Thursday, October 20, 2016

Does a Decentralized Police Department Encourage Criminal Behavior?

 Joel Anguiano
(I will not be taking the second half)

Background
Some cities across the United States have experienced a dramatic spike in violent crime within the past year.  In the month of August alone, Chicago experienced 90 homicides and 384 shootings.  In Los Angeles, robberies and aggravated assaults went up.  In Dallas, during the January – June period, murder rates were up 40 percent over the same period in 2015.  Law enforcement officials have struggled to pinpoint the exact cause of these upticks.  Some posit, rather unsurprisingly, that this increase in crime is due to the increased public scrutiny that police officers have been subjected to in the past couple of years.  Not least among those who point to this “Ferguson Effect” as a possible explanation is FBI Director James Comey, who hypothesized that the increased scrutiny on police officers has made them less aggressive and has emboldened criminals to commit more violent acts.

Among the cities that face this increase in violent crime is Las Vegas.  However, many officials have been quick to point out that the rise in violent crime may have less to do with the “Ferguson Effect” and more with the recent decentralization of the police department’s specialized crime units, particularly its gang unit.  The decentralization was done to place detectives closer to its communities, arguing that it would allow them to respond quicker and more effectively during investigations.  Officials that defend the move also argue that decentralizing its specialized units puts its detectives in positions to strategize, share intelligence, and more directly collaborate with patrol officers.  This increased collaboration would drive crime rates down.  Those that blame the rise in crime on decentralization argue that a lack of specialized units breaks up the flow of information among its specialized detectives and diverts resources from issues that require special focus.  A lack of centralized units, they argue, ultimately undermines detectives’ efforts because they don’t develop the connections necessary to disrupt highly specialized criminal behavior.

Research Question:
Given the debate surrounding the decentralization issue, this analysis would seek to answer the question: to what extent do centralized special police units enhance their ability to curb violent crime?  Should robbery, sex assault, or gang units exist as cohesive units, separate from the broader police precincts/area commands? Or, should detectives be assigned to geographic areas (rather than specific crimes) and focus their energy and resources on all crimes with a less specialized focus where they will have more direct interaction with citizens and patrol officers?

Hypothesis:
Although decentralizing special crime units will put detectives in more direct contact with citizens and patrol officers and help them become more familiar with their geographic areas, ultimately the network will become too scattered to address these issues effectively.  Whereas centralized units allowed for more direct flow and sharing of information pertinent to specific crimes, where information from all geographic areas would converge in one place, the breakup of these police networks will scatter the detectives, effectively making information sharing more difficult than it had been.  Furthermore, with detectives focusing on all crimes, the collaboration that once existed becomes too broad and the intricate communication networks that existed become less effective. 

Data Collection & Methodology:
The data for this project would have to come internally from the police department.  Ideally we would have data for pre- and post-decentralization.  This data would be collected through a combination of employee databases and surveys.  Basic information for detectives would be:
·         What is your rank?
·         What academy graduating class do you belong to?
·         What area command did you last work in before being assigned to your unit?

Questions to ask for Pre-Decentralization data:

  •  Identify 2 people within your unit with whom you collaborated the most. 
  •    Identify 3 people within each area command with whom you collaborated the most. 
  •  Identify one area command you feel you partnered with the most.
  • For each specialized unit, identify how often you turn to it for information?

Questions to ask for Post-Decentralization data:

  • Identify 2 people within your area command with whom you collaborate the most.
  • Identify 5 people within all other area commands with whom you collaborate the most. 
  •   Of the detectives assigned to other area commands, identify who you turn to the most for information.
  • Of the detectives in your area command, identify the one you turn to for information.

After collecting pertinent information, I would use centrality measures such as betweenness, indegrees, and subgroup analysis such as k-core or clique analysis to determine who is connecting detectives across the different area commands (if any) and if any strong “units” have emerged.  If any strong leaders emerge within the different area commands, they could be identified and brought together periodically to collaborate and disseminate information to their respective area commands.
Seeing as how this data would be largely internal, and how historically police departments are wary of disseminating that data, this might be a project best left for internal use either for a contractor or analysts of the department to pick up. 

Conclusion:
Network analysis has been used in law enforcement as an effective tool in identifying criminal networks.  Now is a good time to turn the reflectors inward and answer the question of whether an information flow was disrupted when decentralization of the Las Vegas police department occurred.  Although this analysis would provide insights into the network of the police department pre- and post-decentralization, the larger issue of whether this reconfiguration of the department contributed to the rise in crime would need to be analyzed in conjunction with other research and information. The network analysis would only be a part of the greater whole.  For example, crime trends would need to be looked at for a period that extends beyond pre- & post-decentralization.  Even then, if a correlation did exist, more information would still be necessary to explain causality.  Although this issue of rising crime is a complex one, a network analysis of the organizational design of the police department would provide a great window into how efficient the police department is in using its resources.
 


1 comment:

Christopher Tunnard said...

Nice job. At the beginning, I wondered if you were going to use SNA to support the decentralization efforts of the LVPD, as it wasn't clear from your Question ("...enhance their ability"?) However, your idea of comparing the pre and post collaboration efforts is a good one. What you need to do is connect that more clearly to an outcome--e.g. what network patterns would lead towards enhanced abilities? (I also appreciate the link to the criminal nets article.)