What Works in Policing?
“Standard Model” Policing Tactics
Resources
What is the “Standard Model” of Policing?
- What Can Police Do to Reduce Crime, Disorder, and Fear? (David Weisburd and John E. Eck, Annals of the American Academy of Political and Social Science)
- The Evolving Strategy of Policing (George L. Kelling and Mark H. Moore, Perspectives on Policing, Harvard Executive Session)
What is the Evidence on the “Standard Model” of Policing?
- Fairness and Effectiveness in Policing: The Evidence (National Research Council)
- Knowing What is Effective (and Not) in Policing (Center for Problem-Oriented Policing)
- The Kansas City Preventive Patrol Experiment: Summary Report (George Kelling et al., Police Foundation)
- Calling the Police: Citizen Reporting of Serious Crime (William Spelman and Dale Brown, National Institute of Justice)
Arrests for Domestic Violence
- Practical Implications of Current Domestic Violence Research: For Law Enforcement, Prosecutors and Judges (Andrew R. Klein, National Institute of Justice)
- The Effects of Arrest on Intimate Partner Violence: New Evidence From the Spouse Assault Replication Program (Christopher Maxwell, Joel H. Garner, and Jeffrey A. Fagan, National Institute of Justice Research in Brief)
- States with Mandatory Arrest Provisions (National Institute of Justice)

Seattle Police officer image courtesy of Flickr user Seattle Municipal Archives and used under a Creative Commons license.
What is the “Standard Model” of Policing?
“Standard model” strategies are often seen as traditional police approaches to dealing with crime that developed largely during the reform or professional era beginning around the 1930s (see Kelling & Moore, 1988). While these tactics are now 60 or more years old, they drive much of current police activity. They are seen as the “standard model” for a reason.
Weisburd and Eck (2004: 49) focused on five aspects of the “standard model” although we focus on only the first three on this page because the evidence for these three is the weakest. We cover the fourth and fifth aspects on separate pages under the “What do we need to know more about?” category:
1. Random patrol across all parts of the community
2. Rapid response to emergency calls for service (i.e. 911 calls)
3. Generally applied intensive enforcement and arrest policies.
4. Generalized investigations of crime
5. Increasing the size of police agencies
What is the Evidence on the “Standard Model” of Policing?
“Standard model” strategies are listed under “What doesn’t work?” on our Review of the Research Evidence. We summarize the evidence base below, but note, as Sherman (2013) argues, that more research is needed that compares these strategies directly to more effective approaches.
Random preventive patrol (or random beat patrol) has shown little or no evidence of effectiveness as a crime fighting tool for police. The most influential and well-known study in this area was the Kansas City preventive patrol experiment conducted by Kelling and colleagues (1974). They found no evidence that changes in the amount of preventive patrol across beats had a significant impact on reported crime, reported victimization, or levels of citizen satisfaction. The study presented no evidence that routine preventive patrol is an effective deterrent or way to improve police efficiency and effectiveness.
The finding that police randomly patrolling beats is not an effective crime deterrent makes sense based on the literature on hot spots policing. Hot spots policing is an effective strategy in part because it takes advantage of the fact that crime is strongly concentrated in a small number of places across cities. Since crime is very concentrated across cities, it makes little sense from an effectiveness and efficiency standpoint to respond with a strategy relying on the random distribution of police resources across large geographic areas.
A second standard policing tactic that appears to have little impact on crime is rapid response to 911 calls. Rapid response can sometimes lead to the apprehension of suspects, particularly calls for a “hot” robbery or burglary. However, there is no evidence that rapid response to most calls increases apprehension rates or decreases crime (Spelman & Brown, 1984). The problem is that citizens frequently wait too long after an incident occurs for rapid response to be of much assistance. We do not argue here that police should ignore 911 calls, but instead that they should not expect crime control gains to come simply by decreasing response times to the vast majority of calls.
Finally, more across the board increases in arrests are not particularly effective in reducing crime. As Sherman and Eck (2002: 310) note, “the evidence in support of the reactive arrest hypothesis is remarkably unencouraging at both the community and individual levels of analysis.” It is difficult to reach strong conclusions on the effectiveness of arrest as a crime-control strategy because of mixed evidence on interventions that rely primarily on arrest and the fact that many interventions that include increases in arrest also feature a number of other facets, and disentangling the impacts of various factors can be difficult. We see little reason to believe that more broad-based reactive arrest policies will be very effective in reducing crime, and instead we argue for greater focus, either on high risk offenders, high risk places, or both.
Arrests for Domestic Violence
Arrests for cases of misdemeanor domestic violence are a subset of general arrest policies, but one that has been more extensively studied than almost any other policing tactic. Unfortunately for police practitioners, the evidence base on the benefits of arresting offenders in domestic violence cases is fairly mixed. The initial Minneapolis experiment showed beneficial results in terms of reduced recidivism when comparing arrest to mediation or separation (Sherman & Berk, 1984), but results from the replication studies were more varied and depended in part on the employment status of offenders (e.g. see Pate & Hamilton, 1992; Sherman et al., 1992), a problematic criterion for officers to rely on in street-level decision making. More recently, Maxwell and colleagues (2001), in analysis of data from all of the domestic violence arrest experiments, found an overall deterrent impact of arrest, particularly when examining victim accounts of subsequent abuse. Cho and Wilke (2010) also find some evidence of a deterrent effect of arrest. Thus, it may be the case that arrest has a modest deterrent effect on reducing subsequent offending, at least for certain types of offenders.
Following the Minneapolis study, increased pressure from women’s groups, court cases where departments were sued for negligence in responding to domestic violence, and a federal push in the Violence Against Women Act led to more than half of states (including Washington) adopting mandatory arrest laws, leaving officers little discretion in the decision to arrest in cases where there is probable cause to believe misdemeanor domestic violence has occurred. Because of the unclear implications of the research evidence in this area, it is not clear that mandatory arrest laws or presumptive arrest laws will have a significant impact on crime rates. Maxwell et al. (2001) note that even though their findings overall show that arrests tend to reduce subsequent abuse, many offenders never have a repeat domestic violence incident, regardless of whether they are arrested or not and a small group of offenders are chronic abusers, regardless of how their case is handled. Thus, mandatory arrest may not be the most efficient use of police resources to deal with domestic violence.
Individual Studies from the Evidence-Based Policing Matrix
Not all individual-based studies are domestic violence studies. The 9 domestic violence studies are all in the specific, reactive part of the individual slab of the Matrix (the part where the greatest number of individual-level studies are concentrated).
Evidence-Based Policing Matrix
The Evidence-Based Policing Matrix includes only one standard model evaluation not related to domestic violence (Kelling et al., 1974). The Matrix does include a number of evaluations on arrest for domestic violence:
Domestic Violence Studies from the Evidence-Based Policing Matrix:
Author (Year)
|
Intervention
|
||||
Berk et al. (1992) | Arrest of spousal abusers reduced recidivism |
![]() |
VR
|
F
|
R
|
Cho & Wilke (2010) | Arrest of intimate partner violence perpetrators associated with reduced victimization compared to non-arrest. |
![]() |
M
|
G
|
R
|
Dunford (1990) | Arrest warrant for domestic violence suspect reduced absent offender recidivism 50% |
![]() |
VR
|
F
|
R
|
Jolin et al. (1998) | Domestic violence unit designed to increase arrests and prosecutions of offenders and provide follow-up victim empowerment services associated with decline in victim-reported subsequent violence |
![]() |
VR
|
F
|
R
|
Sherman & Berk (1984) | Arrest condition for domestic violence associated with significantly less offender recidivism compared to separation and mediation. |
![]() |
R
|
F
|
R
|
Pate & Hamilton (1992) | Arrest for domestic violence had a deterrent effect for employed offenders, but increased recidivism among unemployed offenders |
![]() |
R
|
F
|
R
|
Sherman et al. (1992) | Arrest for domestic violence had deterrent effect for married, employed, white high school graduates, but was criminogenic for unemployed, unmarried, black high school drop outs |
![]() |
R
|
F
|
R
|
Dunford (1992) | Arrest for domestic violence increased offense frequency at 12 months |
![]() |
VR
|
F
|
R
|
Hirschel et al. (1990) | Arrest for domestic violence increases official recidivism |
![]() |
VR
|
F
|
R
|
Result: = successful intervention;
= mixed results;
= nonsignificant finding; =
harmful intervention
Rigor: M = moderately rigorous; R = rigorous; VR = very rigorous
Y-axis: F = focused; G= general
Z-axis: R = reactive, P = proactive, HP = highly proactive