It would be valid given those assumptions. If the assumptions are reasonable, then the outcome is reliable.
Ah, so your position is that there is no racial bias in policing, and that claim does not need to be backed up by evidence…
While I don’t take your position as a reasonable prior, given what we know about the history of racial discrimination in the US (including and especially official discrimination by the police), even if we were to take it as a prior, evidence like the study presented, where behavior is controlled for, should make you question your priors. We have evidence that race, when isolated as a causal factor, is playing a role in peoples treatment by the justice system. In light of that, it’s not reasonable to accept, without evidence, that race does not play a role. Everywhere that people are making judgement calls (including when police decide that someone is being aggressive, or resisting, or not cooperating, etc.), race can enter as a factor.
Lefties are minority, are they a race?
Technically, Hispanic is not the same as Latino. While that technical meaning may not always track colloquial use, for the purpose of interpreting this survey and the data on which it relied, I think it’s safe to assume (!) that a study looking at demographic statistics is using the technical and not the colloquial meaning.
I think you’re making unreasonable demands. Studies into these questions are limited by the data we have. We can’t get an infinite amount of data on each case, and we can’t control for an infinite number of variables. But “could have been done differently and with more accuracy” is not the same as “is no good and inaccurate”.
And let me say, I may be coming off as too dismissive of your deep dive into methodology and demand for rigor. I don’t mean to be, it’s a good instinct and I applaud it. But it’s only valuable where it places realistic demands on researchers, where it is applied evenly to reject any study no matter the conclusion, and where the methodological problems you find can reasonably be expected to affect the answer to the question presented.
This study is limited by the data, it makes some assumptions about the nature of crimes based on the statute under which they are prosecuted, and it does not answer any questions about the independent role that Hispanic ethnicity plays in sentencing. But you don’t have a study that goes farther and finds a different answer. You asked for a study that looked at sentencing disparities controlling for criminal history. This study does.
That shouldn’t be the goal. Though I’m not so naive to think that researchers don’t set out with an end in mind, that doesn’t mean that we have to do the same. Look for studies that solve the methodological problems you have with this study, and then see what they find.
Let’s keep going. I think we’ll find that there’s a pattern where you reject reasonable studies for not meeting your impossible standards, present no competing studies – studies that 1) meet your standards, and 2) show that meeting your standards was sufficient to change the finding-- and then claim that you need no evidence for your position. Which is what you’ve done for this study. Let’s see if the pattern holds.
Let’s look at the study listed under “Findings of the Use of Handcuffs”. The link in that article is to a WaPo summary, but the study is here, and the ~300 page technical deep dive into Oakland PD data is here. It has different methodology from the report we were discussing previously, addresses some of your concerns about behavioral differences, and presents similar findings (racial bias in policing) from different evidence. I think we should stick with the shorter report, but I include the latter because I know you appreciate rigor. I’ll refer to the two documents as SfC (‘Strategies for Change’, the summary report) and DfC (‘Data for Change’, the technical report). Page references are to pdf pages, rather than the number written on the page.
A few things that stand out to me:
- Related to your earlier claim, this study found race-bias in the rate of stops, even controlling for the fact that they are more active in areas with a higher black population: “we found no evidence that the OPD was specifically targeting African American neighborhoods. Instead, the OPD does target neighborhoods with higher crime rates. As the first analysis revealed, however, once in a neighborhood, OPD officers tend to stop more African Americans than is proportional to their representation in the neighborhood.” (SfC 11)
- On a related note, the disparity was greater when the officers knew the race of the suspect prior to making the stop, suggesting that race played a role in the interpretation of behavior: “We also found that when officers were able to identify the race of the person before stopping him or her, they were much more likely to be stopping an African American (62%), as compared to when they couldn’t tell the race of the person (48%).” (SfC 11)
- Handcuffing was much more common of black people than white people, even when they weren’t arrested: “We found that African American men were handcuffed in one out of every four stops, as compared to one in every 15 stops for White men. Even after controlling for neighborhood crime rates, demographics, and many other factors, our analyses showed that OPD officers handcuffed significantly more African Americans than Whites. This African American-White handcuffing gap was especially pronounced for vehicle stops and stops made because of traffic violations.” (SfC 11)
- Searches were also more common for black people than white people: “Excluding low-discretion searches, we found that officers searched African American men in one out of every five stops, as compared to one out of every 20 stops for White men. Even after controlling for neighborhood crime rate, racial demographics, and many other factors, our analyses showed that OPD officers were more likely to search African Americans than Whites.” (SfC 12)
- The police body camera language analysis is pretty interesting, showing that police used less formal language when speaking to black people, and used more severe legal language. That’s true even controlling for whether there was a parole-justified stop, and for the severity of the justification for the stop. (SfC 18-19)