What is RIPA data?

 

The California Racial and Identity Profiling Act (RIPA) Board has issued its 2022 Report analyzing police stop activity in 2020, and it’s pretty good.

 

RIPA requires that all police departments in California begin collecting a consistent set of data on every stop that police officers make. The largest departments have already started, and all departments will be collecting the same data in 2023. This data then gets reported to the state, and eventually becomes public.

Understanding RIPA Data

To go any deeper than just a headline understanding of the RIPA data from a police department, we need to know a little bit more about what is in the data, and how it’s collected.


The law requires data to be recorded at each police stop, and in fact lays out in precise detail what pieces of data must be included. Most departments use an app on phones or pads to collect the data. There are a lot of these apps; I don’t yet know which one the Los Altos PD is using. But since the data is specified so well in the RIPA rules, any compliant app should be pretty good. The idea is that during, or immediately after a stop, the officer uses the app to fill in all of the data fields appropriately. A few of those fields are “open text” fields where you can write whatever you want, but most are pull-down menus of choices. That is great for data integrity – it removes a lot of ambiguity about what was intended. One of these, RIPALog, has a convenient if simplified FAQ on the data.


Of course, the data is only as good as it is input. The 2022 RIPA Board report calls for all departments to do audits of their RIPA data, including using bodycam video to confirm that what we recorded was an accurate assessment of what really happened. We must interpret the results of the RIPA data in light of whatever audits may or may not have occurred. 


RIPA requires that each department submit this data to the California Department of Justice. It does not require that the department make it public. However, once it is ingested by the DoJ, it becomes subject to a California Public Records Request; like this one, for example. Los Altos has committed to making public releases of the data on a regular basis.


Many of the data fields collected are demographics. We will list them all at the end, but it’s important to remember that they are recorded as perceived by the officer. If someone identifies as queer, but the officer doesn’t perceive them as queer, then they will not be recorded as such. This is good and bad – it means that the data is not suitable for determining who is actually stopped. But perception is exactly where bias comes in; bias is defined as treating differently people perceived to fall into a particular group. So this should not be thought of as a shortcoming in the data, but it must be remembered in any interpretation of the results. It frames the questions that can be sensibly asked of the data.


Next are the details of where and when the stop took place. The timestamping is pretty easy, and probably automatic. The location is a little bit more challenging. For reasons of privacy, rather than actually using the geolocation services of the device to record a precise location, the officers enter a description of the location. This is often a street name, an intersection, or a block number. But because it is entered by hand, there is no consistency. Spelling errors, hasty entries, and name ambiguities make this data difficult to clean up for analysis. Furthermore, “Foothill Expressway” is not a very useful location; it can refer to points miles apart. Nevertheless, this is a State requirement, so it is not up to the determination of local departments.


Finally, there are the details about the stop. The definitions of these are not crystal clear; here I give my best interpretation. However, there could well be other official or unofficial interpretations.


I’ve listed all the fields below, but there are three main fields of interest: Reason for Stop, Actions Taken, and Result of Stop. The Reason for Stop is the thing that made the officers initiate the stop in the first place. The Actions Taken are what happened during the encounter, and the Result of Stop is what the final resolution is.


So, for example, if an officer stopped someone because of a broken taillight, then ordered the person to the curb and searched the car,  but found nothing, so let the driver go, you would have the following: Reason for Stop is a non-moving violation, Action Taken is Curbside Detention and Search of Property, and Result of Stop is None. By looking at the connections between these fields for each stop, we can learn about how the police operate differently for different racial groups.


  • “Reason for Stop.” This is the primary reason why the officer stopped the person. It doesn’t matter what they later found or learned; this is the justification for the stop. If there is more than one, they are supposed to choose the main reason. The prescribed options are:

    • Traffic violation. This should trigger further division into “Moving Violation,” “Equipment Violation,” or “Non-moving/registration violation.” Furthermore, they are supposed to record the particular vehicle code section that was violated. However, the publicly released Los Altos data does not have anything more specific than “Traffic Violation.”

    • Reasonable Suspicion. This is the legal threshold of believing that criminal activity might be taking place. Again, there is a list of subcategories that are required to be recorded, including things like “Officer witnessed commission of crime,” “Matched suspect description,” “Carrying suspicious object,” etc. Again, the Los Altos data does not contain any of these subcategorizations.

    • Known to be on parole/probation. This is only if the officer knows that before stopping the person.

    • Knowledge of outstanding arrest warrant. 

    • Investigation into whether the person is truant.

    • Consensual encounter resulting in search. This means that the officer is talking to someone who is not being detained. Despite being free to go, the officer asks permission to perform a search, and that permission is granted. Despite universal legal advice to the contrary, 95% of search requests statewide are granted.

  • Description of stop. Officers are required to describe the reason for stop in 250 characters or less. It should add detail on what justified the stop, and is explicitly forbidden to include any personally identifiable information. This field was not included in the Los Altos public data release.

  • Stop made in response for a call for service. This is a yes/no question, and it’s just asking if someone called the police, or if the police initiated the interaction. Note that if someone called the police and it did not result in a “stop” for one of the reasons listed above, it doesn’t count as a stop and does not appear in the data. 

  • Actions Taken. These are the things the officer(s) did in the course of the stop. They can pick more than one, if more than one happened. This is one of the most important fields, because it tells us about the actions of the police. Again, they choose from a list (more detail available in the regs):

    •  Person removed from vehicle by order

    • Person removed from vehicle by physical contact

    • Field sobriety test conducted

    • Curbside detention.

    • Handcuffed or flex cuffed  

    • Patrol car detention 

    • Canine removed from vehicle or used to search 

    • Firearm pointed at person  

    • Firearm discharged or used 1

    • Electronic control device used  

    • Impact projectile discharged or used (e.g., blunt impact projectile, rubber bullets or bean bags)  

    • Canine bit or held person 

    • Baton or other impact weapon used 

    • Chemical spray used 

    • Other physical or vehicle contact. 

    • Person photographed 

    • Asked for consent to search person 

    • Search of person was conducted

    • Asked for consent to search property 

    • Search of property was conducted. 

    • Property was seized 

    • Vehicle impounded 

    • None. None of the previous items occurred.

  • Basis for Search. If a search is conducted, the justification for that search must be provided. It is again a pulldown list, allowing multiple selections. The most common are “Consent Given” or “Incident to Arrest.” The complete list is in the regs.

  • Descriptive Basis for Search. Again, officers are to describe the search justification in more detail, in 250 characters or less, and cannot include any personally identifiable information. This field is also not included in the Los Altos data release.

  • Contraband or Evidence Discovered. Another pulldown list, with items like “firearms” and “drug paraphernalia.” 

  • Basis for property seizure and Property seized. These fields are not included in the Los Altos data disclosure. It’s not clear why; perhaps there was no property seized.

  • Result of Stop. This is another important field – it records the final outcome of the interaction. It’s again a pulldown, with multiple selections allowed:

    • No action. If “No Action” is selected, no other data values can be selected.

    • Warning (verbal or written) 

    • Citation for infraction 

    • In-field cite and release

    • Custodial arrest pursuant to outstanding warrant

    • Custodial arrest without warrant  

    • Field interview card completed 

    • Noncriminal transport or caretaking transport. This includes transport by an officer, transport by ambulance, or transport by another agency.  

    • Contacted parent/legal guardian or other person responsible for the minor 

    • Psychiatric hold 

    • Contacted U.S. Department of Homeland Security (e.g., Immigration and Customs Enforcement, Customs and Border Protection) 

  • Officer ID, Officer’s Years of Experience, Officer’s Type of Assignment. This information was disaggregated from the stop data for public release, to protect the identity of officers. The RIPA law allows additional protection of officer-identifiable data from public release.


The demographics collected are: Perceived Race, Perceived Gender, Perceived to be LBGT, Perceived Age, Perceived English Fluency, and Perceived or Known Disability.



Popular posts from this blog

What's going on in Sunnyvale?

RIPA Update for Los Altos

Is there bias in our policing?