Watsonville City Council
September 22, 2026

VehicleManager -Audit Queries Report
No
1 2
3 4
5 6
7 8
9 10
11 12
13 14
15 16
17 18
19 20
21 22
23 24
25 26
27 28
29 30
31 Query Type
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection Report
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection Report
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Batch Plate Search Detection
Username
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
octavio. tortoleromartinez@associates. hsi .dhs.gov
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
Michael. [email protected]
Michael. [email protected]
Michael. [email protected]
mloriaux
[email protected]
[email protected]
[email protected]
[email protected]
Anthony. [email protected]
Michael. [email protected]
ICE staff and agents have access to search Vigilant's nationwide database. as shown in the above screenshot from public records shared by the Auburn, MA police department.

Is Warrantless, Dragnet Surveillance of Motorists Legal? In 2014, in Commonwealth v. Augustine, the Massachusetts Supreme Judicial Court held that police are required to obtain a search warrant to access cell site locat Declaration of Rights. Four years later, the Supreme Court applied similar protections nationwide in Carpenter v. United States. The Court reasoned that technolog) our public movement s, and to do so both in real time and retroactively, posed a significant threat to our Fourth Amendment rights.

LPRs are analogous to cellphone tracking. These Al-enabled cameras track motorists in real-time and historically, giving the government the means to track peopl information at issue in Augustine and Carpenter.

The Massachusetts Supreme Judicial Court recognized this in 2020, holding in Commonwealth v. McCarthy that "[w)ith enough cameras in enough locations, the h would invade a reasonable expectation of privacy and would constitute a search for constitutional purposes." Yet today, police in Massachusetts are subject to no license plate readers. In the absence of a state statute, police are engaged in mass surveillance of all drivers.

Flock's Data Sharing Model vs. the Massachusetts Shield Law

In August 2024, Massachusetts strengthened its Shield Law, which prohibits Massachusetts law enforcement from providing information or assistance to any othe investigations into reproductive healthcare or gender-affirming healthcare that is lawful in the Commonwealth. The Shield Law was designed to protect people fro access to gender-affirming care, ensuring that people who receive and provide protected healthcare that is lawful in Massachusetts can do so without fear of retri But Massachusetts law enforcement's use of Flock's nationwide data sharing undermines the effectiveness of the Shield Law. Out-of-state officers from thousands information about where and when people are driving in Massachusetts, and there is nothing stopping them from using that information to track the movements protected healthcare.

These concerns are not hypothetical. Earlier this year, police in Johnson County, Texas performed a nationwide search in Flock's database to find a woman who th€ LPR data from many states where abortion is legal and protected, including Massachusetts. We know about this investigation because the Texas officer entered "h Flock's database. Initially, Flock tried to dismiss criticism stemming from the negative press, suggesting that the woman's "family feared she was hurt" and that po subsequent reporting from 404 Media based on public records obtained by the Electronic Frontier Foundation show the police who made this search were pursuir abortion. As 404 Media reporters put it: "In documents created prior to the publication of our article, there is zero mention of concern about the woman's safety. l separate document about the Flock search a week after our article was published, in which they justify the search by saying they were concerned for her safety." Flock claims it protects people's privacy and legal rights by requiring all law enforcement officers to enter a "search reason" before accessing database results. Ace concerns in other states, if an out-of-state officer enters a reason that would violate a state law that protects access to reproductive and gender-affirming healthG LPR data would be excluded from the search results. But the Flock network audits obtained by the ACLU of Massachusetts demonstrate that the company's measu The network audits show police frequently enter vague, tautological terms like "investigation" or "susp" instead of information about the substance of the investig accessing search data unless they provide a substantive reason, police could evade the system's guardrails. As Flock's technology attracts more negative media att its data sharing practices, officers in states that criminalize abortion and gender-affirming healthcare could simply decline to provide specific details about their in death" when investigating an abortion case. The scale of the nationwide searches makes case-by-case oversight impossible. For example, a Flock network audit pr 2025 public records request documents over 450,000 searches of the nationwide database in just a 30-day period in the spring of 2025.

Name Org Name Total Networks Searched Total Devices Searched Time Frame License Plate Reason Case# Fi Clinton IA PD 6193 83023 PM UTC investigation le
Clinton IA PD 6193 83023 PM UTC investigation le
Belpre OH PD 6249 83831 PM UTC Drugs
Sugar Land TX PD 6353 83740 PM UTC Jugging
James City County VA PD 316 6606 PM UTC Abduction Abduction R Coweta County GA SO 6277 84028 PM UTC 37 G
Lynn MA PD 175 3669 PM UTC 94 Invest A
Moweaqua IL PD 6319 85005 PM UTC Drugs
Lynn MA PD 175 3669 PM UTC 94 Invest A
AZ - Department of Public Safety 6227 83470 PM UTC INFO A AZ - Department of Public Safety 6227 83470 PM UTC INFO A Whitfield County GA SO 5726 79091 PM UTC 99 G
Tewksbury MA PD 296 4625 PM UTC Raidor A
Colorado State Patrol 6389 85112 PM UTC CMATT C
Eatonton GA PD 5721 79042 PM UTC stolen G
Gilbert AZ PD 5887 79497 PM UTC 509
Oakland County Ml SO 6334 83538 PM UTC threats rv
Tewksbury MA PD 296 4625 PM UTC Raidor A
Fayette County TN SO 6294 83606 PM UTC investigation G
NYS Crime Analysis Center Network 783 14575 PM UTC HV-25-1767 Lawrence District Attorney PA 1254 16015 PM UTC Investigation N Whitfield County GA SO 5726 79090 PM UTC 99 G
Columbus OH PD 6485 87451 AM UTC 45 A
Houston TX PD 6749 88405 AM UTC INV Tl
The above screenshot shows an actual Massachusetts police department's Flock network audit, with personally identifiable information redacted.

Flock's purported solution to comply with state laws that restrict the sharing of information related to investigations of protected healthcare services is no solutior effectively unrestricted access to Flock's nationwide database, including data about Massachusetts residents and visitors.

Flock's Contract Problem
Police departments that use Flock and seek to protect the rights of Massachusetts residents from out-of-state or federal scrutiny must turn off information sharin~ they collect. But even when departments opt to restrict sharing in Flock's system settings, they may not actually be protecting people's privacy. According to publi template user agreement many law enforcement agencies sign gives Flock "a non-exclusive, worldwide, perpetual, royalty-free right and license (during and after

Aggregated Data to improve and enhance the Services and for other development, diagnostic and corrective purposes, other Flock offerings, and crime preventio any Footage) to enable law enforcement monitoring against law enforcement hotlists as welt as provide Footage search access to law enforcement for investigative pu f i' o c k s a f e t y
systems and technologies (including, without limitation, information concerning Agency Data and data derived therefrom). For the sake of clarity, Aggregated Data is compiled anonymous data which has been stripped of any personal identifying information. Agency acknowledges that Flock will be compiling anonymized and/or aggregated data based on Agency Data input into the Services (the "Aggregated Data"), Agency hereby grants Flock u nonexclusive, worldwide, perpetual, royalty-free right and license (during and after the Service Tenn hereof) to (i) use and distribute such Aggregated Data to improve and enhance the Services and for other development, diagnostic and corrective purposes, other Flock offerings, and crime prevention efforts, and (ii) disclose tho Agency Data {both inclusive of any Footage) to enable law enforcement monitoring against law enforcement hotlists as well as provide Footage search access to law enforcement for investigative purposes only. No rights or licenses are granted except as expressly set forth herein. Flock shall not sell Agency Data or Aggregated Data.

The above screenshot shows an excerpt of Flock's template user agreement, which grants Flock the license to use, distribute, and disclose data collected by a given agency.

What this means in practice is troubling: even when a police department chooses in Flock's application to restrict data access to its own officers, the template agre department's data both to law enforcement nationwide and federal agencies for "investigative purposes." Individual police departments can, and sometimes have, addressed this problem by amending Flock's contract language. For example, when the Boston Police De1 earlier this year, the department elected in its sharing settings not to share data outside the department - but the BPD also appears to have rewritten Flock's tern many police departments have agreed to, the Boston Police Department's agreement does not give Flock the right to disclose its LPR data.

4.2 Customer Generated Data. Flock may provide Customer with the opportunity to post, upload, display, publish, distribute, transmit, broadcast, or otherwise make available, messages, text, illustrations, files, images, graphics, photos, comments, sounds, music, videos, infonnation, content, ratings, reviews, data, questions, suggestions, or other infonnation or materials produced by Customer ("Customer Ge11erated Data'} Customer shall retain whatever legally cognizable right, title, and interest in Customer Generated Data. Customer understands and acknowledges that Flock has no obligation to monitor or enforce Customer's intellectual property rights of Customer Generated Data. Customer grants Flock a non-exclusive, irrevocable, worldwide, royalty-free, license to use the Customer Generated Data for the purpose of providing Flock Services. Flock does not own and shall not sell Customer Generated Data The above screenshot shows an excerpt of the Boston Police Department"s user agreement with Flock, which notably differs from Flock's standard template user agreement in that the Boston Police Department's agreement does not give Flock the right to disclose agency data.

But the vast majority of police departments in Massachusetts lack the legal resources and sophistication of Boston's department, putting them at a serious disadv, from Andreessen Horowitz, has been engaged in an aggressive marketing campaign to win more contracts and secure more taxpayer dollars. This campaign inclu a free TopGolf event and an opportunity for agencies to participate in Flock's "Project Prove It[,]" in which Flock installs LPR cameras on the roads and the agency< offer to work with local Massachusetts departments and other municipal leaders to identify and apply for grant funding for Flock LPR systems.

From: flocksafety.com>
Sent on: Wednesday, May 21, 2025 1:07:20 PM
To: @dovermapd.com>
Subject: [Please Read] Flock Safety TopGolf Event June 4th Caution: This email originated from outside of the Town of Dover Police mail system. Do not click links or open atta and know the content is safe.

We are planning to host a Flock meetup on June 4th from 5-7pm at TopGolf in Canton and I wanted to extend you an inv We are to bring together some of our local customers and prospective customers to take a few swings, have some drinks, Flock for keeping our communities safe.

Space is limited, so if you are interested please REGISTER HERE to secure your spot. Hope to see you there! Best,
Peter
Flock draws police departments into its nationwide data sharing network by offering officers and other staff the opportunity to participate in free events, like at TopGolf.

' '
@flocksafety.com>
8:59-0400
To: police.waltham.ma.us>
Subject: Project Prove It - Waltham, MA
Hello Sgt.

I wanted to follow-up our discussion with the information about the "Project Prove It," which allows for you to Implement Flock Technology with an Out-Clause built in. The details are as follows:

A PPJ allows for you to emplace flock LPR on a trial basis, with an out clause to remove the technology.

The two things required for a PPJ are a signed Deployment Plan (which ls what Sgt. Dean is working on now) and a signed contract. That contract has a build in Out-Clause that reads:

PROJECT PROVE IT: Customer will have a "45" day opt-out period ("Opt-Out Period") after Implementation of the first Flock Hardware to terminate this Agreement without penalty or fees.

After the Opt-Out Period, Customer may not terminate the Agreement, and Customer will pay any lnvolce(s) for the remainder of the Term, Net 30.

How this would work:

1. Waltham Signs the PPI Contract

2. It will take 4-6 weeks to Install LPR

3. Once the Installation takes place, the 45-day PPI starts

4. Anytime with In those 45 days, you can cancel the contract for 2 reasons 4A. PPI Contract - You can cancel at any time In those 45 Days 4B. You can cancel if you do not getthe funds appropriated (All Contracts have this option.

5. Once the 45 day period ends, and you want to continue with Flock, you will be Invoiced and have 30 days to Pay the Invoice.

Please let me know if you have any questions. This is a great way to quickly start a LPR program with Flock without any risk, as we will take the technology back if you cancel the contract.

I lool< forward to hearing from you!
The above screenshot of an email from a Flock salesperson to the Waltham police department shows Flock's typical sales tactics used to convince even departments that are hesitant to invest in its LPR technology to adopt it through Flock's •project Prove 11.•

From:
~
~flock.safety.com>
Sent: 18, 2025 3:28 PM
To:

Subject: Flock Safety in Framingham: Let's find funding together! CAUTION: This email originated from outside your organization. Do not click links or open attachments unless you recognize and verify the sender and know the content is safe.

Hello Deputy Chief -
I just wanted to reach out because I know budgets are being allocated and I wanted to offer my help in any way I can to get the Framingham PD the funding you need. The best way that we have found to get more funding from creative places for Flock, is to bring in key stakeholders from the City/Town and let them know what Flock can do to help enhance public safety. We have had conversations with many different stakeholders that have ended up providing more funding for the PD to get the technology they need, and I would love to discuss how we could do that in Framingham.

Would you be willing to try to get some key stakeholders from Framingham in a room to discuss Flock? I would gladly sponsor a lunch or dinner, bring lunch into the PD, or meet them at a place that is convenient for them.

Some of the stakeholders we have found that can help find/allocate funding are:
• City/Town Manager
• City Council Members
• District Attorneys
• School Departments/School SAU's
• IT Departments
• DPW
• Parks and Recreation
• Housing Authority
• Commercial Businesses
If you have 10-15 minutes for a phone conversation, I would love to explain how setting up a lunch/dinner with these stakeholders could help you obtain more funding for Flock. Please let me know if we could discuss this over the next week. The sooner we include these stakeholders, the more inclusive the project becomes! I look forward to connecting with you!
Flock sales staff even offer to help local police departments obtain funding to contract with Flock.

Faced with this onslaught from a company with seemingly endless resources at its disposal, small- and medium-sized police departments across the country are fa Individual action by sophisticated departments like Boston's isn't enough. Residents in communities with less well-resourced departments deserve protection too.

How Lawmakers and Law Enforcement Can Address Issues with LPRs The scope and severity of this dragnet surveillance make clear that both legislative action and individual police department changes are needed to achieve statew First, state lawmakers must pass comprehensive LPR legislation. H.3755 strikes the right balance, protecting civil rights and civil liberties while allowing police to enforcement from using LPRs to monitor individuals based on First Amendment-protected activities, including political protests or religious gatherings. It requires tied to a specific, ongoing criminal investigation supported by articulable facts. Access to LPR data would be limited to law enforcement for investigative purposes The bill also prohibits buying, selling, renting, or sharing LPR and related location data unless required by judicial process, and bars law enforcement from accessir search warrant. These provisions prevent the accumulation and sharing of data on millions of people not suspected of any wrongdoing, while allowing law enforcE legislation would ensure consistent protections apply for all Massachusetts residents.

Second, police departments must take immediate action. Individual departments should stop voluntarily sharing data with out-of-state and federal agencies. Thi department retains full control of all data they collect. Finally, police departments must adopt internal policies requiring that every Flock network search be justifie clarifying to all officers and staff that a non-descriptive entry like "investigation" will not suffice.

Take Action
Right now, your movements on Massachusetts roads are being tracked and shared with thousands of agencies nationwide. No warrant, no suspicion, no safeguarc Join us in calling on the Massachusetts Legislature to pass common sense checks and balances on police use of license plate readers, to ensure that you retain ba! Emai l your state legislators today to show your support for H.3755, An Act Establishing Driver Privacy Protections. Page 638 of 716 -- The Harvard Crimson

ADVERTISEMENT
NEWS
Cambridge Ends Contract With License Plate Reader
Amid Concerns Over 'Breach of Trust'
The view from Central Square down Massachusetts Avenue. I By Pavan V. Thakkar By Matan H. Josephy and Laurel M. Shugart, Crimson Staff Writers December 9, 2025
Cambridge will end its contract with security technology company Flock Safety after its technicians installed two new cameras less than a month after the Cambridge City Council suspended their use due to privacy concerns.

The city contracted with Flock after the Council voted in February to purchase automated license plate readers. But councilors reversed course in October and ordered Flock's cameras be temporarily shut down amid concerns that data collected by the company could be shared with the federal agencies like Customs and Border Protection.

Despite the Council order, Flock officials informed city leaders that it had put up two new cameras in late November, city spokesperson Jeremy H. Warnick confirmed. The new cameras were erected "without the City's awareness," Warnick wrote, because of an "outstanding work order that should have been canceled when the City originally deactivated the cameras and account." "Due to this material breach of our trust and the agreement, the City is terminating its contract with Flock Safety," Warnick added. A spokesperson for Flock did not respond to multiple requests for comment.

Kevin Kane, a lobbyist for Flock, said at a Tuesday hearing of the City Council's Public Safety Committee that the cameras have been removed and none of their data was accessed.

Still, Councilors and residents have repeatedly called the security of data collected by Flock's cameras into question. Data collected by Flock's 90,000 cameras nationwide enters one database, which can be searched by federal and local law enforcement agencies.

Councilors raised red flags over reports that data collected by Flock cameras could be shared with federal agencies - including Immigration and Customs Enforcement- before deciding to suspend their use in October.

ADVERTISEMENT

Get The Crimson in your inbox.
'--E_n_te_r_y_o_u_r _em_a_n _ _ ___ ___ ____,11 L.. _ _ s_i_g_n_u_p__, Kane said that Flock's database requires law enforcement agencies to provide a justification when they search the database, which is then logged in a network audit. The system excludes data from municipalities with sanctuary ordinances if the search terms relate to immigration, according to Kane.

Gideon Epstein, policy counsel for the American Civil Liberties Union of Massachusetts, presented screenshots of a network audit that had been obtained through a public records request at Tuesday's meeting. Of 2.6 million searches, he said, only about 1,500 searches explicitly mentioned "immigration" or "ICE." Epstein said that this statistic was not comprehensive, as vague justifications like "investigation" or "wanted" could also be used to collect data for immigration enforcement without being flagged in Flock's system.

Councilors also raised concerns that Flock may be sharing data directly with federal agencies when deciding to suspend the use of the cameras in October, not just allowing the data to be accessed.

Flock operated "limited pilot" programs with Customs and Border Protection, all of which were paused in August, according to a statement from chief executive Garrett Langley at the time. Flock has also imposed "strict limitations on federal access," the company announced in October.

Still, councilors worried that any data sharing with federal agencies would violate the city's sanctuary ordinance, which prevents the Cambridge Police Department from assisting ICE operations or voluntarily providing information on a resident's citizenship status.
"At a point in time when the collaboration between local governments and federal law enforcement has been severely strained, the city recognizes and Page 641 of 716 appreciates the significance of these concerns," Warnick wrote.

ADVERTISEMENT
-Staff writer Matan H. J osephy can be reached
[email protected]. Follow him on X @matanjosephy. -Staff writer Laurel M. Shugart can be reached at
[email protected]. Follow them on X @laurelmshugart or on Threads @laurel.shugart.

Want to keep up with breaking news? Subscribe to our email newsletter Have a tip for The Crimson? Share it confidentially
( ~ Share )
SEE MORE ON: CITY POLITICS, CAMBRIDGE CITY COUNCIL, CAMBRIDGE POLICE, METRO,

IMMIGRATION, FRONT BOTTOM FEATURE
More in News
New IOP Director Plans to
Expand Diminished Senior
HGSE Partners With
Hamilton College on Early
SEAS Pilots Client-Facing
Capstone Option for

(!Congre1iii of tbe mtntteb $)tate1i
The Honorable Andrew N. Ferguson
Federal Trade Commission
600 Pennsylvania Avenue, NW
Washington, D.C. 20580
Dear Chair Ferguson:

Wasbtngton, 11BQI: 20515
November 3, 2025
We write to request that the Federal Trade Commission (FTC) investigate Flock Safety ("Flock"), the license plate surveillance camera company, and, where appropriate, hold the company responsible for its negligent cybersecurity practices. Flock's surveillance data can reveal Americans' movements over time, including trips to doctors and therapists, support group meetings for alcohol or drug addiction, as well as places of worship and protests. By not requiring industry-standard multi-factor authentication (MFA) to secure law enforcement accounts, Flock has needlessly exposed Americans' sensitive personal data captured by the company's surveillance cameras to theft by hackers and foreign spies, and unauthorized access through multiple documented instances of unauthorized password sharing.

Flock operates the largest network of surveillance cameras in the United States, reportedly contracting with over 5,000 police departments, 1,000 businesses, and numerous homeowners associations across 49 states. When a vehicle passes by a Flock camera, Flock records plate information, vehicle characteristics, and when and where the vehicle was spotted. Flock's network of surveillance cameras generate and store billions of vehicle scans each month. Flock repo1tedly enables law enforcement to search through this data not just by plate number, but also by make, model, and even bumper stickers.

In addition to being able to search data generated by cameras their agencies pay for, Flock's law enforcement customers can also search data collected by other Flock customers' cameras. Flock customers can allow access by specific law enforcement agencies and can also enroll into the "National Lookup Tool" which allows their data to be searched by all other enrolled Flock customers. Flock informed Congress this August that approximately 75% of Flock's law enforcement customers have opted into the National Lookup Tool, which is likely due to Flock restricting access to nationwide searches to agencies that have agreed to share their data. But because most law enforcement agency customers have access to this tool, hackers, foreign spies, or anyone else that gains access to a Flock customer account can misuse it to track Americans across the countly.

Flock has unnecessarily exposed Americans' sensitive personal data to theft by hackers and foreign spies. While Flock offers support for MFA, a widely recognized cybersecurity best practice, Flock does not require it, which the company confirmed to Congress in October. Moreover, Flock continues to support insecure methods of MF A, such as sending a numeric code to a phone by text message, which is vulnerable to interception and phishing. Flock does not natively support phishing-resistant MF A, which the Cybersecurity and Infrastructure Security Agency calls "the gold standard method of [MF A]." Phishing-resistant MF A is required of federal agencies and mandated by both the Federal Communications Commission and the FTC

Flock's failure to natively support and require phishing-resistant MFA means that if hackers or foreign spies steal, phish, or otherwise learn a law enforcement officer's Flock password, they can gain access to lawenforcement-only areas of Flock's website and search the billions of photos of Americans' license plates collected by taxpayer-funded cameras across the country. This threat is not theoretical. A search by Congressional staff of a public tool operated by the cybersecurity company Hudson Rock documenting accounts compromised by a form of malware known as an "infostealer" reveals that passwords for at least 35 Flock customer accounts have been stolen. Security researcher Benn Jordan also recently provided our offices with a screenshot from a Russian-language cybercrime forum in which Flock accounts appear to be offered for sale.

By not requiring MF A, Flock has also enabled unauthorized access through law enforcement officers sharing their Flock passwords. If Flock required MF A, law enforcement users of Flock would be required to approve each login to their account from a new computer. On August 11, 2025, Unraveled Press revealed that a Drug Enforcement Administration officer used the Flock account of a local police detective in Palos Heights, Illinois, to conduct several searches. Public records obtained by the press revealed that the detective confirmed that it was a common practice for members of a DEA task force that he was part of to use his login credentials to conduct Flock searches. That numerous federal agents were able to access Flock's systems using passwords belonging to other users without being detected or blocked raises serious questions about the effectiveness of Flock's cybersecurity defenses. A July 9, 2025, memo by the Palos Heights Deputy Chief of Police about the incident, which was released to the press, revealed that the department had not enabled the MF A option until after the password sharing was first identified by a reporter at 404 Media.

Flock's failure to require MFA is likely an unfair business practice prohibited by Section 5 of the FTC Act. In at least four cases - against Uber, Chegg, Drizly, and Blackbaud- the FTC held that failure to require MFA is an unfair business practice. Flock's negligent cybersecurity practices may also violate state laws related to data protection, consumer privacy or the use of automatic license plate reader (ALPR) systems; for example, California Civil Code§ 1798.90.53 requires entities operating ALPR systems to "maintain reasonable security procedures and practices, including operational, administrative, technical, and physical safeguards, to protect ALPR information from unauthorized access, destruction, use, modification, or disclosure." Flock has received vast sums of taxpayer money to build a national surveillance network. But Flock's cavalier attitude towards cybersecurity needlessly exposes Americans to the threat of hackers and foreign spies tapping this data. Accordingly, we urge the FTC to hold Flock accountable for its negligent cybersecurity practices. ~~p.,._
Ron Wyden
United States Senator
Sincerely,

More license plate readers are coming to Bay
Area, but do they work?
By Lisa Fernandez and Brooks jarosz
KTVU FOX 2 Piedmont Updated October 19, 2023 9:10 AM PDTPublished September 28, 2023 5:12 AM PDT License plate readers may give false sense of security when it comes to crime As cities around the Bay Area are clamoring for ideas on how to combat crime and curb carjackings, more municipalities are turning to cameras for help. KTVU dug into the data to see if this technology really works.

PIEDMONT, Calif. - Cities around the Bay Area are clamoring for ideas on how to combat crime and curb carjackings, leading more municipalities to turn to licenseplate-reading cameras for help.

Oakland, San Francisco, San Jose, Berkeley and San Leandro are just the latest poised to buy hundreds of automated license plate readers and security cameras.

They would be installed on police cars and street poles in the hopes of deterring criminal activity and catching perpetrators in the act of stealing cars and committing other crimes.

But do these cameras even work?
If you ask Piedmont Police Chief Jeremy Bowers, the answer is an unequivocal yes.

Bowers attributes the arrest of 248 people - about 27 people a year - directly to the help of the automated license plate readers that Piedmont installed in 2014. In that same time period, 375 stolen cars were recovered, also with the help of the readers.

The 39 cameras snap pictures of license plates, similar to the FasTrak cameras used on the region's bridges. The cameras only register a "hit" when a car is reported stolen or on some type of police wanted list.

Roughly 24% of the city's crimes over three years - beyond stolen cars - are solved through the automated license plate readers, he told KTVU.

Bowers said Piedmont's license plate reader system works because of several complementary components: The city has one fixed surveillance camera, which captures movements of people in the city and works in conjunction with the license plate readers. Officers have the ability to respond to stolen car hits in real time. It also helps that Piedmont is a tiny, affluent city of 11,000 residents that hasn't had a homicide since July 1999.

Piedmont is blessed with these fortunes, Bowers acknowledged, which other, larger cities, like neighboring Oakland, are not.
"We're small, but we're fully staffed," Bowers said. "The volume is manageable where we can drive resources to ... locate that vehicle."

Bowers is also astute enough to know that these readers aren't a panacea, but a helpful aid.
"They are a tool," he said. "And what we found in Piedmont is that they're an effective tool. We know that sometimes we're looking for that needle in a haystack so that's exactly what the ALPR [automated license plate readers] technology is there to do - to be a force multiplier, to help us identify those needles in a haystack and deal with them."
And, like other cities, Piedmont police won't chase the driver through residential neighborhoods, but if that driver speeds out of Piedmont into Oakland for example, the chief considers that a win for his city.

Data is mixed on plate-reading technology
Piedmont's enthusiasm for the license plate readers is inspiring cities to want the same technology, as well.

But the data showing whether the technology works en masse and in larger, more crime-ridden cities isn't so clear.

For instance, in the pro-camp, a 2011 study conducted by the Police Executive Research Firm shows that the license plate readers "considerably enhanced" recovering stolen vehicles and arresting car thieves in Arizona.

In the mixed results camp, a 2016 study by the National Policing Institute found that the camera can work, but it all depends on how the technology is implemented.

And in the doesn't-really-do-much camp, a 2010 study conducted by George Mason University researchers indicate that the license plate reader use led to increased stolen vehicle recoveries but not arrests. And the same study noted no significant improvements in clearances for auto theft cases citywide, but the clearance rate did improve in areas with high concentrations of the cameras.

Automated license plate readers, known by their acronym of ALPR, are cameras that scan license plates and take pictures of the rear plate and the rear part of the car.

They can be attached to patrol cars and move around the city. Or, they can be stationary, if they are fixed to a light pole or traffic light. Piedmont has both.

The technology can only determine if the plate is stolen or wanted; the cameras do not take pictures of people's faces.

Various cities have various policies on retaining the data. And there are laws in California regulating the use and storage of the information.

Piedmont purges its ALPR data every 60 days.

Some cities use the technology to determine if a driver has parked in a spot too long, as well. Other cities, like Piedmont and Oakland, want to use the cameras as a crime-fighting tool.

The cost of the cameras
A company called Flock Safety makes a popular version of the cameras, which sell for about $2,500 to $3,000, with an additional one-time installation cost, as well as ongoing maintenance costs.

Piedmont's maintenance costs are about $100,000 a year, the chief said.

Flock Safety touts that police can "accelerate case clearance" by using its "intelligence solution."
However, to date, there is no agency that tracks whether the ALPR systems actually help clearance rates, which is calculated by dividing the numbers of crimes that are charged by the total number of crimes recorded.

Piedmont police do not track clearance rates to know if the ALPR-arrests actually lead to charges and convictions.

Analyst says anecdotal success is not statistically significant Locally, Jonathan Hofer, a policy analyst with self-described "libertarian-ish" leanings, conducted his own analysis of Piedmont by looking at data before the city bought the license plate readers and afterward.

His conclusions?
Piedmont's successes are anecdotal but not statistically significant. "What we know right now is the results are pretty poor," Hofer said. "They're making a negligible difference."
Specifically, Hofer found that the license plate hits-to-investigative leads are low - less than .3%.

His data analysis also shows there is no real deterrent effect produced by having the cameras.

For example, the number of stolen cars dropped initially after the installation of the technology - from about 60 a year to 35. But since 2019, the number of stolen cars has been climbing back up to the 60-a-year-level.

Hofer isn't necessarily against the technology.

But he wants the public to be aware of the costs - both financially and in terms of civil liberties - before signing onto any program.

He said if the license plate technology costs $100,000 and the arrests, conviction rates and deterrence are negligible, maybe that money could better be used on a human detective, who might get the same, if not better results. "You have to factor in installation costs and maintenance costs," said Hofer. "This stuff's not cheap, and so we have to ask ourselves do we want to divest our police resources into tools that may or may not help."
Jonathan Hofer is the brother of Brian Hofer, chair of Oakland's Privacy Commission and Secure Justice, both organizations that strongly oppose mass surveillance.

Coincidentally, both Hofer brothers also won a lawsuit against the Contra Costa County Sheriff's Office because they were held up at gunpoint when an automated license plate reader in San Pablo pinged police that they were in a stolen rental car, when in fact, it had been returned, but not properly inputted into the system.

In Bay Area, license-plate-reader systems show minimal results KTVU surveyed several cities that have, or will have, license-plate-reader program. The survey found piecemeal data on whether they're effective or not.

The results were not conclusive, and if anything, showed a grab-bag of results for arrests and recovered cars, and a dearth of investigative leads.

For example:
• Oakland police have had automated license plate readers on 36 patrol cars since 2008 and acknowledged that they got no investigative leads from the license plate readers in 2022, according to the most recent annual report. In that same time period, 34 stolen cars were recovered but no arrests were made.
• Vallejo has been using the license plate readers since November 2021 and has recovered 113 cars and made 116 arrests directly as a result of the technology, the police department's website said.
• BART launched a pilot program of license plate readers - 7 mobile and 2 fixed - at the MacArthur BART parking garage in May. The goal was to catch people breaking into cars and stealing catalytic converters. To date, no arrests have been made. However, BART noted in its annual report that 288 parking citations have been handed out during the pilot phase.
• San Leandro's Public Safety camera program has only been in place since this year so there is no historical data. The police department is approved for 43 total cameras with installations nearing completion. • Berkeley has been approved to buy 52 fixed automated license plate readers at a startup cost of $250,000 and an annual fee of $175,000, but

the cameras have yet to be installed. Therefore, there is no data to support the system's effectiveness.
• San Jose has 11 mobile cameras since 2013 costing $200,000. Police have not tracked whether they can link any arrests or clearance rates to the license plate readers.
• In response to a public records request, Tiburon said it had no records to disclose regarding arrests or costs relating to its automated license plate readers. However, Tiburon did provide numbers of cars recovered since 2010, when the town installed the cameras: 57 cars were recovered in 13 years - roughly four cars were recovered annually.

Since February, Oakland police disabled all their automated license plate readers on their patrol cars because the department said the reporting policies were too onerous, especially in light of the aging technology.

In their annual report to the city, Oakland police also cited that several members of the public complained about the negative impacts to privacy and mass surveillance.

Oakland poised to install 300 license plate readers
Despite that, Oakland City Councilman Kevin Jenkins and Mayor Sheng Thao are pushing for more.

Earlier this month, Thao said she wants to use a $1.2 million state grant to install 300 automated license plate reader cameras across the city, an effort that Jenkins touted on social media earlier this summer.

To date, no formal plans have been submitted to Oakland's Privacy Commission to review.

In an interview with KTVU, Jenkins said he wants the automated license plate readers to be installed in crime hotspots so that they can help solve cases and "put these criminals behind bars who are creating chaos in our community." The George Mason University study does indicate that targeting the cameras in these spots yields better results.

But Oakland has an uphill battle in terms of curbing rampant car thefts in a city that's also grappling with nearly 100 homicides this year.

So far, there have been nearly 11,000 auto thefts in Oakland this year, averaging about 40 a day. That's the highest number in 15 years.

With crime spiking and car thefts skyrocketing, Jenkins said he wants to give police any extra tool in their toolbox, even if the evidence isn't there yet to prove that the license plate readers work on a large scale.

There is other evidence to consider, Jenkins said.
"I have evidence that what's going on right now, does not work," Jenkins said. "I have evidence that neighbors are scared. We have to be willing to try something different. We have to be willing to give technology a chance." He also knows that stolen cars won't take precedence over homicides, that Oakland police have a strict no-pursuit policy, and there's a shortage of officers to follow up on the stolen car hits and leads.

"But we have to walk and chew gum at the same time. We have to boost our technology and we also have to recruit," he said. "I don't think it's an either-or. You have to have technologies and you have to have competent law enforcement and do both at the same time."
Lisa Fernandez is a reporter for KTVU. Email Lisa at [email protected] or call her at 510-87 4-0139. Or follow her on Twitter@ljfernandez Brooks Jarosz is a reporter for KTVU. Email him at [email protected] and follow him on Facebook and Twitter @BrooksKTVU

I .

Journal of Experimental Criminology (2021) 18:725-746
https://doi.org/10.1007 /s11292-021-09473-y
Do license plate readers enhance the initial
and residual deterrent effects of police patrol?
A quasi-randomized test
Christopher S. Koper 1 G> • Cynthia Lum 1 • Xiaoyun Wu 1 • William Johnson 1 • Megan Stoltz1
Accepted: 10 May 2021 / Published online: 25 May 2021
©The Author(s) 2021, modified publication 2025
Abstract
Objectives To determine whether the use and display of license plate readers (LPRs) enhance the crime prevention effects of police patrol, particularly by increasing initial and residual deterrence from patrol presence.

Methods Crime hot spots in a large suburban jurisdiction were randomly assigned to receive intermittent patrols (15-30 min each) by officers with or without LPRs on their vehicles for 4.5 months. Data on 785 patrols at 33 hot spots were used to compare initial and residual outcomes of LPR and non-LPR patrols. Outcomes analyzed for each patrol visit via chi-square tests, survival analyses, and/or logistic regression included the following: the likelihood of a vehicle recovery or arrest; the likelihood of a new crime or disorder call in the location while the officer was present (measuring initial deterrence); time until a new crime or disorder call following the officer's departure (measuring residual deterrence); and first type of call (crime or disorder) following the officer's departure (measuring residual deterrence).

Results LPR use increased stolen vehicle recoveries but not arrests. LPR use did not reduce the likelihood of a new call in the hot spot while an officer was present, nor did it affect the timing or seriousness of the next call following a patrol.

Conclusions This study provides little clear evidence for the crime prevention efficacy of using LPRs in general patrol, which is a common practice in the USA.

Police and researchers should give more attention to testing the relative benefits of different LPR uses and modes of deployment.

Keywords License plate readers • Police patrol • Hot spots • Deterrence l:8J Christopher S. Koper
[email protected]
Extended author information available on the last page of the article ~ Springer

726 C. S. Koper et al.

Introduction
License plate recognition technology has spread rapidly in policing during recent decades.

Automated license plate readers (LPRs) are high-speed camera and information systems that read vehicle license plates in real-time, check them instantaneously against databases with license plates linked to vehicles and persons of interest (e.g., stolen vehicles, stolen license plates, and vehicles linked to wanted persons), and give automatic alerts when matches are made. LPRs also record information and images for scanned plates and vehicles, which police value for investigative purposes (e.g., Willis et al., 2018). LPRs can be installed in police cars for mobile surveillance or positioned at fixed locations.

Given these capabilities, roughly two-thirds of large police agencies in the USA and all agencies in England and Wales have acquired LPRs and use them to address a wide variety of crime and vehicle violations (Lum et al., 2019; PA Consulting Group, 2006). Yet, despite this large investment, research on the uses and impacts of LPRs is very limited. Prior studies of LPRs, conducted in the USA and abroad, have focused largely on demonstrating the accuracy and efficiency of the devices in scanning license plates and assessing their utility for increasing arrests, recoveries of stolen vehicles, and seizure of other contraband (e.g., Cohen et al., 2007; Koper et al., 2019; Ohio State Highway Patrol, 2005; Ozer, 2010; PA Consulting Group, 2004; Potts, 2018; Taylor et al., 2012). Some research also suggests that large-scale LPR deployment can improve investigative clearance rates for crimes like auto theft and robbery (Koper & Lum, 2019). In contrast, there is scant evidence on whether LPR use can reduce crime through deterrence and/or incapacitation (Lum & Koper, 2017, pp. 120-124).

In practice, crime prevention outcomes from LPR use likely depend on several factors, including the volume of LPR deployment, the manner in which LPRs are deployed (mobile versus fixed), the types of data accessed by LPR systems, how officers use LPRs in the field, how LPR data are saved and used for investigations, and the public's reaction to (and influence upon) LPR use. Given these complexities and the resources spent on LPRs, there is considerable need to build a stronger and broader evidence base on LPR applications and their effects to inform police decisions about LPR adoption and uses. To that end, this study examines the potential crime prevention effects of LPR use in general patrnl.

Background: the uses and impacts of license plate readers in patrol In the USA, agencies using LPRs deploy more than half of their LPR units in patrol, making it the most common way that police use this technology (Lum et al., 2019).

In principle, deployment of LPRs in patrol may help reduce crime through a combination of incapacitation and deterrence. When using LPRs, officers should have a greater chance of detecting and apprehending auto thieves and other wanted persons in the field. It is also conceivable that the display of LPR cameras on patrol vehicles increases offenders' perceptions of surveillance and apprehension risk, thus potentially creating deterrent effects akin to those stemming from the use of other surveillance technologies like closed-circuit television (CCTV) systems (Piza et al., 2019).

If true, patrols using LPRs could have greater deterrent effects than patrols without ~ Springer

Do license plate readers enhance the initial and residual ... 727 LPRs. This issue has not received much explicit consideration, but it is arguably important to making a comprehensive assessment of the value of LPRs.

To date, very few studies have attempted to measure the crime prevention value of LPRs in patrol, and their results have not provided clear indications of added deterrence from LPRs. For example, Lum and colleagues (Lum et al., 2011) found no evidence of crime reductions associated with LPR patrols in crime hot spots in two Northern Virginia jurisdictions. Similarly, Wheeler and Phillips (2018) reported inconsistent and inconclusive results when they tested the effects of LPR use at roadblocks in Buffalo on crime and vehicle accidents. However, both studies tested low-dose interventions, and neither provided a basis for identifying the unique effect of LPRs beyond that of patrol presence.

A randomized experiment conducted in Mesa, AZ, attempted to measure the effects of LPRs more specifically by contrasting the results of patrol operations conducted with and without LPRs at high-risk, hot spot road segments. As expected, officers were more likely to detect and recover stolen vehicles and apprehend auto thieves when using the LPR devices, though the numbers of vehicle recoveries and arrests were small (faylor et al., 2012; also see Koper et al., 2019). With regard to crime prevention (Koper et al., 2013), the locations that received LPR patrols experienced reductions in calls for service for drug crimes that lasted for at least several weeks beyond the intervention. The other routes, which were patrolled by the same officers but with their LPRs deactivated, experienced short-term (i.e., 2-week) reductions in calls regarding person offenses and auto theft. Interpretation of the results was complicated by the fact that LPRs were mounted on the vehicles at all times in both the LPR and non-LPR conditions; hence, any deterrent effects stemming from the display of LPRs would have affected both sets of hot spots. The researchers thus attributed the differing crime prevention results to differences in the style of the patrols (Koper et al., 2013).

When the officers used their LPRs, they spent more time conducting fixed surveillance in prominent locations, which may have deterred people who were driving to the area for drug purchases. In contrast, the officers spent more time roaming streets, parking lots, and apartments of the hot routes at slow speeds and with frequent pauses when they were operating without LPRs and focusing on manual checks of vehicle license plates.

In short, the limited evidence currently available on the effects of LPR patrol leaves much ambiguity about the value of using LPRs in this common fashion, particularly in regard to crime prevention. In this study, we attempt to extend LPR research by more explicitly testing whether LPR-enhanced patrol creates deterrent and apprehension effects that go beyond those of patrol without LPRs in high-risk, crime hot spot locations.1 1 Like prior LPR patrol studies, this study examines LPR use in the context of geographically focused, "hot spot" policing-i.e., police efforts focused on small areas or very specific places (e.g., addresses, intersections, and street blocks) where crime is concentrated (e.g., Sherman et al., 1989; Weisburd et al., 2004). Evidence from numerous studies indicates that directed patrol and other police interventions reduce crime and disorder at such locations without obvious signs of displacement (for reviews, see Braga et al., 2019a; Lum & Koper, 2017; National Academies of Sciences, Engineering, and Medicine, 2017; Telep & Weisburd, 2012). Hot spots are logical places for deployment of LPRs, whether for auto theft detection or other proactive activities. Like many crimes, auto theft is concentrated geographically in a manner linked to various environmental characteristics (Barclay et al., 1995; Fleming et al., 1994; Plouffe & Sampson, 2004; Potchak et al., 2002; Rengert, 1996; Rice & Smith, 2002). Hence, the likelihood of detecting both stolen vehicles and wanted persons through the use of LPRs would seem to be ~ Springer

728 C. S. Koper et al.

Initial and residual deterrence from patrol with license plate readers In addition to examining vehicle recoveries and arrests with LPR patrols (i.e., recovery and apprehension effects), we assess LPR-related deterrence by comparing initial and residual deterrence effects associated with specific LPR and non-LPR patrol visits to crime hot spots. The concepts of initial and residual deterrence were originally developed by Sherman (1990) to explain the effects of police crackdowns, which he defined as operations involving a sudden increase in police presence and/or enforcement. As described by Sherman (1990), a crackdown can produce initial deterrence while it is in operation and residual deterrence that lasts for some time after it has ended. To explain the latter, Sherman postulated that crackdowns increase perceptions of risk among offenders and that it may take some time for those heightened risk perceptions to diminish even after a crackdown has ended. In subsequent work, Koper (1995) applied these concepts to individual instances of police presence at hot spots, hypothesizing that each patrol appearance produces both initial and residual deterrence on a smaller scale. We use that framework for this analysis.

In this context, initial (or direct) deterrence refers to the deterrent effects that police have on behavior in a location when they are physically present at that location. Residual deterrence, in contrast, refers to the deterrence aftereffect that a police visit has on crime and disorder after an officer leaves the location. If a patrol visit to a hot spot generates a deterrent effect, it may drive troublesome people away from the location or otherwise prompt them to be more cautious in their behavior after the officer leaves, thus reducing the risk of crime and disorder at the location for some time after the officer's departure. Although the mechanisms underlying this effect are not explicitly understood, it seems that it could occur in a number of ways.

For example, people may experience or fear unwanted (and potentially threatening) contacts with police, they may feel that they have been under surveillance and potentially noticed by police, and/or they may fear that officers are interested in the location and likely to return (for further discussion, see Koper, 1995). A police presence that has a stronger deterrent effect can thereby be expected to result in a longer period without crime and disorder following the visit.

Koper (1995) used this framework to examine residual deterrence in relation to the length of a police visit to a hot spot. This produced findings on the optimal length of hot spot patrols that have substantially influenced subsequent research and practice (e.g., Ariel et al., 2016, 2020; Barnes et al., 2020; Kochel & Weisburd, 2018; Koper et al., 2015, 2020; Rosenfeld et al., 2014; Telep et al., 2014; Williams & Coupe, 2017). Here, we examine initial and residual deterrence in reference to LPR versus non-LPR patrols. In theory, patrols with LPRs may generate stronger deterrence if the display of the cameras causes offenders to be more fearful of immediate detection (e.g., if they are in a stolen vehicle, have open warrants, or are being sought for questioning) or surveillance (i.e., having their presence in an area Footnote I ( continued)
greatest in hot spot locations. In addition, focusing LPR use in hot spots should also maximize the deterrent value of displaying the cameras. ~ Springer

Do license plate readers enhance the initial and residual. .. 729 recorded). (This may occur even if people do not understand what type of camera is being used on the police vehicle.) If true, then stronger direct and residual deterrent effects from LPR patrol visits should arguably represent the most elemental manifestations of LPR-related deterrence. As described below, we investigate these hypotheses using data from a field test conducted with a large suburban police agency.

Deterrence and other surveillance technologies
These hypotheses about the perceptual impacts of LPRs are also grounded in the broader body of research on police surveillance technologies. Most notably, studies of CCTV have produced consistent evidence that these camera systems deter property, vehicle, and drug-related offenses (Piza et al., 2019). In some contexts, they have also been associated with reductions in violent crime and disorder (e.g., La Vigne et al., 2011; Mazerolle et al., 2002). CCTV systems are particularly effective in reducing crime when actively monitored, thus facilitating rapid police responses to events captured by the cameras (Piza et al., 2019; also see La Vigne et al., 2011; Piza et al., 2015).2 This would seem to strengthen the possibility of LPR patrols having enhanced deterrent effects since LPRs are also actively monitored (and obviously so to observers) when used by patrol officers.

Along similar lines, research on police body-worn cameras (BWCs) suggests that self-awareness generated by surveillance devices can deter misconduct by both police and citizens (see discussions in Ariel et al., 2015; Ariel et al., 2017; Ariel et al., 2018). For example, several studies have found that police use of BWCs reduces police use of force and citizen complaints against police (e.g., Ariel et al., 2015; Braga, Turchan, et al., 2019b; Jennings et al., 2015; Stolzenberg et al., 2019).

Despite some inconsistencies and ambiguities in the BWC evidence overall (see Lum et al., 2020), these findings imply that BWCs can discourage misbehavior by police and/or citizens in the course of their interactions (at least in some contexts).

Some experts have also hypothesized that police use of BWCs may enhance the deterrent effects of police presence more generally, particularly for offenses involving public disorder, domestic violence, and traffic violations, because the cameras can strengthen the evidence gathered against such perpetrators, thus increasing their perceptions of the certainty of punishment (Ariel, 2016; Ellis et al., 2015; Goodall, 2007). A few studies with weak designs have tentatively supported this notion, suggesting that police use of BWCs may reduce a variety of criminal and disorderly behaviors (Ellis et al., 2015; Goodall, 2007; ODS, 2011; but see Ariel, 2016 for a contrary finding from a more rigorous study). Finally, studies of speeding and redlight cameras provide further evidence of the deterrent impacts of police surveillance technologies in showing that camera systems can reduce vehicle crashes by discouraging speeding and disobedience of traffic signals (for reviews, see Blais & Dupont, 2005; Elliott & Broughton, 2005; Wilson et al., 2010). 2 To the eKtent that CCTV helps police solve a greater share of criminal investigations, crime prevention impacts associated with CCTV could reflect incapacitation as well as deterrence effects. However, evidence on how CCTV affects investigations is very limited (Ashby, 2017). ~ Springer

730 C. S. Koper et al.

Study location and context
Our study of the impacts of LPR patrols was conducted in a large metropolitan urban-suburban county in the South Atlantic region of the USA, referenced anonymously here as "Greenvale County," with the county's primary police department (hereafter, "Greenvale PD").3 This county, which shares a border with the metropolitan area's core city, has a population of almost 1 million people and has high crime rates compared to typical suburban jurisdictions. In the year prior to initiating this study, Greenvale experienced over 4,000 UCR (Uniform Crime Reports) Part 1 violent crimes and over 20,000 UCR Part 1 property crimes, and its rates per population of violent and property crimes were roughly 75% and 25-50% higher, respectively, than those of other metropolitan and suburban counties.

The Greenvale PD, an agency with close to 1700 sworn officers and 1900 total employees, has jurisdiction over most of the county. At the time of this study, the Greenvale PD had more than 50 LPRs in operation, which were divided among patrol officers, other specialized or investigative units, and fixed locations. The agency's LPR systems included multi-jurisdictional data on stolen vehicles and license plates as well as license plates linked to motor vehicle violations (e.g., parking tickets and registration violations), open warrants, sex offenders, AMBER alerts, gang members, and U.S. Department of Homeland Security watch lists.4 Officers in the agency could also upload ad hoc entries from the field for vehicles of interest. The agency's LPR scans were saved for 1 year unless needed longer for a specific investigation.

The Greenvale PD committed to working with the research team on a randomized experiment to test whether hot spot patrols conducted with LPRs have greater crime prevention effects (via deterrence and/or apprehensions) than hot spot patrols conducted without LPRs. The agency initially committed nine LPR patrol officers to the project who worked in three of the agency's six districts.5 In addition, nine non-LPR officers were selected who worked in the same patrol squads as the LPR officers.
3 We reference the agency anonymously based on an agreement made with the agency at the start of the project. 4 These capabilities are relatively advanced compared to those of other American agencies with LPRs (see Lum et al., 2019). 5 This level of LPR deployment is typical of what would be available in many American jurisdictions.

Among large US police agencies that own LPRs, three-quarters own fewer than 8 units, 90% own less than 15, and only 5% own more than 25 (Lum et al., 2019). Further, US agencies use half of their LPRs for purposes other than general patrol. For further discussion of the potential benefits and limitations of large-scale LPR deployment, see Koper and Lum et al. (2019) and Willis et al. (2018), as well as LPR evaluations conducted in the UK (PA Consulting Group, 2004, 2006). ~ Springer

Do license plate readers enhance the initial and residual. .. 731 Research methods
Selection and random assignment of hot spots
At the outset, 42 hot spots were selected for the project across the three study districts. The hot spots were chosen based on an analysis of 2 years of Greenvale data on violent and property crimes.6 Kernel density techniques were used to identify small areas of crime concentration as candidate locations.7 We then worked with Greenvale PD crime analysts and the participating officers in confirming the validity of the locations as hot spots, learning more about their nature, delineating their operational boundaries, and selecting the final study locations.8 In determining the final sample, we considered the base rates of incidents at candidate locations (almost all of the final locations had at least 200 criminal incidents during the 2-year selection period) and the number of officers available for the study (as discussed below, each officer was assigned to 1 to 3 hot spots). The final selected hot spots, which had an average of 229 criminal incidents per year, were approximately 0.1 square miles in size on average and contained a roughly even mix of residential and commercial locations.
9 A randomized block design was used to assign the locations to LPR or non-LPR patrols (e.g., see Weisburd, 1993). Hot spots within each study district were ranked based on their crime counts and divided into pairs based on these rankings (i.e., the top two ranked hot spots constituted 01:1e pair, the third and fourth ranked hot spots constituted the next pair, and so on). This produced 21 pairs of hot spots across the 3 districts. Using a random numbers generator, one hot spot from each pair was assigned to LPR patrol and the other to non-LPR patrol. This procedure produced groups of LPR and non-LPR hot spots that were statistically equivalent with respect 6 Types of incidents analyzed included homicides, robberies, assaults, domestic violence incidents, sex offenses, juvenile-related offenses, thefts, auto thefts, burglaries, and other criminal incidents. 7 Parameters were specified as 0.1 square miles for cell size and 0.1 mile for search radius. In deciding on the desired size of the hot spots, we considered multiple factors. Selecting smaller locations in general maximizes the effects of patrols by targeting them more precisely on the highest risk locations and maximizing officers' ability to affect perceptions through visibility. Also, we sought to define hot spots that were small enough for officers to cruise them multiple times in their entirety during a short patrol visit (e.g., 15 min). At the same time, we selected hot spots larger than the sorts of micro places used in some hot spot studies (e.g., specific blocks or intersection areas). This was based in part on the desire to maximize the base rates of crime in the chosen hot spots (for more robust statistical analyses). In addition, it was felt that making the hot spots somewhat larger than a micro place would enable the LPR officers to better capitalize on the value of their LPRs by giving them the opportunity to scan a wider range of vehicles in and around specific micro hot spot street segments and intersections. This style LPR hot spot patrol has appeared more effective for detecting problematic vehicles in other settings (Koper et al., 2019).
8 The project team sought to draw boundaries that closely corresponded to the natural layout of each location and to avoid creating multiple artificial hot spots out of one connected area. We say more below about the proximity of some hot spots and the potential for spillover effects. 9 Most of the commercial locations were mixed-use locations with businesses and residences. ~ Springer

732 C. S. Koper et al.
to their crime levels, size, and percentage that were residential only (as opposed to commercial or mixed-use locations). 10
Implementation of hot spot patrols
Each pair of hot spots created during the randomization procedure was assigned to a pair of officers-one LPR and one non-LPR-from the same patrol squad. The assignments were made in this manner to balance patrol delivery schedules across LPR and non-LPR locations. The specific days of the week that each patrol squad worked varied from one week to the next, 11 as did their shifts. Squads mostly alternated week to week between working the agency's day shift (6:30am to 4:30pm) and evening shift (3:30pm to 1:30am). Periodically, they were assigned to work the midnight shift (11:30pm to 7:30am) for 4 weeks in a row. Hence, the patrols were delivered during days and times that varied continuously both within and across locations. In assigning squads to hot spots, we sought to assign officers to hot spots that were in or as near as possible to their normal patrol areas.

Each participating officer was assigned to 1 to 3 hot spots (and had sole responsibility for those locations), depending on the number of hot spots in the officer's district and the number of officers from that district who were participating in the experiment. The officers were asked to patrol their assigned hot spots as frequently as practical (ideally 1 to 3 times each day they worked) for 15-30 min per visit.12 They were also asked to record each visit in a project log developed by the research team. The log required officers to record the location, time, and duration of each visit. Officers were also asked to provide a general description of how they patrolled the hot spot (e.g., drove around the location, parked at one spot) and to record any notable events that occurred, namely, LPR hits, vehicle stops, arrests, or other citizen contacts. Otherwise, officers were given discretion over how they patrolled the hot spots and (for LPR officers) how they used their LPRs.

The experiment began in the latter part of May 2017 and continued through September 2017. In total, the officers recorded 785 patrols they conducted at 33 of the 42 hot spots (9 locations could not be patrolled as planned, primarily due to officer reassignments or vehicle losses). These visits averaged 22-25 min and were roughly evenly divided between the LPR and non-LPR locations (further information about the visits is discussed below). Because the treated locations received only 1-2 visits 10 To gain some further sense of the comparability of the LPR and non-LPR hot spots, we also examined the social characteristics of the Census tracts in which they were located. Tracts containing LPR locations did not differ from those containing only non-LPR locations with respect to unemployment, household income, poverty levels, racial and ethnic diversity, and residential stability. Census tracts with LPR hot spots as initially assigned had lower levels of home ownership. However, this difference was diminished and not statistically significant after adjusting for the final sample of hot spots used in the analysis (see below on final sample adjustments). 11 Patrol officers in this agency typically worked 4-5 days in a row followed by 3-4-days off. 12 The 15- 30-min guideline is grounded in research on the optimal length of time for police to spend when visiting hot spots (Koper, 1995), adjusted somewhat for the size of the hot spots and the focus on LPR use (for a similar approach, see Lum et al., 2011). ~ Springer

Do license plate readers enhance the Initial and residual. .. 733 per week on average, we did not analyze changes in crime at the LPR and non-LPR locations (as in a typical randomized experiment) out of concern that the added patrol dosage was insufficient to reduce crime in either set of locations. 13 Alternatively, we addressed the study's central question-understanding the added crime prevention effects (if any) of using LPRs in patrol-in a more direct manner by conducting a quasi-randomized analysis of the initial and residual effects of each of the hundreds of LPR and non-LPR patrol visits to the hot spots using methods described below.

Statistical methods
Multiple outcomes were examined for the LPR and non-LPR hot spot patrols. For starters, stolen vehicle recoveries and arrests per visit were analyzed to determine if officers were more likely to recover stolen vehicles and apprehend offenders when using LPRs. These constitute important immediate outcomes that may arise from LPR use. Apprehensions made by LPR officers are also relevant to judging possible incapacitation and deterrence effects generated by LPR use.

Analyses were also conducted to assess initial and residual deterrence effects associated with each patrol visit. This involved analyzing the occurrence of crime and disorder at the hot spots relative to the comings and goings of the experimental patrols (see Koper, 1992, 1995). The initial deterrent effects of the LPR and nonLPR patrols were evaluated by examining the likelihood of a new crime or disorder occurring in a given hot spot, as measured by calls for service, while a project officer was present in that location. This likelihood was then compared across LPR and non-LPR patrols using chi-square tests for differences in the proportion of visits during which a new crime or disorder occurred.

Residual deterrence from the patrols was operationalized based on the time until the next occurrence of crime or disorder at the location (Koper, 1995) as measured through calls for service. 14 Tests for differences in residual deterrence between LPR and non-LPR patrols were conducted using methods of survival analysis, a group 13 Officers were unable to conduct the patrols as frequently as planned due to several factors, namely, temporary or permanent reassignments, vehicle breakdowns or damage, periodic in-service training, annual leave, and other service demands (e.g., heavy call volume). 14 As noted, the hot spots were identified based on more serious criminal incidents in order to ensure that the study locations would be places with both serious and minor crime and disorder issues. For the outcome analyses, we examined measures of overall crime and disorder as well as measures of more serious offending based on previously discussed literature on the effects of LPRs and other surveillance technologies. For this purpose, we utilized calls for service data, which provide the widest net of problem behaviors that come to the attention of police and are especially useful for measuring minor offenses and disorderly behaviors that are less likely to result in official reports or arrests (Sherman et al., 1989; Warner & Pierce, 1993). Calls for service are thus widely used in studies of crime and policing (e.g., Braga & Bond, 2008; Braga et al., 1999; Kochel et al., 2015; Koper et al., 2013; Mazerolle et al., 2000; Sherman & Weisburd, 1995; Weisburd & Green, 1995; Wu & Lum, 2017; but see Klinger & Bridges, 1997 on potential sources of error in calls for service). The use of calls for service was also necessary for this study in order to measure each recorded police stop at the study locations, including proactive stops as well as stops in response lo citizen calls (see below on the measurement of survival times lo the next police presence or the next criminal or disorder event). ~ Springer

734 C. S. Koper et al.
of statistical techniques for analyzing the occurrence and timing of events (e.g., see Allison, 1995). These methods were used to analyze the time from each experimental patrol visit at a hot spot until the next occurrence of a criminal/disorderly event at that location or the next police visit of any sort to the location (by project officers or others), whichever came first. If no new crimes or disorders occurred prior to the next police visit, then the observation is censored, meaning that we know only that the location "survived" without a new crime or disorder during the time that elapsed from the end of the experimental patrol visit to the next police presence. 15 Survival analysis techniques were developed specifically for the study of censored data. These methods allow one to control for the time during which each case was observed and to control for the simultaneous influence of other case characteristics on the likelihood and timing of a new event of interest.

In light of its operational significance for calibrating hot spot patrols (Koper, 1995), our analyses of residual deterrence are more extensive. Differences in residual deterrence between the LPR and non-LPR patrols were first examined using the life table method, which shows the likelihood that a new crime or disorder occurred within selected time frames following LPR and non-LPR patrols, adjusted for censoring. 16 In addition, multivariate Cox proportional hazards (CPH) models were estimated to gauge whether residual deterrence from LPR and non-LPR patrols differed while controlling for several characteristics of the hot spots and other aspects of the patrol visits. The CPH models provide estimates of changes in the "hazard rate" for crime or disorder, which represents the risk that a new crime or disorder occurred at the location at a given point in time, conditional on the event not having occurred before that point. 17
15 Although residual deterrence from a police patrol might conceivably last beyond the occurrence of the next crime, disorder, or police presence event, the immediate period from the end of the patrol until the next event should provide the clearest measure of this residual effect. Afterwards, the measurement of a patrol's residual deterrence will be confounded by the impacts that subsequent crimes, disorders, and/or police actions have on future events at the location. This method of measuring residual deterrence also has practical and operational significance, as it reflects time until the next event that will require a police response.
16 In the life table method, the analyst groups the event times into intervals of a chosen length- in this application, hours-and calculates S1, which is the probability that the case "survived" (i.e., did not experience the event of interest) to the start of interval t. For each interval, the value of S1 is based on the probabilities of events occurring in prior intervals. For example, the probability of surviving lo the third interval or beyond would be the product of (l-q1)(1-q2), where q1 and q2 represent the probabilities of events occurring during intervals 1 and 2, respectively. For a given interval, the probability of an event ( conditional on survival to the start of the interval) is denoted as q = d / (n - rn/2), where d equals the number of events occurring during the interval, n refers to the sample al risk al the start of the interval (i.e., the number of cases that have not experienced an event or been censored by the start of the interval), and m is the number of cases censored during the interval (Teachman, 1983:270). For further discussion of the life table method, see Allison (1995) and Teachman (1983). 17 The Cox proportional hazards model is often expressed as: h;(l) = )-0(t)exp(B1x;1+ ... +Bkxik), where h;(t) represents the hazard for subject i at time t, AQ(l) represents a baseline hazard function (which can be regarded as the hazard function for a subject whose covariates all have values of zero), X;1 through X;k represent a set of fixed covariates, and BI through Bk represent the effects of those covariates (these effects are then exponentiated) (Allison, 1995: pp. 113-114). The model assumes that the ratio of the ~ Springer

Do license plate readers enhance the initial and residual ... 735 Greenvale PD calls for service data for the intervention period were merged with the officer log data to determine whether criminal or disorder events occurred during each experimental patrol visit and to measure the time to the next crime or disorder event following the patrol. Crime and disorder events were defined based on citizen-initiated calls for crimes (such as violent or property crimes or drug-related offenses) and other disorderly or potentially criminal incidents (such as noise complaints, suspicious people, juvenile-related complaints, disputes, and alarms). We examined the effects of LPR patrols on crime and disorder calls broadly because, as discussed, prior research suggests that LPRs and other surveillance technologies may reduce a wide variety of criminal and disorderly behaviors. The same has also been shown for hot spot policing more generally (Braga et al., 2019b).18 However, as described below, we also conducted analyses differentiating between LPRs' impacts on more and less serious misbehaviors.

Identification of censored cases and times was made by determining whether each experimental patrol was followed by another police presence at that location prior to the occurrence of the next crime or disorder. Such police presences were identified using officer-initiated calls for proactive activities (e.g., premise checks, traffic stops, and subject stops) as well as patrol responses to citizen calls for noncriminal, non-disorder events (these consisted of miscellaneous service and administrative calls). 19 Likewise, cases were treated as censored if the next recorded event at the hot spot was a visit by the assigned project officer. A caveat is that the analysis cannot account for police presences that were not registered in calls for service (like a simple drive through an area) or crime and disorder events that did not result in citizen calls. The analysis assumes that any bias from these unrecorded events works equally for LPR and non-LPR visits (and thus should not affect inferences about the effects of LPR use), conditional on the random assignment of the patrols to hot spots and other covariate adjustments described below.

Hot spot characteristics incorporated into the multivariate CPH models as covariates included the location's size, type of land use (residential, business, or mixed), police district, and level of crime-related calls during the 3 months prior to the experiment. Items measured for each patrol visit included the day of the week, patrol shift (day, evening, or midnight), month, and length (in minutes) of the patrol.

Using the officer logs, two binary indicators were also created to denote whether the officer made an arrest or engaged in any other type of activity during the visit besides patrolling or stationary surveillance. Activities in the latter category include making traffic stops, issuing tickets or citations, investigating suspicious persons or vehicles, conducting premise checks, or engaging in other miscellaneous activities Footnote 17 ( continued)
hazards for any two subjects remains constant over time (i.e., that they remain proportional to one another), but makes no assumption about the distribution, or shape, of the baseline hazard rate. 18 It is commonplace for criminal offenders to engage in a variety of criminal and other anti-social behaviors (e.g., see Gottfredson & Hirschi, 1990). Accordingly, driving troublesome people away from a location or otherwise moderating their conduct could conceivably affect various problem behaviors at that location. 19 The officer's presence and time of arrival were determined using the on-scene time for these calls. ~ Springer

736 C. S. Koper et al.
(e.g., community policing activities). For each patrol visit, the count of prior project visits to that hot spot as of that date and time was also calculated to measure any cumulative effects the patrols may have had as the study progressed. Although many of these factors were balanced across the LPR and non-LPR conditions via randomization, including them can add further precision to the estimated LPR effects and correct for any imbalances that developed due to deviations from the study protocol (e.g., preliminary analysis revealed that LPR patrols were more likely to be carried out during the night shift relative to non-LPR patrols). Finally, indicators were added to the models for each blocked pair of hot spots that was created during the randomization procedure (i.e., blocking effects) to provide further control in the analysis for stable differences in crime risk across the pairs of hot spots used in the study. The models were also estimated with standard errors clustered by hot spot (Lin & Wei, 1989).

As an additional step to the residual deterrence analysis, a logistic regression model was estimated to determine whether LPR use was related to the type of event---crime or disorder-that first occurred following a police presence. This analysis was done to test whether LPR use reduces the likelihood of a more serious problem occurring soon after a police patrol. The logistic model was estimated with only those cases that resulted in a crime or disorder failure and used the patrol type and other covariates discussed above to estimate the likelihood that an observed failure was a criminal event (versus disorder).20 As applied to this study, using logistic regression to examine different failure types assumes there is one process governing the timing of residual deterrence decay but a different mechanism determining the type of event (crime or disorder) that occurs (see Allison, 1995: pp. 206-208). This approach seems preferable in this application to a competing risk survival analysis, which assumes that the occurrence of one event type precludes the other from occurring (Allison, 1995: pp. 185-209). (Nevertheless, an alternative competing risk model treating disorder failures as censored cases produced the same inference as the model discussed below.)
In creating the analytical database, 12 of the 785 patrols were combined with other immediately preceding visits because they appeared to represent continuous instances of police presence (i.e., an officer logging a new visit immediately after a previous one ended). In addition, 4 LPR visits were excluded from analysis because the officer did not have a working LPR, and 49 visits were excluded because nonproject officers arrived and engaged in proactive activities while a project officer was present, thus complicating assessments of direct and residual deterrence stemming from those experimental patrols. Other adjustments to the data are noted below. 20 In a logistic regression model for a binary outcome measure, the model variables are used to predict the natural logarithm of the odds of the outcome of interest. The odds are defined asp/ (l- p), where pis the probability of the outcome of interest. Exponentiating the model coefficients reveals each variable's multiplicative effect on the odds of the outcome. Here, the outcome of interest is a failure due to a criminal event as opposed to a disorder event. ~ Springer

Do license plate readers enhance the initial and residual. .. 737 Results
Immediate outcomes: initial deterrence, recoveries of stolen vehicles, and arrests As shown in Table 1, there were 35 instances in which a new crime or disorder occurred in a study location while an experimental officer was present. These failures of initial deterrence occurred in about 5% of patrol visits overall. Although they were slightly more common during LPR patrols than during non-LPR patrols, the difference was not statistically significant. These cases were excluded from subsequent analysis to provide clearer inferences about LPR versus non-LPR patrol activities and outcomes under normal conditions. (All subsequent analyses are thus based on a sample of 685 patrols.)
With regard to other immediate outcomes, LPR officers recovered stolen vehicles in 3% of their visits (amounting to 11 recoveries), in contrast to non-LPR officers who did not recover any (Table I). Arrests did not differ between LPR and non-LPR officers and were rare ( <l % of visits) among both groups (Table 1).

Residual deterrence
For the full sample of cases used in the residual deterrence analysis (n=685), 41 % were cases in which the police patrol was first followed by a crime or disorder (i.e., failure cases) and 59% were cases in which the patrol was first followed by another instance of police presence (i.e., censored cases). The average time to a crime/disorder failure was 534 min (about 9 h), and the average time to censoring was 352 min (nearly 6 h).

Table 2 provides life table estimates of the likelihood that a new crime or disorder call would occur following a hot spot patrol by the LPR or non-LPR officers. Estimates are provided by hour for up to 8 h following a patrol.21 Overall, the likelihood of a new crime or disorder following a patrol was approximately 13-15% within 1 h and 22- 25% within 4 h. By 8 h, this chance increased to 40-45%. The likelihood of a new incident was consistently higher following an LPR patrol, but the overall difference in survival distributions between the two patrol modes was not statistically significant.

Results from a multivariate CPH model are shown in Table 3, where the effect of each variable is presented as a hazard ratio that shows its multiplicative relationship to the hazard rate of a new crime or disorder. A ratio greater than one indicates that the characteristic in question increased the likelihood of a new crime or disorder, while a ratio less than one indicates that the characteristic reduced it. Controlling for several characteristics of the hot spots and patrols, the likelihood of a new crime or disorder was 1.36 times higher following an LPR patrol. This result can be expressed in percentage terms by subtracting one from the hazard ratio and 2 1 Results are based on life tables constructed with 1-h intervals. The estimates of a crime or disorder failure are based on I minus the likelihood that the case would "survive" without a crime or disorder to the selected hour, with the difference then converted to a percentage. ~ Springer

738 C. S. Koper et al.
multiplying the difference by 100. Thus, the likelihood of a new incident increased by (1.36-l)xlO0 = 36% following an LPR visit. However, as in the life table analysis, the difference in residual deterrence between the LPR and non-LPR patrols was not statistically significant.

Other notable results indicate that residual deterrence times were impacted by the land use, crime levels, and district of the hot spots. Other characteristics of the locations and patrols had weaker or statistically non-significant relationships to residual deterrence times.

Further, the use of LPR was not significantly related to the next type of event that occurred in the hot spot as residual deterrence decayed. Logistic regression results in Table 4 show that the odds of the next event being a crime were 36% lower following an LPR patrol visit (for this model, the displayed coefficients show the multiplicative relationship of each variable to the odds that a new event was a crime rather than a disorder); however, this association was not statistically significant, controlling for characteristics of the hot spots and patrols. (Note that arrest is not included in the model given that it occurred in only one case. The model also includes a variable for the length of time that elapsed between the patrol visit and the event's occurrence.)
As a final step, we considered the possibility that the proximity of some hot spots could have affected the observed results. The patterns of crime and geography in the study districts necessitated the creation of several adjacent hot spots as shown in Fig. 1. In designing the study, efforts were made to separate adjacent hot spots by a few blocks or by natural boundaries (e.g., wooded areas) that impeded the flow of persons and vehicles between them. Nonetheless, if LPRs do generate additional deterrence, it is possible that non-LPR hot spots near an LPR location benefitted from spillover LPR effects that may have accumulated over time. If so, this could have minimized the apparent differences between LPR and non-LPR patrols, thus masking LPR impacts. Although this seemed unlikely given the low patrol dosages delivered during the study, we ran additional tests differentiating between non-LPR spots located nearby or distant from LPR locations. For this purpose, non-LPR locations within 2-3 blocks (or an equivalent distance) of an LPR location were treated as LPR-adjacent; the remaining non-LPR locations were substantially farther from LPR locations. The models in Tables 3 and 4 were then re-estimated with distinctions between non-LPR patrols that occurred in locations adjacent to or farther from LPR locations. These results (not shown) also provided no evidence of LPR effects.

Most notably, LPR patrols did not produce greater residual deterrence than non-LPR patrols in either nearby or distant non-LPR locations.

Discussion and conclusions
In summary, the results of the Greenvale PD study provide little clear evidence for the efficacy ofusing LPRs in general patrol. Although officers in this study were more likely to recover stolen vehicles when using LPRs, they did not make significantly more arrests, nor did they generate stronger deterrence of crime and disorder. Recovering more stolen vehicles is a benefit of LPRs, but it may not alone be worth the cost of LPRs, particularly ~ Springer

Do license plate readers enhance the initial and residual ...

Table 1 Direct outcomes during LPR and non-LPR patrols in project hot spots Outcome
New crime or disorder
Recovery of stolen vehicle
Arrest
LPR patrols
22 of 385 (5.7%)
11 of 363 (3.0%)*
3 of 363 (0.8%)
*Difference statistically significant at p<=.01 with chi-square test 739
Non-LPR patrols
13 of 335 (3.9%)
0 of 322 (0%)
2 of 322 (0.6%)
if LPR use primarily affects the speed rather than the likelihood of recovering stolen vehicles (an issue not examined here). The findings from this study are consistent with others that have failed to produce evidence of crime prevention stemming from LPR use in patrol (Lum et al., 2011; Wheeler & Phillips, 2018). These findings are noteworthy because mounting LPRs on general patrol vehicles is the most common way that police currently deploy this technology (Lum et al., 2019). By and large, therefore, police may not be using LPRs in the most optimal ways. Further, this study focused LPR patrol use strategically in crime hot spots (thus increasing the likelihood of positive results), which is something that agencies with LPRs often fail to do in practice (Lum et al., 2019).

Nonetheless, given police expenditures on LPRs and the continued interest in this technology among police and citizens, further studies are needed to bolster the evidence base on LPRs. It would be particularly useful to see additional LPR patrol studies in jurisdictions with higher base rates of crime and disorder. Some studies suggest, for example, that the detection and apprehension value of LPRs may be greater in high-crime urban locations (Ozer, 2010). Also, while this study focused on the effects of individual instances of LPR use at hot spots, intensive LPR patrols maintained over long periods might conceivably produce cumulative area-level effects that could not be examined here.

Results found here may have also been conditioned by how officers used LPRs.

Interviews and observations of the participating officers revealed that the LPR officers often deactivated or ignored LPR alerts for minor issues (e.g., parking and registration violations) because they found such alerts to be too distracting and timeconsurning, particularly given other demands on their time created by call volume Table 2 Likelihood of new crime or disorder following LPR and non-LPR patrols in project hot spots for selected follow-up periods
Follow-up time LPR patrols Non-LPR patrols
I h 15.2% 12.7%
2h 21.0% 17.3%
3h 25.4% 22.6%
4h 30.2% 27.7%
Sh 33.8% 31.8%
6h 39.1% 35.4%
7h 40.6% 37.8%
8h 44.6% 39.5%
Life table estimates (n=685). Differences were not statistically significant (p>.10) with a log-rank test ~ Springer

740 C. S. Koper et al.

Table 3 Effects of patrol type on the likelihood of new crime and disorder events in project hot spots (Cox proportional hazard model estimates)
Hazard ratio
LPRpatrol 1.36
Size (in squared miles) 13.37
Land use
Business 1.26
Mixed 1.88**
District 3 0.59*
District 4 0.31#
Crime in prior 3months I.OD*
Shift
Evening 0.92
Night 0.91
Arrest made 0.35
Police activity 0.82
Length of patrol I.OJ#
Prior visits I.OD
Standard error

0.28 35.42

0.31 0.40

0.15 0.20

0.00 0.11

0.16 0.55

0.11 0.00

0.01 Z-score
test
statistic

1.46 0.98

0.95 2.96
-2.05
- 1.81

2.32 -0.67
-0.53
-0.67
- 1.48

1.87 -0.49
N=685. Day of week, month, and block random assignment indicators are not shown. Robust standard errors were used, allowing for clustering by hot spot. Land use indicators are interpreted relative to residential locations, district indicators relative to District I, and shift indicators relative to day shift # Statistically significant at p<=O.l; * statistically significant at p<=0.05; ** statistically significant at p<=0.01
and staffing limitations. Activating and following through on such alerts may generate additional citizen contacts that have added deterrent effects and result in more detection of problem persons (on the latter point, see Cohen et al., 2007). On the other hand, this may also increase intrusive and negative police-citizen contacts, thus eroding citizens' views of police legitimacy (e.g., see Lum et al., 2010; Merola et al., 2014), and potentially exacerbate racial bias in policing actions. This quandary (along with the general difficulty of acting on an LPR alert when answering a service call) highlights some of the limits to using LPRs in general patrol. Future studies might examine this issue more closely, and perhaps even test the use of different LPR alert settings, to examine how this affects officers' activities and results with regard to crime-related matters as well as traffic and vehicle violations.

A related point is that use of LPRs may also affect officer behavior in other ways that enhance or reduce the effectiveness of patrol. For instance, the logs completed by the officers in this study revealed that LPR officers were more likely to issue citations and tickets, while non-LPR officers were more likely to do premise checks and investigate suspicious persons and vehicles.22 Observations during ride-alongs with 22 LPR officers reported issuing citations or tickets in 15% of their hot spot patrols, in contrast to nonLPR officers who reported doing so in 7% of their patrols. Non-LPR officers, on the other hand, reported ~ Springer

Do license plate readers enhance the initial and residual. .. 741 Table 4 Effects of patrol type on whether the next event in a project hot spot was a criminal rather than disorder event (logistic regression estimates)
LPR
Size (in squared miles)
Land use
Business
Mixed
District 3
District 4
Crime in prior 3months
Shift
Evening
Night
Police activity
Length of patrol
Prior visits
Time to event
Odds ratio

0.64 1000.19

3.22*

0.83 0.61

2.16e-06**

1.00 1.52

1.21 0.71

1.00 1.00

1.00#
Standard error

0.28 6459.68

1.61 0.33

0.38 4.4le-06

0.00 0.53

0.39 0.18

0.01 0.02

0.00 Z-score
test
statistic
-1.03

1.07 2.34
-0.46
-0.79
-6.39

0.03 1.22

0.60 -1.37

0.42 0.17

1.84 N=279. Day of week, month, and block random assignment indicators are not shown. Robust standard errors were used, allowing for clustering by hot spot. Land use indicators are interpreted relative to residential locations, district indicators relative to District 1, and shift indicators relative to day shift # Statistically significant at p<=O.l; * statistically significant at p<=0.05; ** statistically significant at p<=0.01
the project officers also suggested that LPR and non-LPR officers had somewhat different patrol styles. LPR officers tended to focus more on scanning license plates in places and along roads where they thought they would be more likely to find stolen vehicles or possibly persons with warrants (relatedly, LPR officers were also more likely to note driving around and canvassing the area as a major activity in their logs). Non-LPR officers, on the other hand, seemed more focused on looking for suspicious persons and vehicles. These behavioral differences did not produce clear differences in apprehensions or deterrence in this study, but similar issues with LPR use and patrol style have been observed by others (Koper et al., 2013, 2019) and should perhaps receive further study to determine if there are specific ways to optimize LPR use in patrol.

Notwithstanding these caveats, this study and others suggest that police and researchers should give more emphasis to testing alternative uses of LPRs. For example, deploying stationary LPRs at fixed locations or on mobile trailers may Footnote 22 ( continued)
making premise checks during 11 % of their patrols and investigating suspicious persons or vehicles during 15%. The comparable figures for LPR officers were approximately 1-2%. ~ Springer

742
I
)•
'-, I l , ( ,, ., ~ '
I ,.
i
J
' I '
: ~ ;,·
~ _;~~-.-
' ,. \<\ ... '\;
Soorce : prOY!ded by lhe pol.Ce agency
I
1,, <' I '
C. S. Koper et al.

I
' l/
)
,/
[:=I Non-LPR ,
LPR
-, A l • ' ·,. '.l ') 0 0,5 1 2
- - Miles
Fig. 1 Map of study hot spots. Notes: Map depicts 33 hot spots used in the analysis. Treatment hot spots (in red) denote LPR patrol locations and control hot spots (in blue) denote non-LPR patrol locations have greater advantages for investigative applications because it provides constant surveillance and a permanent record of vehicles using selected roadways, Placed at strategic roads and intersections (e.g., roads with high volume traffic in highcrime areas), stationary LPRs can increase the chance of detecting suspect vehicles, including those traveling to or from crime scenes. This can prove valuable for realtime apprehension, follow-up investigations, and other types of special operations (e.g., Koper & Lum, 2019; PA Consulting Group, 2004; Potts, 2018; Willis et al., 2018). Other uses of LPRs by special mobile or investigative units or to address driving and vehicle violations may also prove to be more effective and cost efficient than using LPRs in general patrol (e.g., Eberline, 2008; Gierlack et al., 2014; Koper et al., 2013; PA Consulting Group, 2004).23 American police agencies currently deploy about one-quarter of their LPRs at fixed locations and about 10% with specialized units (Lum et al., 2019), which should afford opportunities for comparative research on different LPR deployments. Further study along these lines can help police make informed decisions about whether and how extensively to invest in LPRs and how best to utilize LPRs they possess.
23 With regard to cost efficiency considerations, reported LPR costs range from $10,000 to $25,000 per unit for mobile units, with fixed costs as high as $100,000 depending on where they are placed (Gierlack et al., 2014). Additional long-term operational and maintenance costs must also be considered. ~ Springer

Do license plate readers enhance the initial and residual. .. 743 As final points in closing, we note that this study also contributes in two broader ways to research on policing. First, it adds to a larger body of research on police technology that questions whether and to what degree technological advancements make police more effective (e.g., Brown, 2015; Byrne & Marx, 2011; Chan, 2011; Garicano & Heaton, 2010; Harris, 2007; Lum & Koper, 2017, pp. 111-130; Piza et al., 2019). Indeed, studies suggest that the impacts of new technologies in policing can be quite variable depending on a variety of technical, contextual, and organizational factors. This study of LPRs adds to the complexity and mixed findings of this broader body of work and underscores the need for a strong commitment to evaluation research that can help police optimize their decisions regarding technology acquisition and uses. Second, this study complements others that have investigated the crime prevention effects of police activities at micro levels of place and time (e.g., Santos & Santos, 2015; Sherman & Rogan, 1995; Weisburd et al., 2016; Wyant et al., 2012), and it can particularly inform future research on placebased patrol. While the focus of this study has been on the deterrent effects of LPRs, analyses of the initial and residual effects of specific instances of police patrol have implications for the calibration of hot spot patrols more generally (Koper, 1995).

Studying the timing of crime and disorder events following police visits, and the characteristics of hot spots and patrol actions that affect that timing, could potentially be used to plan preventive patrol visits and activities for hot spots, adjusted for their risk levels. Moving forward, such research might help to better illuminate optimal dosages, activities, and schedules for hot spot patrols.

Acknowledgments This project was supported by Grant 2013-U-CX-0017 from the National Institute of Justice (Office of Justice Programs, U.S. Department of Justice). The authors are also grateful to the agency and individual officers that took part in this project.

Declarations
Disclaimer The views expressed are those of the authors and should not be attributed to the U.S. Department of Justice or George Mason University.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

References
Allison, P. D. (1995). Survival analysis using SAS: A practical guide. SAS Institute. Inc ..

Ariel, B. (2016). Increasing cooperation with the police using body worn cameras. Police Quarterly, 19(3), 326-362.
~ Springer

744 C. S. Koper et al.

Ariel, B., Farrar, W. A., & Sutherland, A. (2015). The effect of police body-worn cameras on use of force and citizens' complaints against the police: A randomized controlled trial. Jo11mal of Quantitative Criminology, 31(3), 509-535.

Ariel, B., Sherman, L. W., & Newton, M. (2020). Testing hot-spots police patrols against no-treatment controls: Temporal and spatial deterrence effects in the London Underground. Criminology, 58(1), 101-128.

Ariel, B., Sutherland, A., Henstock, D., Young, J., Drover, P., Sykes, J., Megicks, S., & Henderson, R. (2017). "Contagious Accountability": A global multisite randomized controlled trial on the effect of police body worn cameras on citizens' complaints against the police. Criminal Justice and Behavior, 44(2), 293-316.

Ariel, B., Sutherland, A., Henstock, D., Young, J., Drover, P., Sykes, J., Megicks, S., & Henderson, R. (2018). Paradoxical effects of self-awareness of being observed: Testing the effect of police bodyworn cameras on assaults and aggression against officers. Journal of Experimental Criminology, 14(1), 19-47.

Ariel, B., Weinborn, C., & Sherman, L. W. (2016). "Soft" policing at hot spots-Do police community support officers work? A randomized controlled trial. Joumal of Experimental Criminology, 12(3), 277-317.

Ashby, M. P. J. (2017). The value of CCTV surveillance cameras as an investigative tool: An empirical analysis. European Journal 011 Criminal Policy and Research, 23, 441-459.

Barclay, P., Buckley, J., Brantingham, P. L., Brantingham, P. J., & Whinn-Yates, T. (1995). Preventing auto theft in suburban Vancouver commuter lots: Effects of a bike patrol. In R. Clarke (Ed.), Crime prevention studies (6th ed.). Criminal Justice Press.

Barnes, G. C., Williams, S., Sherman, L. W., Parmar, J., House, P., & Brown, S. A. (2020). Sweet spots of residual deterrence: A randomized crossover experiment in mi11imalist police patrol. SocArXiv Papers. https://doi.org/l0.31235/osf.io/kwf98.

Blais, E., & Dupont, B. (2005). Assessing the capability of intensive police programmes to prevent severe road accidents: A systematic review. British Journal of Criminology, 45(6), 914-937.

Braga, A. A., Barao, L. M., Zimmerman, G. M., Douglas, S., & Sheppard, K. (2019a). Measuring the direct and spillover effects of body worn cameras on the civility of police-citizen encounters and police work activities. Journal of Quantitative Crimi110/ogy, 36, 851- 876.

Braga, A. A., & Bond, B. J. (2008). Policing crime and disorder hot spots: A randomized controlled trial.

Crimi110/ogy, 46(3), 577--607.

Braga, A. A., Weisburd, D. L., Waring, E. J., Mazerolle, L. G., Spelman, W., & Gajewski, F. (1999).

Problem-oriented policing in violent crime places: A randomized controlled experiment. Criminology, 37(3), 541-580.

Braga, A. A., Turchan, B. S., Papachristos, A. V., & Hureau, D. M. (2019b). Hot spots policing and crime reduction: An update of an ongoing systematic review and meta-analysis. Joumal of Experimental Crimi110/ogy, 15(3), 289-311.

Brown, M. M. (2015). Revisiting the IT productivity paradox. The A111erica11 Review of Public Administration, 45(5), 565-583.

Byrne, J., & Marx, G. (2011). Technological innovations in crime prevention and policing: A review of the research on implementation and impact. Journal of Police Studies, 20(3), 17-40.

Chan, J.B. L. (2011). The technological game: How information technology is transforming police practice. Cri111inology and Criminal Justice, 1(2), 139- 159.

Cohen, I., Plecas, D., & McCormick, A. (2007). A report 011 the utility of the au/0111ated license plate recognition system in British Colu111bia. Unpublished report. University College of the Fraser Valley.

Eberline, A. (2008). Cost/be11efit analysis of electronic license plates. Final report 637. Arizona Department of Transportation.

Elliott, M. A., & Broughton, J. (2005). How methods and levels of policing affect road casualty rates.

TRL Limited.

Ellis, T., Jenkins, C., & Smith, P. (2015). Evaillatio11 of the introduction of personal issue body worn video ca111eras (Operation Hyperion) 011 the Isle of Wight: Final report to Hampshire Constabulatory. Institute of Criminal Justice Studies, University of Portsmouth.

Fleming, Z., Brantingham, P. L., & Brantingham, P. J. (1994). Exploring vehicle theft in British Columbia. In R. V. Clarke (Ed.), Crime prevention studies (6th ed.). Criminal Justice Press.

Garicano, L., & Heaton, P. (2010). Information technology, organization, and productivity in the public sector: Evidence from police departments. Journal of Labor Economics, 28(1), 167- 20 I. ~ Springer

Do license plate readers enhance the initial and residual. .. 745 Gierlack, K., Williams, S., LaTourrette, T., Anderson, J. M., Mayer, L. A., & Zmud, J. (2014). License plate readers for law enforcement: Opporlllnities and obstacles. RAND Corporation.

Goodall, M. (2007). G11idance for the police use of body-wom video devices: Police a11d crime standards directorate. Home Office.

Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime. Stanford University Press.

Harris, C. J. (2007). The police and soft technology: How information technology contributes to police decision-making. In J. M. Byrne & D. J. Rebovich (Eds.), The 11ew teclr11ology of crime, law and social co11trol. Criminal Justice Press.

Jennings, W. G., Lynch, M. D., & Fridell, L. A. (2015). Evaluating the impact of police officer bodyworn cameras (BWCs) on response-to-resistance and serious external complaints: Evidence from the Orlando Police Department (OPD) experience utilizing a randomized controlled experiment.

Journal of Criminal J11stice, 43(6), 480-486.

Klinger, D. A., & Bridges, G. S. (1997). Measurement error in calls-for-service as an indicator of crime.

Criminology, 35(4), 705-726.

Kochel, T. R., Burruss, G., & Weisburd, D. (2015). St. Louis Co11nty /rot spots in residential areas (SCH/RA) final report: Assessing the effects of /rot spots policing strategies 011 police legitimacy, crime, and collective efficacy. Southern Illinois University.

Kochel, T. R., & Weisburd, D. (2018). The impact of hot spots policing on collective efficacy: Findings from a randomized field trial. Justice Q11arterly, 36(5), 900--928.

Koper, C. S. (1992). The deterrent effects of police patrol presence upon criminal and disorderly behavior at /rot spots of crime. M.A. Thesis. College Park, MD: Department of Criminology and Criminal Justice, University of Maryland.

Koper, C. S. (1995). Just enough police presence: Reducing crime and disorderly behavior by optimizing patrol time in crime hot spots. J11stice Quarterly, 12, 649-672.

Koper, C. S., & Lum, C. (2019). The impacts of large-scale license plate reader deployment on criminal investigations. Police Quarterly, 22(3), 305-329.

Koper, C. S., Lum, C., & Hibdon, J. (2015). The uses and impacts of mobile computing technology in hot spots policing. Evaluation Review, 39(6), 587-624.

Koper, C.S., Lum, C., Wu, X., & Hegarty, T. (2020). The long-term and system-level impacts of institutionalizing hot spot policing in a small city. Forthcoming in Policing: A Joumal of Policy and Practice.

Koper, C. S., Taylor, B. G., & Park, S. (2019). Optimizing the geographic deployment of hot spot patrols with license plate readers. Journal of Experimelllal Criminology, 15(4), 641-650.

Koper, C. S., Taylor, B. G., & Woods, D. J. (2013). A randomized test of initial and residual deterrence from directed patrol and use of license plate readers at crime hot spots. Journal of Experimental Criminology, 9(2), 213-244.

La Vigne, N. G., Lowry, S.S., Markman, J. A., & Dwyer, A. M. (2011). Evaluating tire use ofp11blic stll'- veillance cameras for crime control and prevention. The Urban Institute.

Lin, D. Y., & Wei, L. J. (1989). The robust inference for the Cox proportional hazards model. Jou/'/lal of tire American Statistical Associatio,1, 84(408), 1074-1078.

Lum, C., Hibdon, J., Cave, B., Koper, C. S., & Merola, L. (2011). License plate reader (LPR) police patrols in crime hot spots: An experimental evaluation in two adjacent jurisdictions. Journal of Experimental Criminology, 7(4), 321-345.

Lum, C., & Koper, C. S. (2017). Evidence-based policing: Tra11slating research illlo practice. Oxford University Press.

Lum, C., Koper, C. S., Willis, J. J., Happeny, S., Vovak, H., & Nichols, J. (2019). The rapid diffusion of license plate readers in US law enforcement agencies. Policing: A11 /11ternational Joul'llal of Police Strategies and Management, 42(3), 376-393.

Lum, C., Koper, C. S., Wilson, D. B., Stoltz, M., Goodier, M., Eggins, E., Higginson, A., & Mazerolle, L. (2020). Body-worn cameras' effects on police officers and citizens' behavior: A systematic review.

Campbell Systematic Reviews, 16(3), e 1112. https://doi.org/10.1002/cl2. l l l 2.

Lum, C., Merola, L., Willis, J., & Cave, B. (2010), License plate recognitio11 technologies for law enforcement: An outcome and legitimacy evaluatio11. Report to SPAWAR and National Institute of Justice. Fairfax, VA: Center for Evidence-Based Crime Policy, George Mason University.

Mazerolle, L., Hurley, D., & Chamlin, M. (2002). Social behavior in public space: An analysis of behavioral adaptations to CCTV. Security Journal, 15, 59-75.

Mazerolle, L. G., Ready, J., Terrill, W., & Waring, E. (2000). Problem-oriented policing in public housing: The Jersey City evaluation. Justice Quarterly, 17, 129-158. %) Springer

746 C. S. Koper et al.

Merola, L. M., Lum, C., Cave, B., & Hibdon, J. (2014). Community support for the use of license plate recognition by police. Polici11g: An International Jo11mal of Police Strategies and Ma11agement, 37(1), 30-51.

National Academies of Sciences, Engineering, and Medicine. (2017). Proactive policing: Effects 011 crime and commrmities. The National Academies Press.

ODS Consulting. (2011). Body worn video projects irr Paisley and Aberdee11, self evaluation. Author.

Ohio State Highway Patrol. (2005), A11tomatic plate reader technology. Planning Services Section, Research and Development Unit.

Ozer, M.M. (2010). Assessi11g the effectiveness of the Cincinnati Police Department's automatic license plate reader system within the framewor* of i111e/lige11ce-led polici11g a11d crime prevelllio11 Theory.

Ph.D. dissertation. School of Criminal Justice, University of Cincinnati.

PA Consulting Group. (2004). Driving dow11 crime: De11ying crimi11a/s the use of the road. Author.

PA Consulting Group. (2006). Police standards emit: Thematic review of the 11se of automatic 11umber plate recog11itio11 withi11 police forces. PA Consulting Group.

Piza, E. L., Caplan, J. M., Kennedy, L. W., & Gilchrist, A. M. (2015). The effects of merging proactive CCTV monitoring with directed patrol: A randomized controlled trial. Journal of Experime11ta/ Crimi11ology, 11(1), 43-69.

Piza, E. L., Welsh, B. C., Farrington, D. P., & Thomas, A. L. (2019). CCTV surveillance for crime prevention: A 40-year systematic review with meta-analysis. Criminology & P11blic Policy, 18, 135-159.

Plouffe, N., & Sampson, R. (2004). Vehicle theft and theft from autos in parking lots in Chula Vista, CA.

In M. G. Maxfield & R. V. Clarke (Eds.), Understa11di11g a11d preve11ti11g car theft, crime prevention sllldies (17th ed.). Criminal Justice Press.

Potchak, M., McGloin, J. , & Zgoba, K. (2002). A spatial analysis of criminal effort: Vehicle thefl in Newark, New Jersey. Crimi11al Justice Policy Review, 133, 257-285.

Potts, J. (2018). Assessing the benefits of automated license plate readers. The Police Chief. 2018, 14-15.

Rengert, G. (1996). Vehicle theft in central Philadelphia. In R. Home) (Ed.), Policing for prevelllio11:

Red11cing crime, public intoxication and injury, crime preve11tion studies (7th ed.). Criminal Justice Press.

Rice, K. J., & Smith, W.R. (2002). Socioecological models of automotive theft: Integrating routine activity and social disorganization approaches. Journal of Research in Crime and Delinq11ency, 393, 304-336.

Rosenfeld, R., Deckard, M. J., & Blackburn, E. (2014). The effects of directed patrol and self-initiated enforcement on firearm violence: A randomized controlled study of hot spot policing. Criminology, 52(3), 428-449.

Santos, R. G., & Santos, R. B. (2015). An ex post facto evaluation of tactical police response in residential theft from vehicle micro-time hot spots. Journal of Quantitative Criminology, 31, 679-698.

Sherman, L. (1990). Police crackdowns: Initial and residual deterrence. In M. Tonry & N. Morris (Eds.), Crime and justice: An an11ual review of research (12th ed.). University of Chicago Press.

Sherman, L., Gartin, P., & Buerger, M. (1989). Hot spots of predatory crime: Routine activities and the criminology of place. Criminology, 271, 27- 56.

Sherman, L. W., & Rogan, D. P. (1995). Deterrent effects of police raids on crack houses: A randomized, controlled experiment. Justice Quarterly, 12(4), 755-781.

Sherman, L. W., & Weisburd, D. (1995). General deterrent effects of police patrol in crime "hot spots": A randomized, controlled trial. Justice Quarterly, /2(4), 625-648.

Stolzenberg, L., D'Alessio, S.J., Flexon, J.L. (2019). Eyes 011 the street: Police use of body-wom cameras in Miami-Dade Co1111ty. Weston Publishing, LLC.

Taylor, B., Koper, C. S., & Woods, D. J. (2012). Combating vehicle theft in Arizona: A randomized experiment with license plate recognition technology. Criminal Justice Review, 37(1), 24-50.

Teachman, J. D. (1983). Analyzing social processes: Life tables and proportional hazards models. Social Science Research, 12(3), 263-301.

Telep, C. W., Mitchell, R. J., & Weisburd, D. (2014). How much time should the police spend at crime hot spots? Answers from a police agency directed randomized field trial in Sacramento, California.

Justice Quarterly, 31(5), 905- 933.

Telep, C. W., & Weisburd, D. (2012). What is known about the effectiveness of police practices in reducing crime and disorder? Police Quarterly, 15(4), 331- 357.

Warner, B. D., & Pierce, G. L. (1993). Reexamining social disorganization theory using calls to the police as a measure of crime. Criminology, 31(4), 493-517.
~ Springer

Do license plate readers enhance the initial and residual. .. 747 Weisburd, D. (1993). Design sensitivity in criminal justice experiments. In M. Tonry (Ed.), Crime and justice: A review of research (17th ed., pp. 337-380). University of Chicago Press.

Weisburd, D., Bushway, S., Lum, C., & Yang, S. M. (2004). Trajectories of crime at places: A longitudinal study of street segments in the city of Seattle. Crimil10/ogy, 42, 283-321.

Weisburd, D., & Green, L. (1995). Policing drug hot spots: The Jersey City drug market analysis experiment. Justice Quarterly, 12(4), 711-735.

Weisburd, D., Wooditch, A., Weisburd, S., & Yang, S. M. (2016). Do stop, question, and frisk practices deter crime? Evidence at micro units of space and time. Criminology & Public Policy, I 5(1), 31- 56.

Wheeler, A. P., & Phillips, S. W. (2018). A quasi-experimental evaluation using roadblocks and automatic license plate readers to reduce crime in Buffalo, NY. Security Journal, 31, 190--207.

Williams, S., & Coupe, T. (2017). Frequency vs. length of hot spots patrols: A randomised controlled trial. Cambridge Journal of Evidence-Based Policing, 1, 5-21.

Willis, J. J., Koper, C., & Lum, C. (2018). The adaptation of license plate readers for investigative purposes: Police technology and innovation re-invention. Justice Quarterly, 35(4), 614-638.

Wilson, C., Willis, C., Hendrikz, J. K., Le Brocque, R., & Bellamy, N. (2010). Speed cameras for the prevention of road traffic injuries and deaths. The Cochrane Database of Systematic Reviews, JO, CD004607.

Wu, X., & Lum, C. (2017). Measuring the spatial and temporal patterns of police proactivity. Journal of Quantitative Criminology, 33(4), 915- 934.

Wyant, B. R., Taylor, R. B., Ratcliffe, J. H., & Wood, J. (2012). Deterrence, firearm arrests, and subsequent shootings: A micro-level spatio-temporal analysis. Justice Quarterly, 29(4), 524-545.

Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Dr. Christopher S. Koper is an associate professor in the Department of Criminology, Law and Society at George Mason University and the principal fellow of Mason's Center for Evidence-Based Crime Policy. He specializes in issues related to policing, firearms policy, program evaluation, and evidence-based crime policy.

Dr. Cynthia Lum is a professor in the Department of Criminology, Law and Society at George Mason University and director of its Center for Evidence-Based Crime Policy. Her research focuses on policing and police organizations, including patrol and investigative deployment, police use of technologies, crime prevention, and translating and institutionalizing the use of research into practice.

Dr. Xiaoyun Wu is a research associate in the Department of Criminology, Law and Society at George Mason University. Her research interests include policing, deterrence, and crime prevention. Her current work focuses on understanding everyday police proactive activities and deployment and its dynamic relationship with crime and disorder, using various data sources, measurements, and statistical methods.

William Johnson is a doctoral student in the Department of Criminology, Law and Society at George Mason University and a graduate research assistant in the Center for Evidence-based Crime Policy. His research interests include hate crimes, firearms violence prevention, and policing.

Megan Stoltz is a doctoral student in the Department of Criminology, Law and Society at George Mason University and a graduate research assistant in the Center for Evidence-based Crime Policy. Her current research focuses on police officer decision-making during their interactions with persons with mental disorders. Her additional research interests include evidence-based policing, translational criminology, and the intersection of mental health issues and the criminal justice system. ~ Springer

748 C. S. Koper et al.

Authors and Affiliations
Christopher S. Koper 1 tD> • Cynthia Lum 1 • Xiaoyun Wu 1 • William Johnson 1 • Megan Stoltz 1
Cynthia Lum
[email protected]
Xiaoyun Wu
x [email protected]
William Johnson
[email protected]
Megan Stoltz
[email protected]
Department of Criminology, Law and Society, George Mason University, 4400 University Drive, MS 6D 12, Fairfax, VA 22030, USA
~ Springer

LAP Regularly Pulled Over Innocent
People Because License Plate
Readers Flagged Their Cars As Stolen
re JASONKOEBLER • JUL13,2026AT1:18PM
The Los Angeles Police Department let its Flock contract expire over the weekend in part because it was regularly "investigating" and surveilling innocent people.

IMAGE: LAPD

The Los Angeles Police Depaiiment (LAPD) announced it will let its surveillance contract with automated license plate reader company (ALPR) Flock expire, becoming the largest police department in the country to drop its contract. Notably, the decision came after an audit of ALPR technology found that, in a two-month period, the LAPD had improperly "investigated" 161 people whose cars were flagged as stolen in the LAPD 's ALPR system but were not actually stolen.

The news that LAPD pulled over 161 innocent people in two months because of improper tagging in the department's system comes after several high-profile incidents in which people in other states were accosted by police because of data entry or clerical en-ors in ALPR systems. Joel Feder, an editor of the car journalism website The Driv~, detailed a ha1Towing tale in which he was tracked for days and ultimately pulled over by police in Minnesota because the license plate of the car he was reviewing for the website had been entered into the Flock system as stolen by a police depatiment in California. Monday, the website MotorBiscuit wrot~ about an innocent woman who was jailed for 13 days because she drove a black Dodge Durango and police searched the Flock system for a Black Dodge Durango suspected of being involved in a fatal hit-and-run accident.

C~ Determined Not Stolen
When records are not updated in a timely manner, officers may act on inaccurate or outdated information, increasing the risk of \lnnecessary enforcement actjons, including vehicle s·tops and wrongful detentions, or a co-nfrontation with serious cons.equence.q_ Department policy dictates that officers attempt to verify all records prior to conducting an investigative stop; however, conditions outside the officer's control can also impact a stop. For example, delays in updating records may cause a license plate to remain on a Hot List after the vehicle has been recovered or is no longer considered stol~n.

During the review period, officers acknowledged 161 alerts as accurate license plat-e matches; however~ subsequent investigations det,ermined 1he vehlcles were not stolen.

In addition to creating an inconvenience for vehicle owners, these inaccuracies can affect individual lib-erty interests1 erode public trust, and potentially create substantial legal and financial liability concerns.

IMAGE: LAPD OIG

A new report by the LAPD Office of the Inspector General (OIG) suggests that instances of people being falsely pulled over because their license plates have shown up on an ALPR "hot list" are very common, and that the surveillance of people on hot lists that ultimately result in no action from police is staggering. Many ALPR systems have this "hot list" feature, which is where police enter a license plate and get a ping or notification about the vehicle's whereabouts whenever it passes a connected ALPR camera. In a two-month period between August 1 and September 30, 2025, the LAPD's cameras generated more than 210.5 million license plate reads, according to the report.
"During the review period, officers acknowledged 161 alerts as accurate license plate matches; however, subsequent investigations determined the vehicles were not stolen," the report reads. "In addition to creating an inconvenience for vehicle owners, these inaccuracies can affect individual libe1ty interests, erode public trust, and potentially create substantial legal and financial liability concerns."
The report notes that this happened because of "inaccurate or outdated information, increasing the risk of unnecessary enforcement actions, including vehicle stops and wrongful detentions, or a confrontation with serious consequences," and that in many cases, license plates remained on a hot list after a stolen vehicle had already been recovered or was reported as not stolen, meaning the cops are in some cases pulling over the lawful owner of the vehicle.

Notably, the report states that when police get an ALPR hot list hit, the department generally considers any subsequent action to be a "high-risk" stop, meaning the risk of confrontation or potential danger is greatly increased from routine traffic stops for running a red light or speeding.
"When a license plate matches with a vehicle of interest on a Hot List, an ale1t will appear on the police vehicle's Mobile Digital Computer," the report reads. "Often, officers will approach the vehicle with extreme caution or conduct a 'high-risk' stop. This involves calling for back up, air suppott, and a supervisor and ordering the suspect out of their vehicle." The repoti says, "department policy requires officers to attempt to verify the accuracy of the ALPR alert prior to conducting a stop," but that often does not happen. The report also states that, on the vast Page 681 of 716 majority of hot list hits, no action is taken by police meaning that specific people are being subjected to tracking and surveillance for no readily discernible reason. In the two-month audit period, 5,911 different license plates were tracked. No action was taken against 4,575 of those cars.

The LAPD said in response to the report that cars improperly flagged as stolen "generally result from the timing of record updates outside of the Department's control, such as delays by another jurisdiction or a vehicle owner in clearing a plate from a Hot List after a vehicle has been recovered or is no longer wanted." In other words, LAPD is often relying on other police departments to remove license plates from a hot list, highlighting the problems with networking different surveillance systems together.

The LAPD OIG report, which appears to have directly led the LAPD to allow its Flock contract to expire, studied the use of three different ALPR systems the department has been using, including static, pole-mounted cameras from Motorola and Flock and cameras in police cruisers made by Axon. In total, the department has nearly 2,000 ALPR cameras; LAPD accesses data for both Flock and Axon systems through Flock's backend thanks to a data sharing partnership between Axon and Flock, according to the report. The report said the department was able to recover 3 3 7 stolen cars during the two months and that ALPR data led to 7 4 arrests total.

Both the OIG and the LAPD determined that the ALPR system needs to be reconsidered. The OIG suggested that the LAPD "suspend the deployment of new ALPR cameras and the execution of new ALPR-related contracts pending public input and a broader reassessment of vendors and data practices" and "strengthen oversight of ALPR data access." The LAPD allowed its Flock contract to expire over the weekend, and said it would not enter into new contracts until going through a full audit process.

She Spent 7 Months in Legal Hell After Cops
Used Flock Surveillance To Identify the Wrong
Car
From wrongful arrests to officers stalking exes, license plate cameras keep exposing police failures.

JOE LANCASTER I 8.13.2026 10:00 AM
( Add Reason to Google)
Olisten to this article
18 min
On October 4, 2025, Joaquin Deno attended his
granddaughter's birthday party in Sanford, Florida. At
the end of the night, he headed to his daughter's
house.

Eight miles down Interstate 4, a speeding car swerved
(The Law Offices of Patrick J. McGeehan, P.A./yourdaytonalawyer.com) across multiple lanes at over 100 mph. It swiped Deno's motorcycle, then hit a Ford Focus, causing a pileup, before driving away. Deno died at the scene, as did a county official and his wife.

Witnesses said the speeding car was a Dodge Durango. Police found that a Flock camera-a type of automated license plate reader (ALPR)- had spotted a Durango nearby at the time, and they arrested the owner, 23-year-old Lindsey Isaacs.

But Isaacs wasn't involved in the crash, and officers had enough information at the time to exclude her as a suspect. Yet she still spent two weeks in jail and faced the threat of life behind bars for months because police didn't use basic scrutiny to be sure their cameras implicated the right person.

Much has been written about Flock cameras making mistakes- including cases in which innocent people were detained at gunpoint because a camera mistook a 2 for a 7, or misread the state on a license plate.

But it's worse than just imperfect software: All too often, police pull people over, anest them, and even jail them based on camera misidentifications that officers could have caught with basic scrutiny. There are also numerous examples of officers with access to Flock's database using it to stalk and surveil people in their personal lives.

While this technology has legitimate police uses, it's worth asking whether the police can be trnsted with it. ( O"' Listen to this articl~
t.~ 18 min
I:)""' ........... - ~
X

"The first license plate reader was invented in 1976, but the devices became common tools for law enforcement over the last 20 years or so," reports The Marshall Project. "Fundamentally, they are cameras that capture pointin-time images of license plates on public roads and store those sightings in a searchable database." "ALPR technology can also detect and capture additional, related infotmation, including vehicle type and color, global positioning system (GPS) location data, and date and time," adds a Congressional Research Service reP.Ort.

Often mounted on freestanding poles along public roads, ALPRs can record thousands of tags per minute.

Investigators can get alerts when vehicles are spotted or search the database for past sightings.

Even when used appropriately, that level of surveillance is troubling. "What makes ALPR[s] so powerful is that drivers are required by law to install license plates on their vehicles," explains the Electronic Frontier Foundation. "In essence, our license plates have become tracking beacons." The problem has only grown over time. "Tens of thousands of LPRs are being used by law enforcement agencies all over the country," Ars Technica wrote in 2012. "But the system's unchecked and largely unmonitored use raises significant privacy concerns.
11 That same year, internal emails obtained by the American Civil Liberties Union revealed the FBI had paused camera purchases because the Office of General Counsel was "wrestling with LPR privacy issues." (The author of that email expected the issue to clear up soon, noting, "The program is still urrm,ino- :::inl'i WP Pninv trPmPnnnm: TIP.In imnnnrt 11'1 Listen to this article
18 min

There are now ALPRs taking photos of cars all over the country, every second, and in nearly every major American city. Flock Safety, by far the largest ALPR supplier, touts "more than 120,000 cameras across 49 states" that cap~ "more than 20 billion vehicle reads each month." Flock say..s. its cameras read plate characters with 96 percent accuracy-but with six or seven characters in each license plate, that translates to a much lower accuracy rate for identifying a specific car. It's not unreasonable to conclude that Flock misreads billions of license plates every month. Consider this real-world example: The police department in Roseville, California, found that 71 percent of its Flock alerts were incorrect.

But even when ALPRs provide accurate readings, their scope is shockingly broad, as is their potential for misuse or abuse.
"It's much bigger than a camera owned by a police department taking pictures of the license plate," says Cris van Pelt, founder of hayeibeenflocked.com, which lets users see if their plate has been searched. "It's a national network. It doesn't matter how small your town is, you're probably running a million searches in a month, from everywhere. A lot of people are just not aware of how networked and interconnected all of this is." Investigators used Flock to find Lindsey Isaacs, but plenty of shoddy police work went along with it.

Lindsey Isaacs' 7 Months in Legal Hell
For weeks after the deadly crash, the Florida Highway Patrol (FHP) told the nublic that troopers had located the Durango in a parking lot and were still looking for the driver.

But police had found Isaacs right away. The "parking lot" where they found her car was the one outside her apartment, where, according to a search warrant .l!P-P-lication, an FHP investigator observed "damage ... consistent with" the crash.

The FHP seized the Durango within hours of the accident but then just kept it impounded. Police did not arrest Isaacs until six months later, after she sued seeking the car's return. (The court then dismissed her request, since the car was now evidence.)
Isaacs faced eight felony counts, including three counts of vehicular homicide. Conviction would mean life in pnson.
"I was terrified," Isaacs tells Reason. "I didn't know ifI was going to get hurt in there or how I was going to get treated, where I was going to be, how long I would be in there." "I didn't want to be alive," she recalls. "It was a nightmare." Isaacs was originally held without bond, and she ~pent 13 days in jail before the state granted her release on $250 (l(l() h<:111 nnh, ,:,-ftp1• l,pr ,:,rrPd rlirl inuPdiCT<:>tnr" -fn1· ""r""l, r P f'nn.,tn,f'tinn <inrl inuPdiCT<:>tiuP """""rt" folrp Listen to this article
Q.
18 min

Immediately, the case fell apart: Isaacs' Durango, purchased two months before the accident, still looked new, with no sign of the "damage" the investigator reported. The Ford Focus, meanwhile, was heavily damaged, with "red or maroon paint transfer"- not a match for Isaacs' black car. One witness to the accident said the driver's airbag went off, which Isaacs' had not. A 911 caller also reported a partial license plate number that did not match Isaacs' tag.

Listen to this article
18 min

Listen to this article
18min
.7<1'.-.~~~T.--,;u~~ff""r~,a.~,._..,-,.~.-.,r...1,.,.,...,. _.,....__, ___ J..-s.r~-..............-~.-- w-.~fl'\~~~..:.___....,;

All of this information was available on the day of the accident. Instead, investigators simply searched Flock's cameras for a Dodge Durango and seemingly went with the first one they found, ignoring or not pursuing contradictory evidence and claiming "damage" on a car that had none.

Seven months after the accident, prosecutors would dismiss charges against Isaacs when police arrested Alisa Montalvo, a Deno family friend who had been at the birthday party. When investigators searched Montalvo's car -a red Durango-they found "significant evidence of concealment and damage consistent with this crash." The paint and window tint on the driver's side doors didn't match the rest of the car, one airbag had been removed, and another was "stufed [sic] back into the seat."
Isaacs' arrest was a particularly egregious case, but ultimately, the Flock camera performed as it was asked to; officers' carelessness caused the trouble. Unfortunately, it's far from the first time.

Police Incompetence, Not Flock Mistakes, Is
Getting Law-Abiding Citizens in Trouble
In June, Joel Feder and his wife were running weekend errands in their Minneapolis suburb in a brand-new $155,000 Range Rover. As he shifted into reverse to leave a parking space, the car started making noise. "We maybe backed up two feet," he tells Reason. "All of a sudden, all the sensors go off like [I'm] about to run over something."
Four police cars had converged, pinning them in. Officers, with hands on their guns, ordered them out of the car and patted them for weapons. They said the Range Rover was reported stolen in Los Angeles; Flock spotted it in town two days earlier, and police raced over when the cameras caught it at the department store.

When an officer asked Feder if the car was his, he hesitated: "It's complicated," he replied.

In fact, the car was not his, but it also wasn't stolen: Feder is a journalist who reviews cars for the automotive website The Drive. He had documentation showing the Range Rover was legally in his possession, but that still didn't explain why it was reported stolen.

It turned out the Los Angeles officer who took the stolen vehicle report had only keyed in part of the plate number. Those numbers were also on Feder's tag- along with, presumably, many others across the country- so Flock flagged it. (Worse, the original car wasn't even stolen: The dealer simply lost the license plate and had to renort it as a precaution.)
Incidentally, Feder recognized what was happening because The Drive had renorted onALPRs, and he hadjust interviewed security expert Matt Hurewitz about their privacy implications. "I think broadly, privacy only becomes a consumer concern when something bad happens and they get caught up in it," Hurewitz cautioned.

Listen to this article
18 min

The Institute for Justice (1.J.) has identified "at least 27 such cases since 2018, with the majority of those happening since 2023." While some resulted from cameras misreading plates, most stemmed from police incompetence.

Chrisanna Elser, a Denver financial planner, received a court summons in September 2025. The officer said she was on camera stealing a package off a front porch. What ensued was a Kafkaesque scene in which Elser insisted it wasn't her, and the officer said he had indisputable proof but refused to show her unless she admitted she stole the package.

In fact, the "evidence" consisted of Flock spotting her truck in town around the time of the theft, and doorbell camera footage of a woman stealing a package and running away. Police only dropped the charge once Elser provided her cellphone tracking data, camera footage from her truck, and video placing her elsewhere at the time of the theft.

In each of these cases, plus numerous others, the cameras did not make the mistake; officers either keyed in a number incorrectly or spent too little effort confirming their first hunch. But they pale in comparison to intentional officer misconduct.

Officers Are Using Flock To Stalk Their Exes
Earlier this year, Milwaukee police officer Josue Ayala resigned after he was charged with attempted misconduct in public office. Over a two-month period, P-rosecutors say Ayala searched his girlfriend's license plate 124 times, and her ex-boyfriend's plate 55 times, in the Flock database. Each time, Ayala wrote "investigation" as the reason for the search.
"This case is exactly why Flock builds our technology to include immutable, transparent audit trails, so rare cases of potential abuse can be detected, investigated, and addressed," Flock said in a statement provided to local news.

But Flock's safeguards didn't catch Ayala. The ex-boyfriend searched his own plate on haveibeenflocked.com and discovered Ayala had searched it "on numerous occasions," according to court records.

Officers have been caught using ALPRs to stalk or surveil intimate partners at least 39 times in recent years, most of which have occurred since 2024, according to I.J. How many more abuse the system and aren't caught? Michael Steffman, chief of police in Braselton, Georgia, stalked his ex-girlfriend by searching both her and her daughter's tags hundreds of times. In October 2023, Lee Nygaard resigned as the police chief in Sedgwick, Kansas, amid an investigation into misconduct: From June to October that year, Nygaard searched his exgirlfriend's plate 164 times-more than once per day, on average-and her new boyfriend's plate 64 times in the Flock database. When conducting the searches, Nygaard listed reasons such as "missing child" and "narcotics ir Listen to this article
18 min
rec... Page 689 of 716
"It's good that people are finding out that these incidents are happening, because I think for a long time that has been completely inaccessible," van Pelt, haveibeenflocked.com's creator, tells Reason. "Dollars to donuts, this has happened many, many times before with other tools. It just hasn't been transparent in any way." Indeed, 404 Media reported that an internal memo from a Houston-area police intelligence task force said haveibeenflocked.com "poses a significant officer safety risk," and officers should "be as vague as permissible" when listing reasons for their searches- for example, by simply writing "investigation." Cops Say Flock Fights Crime. Critics Say It's a
Surveillance Dragnet.

Clearly, Flock misuse and abuse takes a toll on its targets. "Every day I do some sort of therapy, whether it's physical activity or with a therapist," Isaacs says. "I have a wonderful family, wonderful attorneys .. . .I wouldn't be here without them." She is suing the FHP over her experience.

Flock's defenders say the system is being used to catch more criminals and reduce crime rates. In an interview with Feder, Flock CEO Garrett Langley said "any level of abuse is unacceptable." But, he added, "over a million cases were solved last year with this technology, and the level of abuse is incredibly low relative to that." Flock also claims its cameras helped locate "more than 10,000 missing persons" in 2025, "totaling more than 27 per day."
Flock say~ its cameras "are not mass surveillance tools" and "do not and cannot track vehicles, much less individual people." Rafael A. Mangual of the Manhattan Institute similarly dismisses "the notion that local governments are using these cameras to compile long-term logs of Americans' movements." "I think those talking points are just pure nonsense," says Michael Soyfer, an I.J. attorney. "Cars don't commit crimes, people commit crimes. That's why the cameras are taking pictures of cars and noting the time, location, all of that."
Even if Flock cameras and ALPRs don't literally "track" vehicles, they capture enough information that the distinction becomes meaningless. "The coverage is extensive enough that you can connect the dots and spot patterns in people's movements or even spot when people are going to intensely private, sensitive places," Soyfer adds, "because many cameras are near sensitive places like immigration lawyers' offices, gun stores, places of worship, [ and] reproductive health care facilities."
In recent years, departments have shared Flock information with other states and federal agencies. Police in Illinois conducted thousands of searches seemingly on behalf of federal immigration authorities, while Sacramento shared ALPR data with other states, including those that criminalize abortion-even though both places forbid such sharing.

F Listen to this article
18min
only \.YYV .:>Vl\ .. U ,1V1.1.:) . UVL.\.VJ. 5 ua.1.u1.a.u.:, V.l J.\.I YY\.11 V U .11.1\,,,.lc.&L).

Last year, I.J. launched the Plate Privacy Project. So far, it has challenged ALPRs in court and proposed model legislation on their use, which, among other things, requires officers to get a warrant to access any data more than 24 hours old.

The Los Angeles Police Department recently dro1wed its Flock contract when an audit found that in just two months, officers pulled over 161 cars incorrectly flagged as stolen. Other police departments in California, Oregon, and Ohio have also dropped or suspended use of Flock.

Some people have taken matters into their own hands, destroY.ing, sabotaging, and even stealing Flock cameras in their cities. Short of that, citizens can be diligent about the surveillance they're subjected to on a daily basis.

Websites such as DeFlock and Atlas of Surveillance let users see what forms of surveillance are in use around them. Haveibeenflocked.com lets users search police data, but it only contains what agencies have seen fit to release in response to specific records requests.

Are legal guardrails enough, or should the cameras come down altogether? "We have a substantial amount of safeguards and guardrails already in place," van Pelt says. "At every single level, every single safeguard that we already have is being ignored and it's failing. I don't know that adding more is going to help." Listen to this article
18 min

Flock Safety
Public safety and accountability

CITY COUNCIL PRESENTATION
Brian Shab, Chief of Police
David Rodriguez, Assistant Chief of Police

02 Our Deployment
37 CAMERAS
$124, 650K]

ANNUAL COST
[# 2 YRS]
TERM
30 / 7 DAYS]

RETENTION

03 What It Sees
A vehicle observation

04 A Search

05 Every Search Is Logged

06 Our Policy
Who can use it. Why. How we verify.

07 California Law
Clear limits and required safeguards

08 The Legal Landscape

CARPENTER
2018

NORFOLK
2026

SAN JOSE

PENDING

09 Carpenter
Important, but a different technology

10 Immigration
Policy, law and system controls

11 Security
Layers of protection

12 Misuse
A human problem

13 Stronger Safeguards
The system has changed

14 Our Results
[7818]

SEARCHES
[894]
CASES

ASSISTED
[-45.5%]
Auto Theft
[~$500,000]
VALUE

RECOVERED

13 Behind the Numbers

COUNCIL
DISTRICT 1

FLOCK-ASSISTED

INVESTIGATIONS
9/15/2025 – 9/15/2026
20 INVESTIGATIONS

SERIOUS CRIMES INVESTIGATED

WEAPONS / SHOOTING
25WHOMICIDE
25WAGGRAVATED ASSAULT / PC 245
25WASSAULT / BATTERY
25WWEAPONS
26WRAPE / SEX OFFENSE
26WDOMESTIC VIOLENCE / ASSAULT
26WRAPE / SEX OFFENSE / DOMESTIC ASSAULT
26WRAPE / SEX OFFENSE
26WASSAULT / BATTERY
26WRAPE / SEX OFFENSE / ASSAULT
26WDOMESTIC VIOLENCE / ASSAULT / SCHOOL
26WWEAPONS / TERRORISTIC THREATS
26WRAPE / SEX OFFENSE (MINOR)
26WASSAULT / BATTERY
26WWEAPONS / SHOOTING
26WASSAULT / BATTERY
26WWEAPONS / SHOOTING
26WSTALKING
26WWEAPONS / SHOOTING
26WCase numbers redacted Page 708 of 716

COUNCIL
DISTRICT 2

FLOCK-ASSISTED

INVESTIGATIONS
9/15/2025 – 9/15/2026
10 INVESTIGATIONS

SERIOUS CRIMES INVESTIGATED

WEAPONS / SHOOTING
25WINDECENT EXPOSURE / LEWD
25WKIDNAPPING / ASSAULT
26WASSAULT / BATTERY
26WWEAPONS / ASSAULT / ATTEMPT MURDER / CHILD ABUSE
26WDOMESTIC VIOLENCE / ASSAULT
26WASSAULT / BATTERY
26WCHILD ABUSE / NEGLECT
26WASSAULT / BATTERY
26WWEAPONS / SHOOTING
26WCase numbers redacted Page 709 of 716

COUNCIL
DISTRICT 3

FLOCK-ASSISTED

INVESTIGATIONS
9/15/2025 – 9/15/2026
5 INVESTIGATIONS

SERIOUS CRIMES INVESTIGATED

WEAPONS / SHOOTING
26WHUMAN TRAFFICKING / SEXUAL ASSAULT MINOR
26WROBBERY
26WRAPE / SEX OFFENSE
26WASSAULT / BATTERY
26WCase numbers redacted Page 710 of 716

COUNCIL
DISTRICT 4

FLOCK-ASSISTED

INVESTIGATIONS
9/15/2025 – 9/15/2026
7 INVESTIGATIONS

SERIOUS CRIMES INVESTIGATED

RAPE / SEX OFFENSE
25WDOMESTIC VIOLENCE
25WWEAPONS / SHOOTING
26WTHREATS / HARASSMENT
26WROBBERY
26WWEAPONS / SHOOTING
26WSTALKING
26WCase numbers redacted Page 711 of 716

COUNCIL
DISTRICT 5

FLOCK-ASSISTED

INVESTIGATIONS
9/15/2025 – 9/15/2026
11 INVESTIGATIONS

SERIOUS CRIMES INVESTIGATED

DOMESTIC VIOLENCE / ASSAULT
25WRAPE / SEX OFFENSE
25WATTEMPTED HOMICIDE / SHOOTING
25WRAPE / SEX OFFENSE
26WRAPE
26WWEAPONS
26WWEAPONS / SHOOTING
26WDOMESTIC VIOLENCE / ASSAULT / CHILD ABUSE
26WDOMESTIC VIOLENCE / ASSAULT / KIDNAPPING
26WDOMESTIC VIOLENCE / ASSAULT
26WRAPE / SEX OFFENSE / EXTORTION
26WCase numbers redacted Page 712 of 716

COUNCIL
DISTRICT 6

FLOCK-ASSISTED

INVESTIGATIONS
9/15/2025 – 9/15/2026
12 INVESTIGATIONS

SERIOUS CRIMES INVESTIGATED
AGGRAVATED ASSAULT / PC 245
25WDOMESTIC VIOLENCE / PC 273.5
25WWEAPONS / SHOOTING
26WATTEMPTED HOMICIDE / WEAPONS
26WRAPE / SEX OFFENSE MINORS (MULTIPLE VICTIMS)
26WHOMICIDE / WEAPONS / ASSAULT
26WASSAULT / BATTERY
26WASSAULT / BATTERY
26WRAPE / SEX OFFENSE / DOMESTIC ASSAULT
26WRAPE / SEX OFFENSE
26WWEAPONS / SHOOTING
26WROBBERY
26WCase numbers redacted Page 713 of 716

COUNCIL
DISTRICT 7

FLOCK-ASSISTED

INVESTIGATIONS
9/15/2025 – 9/15/2026
5 INVESTIGATIONS

SERIOUS CRIMES INVESTIGATED
RAPE / SEX OFFENSE / PC 261
25WRAPE / PC 261
25WDOMESTIC VIOLENCE / ASSAULT
26WSTALKING
26WASSAULT / SEXUAL BATTERY
26WCase numbers redacted Page 714 of 716
15 Cost and Value
Some outcomes resist a price tag

16 Two Families