) and therefore may be more (or less) receptive to various crime control strategies. Additionally, the non-White respondents may have differential perceptions of the police (Boehme et al. 2022). Therefore, we examine treatment heterogeneity across the non-White and White respondents to assess whether these racial subgroups are differentially impacted by this specific police strategy. To test heterogeneity of the treatments between the non-White and White respondents, Stata(Version 19.5)’s margins command is used to estimate marginal effects. Point estimates and p-values in the presence of an interaction may be misleading (Busenbark et al. 2022); therefore, marginal effects and graphical methods (Figure 1 below) help tease out the nature of the interactions (Leeper 2017; Williams 2012). The non-White respondents assigned to the door hanger treatment (compared to White respondents) held significantly higher support for this strategy. Therefore, within the treatment condition, the non-White respondents held higher support than the White respondents, indicating that the treatment had differential positive effects on the non-White respondents than the White respondents (

Table 3. Heterogeneity tests of treatments across non-White respondents.

	AME (S.E.)	p	C.I.
Treatment X Non-White	0.150 (0.050)	0.003	0.051–0.248
Control X Non-White	0.019 (0.054)	0.721	−0.087–0.125
Trust/Confidence	0.180 (0.003)	0.000	0.115–0.245
Legitimacy	0.195 (0.029)	0.000	0.139–0.252
Obey	0.117 (0.023)	0.000	0.072–0.162
Liberal	0.000 (0.044)	0.994	−0.087–0.087
Moderate	0.024 (0.029)	0.414	−0.034–0.082
Male	−0.129 (0.028)	0.000	−0.184–−0.074
Education	−0.026 (0.010)	0.012	−0.046–−0.006
Age	0.001 (0.014)	0.932	−0.025–0.028
Married	−0.004 (0.029)	0.891	−0.062–0.054
Employed	−0.017 (0.032)	0.580	−0.079–0.044
Crime Victim	−0.008 (0.055)	0.877	−0.115–0.099
Police Contact	−0.005 (0.030)	0.878	−0.063–0.054
N	2247
R-Squared	0.290
RMSE	0.624

Notes: AME = average marginal effects; S.E. = robust standard errors; p = p-value; C.I. = 95% confidence intervals; RMSE = root mean squared error. The White respondents in the respective condition serve as the reference category.

below). At baseline levels, the respondents holding high trust/confidence in police, more willingness to obey, and higher perceptions of police legitimacy were more likely to support these strategies. Males (compared to females) and those with higher educational attainment were less likely to support these strategies. Sensitivity analyses executing these interactions, comparing the disaggregated Black race category to the White respondents, confirmed these findings that the Black respondents assigned to the treatment held significantly higher support for this strategy.

4. Discussion

This study examined whether information provision about super-cocooning strategies shaped public support for such strategies. As police are experiencing a staffing shortage (Mourtgos et al. 2022) and Americans are more likely to fall victim to property crimes (Thompson and Tapp 2022) (relative to violent crimes), agencies may consider super-cocooning strategies as a property crime control strategy (Boehme et al. 2025). However, no known study has assessed whether the general public approves of such strategies, and whether informing the public influences such support. Understanding public support of these strategies may inform whether the public believes agencies are proactively addressing crime, ultimately contributing to public perceptions of police effectiveness, agency support by the public, and effective strategy implementation. We detail the implications of our findings below.

Descriptively, the entire sample largely supported the strategy introduced in this study. Over 60% of the sample supported that this strategy will be effective in preventing future burglaries, will increase their trust in the police, and will make the public feel safer within their neighborhood. Even greater support (over 87% of the sample) agreed that this strategy will make the police more visible within their neighborhood, that the public would support police implementing this strategy in their neighborhood, and that this strategy will make the public feel that the police are aware of neighborhood crime. Specific to this strategy, public “buy-in” is vital, as super-cocooning strategies rely on public engagement to co-productively reduce crime. That is, these strategies rely on public action, which may be driven by their support of this strategy, to be effective. A concern of such a strategy is whether alerting the public to recent crime insights increases fear of crime among the public (Johnson et al. 2017), which may have deleterious consequences and be counter-productive for police crime reduction efforts. Lasting fear of crime may erode trust in police, prevent cooperation with the police, and affect the public’s everyday behavior. However, these descriptive findings show that the public is generally in support of these strategies, and specifically that it will “make them feel safer” if implemented.

The respondents assigned the treatment, which provided greater detail and the goals of the strategy, and a photo of the door hanger used by a South Carolina police agency did not report significantly different support of respondents in the control condition. In fact, there was a non-significant decrease in support in the treatment group relative to the control group. Those in the treatment group may have seen the photo of the door hanger, read the material on the door hanger, and felt that this is not an effective strategy because they already execute these target hardening strategies. Additionally, the respondents in the treatment group may have felt that simply hanging this door hanger was not “enough” by the police to combat property crime to differentially sway support from the control group. As stated earlier, there was already high support for these strategies across the sample, and this additional information did not move the “needle” to encourage public support. That is, the public may not feel that door hangers address the underlying crime problem or have lasting positive effects. Alternatively, a single stationary image may lack the salience of an in-person visit, police patrol, or multipronged publicity campaign. Future research should test variations in this strategy with varying degrees of dosage.

While the main effects of the treatment produced non-significant effects on public support of super-cocooning strategies, we found treatment heterogeneity across the non-White respondents. That is, within those assigned the treatment, the non-White (and Black) respondents held higher support of these strategies relative to the White respondents. This finding provides several discussion points: First, non-White/Black Americans disproportionately experience higher rates of criminal victimization in general (Thompson and Tapp 2022), and specifically, property crime. Therefore, property crime intervention strategies may be more likely to be supported by racial groups that are more likely to be victims of property crimes. Second, we tested heterogeneity due to the historically strained relationship between non-White/Black Americans with the police (Boehme et al. 2022) to see who may be less likely to support police crime control strategies. We found that the non-White/Black respondents, relative to the White respondents, were differentially and positively impacted by the treatments, which is a novel finding for future research.

Third, and related to the second point, this specific strategy may be supported by the non-White/Black respondents for several reasons. Prior research demonstrates that relationships are strained between police and minority communities, especially when police engage in high-contact, heavy-handed enforcement strategies like stop and frisk, aggressive traffic enforcement, and crackdowns (Boehme et al. 2022; Gau and Brunson 2010; Geller et al. 2014). Conversely, the strategy tested here is a light-footprint, non-contact strategy intended to reduce crime. It runs contrary to other crime control strategies that may require greater public–police contact, such as hot spots, problem-oriented policing, and discretionary traffic stops (Boehme and Mourtgos 2024; Braga et al. 2019; Weisburd 2011). While these strategies are certainly effective and do not always erode police trust (Koper et al. 2024), more contact between the public and police may not be as supported by non-White/Black civilians. This speaks to “just right” and “soft” policing (Ariel et al. 2016; Sherman 2022) by using light footprint crime reduction strategies emphasizing reduced police contact while still being present (Smith et al. 2024; Tregle et al. 2024). Super-cocooning strategies may contribute to this movement, whereby police are demonstrating presence without excessive police contact. Due to this, these strategies may be more likely to be supported by non-White/Black residents who have been disproportionately subjected to over-policing. They may have optimism toward lower contact policing as opposed to traditional methods.

Although we used an experimental design with a large sample, exerting a high degree of internal validity, there remain some limitations worth discussing. The sample surveyed South Carolina heads of households and did not produce a nationally representative sample. While there were some characteristics of the sample that mimic national representation (e.g., percent Black and sex assigned at birth), caution should be taken before making external claims to the rest of the United States. Our survey may have similarly excluded some households with limited internet access. Although we presented a photo of the door hanger and asked for perceptions in the moment, we cannot disentangle behavioral changes. That is, while the treated respondents did not have differential support of these strategies, we cannot parse out what the respondents would do if we individually distributed a door hanger after a recent victimization. Information provision survey experiments can only provide a certain amount of information; however, there may be other mechanisms that sway public support of various strategies that researchers should parse out.

Notwithstanding these limitations, this is the first known study to experimentally test whether information provision of super-cocooning strategies to the general public sways public support of these strategies. While the treatment showed non-significant effects, the treatments differentially swayed public support among the non-White/Black respondents. Police strategies should be tailored to address crime issues while simultaneously considering the impact on public perceptions of the police, particularly among populations that are more affected by crime. Further, agencies should consider leveraging the public to help with combating crime, such as third-party policing (Mazerolle and Ransley 2006) and relevant super-cocooning strategies.

5. Conclusions

This study assesses whether an information provision strategy modeled on super-cocooning efforts shaped public support for property crime prevention initiatives. While the treatment condition, exposing respondents to a visual and description of a police-issued door hanger, did not achieve a statistically significant change in overall support, the descriptive findings showed high baseline levels of public approval for the strategy. This suggests that the general public may already view light footprint crime control efforts positively, especially in the context of property crime and limited police staffing. Importantly, we found treatment heterogeneity by race. The non-White and Black respondents assigned to the treatment condition expressed significantly greater support for the strategy relative to the White respondents. These findings suggest that light contact strategies, like super-cocooning, may resonate more with groups who have been historically overpoliced or underserved by traditional models. Overall, police departments seeking to address property crime during staffing shortages may benefit from considering low-cost, non-intrusive interventions that both foster public awareness and secure support across diverse communities.