This Technical Appendix is a supplement to the report Policing Veterans: An Analysis of Veterans Affairs Police Department `Incidents,', UCLA Veterans Legal Clinic and National Association of Minority Veterans of America (NAMVETS), March 2024, (hereinafter, the "Report"). It provides additional detail about the electronic incident type data used in the Report and the methodology for processing that data. 

In November 2021, we published a report that examined the types of incidents in which the VAPD officers were involved (hereinafter "November 2021 report").1 That report covered VAPD police encounters from the period May 1, 2019, through June 27, 2021, in four VA health system locations. The current report follows the same methodology, covering the same locations, but utilizes incident data extending through September 30, 2022.2

Methodology

The VHA produced the data relied upon in this report pursuant to negotiations following a Freedom of Information Act (FOIA) request between the federal government and UCLA Veterans Legal Clinic on behalf of NAMVETS.

We requested reports on general incidents involving the VAPD covering four VHA locations: Greater Los Angeles Health Care System, Tampa Health Care System, Central Ohio Health Care System, and St. Albans VA Medical Center in Queens, NY.3

Pursuant to this request, the VHA produced Excel spreadsheets for each of the four requested locations that provide data identifying the type of incidents recorded by VAPD officers following policing encounters. These data record incidents from encounters occurring between May 1, 2019 and September 30, 2022.4 5 

Following an encounter, which may involve one or more individuals, each VAPD officer involved prepares an "Incident Report" in an electronic database to record events of that encounter. Each Incident Report has a unique report reference number.

As part of the Incident Report, for each individual involved, the officer will record one or more “incident types."

The incident type is entered into the VA’s database through a “tree” of descriptions from which the VAPD officer can select from a drop-down menu when filling out an Incident Report on the VHA incident reporting system.6  For example, in incident type INCIDENTS : 13A - 13C ASSAULT OFFENSES : 13B SIMPLE ASSAULT / BATTERY : BATTERY (MINOR INJURIES), the colons separate different levels of the tree.7

The data files produced by the VHA include, for each Incident Report, the Incident Report reference number, the date the Incident Report was created by the VA police officer, and one or more incident types.8 Because incident types are recorded for each individual associated with the encounter, a particular incident type could appear multiple times under a given Incident Report reference number if that incident type applied to more than one individual involved in an encounter.9

The data provided does not contain sufficient information to ascertain the numbers of individuals involved in each encounter, or to associate incident types with particular individuals within an Incident Report. Therefore, for the analysis in this Report, and following the approach taken in the November 2021 report, we treat each record of an incident type as a unique "incident" in our analyses. That is, we assume these are no erroneous duplicates of reported incident types.

The Greater Los Angeles Health Care System and the Tampa Health Care System each comprise more than one facility; the Queens location and Central Ohio Health Care System each comprise only a single hospital (St. Albans VA Medical Center and Columbus Chalmers P. Wiley Ambulatory Care Center, respectively). Thus, the incidents data from Los Angeles and Tampa included thousands of documented incidents, while Queens and Columbus each reported less than 800 documented incidents. 

For the November 2021 report, the VHA produced a list of 454 “Key Incident Types,” from which the police officer could select for recording their interactions into their Incident Reports.10 For that analysis, we reduced these 454 types into 18 incident categories and 68 subcategories.

To do so, we examined the language of each incident type and, when applicable, the text of the respective law. For example, when the incident type language included federal criminal statutes, federal regulatory crimes or Uniform Crime Reporting codes, we examined the natural language of the statute, regulation, or code.11 We then weighed two factors while categorizing incident types. The first factor grouped incident types based on behavior. For example, trespass and loitering incident types were grouped into the same category ("Spatial Control") because they concern similar behavior (unlawfully remaining on property). The second factor grouped incident types based on policing goals.12 For example, disorderly conduct and traffic violation incident types were grouped together within the same category (also "Spatial Control") because these incidents describe common ways police regulate the campus physical space itself and its surrounding areas. 

For the current Report, we use the same methodology for categorizing incident types.

At various times, and in particular, starting with the 2022 fiscal year (October 2021 through September 2022), the VHA changed the language of incident types among which the officers can select. In most cases, this Report captures the change through a simple categorization remapping.13 In other cases, where the VHA created new or more detailed incident types, we assigned the incident types to already-existing categories and subcategories if appropriate.14 

Given the expanded data and revisions to incident type language, for the current Report, we have grouped the incidents into 18 categories and 73 subcategories. The complete list of unique incident types that were applied by police officers as reported in the provided incidents data, along with the categories and subcategories to which we assigned them are available here.

After creating the categories and subcategories, we calculated the number of incidents within each category and subcategory for each of the four locations. These tabulations are presented in the Report.

(Return to Main Report)