Nov 28, 2016 (San Diego) The Report was released last Wendesday before Thanksiging. So we decided to take some time before running a breathless story on this that would quickly disappear. The report can be found here and I recommend reading it.
This is not the first time an outside agency is asked to review San Diego Police Stop data. The first study was done in 2000-2001 with a University of Kentucky criminologist. He, like the current study done with San Diego State researchers, found some problems with the data, that would bring it into question. In our mind, this is actually in the number of tickets issued and followed by the courts, and the self reported by officers.
The study states that between 2014-5 the span of the study:
the SDPD issued 183,402 citations over this two-year period, a sum 26.1 percent greater than the 145,490 citations logged by officers via the traffic stop data card. As is shown in Table 3.4, we used stop card citation rates for each racial/ethnic group to generate rough estimates of unreported traffic stops. All told, we estimate that the SDPD conducted somewhere between 60,000 and 70,000 traffic stops for which no stop card information was submitted.36 We do note that the racial/ethnic composition of the stop card citation records largely reflects the composition of the actual citations issued, which suggests that the under-reporting was not race-determinative.
There is a hole in the dataset. This matters, The authors of the study also found that many of those holes are found in majority minority communities, such sa South East. The study authors do ask the obvious question. Was this an attempt at data manipulation? To be fair, this is not uncommon with police agencies nationwide. There is a certain level of resistance by officers to oversight by those outside police agencies. This will not help with community relations. The study authors do ask if a downward trend in recorded stops :indicate overt race-driven data manipulation.”
Now there were other findings that are equally troubling and point to a lack of compliance with internal department policy.
In 2015, 21.1 percent of stop cards were missing at least one piece of information, with nearly half of those missing both demographic and post-stop information. A significant spike of stop cards missing both field interview and citation data occurred between March and August of that year, raising questions about the quality of these data during that period. We also note that the volume of missing data increased as monthly stop totals reached their lowest levels. In other words, the quality of the stop card data declined across the year along with the number of both recorded stops and searches.
We have to wonder if the department will enact policies with corrective goals. This is not the first study to find this.
The study also found that same tension that earlier studies have found. This includes the Police Executive Research Forum (PERF) requested after the sexual abuse misconduct emerged a coupipe years ago. It was requested by then Chief William Landsdowne We also linked to that study in our coverage when it came out. You can find the link in the article.
One of the recommendations from that report seem consistent with the findings in this stop data, and not filling up the forms. There is no consisnstent disicplining and this is straight from the PERF report:
During PERF’s interviews with SDPD personnel for this assessment, one employee said that “there are inconsistencies in both discipline and rewards [within the department].”
So back to the stop data report.
It did found a greater incidence of stops for minority drivers, south of the 8, with the exception of Northern Division, covering La Jolla. This division stands to have higher stop data for Blacks. It is critical to explain the technique used by the researchers, to try to even out the data, and actually get a better data. This is what is called the veil of darkness
The “veil of darkness” technique allows researchers to avoid the difficulty of identifying and applying a benchmark – a point of reference, such as Census data – against which to compare traffic stop data. This is the central challenge in the review of such data, as the driving population of a given area may look quite different from the residents of that area, as counted by the Census. Instead, using the veil technique, analysts can examine the likelihood that, for example, Black drivers will be stopped during the day versus at night, and compare that likelihood with the day-versus-night likelihood of White drivers being stopped.
The researchers also used broken equipment, such as a non functioning tail light, since they also comprise the makjority of stops, and are highly discretionary.
The study found that:
Table 4.3 displays the results of our analysis of discretionary traffic stops conducted in the City of San Diego between January 1, 2014 and December 31, 2015 involving Black and White drivers. The data show that in 2014, when driver race/ethnicity was visible, Black drivers were nearly 20 percent more likely to be the subject of a discretionary traffic stop than were White drivers. When confined to drivers aged 25 and under, young Black drivers in 2014 were 43.8 percent more likely to be stopped in daylight than after dark, compared to young Whites. These findings are statistically significant at the 0.01 level and thus indicate racial/ethnic disparity in the distribution of traffic stops.
Curiously that same disparity did not exist in the 2015 data. When combined there is no statistical change that can be found in this. As a reporter I need to ask if police officers changed their behavior knowing that there was a study under way? Or if they are adjusticing their behavior due to awareness of bias in policing?
Then there are differences between divisions. Central division, where the highest crime rates exists, have a higher stop rate. Hispanic drivers face less of a chance of getting stopped, according to the data. Nor are there any differences between Asians and Pacific Islanders when compared to whites.
The conclusion also goes against the perceptions.
Data on stops conducted below Interstate 8 reveal a different set of results. We find substantial evidence to suggest that in the aggregate, both Black and Hispanic drivers were less likely be stopped during daylight hours than they were after dark, compared to stops involving White drivers. In other words, when the police were able to see a driver’s race, they were more likely to stop a White driver than they were a Black or Hispanic driver. At the division level, these results were evident in stops occurring in the Central division and among Hispanic (but not Black) drivers stopped in the Mid-City division.
Where things also get more radicalized if you will is with the outcomes of the stops. According to the report:
Table 5.1 lists by police division both vehicle stop totals and the incidence rates of key post-stop outcomes. In the Northern division, police conducted a search in 3.3 percent of 37,203 vehicle stops, or 1 in 30. Contrast that with the Southeastern division, where 1 in 10 stops resulted in a formal search – three times the rate in the Northern division. The same kind of variance is present in other raw post-stop data. Drivers stopped in the Western division are more than twice as likely to face a field interview (FI) than are drivers stopped in the Eastern division. A similar pattern is visible in citation rates: 45.1 percent of stops conducted in the Northwestern division resulted in the issuance of a ticket, compared to almost 70 percent of stops in the Southern division.
While the study does not mention it, we must ask the obvious question. Were many of the stops in the Sourthern divisions 4th Wavers? That is people who are out on parole and have no 4th ammendment protections? This could account for some, but not all of the variance. It would match the era of mass incarceration toughl
They also match areas of the city with much higher rates of patrol officers, to match the allocation of resources to higher crime areas.
We also must ask if some of the policing choices, searches and ticketing is due to broken windows policing in higher crime areas? As well as assumptions made about young black men versus an elderly white woman.
What is true is that city wide the study did challenge preconceptions that black young men are going to be stopped more often across the city. That does not apply by divisions though. Those south of interstate 8 will match national observations far better. We would hope the city would commission a longer study. The data set has issues, and the recommmendations from the researchers are solid. The data set should include the race of the officer as well. Chiefly, we must see these cards turned in every shift and for every stop, and in complete form.