The Perceived Effectiveness of the Broken Windows Theory as it is Applied by Police Departments

The Perceived Effectiveness of the Broken Windows Theory as it is Applied by Police Departments

Introduction
In the early 1980’s, two criminologists named James Wilson and George Kelling proposed that policing minor disorder in a neighborhood could reduce serious crime. This theory, called the Broken Windows theory, is defined as “the idea that disorderly conditions left untended will lead to serious crime, much like a broken window left unfixed will lead to the breaking of more windows, and eventually to the demise of the entire building” (Gault & Silver, p. 240). Because of its simple logic, it is a popular theory. It was first adopted by the police departments of New York and Chicago and has since become the main theory behind the zero-tolerance police policies that are implemented in many police departments around the U.S. (Harcourt & Ludwig, p. 271). It has emerged as one of “the most important policing strategies of the past two decades” (Hinkle & Weisburd, p. 500).

The Broken Windows theory is a controversial theory because of its extensive use in police departments today and its apparent lack of empirically based support for its validity. Given the widespread use of zero-tolerance police policy, there are “relatively few empirical studies of the impacts of broken windows policing efforts” (Hinkle & Weisburd, p.505). Many criminologists think that the link between disorder and crime that the theory promotes has yet to be shown, and all evidence that zero-tolerance policing does reduce serious crime is merely spurious. While both criminologists and police officers are involved in the same interests, the theories that criminologists create and study and the policies that police officers subsequently write and implement do not always complement each other. While the Broken Windows theory is a logical approach to neighborhood disorder, using this largely untested theory in nation-wide zero-tolerance policies could be seen as an impulsive move to satiate the public perception of a bad neighborhood rather than to effectively reduce crime.

The purpose of this research paper is to look at two groups of people that have a stake in the argument about the effectiveness of zero-tolerance policing: criminal justice students and police officers. Both groups were asked “Does the implementation of the Broken Windows theory in zero-tolerance policing directly and effectively reduce crime?” This was the main research question and the dependent variable was this opinion about the Broken Windows theory and whether when implemented by police departments it effectively reduces crime. The second research question and dependent variable was “Does reducing disorder reduce crime”. For both research questions, the independent variable was the participants’ identity as criminal justice students or as police officers. These research questions are important because they recognize the different agendas of criminologists and police officers in respect to criminal theory and police policy. The hypothesis was that criminal justice students would be less likely than police officers to think that zero-tolerance policing directly and effectively reduces crime, and that reducing disorder does not necessarily reduce crime, because they would be more familiar with the broken windows theory and its implications.

The data for this research paper was collected through a non-experimental research design. The researcher used questionnaires containing 10 research questions to criminal justice students at the University of Missouri-St. Louis and to police officers serving in the greater St. Louis area. The participants were collected through a quota sampling method with the end goal of achieving 100 participants. The researcher used both email and ground distribution of the surveys to achieve the quota. The Broken Windows controversy has generated much thought and many critiques from criminologists around the U.S. To gain more context for this research topic, the following is a review of some of the current literature on the Broken Windows theory and the zero-tolerance police policies.

Literature Review
“Broken Windows: New Evidence from New York City and a Five-City Social Experiment”
Bernard E. Harcourt and Jens Ludwig
The University of Chicago Law Review, Vol. 73, no. 1, 2006, 271-320

The purpose of this research is to examine the broken windows theory and the link between social disorder and crime. The researchers used a previously conducted study by Kelling and Sousa (KS study) and their own study of residential rates of crime to test the effect that social disorder has on neighborhood crime rates. They “set out to reanalyze and assess the best available evidence from New York City on the effects of broken windows policing” (Harcourt and Ludwig, 275). Harcourt and Ludwig examined the findings of the 2001 KS study that examined crime trends in New York City. The KS study was one of the few empirical studies that supported the broken windows theory. It showed that there was a strong tie between minor crime arrest rates and serious crime rates. It was quoted as showing that “aggressive misdemeanor arrest policies in New York City account for the significant drop in crime during the mid-to-late 1990’s” (Harcourt and Ludwig, 274). This conclusion was very satisfying to the public and to law enforcement because it demonstrated that police presence really does matter and is an effective method of reducing crime. However, Harcourt and Ludwig argue that the decrease in serious crime rates was due not to the zero-tolerance policing efforts but to the fact that crime levels in New York City had been elevated for some time due to the crack cocaine epidemic, and the decline in crime was more natural and cyclical than based on the efforts of the police. Harcourt and Ludwig go on to state that this study did not in fact prove the efficacy of broken windows policing. They also did their own social experiment called Moving To Opportunity, which was conducted in five U.S. cities. This experiment was conducted because the researchers had a rare opportunity to analyze the direct effect that social disorder has on resident crime rates. The Department of Housing and Urban Development (HUD) decided to relocate 4,600 families on random assignment from high crime buildings to areas that had much lower levels of crime. Harcourt and Ludwig used this decision to study the effects of the change in social disorder levels on the levels of crime among these residents. The primary research question asked if moving residents from a high crime neighborhood to a low crime neighborhood would have an impact on their crime rates. The independent variable was neighborhood social disorder and the dependent variable was crime levels. They used official data from the residential areas to determine disorder levels, which would be their independent variable, and used arrest rates for the residents to find their dependent variable.

Harcourt and Ludwig found no support for broken windows theory based on their analysis of the KS study. They also found that relocating the residents from high crime to low crime areas does not reduce the level of crime that group was participating in. They found no support for the idea that policing social disorder or changing the level of social disorder had a direct impact on crime rates. The evidence from New York City and the HUD social experiment showed no support for a direct link between disorder and crime. The implications of this study are tremendous. Harcourt and Ludwig not only cast into doubt the relationship disorder has on crime levels, but also the effectiveness of a police force in reducing the levels of crime. Their conclusion that focusing a police force on minor crimes and misdemeanors has no effect on violent crime rates is monumental. Many police departments use zero-tolerance policing as a means to attempt to reduce serious crime. If Harcourt and Ludwig’s study is sound, then police departments across the country will have to rethink their policing strategies. Harcourt and Ludwig’s study has direct relevance to this current study, because it demonstrates the need for criminologists and police departments to take an honest approach to what works in reducing crime, and what doesn’t. Rather than follow a policing strategy that is popular because it validates the need for police presence and seems logical to the public, police departments and criminologists alike should study and research key crime theories before putting them into practice. Criminologists are in a position to study policing and offer theories that might reduce crime rates, and they have a responsibility to educate law enforcement to act more effectively. Police departments should respect criminal theory and encourage only empirically tested policing strategies to be implemented on a wide scale.

“The Irony of Broken Windows Policing: A micro-place study of the relationship between disorder, focused police crackdowns and the fear of crime”
Joshua C. Hinkle and David Weisburd
Journal of Criminal Justice, Vol. 36, 2008, 503-512

This article looked specifically at the place that fear of crime has in the Broken Windows theory. It stated that the tie between neighborhood disorder and fear of crime is an important part of the theory. The purpose of this study is to discuss the thought that the fear of crime is a reaction to neighborhood disorder that leads to other reactions that then foster criminal activity. The research specifically examined the tie between fear of crime and neighborhood disorder in the context of a zero-tolerance police program. The primary dependent variable was resident fear of crime. The independent variables were observed social disorder, perceived social disorder, observed physical disorder, extra police presence, crime, pre-intervention average fear, and direct and indirect victimization. Data for this study was collected from data from a police intervention on several high crime areas in Jersey City, New Jersey. The study gathered data on how the intense police intervention to reduce disorder effected the residents’ fear of crime. The participants were residents in these high crime areas that were over the age of 18 and willing to be interviewed after the intervention took place. 733 interviews were conducted.The study found that resident perceptions of social disorder and observations of physical disorder were significantly related to the fear of crime. This supported the hypothesized tie between disorder and fear of crime. The study also found that the multiplied police presence in the neighborhood contributed to a high level of fear in the residents. The implications of this study suggest that police departments should be cautious of using zero-tolerance policies to reduce neighborhood disorder, because the police presence could be counter-productive.The study conducted by these researchers is very relevant to the Broken Windows debate because it directly examines the influence this theory has on policing measures. It looks at the effect that Broken Windows policing has on a neighborhood’s level of fear, which is an important component of this theory. It also shows that the very police presence used to reduce disorder can in fact raise the levels of neighborhood fear and start a chain of events that leads to more criminal activity. This point should be considered and weighed carefully as more police departments adopt zero-tolerance policies.

“Spuriousness or Mediation? Broken windows according to Sampson and Raudenbush”
Martha Gault and Eric Silver
Journal of Criminal Justice, Vol. 36, 2008, 240-243

In their article on the theory of broken windows, Gault and Silver re-examined Sampson and Raudenbush’s assertion that because they found no direct link between crime and disorder, the broken windows theory was invalid. The purpose of their study was to reinterpret Sampson and Raudenbush’s results to show that they were not contrary to the Broken Windows theory. The study looked out how the link between disorder and crime was first understood. Many people thought that Wilson and Kelling were hypothesizing a direct link between the two, to the effect that neighborhood disorder causes crime. Gault and Silver argued that this interpretation is imprecise. They stated that neighborhood disorder leads to a sequence of events that end in crime. They specifically emphasized that disorder reduces informal social control which in turn creates a neighborhood atmosphere that invites serious crime. As a non-experimental study, Gault and Silver identified their position on the original meaning of the broken windows theory and then used it to re-explain Sampson and Raudenbush’s landmark study. In 1999, Sampson and Raudenbush studied 196 Chicago neighborhoods by driving down both sides of their streets and video-taping them to collect physical and social disorder and crime data. They combined this data with survey and official crime data. Their research question was whether disorder was a cause of crime of whether disorder and crime were both caused by low collective efficacy. Sampson and Raudenbush then concluded that the relationship between disorder and crime is spurious and not mediated by the collective efficacy of a neighborhood, and therefore the entire broken windows theory is fundamentally flawed. Gault and Silver disagreed with the conclusions of Sampson and Raudenbush’s study, because they argue that the Broken Windows theory does not state that disorder has a direct effect on crime, but that disorder has an effect on social control, which then has an effect on crime levels. This study implies that there is reason for debate over the Broken Windows topic as a theory just as much as there is as a policing strategy. Many criminologists disagree on the relationship between disorder and crime and that should influence the way police departments handle this approach to crime. Because this topic is so popular both in the criminal justice field and in the law enforcement field, Gault and Silver feel it is important to point out that in their interpretation, Sampson and Raudenbush’s study could be seen as supporting the broken windows hypothesis rather than contradicting it.This article is an insightful look into both the Broken Windows theory and Sampson and Raudenbush’s famous study. Determining the true effect that neighborhood disorder has on crime is the key to effective policing of high crime neighborhoods.

“Effect of Broken Windows Enforcement on Clearance Rates”
Hyunseok Jang, Larry T. Hoover and Brian A. Lawton
Journal of Criminal Justice, Vol. 36, 2008, 529-538

The purpose of this research was to look at how Broken Windows law enforcement strategies effect crime clearance rates using the Uniformed Crime Control (UCR) data. No other study has been conducted with this goal, giving this study a particularly significant voice in broken windows literature. The primary research question focused on how broken windows policing effects crime clearance rates. The independent variable was Broken Windows policing and the dependent variable was crime clearance rates. The dependent variable was broken down into 8 separate clearance rates for: robbery, aggravated assault, burglary, larceny, auto theft, violent offense, property offense, and index offenses. This study analyzed UCR data from the 35 largest cities in Texas from 1990-2004 to get their clearance rate data. They measured the level of broken windows law enforcement by the number of misdemeanor arrests that occurred in a given time period. Comparing the misdemeanor arrest data of the 35 police agencies allowed them to indentify strong or weak zero-tolerance policies. The research found that burglary and auto-theft clearance rates had a positive significant relationship with zero-tolerance policies. Interestingly, increased zero-tolerance policies decreased the clearance rates for larceny. For the other 5 variables, there was no significant relationship between the zero-tolerance policing and their clearance rates. The implications of this study are that zero-tolerance policing is not necessarily as effective as many people believe. This study was done over a large collection of data with very conclusive results that cast doubt on the continued wide-spread use of zero-tolerance policing as an effective means to reduce crime. An analysis of this study shows that their methodology and sampling data are sound. Their interpretation of their data could be over-simplified however. When dealing with 35 police municipalities the reasons why zero-tolerance policing might not work in certain areas would vary widely and can’t be immediately attributed to the ineffectiveness of the police policy. Having said that, more studies like this one should be conducted to gain an accurate empirical review on the realistic impact zero-tolerance policing has on reducing neighborhood crime. As demonstrated by this study, having a high arrest rate for misdemeanors does not guarantee a low arrest rate for more serious crimes.

Methods
In my research I decided to examine the zero-tolerance police policies that came into effect after the popularity of the broken windows theory to determine her research question. I asked “Does the Broken Windows theory if it’s implemented through the zero-tolerance police policies directly reduce crime?” My hypothesis was that criminal justice students would be less likely than police officers to think that zero-tolerance policing directly and effectively reduces crime, because they would be more familiar with the Broken Windows theory and its implications. The dependent variable (DV1) was the opinion of the effectiveness of the broken windows theory, and the independent variable (IV) was the identity of the participants, as criminal justice students or as police officers. I also examined a second bi-variate analysis with the dependent variable (DV2) as the opinion on whether disorder causes crime, and the independent variable (IV) the same, as the identity of the participants. I employed a non-experimental research design. The sample number was determined using a quota sampling strategy. No true sampling method was used. The quota was set at 100 participants. The sample population was determined by the nature of the research topic and the research question.

Because I was interested in the theoretical nature of her topic as well as the applied aspect of her topic, I decided that students of criminal justice theory and police officers who by their job description implement criminal justice theories would be the appropriate participants for this study. The quota of total participants was set at 100 (n=85), with a goal of 50 criminal justice students (n=45) and 50 police officers (40). The criminal justice student population was drawn from the researcher’s university, the University of Missouri-St. Louis. This decision was made because it was the most convenient body of criminal justice students accessible to the researcher. The police officer population was drawn from four police stations in close proximity to the researcher’s house, and from the University of Missouri-St. Louis’ police department. This made a total of five police stations surveyed. The data was collected by the distribution of a questionnaire that asked ten yes or no questions. The questionnaires had the researcher’s name, university, and instructions at the top of the page. The questionnaires were distributed in two different ways. For the University of Missouri-St. Louis students, it was created using an internet sight called surveymonkey.com and a link to the questionnaire was sent as an email attached to the class rosters of six criminal justice classes the researcher had taken. For the police officers, the researcher drove to the five police stations and handed a stack of surveys and a cover letter to each front desk officer. The online website for the student survey compiled the responses as they came in, and the researcher collected the police officer’s surveys eight days after she had distributed them.

Results
The frequency distributions of my survey are shown in Table 1 in the appendix, in my univariate test table. It shows that 47.6 % of my participants were police officers, and 52.4% were criminal justice students. 78% were familiar with the broken windows theory, 22% were not. 31.7 % were both police officers and students of criminal justice students, 68.3 % were not. 79.3% thought there was a link between disorder and crime, 20.7% did not. 74.4 % thought that reducing disorder does reduce crime, 20.7% did not. 47.6% thought that the Broken Windows theory is behind zero-tolerance police policy, 52.4% did not. 57.3 percent thought that police implementation of the broken windows theory reduced crime, 42.7% did not. 65.9% of my participants thought that zero-tolerance policing effectively reduces crime, 34.1% did not. I ran two bi-variate analyses. The first cross-tabulation’s DV was the opinion on if police implementation of the broken windows theory effectively reduces crime. The IV was the participant’s identity as police officers or as criminal justice students. If the participant answered yes, he thought that police implementation of the broken windows theory effectively reduces crime. If he answered no, he thought that police implementation of the broken windows theory does not effectively reduce crime. The cross-tabulation in Table 2 A shows that in answer to the question, 57.3% of my participants said yes, and 42.7% said no. Of police officers, 64.1% said yes, and 35.9% said no. Of criminal justice students, 51.2% said yes, and 48.8% said no. The results for this cross-tabulation were not significant. The chi-square value was 1.40 with a significance value of .237, and in order for it to be significant that value must be less than .10. The value of chi-square for this test was 1.400, which is a relatively small number, showing that the independent variable results for identity were very similar to each other. 57.3% within identity answered yes: 64.1% of the police officers, 51.2% of the criminal justice students. Looking at it from the negative perspective, 42.7% within identity answered no to the test question: 35.9% of the police officers, 48.8% of the criminal justice students.

The second cross-tabulation’s DV was the opinion on if reducing disorder reduces crime. The IV was again the participant’s identity as police officers or as criminal justice students. If the participant answered yes, he thought that reducing disorder would reduce crime. If he answered no, he thought that reducing disorder would not reduce crime. The cross-tabulation in Table 3 A shows that in answer to the question, 74.4% of my participants said yes, and 25.6% said no. Of police officers, 84.6% said yes, and 15.4% said no. Of criminal justice students, 65.1% said yes, 34.9% said no. The results for this cross-tabulation were statistically significant with chi-square equal to 4.082 and a significance level equal to .043. The value of chi-square was 4.082, which is a relatively large number for this study. 74.4% within identity answered yes: 84.6% of police officers and 65.1% of the criminal justice students. Looking at it negatively, 25.6% within identity answered no: 15.4% of the police officers, and 34.9% of the criminal justice students.

First Bi-Variate Analysis
Table 2A
does police implementation of the Broken Windows theory effectively reduce crime * Identity Cross-tabulation
Identity
police Criminology Student Total
does police implementation of the Broken Windows theory effectively reduce crime Yes Count 25 22 47
% within Identity 64.1% 51.2% 57.3%
No Count 14 21 35
% within Identity 35.9% 48.8% 42.7%
Total Count 39 43 82
% within Identity 100.0% 100.0% 100.0%

Table 2B
Chi-Square Tests
Value df Asymp. Sig. (2-sided) Exact Sig. (2-sided) Exact Sig. (1-sided)
Pearson Chi-Square 1.400a 1 .237
Continuity Correctionb .921 1 .337
Likelihood Ratio 1.406 1 .236
Fisher's Exact Test .270 .169
Linear-by-Linear Association 1.383 1 .240
N of Valid Cases 82

Second Bi-Variate Analysis
Table 3A
does reducing disorder reduce crime * Identity Crosstabulation
Identity
police Criminology Student Total
does reducing disorder reduce crime Yes Count 33 28 61
% within Identity 84.6% 65.1% 74.4%
No Count 6 15 21
% within Identity 15.4% 34.9% 25.6%
Total Count 39 43 82
% within Identity 100.0% 100.0% 100.0%

Table 3B
Chi-Square Tests
Value df Asymp. Sig. (2-sided) Exact Sig. (2-sided) Exact Sig. (1-sided)
Pearson Chi-Square 4.082a 1 .043
Continuity Correctionb 3.122 1 .077
Likelihood Ratio 4.200 1 .040
Fisher's Exact Test .075 .038
Linear-by-Linear Association 4.032 1 .045
N of Valid Cases 82

There were several limitations to my paper. The first was the limited number of participants. Having more participants would have made my tests stronger and possibly emphasized any differences of opinion between identities. Secondly, I was using the results of “are you familiar with the broken windows theory” as my test variable, however an interesting problem came up. While I knew that the a higher percentage of criminal justice students would answer yes, I thought that a lower percentage of police officers would answer no, giving me a statistically significant result to my multivariate analysis. As I collected the data, it turned out that 78% of my entire participant body was familiar with the broken windows theory, meaning that many more police officers were familiar with it than I had anticipated. This data made my running a multivariate test problematic, and instead I ran two bi-variate analyses. Thirdly, in doing further research, I should be more specific when asking about familiarity with the broken windows theory. I should use a scale rather than a yes or no question, and then hypothesize that criminal justice students would score higher on the scale than police officers, and thus be able to use that as a test variable in further logic analyses.

Implications
The results for the first cross-tabulation were not significant, chi-square was .237, and in order for it to be significant it must be less than .10. The value of chi-square for this test was 1.400, which is a relatively small number, showing that the independent variable results for identity were very similar to each other. This means that police and criminal justice students did not differ much in their opinion of if police implementation of the broken windows theory effectively reduces crime. 57.3% within identity answered yes: 64.1% of the police officers, 51.2% of the criminal justice students. Because neither percentages are 7 full percentage points away from the 57.3%, we can conclude that there is little variance in the results between identity. Looking at it from the negative perspective, 42.7% within identity answered no to the test question: 35.9% of the police officers, 48.8% of the criminal justice students. Again, both percentages are very close to the percent across the independent variable, showing in response to this test question that police officers and criminal justice students have very similar opinions. These test results show that my hypothesis that criminal justice students would be less likely than police officers to believe in the effectiveness of zero-tolerance policing was a null hypothesis.

The results for the second cross-tabulation were statistically significant with chi-square at .043, which is less than .10. The value of chi-square was 4.082, which is a relatively large number for this study. This means that police offices and criminal justice students differed significantly in their answers to the test question “does reducing disorder reduce crime”. 74.4% within identity answered yes: 84.6% of police officers and 65.1% of the criminal justice students. This means that many more police officers than criminal justice students think that reducing disorder reduces crime. That difference in opinion is statistically significant. Looking at it negatively, 25.6% within identity answered no: 15.4% of the police officers, and 34.9% of the criminal justice students. This means that many more criminal justice students than police officers thought that reducing disorder does not reduce crime. Although this question was not part of my hypothesis, it does show that while police and criminal justice students’ opinions on the police implementation of the broken windows theory are very similar, that they differ substantially when discussing disorder as a cause of crime.

Conclusion
This paper was written to examine if police officers and criminal justice students differ significantly in their opinions of the broken windows theory and zero-tolerance policing. My first research question “Does the police implementation of the Broken Windows theory effectively reduce crime” found that most participants in both groups agreed that the answer is yes. This finding directly contradicts my hypothesis that criminal justice students would be less likely to say yes because they were more familiar with the broken windows theory. My test question was going to be “Are you familiar with the Broken Windows theory”, but most of the police officers as well as almost all of the students answered yes, they were familiar with it. This again shows that my hypothesis was not supported with the data that I collected. My second research question “Does reducing disorder reduce crime” showed the only statistically significant difference in my test analyses. Criminal justice students were almost twice a likely as police officers to say no to this answer. This answer does support part of my hypothesis, because I took into account that criminal justice students would have studied the Broken Windows theory in greater depth, be more familiar with it, and thus be less likely to agree with face-value questions about the theory. For further research, more popular police policies that come directly out of criminal theory need to be examined, rather than used because they appeal to the masses. The world of police policy is far different from that of criminological theory. The point where the two connect, in policies such as the zero-tolerance policy, should be the subject of continued evaluation to see how crime theories hold up or are invalidated when used pragmatically in the nuanced world of crime.

References
Gault, Martha and Silver, Eric. “Spuriousness or Mediation? Broken Windows according to Sampson and Raudenbush”. Journal of Criminal Justice, vol. 36, pp 240-243. 2008.

Harcourt, Bernard E. and Ludwig, Jens. “Broken Windows: New Evidence from New York City and a Five-City Social Experiment”. The University of Chicago Law Review, vol. 73, no. 1, pp 271-320. 2006.

Hinkle, Joshua C. and Weisburd, David. “The Irony of Broken Windows Policing: A micro-place study of the relationship between disorder, focused police crackdowns and the fear of crime”. Journal of Criminal Justice, vol. 36 pp 503-512. 2008.

Hyunseok, Jang; Hoover, Larry T, and Lawton, Brian A. “The Effect of Broken Windows Enforcement on Clearance Rates”. Journal of Criminal Justice, vol. 36, pp 529-538. 2008.

APPENDIX
Uni-variate Analysis
Table 1

Statistics

Frequency Table

Identity
Frequency Percent Valid Percent Cumulative Percent
Valid police 39 45.9 47.6 47.6
Criminology Student 43 50.6 52.4 100.0
Total 82 96.5 100.0
Missing System 3 3.5
Total 85 100.0

Familiarity
Frequency Percent Valid Percent Cumulative Percent
Valid Yes 64 75.3 78.0 78.0
No 18 21.2 22.0 100.0
Total 82 96.5 100.0
Missing System 3 3.5
Total 85 100.0

Are you both a police officer and a Criminology Student
Frequency Percent Valid Percent Cumulative Percent
Valid Yes 26 30.6 31.7 31.7
No 56 65.9 68.3 100.0
Total 82 96.5 100.0
Missing System 3 3.5
Total 85 100.0

does the broken windows theory suggest a direct link between disorder and crime
Frequency Percent Valid Percent Cumulative Percent
Valid Yes 65 76.5 79.3 79.3
No 17 20.0 20.7 100.0
Total 82 96.5 100.0
Missing System 3 3.5
Total 85 100.0

does reducing disorder reduce crime
Frequency Percent Valid Percent Cumulative Percent
Valid Yes 61 71.8 74.4 74.4
No 21 24.7 25.6 100.0
Total 82 96.5 100.0
Missing System 3 3.5
Total 85 100.0

is the broken windows theory behind zero tolerance police policies
Frequency Percent Valid Percent Cumulative Percent
Valid Yes 39 45.9 47.6 47.6
No 43 50.6 52.4 100.0
Total 82 96.5 100.0
Missing System 3 3.5
Total 85 100.0

does police implementation of the broken windows theory effectively reduce crime
Frequency Percent Valid Percent Cumulative Percent
Valid Yes 47 55.3 57.3 57.3
No 35 41.2 42.7 100.0
Total 82 96.5 100.0
Missing System 3 3.5
Total 85 100.0

do zero-tolerance policing policies effectively reduce crime
Frequency Percent Valid Percent Cumulative Percent
Valid Yes 54 63.5 65.9 65.9
No 28 32.9 34.1 100.0
Total 82 96.5 100.0
Missing System 3 3.5
Total 85 100.0