Racial Profiling: Raw Numbers Mean Nothing

[IzzyR]

Report: Many Massachusetts Police Departments Disproportionately Ticket Blacks

BOSTON (AP) - A study of racial profiling, based on state statistics, shows minorities receive a disproportionate amount of traffic tickets in some Massachusetts communities.
The study, conducted by Northeastern University’s Institute on Race and Justice and released Tuesday, shows that 247 of the 341 agencies examined gave a greater proportion of tickets to minorities than expected given the racial makeup of the driving population.

A state law passed in 2000 requires police departments to report a variety of data on people given traffic tickets. The state began collecting the data in April 2001 and the study covered the period from then through June 2003.

State Public Safety Secretary Edward Flynn called the release of the data an “absolutely essential first step” in starting a dialogue on racial profiling in Massachusetts.

The data doesn’t prove that racial profiling is occurring in the state, Flynn said, but “the data provides us a basis from which to ask intelligent questions.”

I don’t think people should be stopped and ticketed simply for DWB. But I think it is a serious mistake to just assume that the number of traffic tickets has to reflect the same racial proportion as the general population. (Even the cop trying to defend his department here relied on the assertion that they get a lot of drivers from other communities). Unless there is actual evidence that black people are being ticketed at a disproportionate rate as compared to the number of traffic offences that they actually commit, it is utterly meaningless to just point to the number of total tickets.

Blacks are also arrested for murder and other violent crimes at rates far exceeding the general population. Does this show racial profiling in murder arrests? (I seem to recall various liberals on this board claiming that in fact it does.) I think this attitude towards traffic tickets is as silly as it is towards murder cases.

Which is not to preclude the possibility that in fact it does represent some form of racial profiling. But it cannot be used as evidence of such.

[MMI]

I would say that what the raw numbers “mean” is “look further”. I agree that the raw numbers of tickets issued may or may not indicate a problem with profiling. However I do feel that a substantial difference in tickets between two hardly monolithic groups across an entire state does indicate a need for further investigation.

[Cheesesteak]

@MMI

I would say that what the raw numbers “mean” is “look further”. I agree that the raw numbers of tickets issued may or may not indicate a problem with profiling. However I do feel that a substantial difference in tickets between two hardly monolithic groups across an entire state does indicate a need for further investigation.

It’s really tough to look further. New Jersey tried it by setting up radar vans with cameras. Basically, they randomly took pictures of speeders (>15mph over the limit), then had a panel conclude the race of the driver via the picture. When the data said that blacks sped more often than whites, the conclusions were derided, and more people were called racists.

I spent a lot of time looking at the NJ data and methodology, and couldn’t come up with any siginificant problems or a better way to do it. I’ll see if I can find a link to that original study.

[mhendo]

It’s certainly true that raw numbers don’t tell the whole story. Unfortunately, neither does the link provided by the OP.

Jack McDevitt and Lisa Bailey, who both work for the Northeastern University Institute on Race and Justice and who were involved in producing the study, wrote a piece in the Boston Globe in which they expanded on the resuts of their study. In particular, they noted:

There can also be costs associated with studies of racial disparities. Not all racial disparities are the result of profiling; some may be the result of legitimate enforcement practices. If, for example, members of a local neighborhood group from a predominantly African-American neighborhood ask the police to enforce traffic laws strictly because they are concerned about speeding cars and the safety of the youth who live there, the subsequent enforcement actions may disproportionately target African-Americans because they are the most common drivers in that neighborhood.

The disparity might look like racial profiling in the statistics, but it might be a legitimate law enforcement reaction to the traffic safety concerns of a community.

If studies jump to conclusions that are not supported by the data, the costs to the local police agency can be very high. Studies of racial profiling that incorrectly or inaccurately label law enforcement agencies as engaging in racial profiling can seriously damage the reputation of a department and consequently the ability of the agency to serve their community.

(Boston Globe, 2 August 2003, p. 19)

McDevitt and Bailey’s piece was designed not just to offer some general observations about the issue of racial profiling, but was specifically in response to an investigation that was carried out by the Globe on disparities in traffic tickets issued in the Boston area.

The newspaper examined 166,000 tickets and warnings issued in Massachusetts over a two-month period. Some of the findings are instructive, and give much more food for thought than a simple examination of tickets issued:

In the subjective encounter between officer and driver, the difference between a costly ticket and a pain-free warning is closely linked with the race, sex, and age of the driver.

<snip>

On city boulevards and rural lanes, whites are far more likely than minorities to receive written warnings instead of tickets when stopped for identical traffic offenses, according to a Boston Globe study of newly released state records. And women, especially young women, get breaks that aren’t afforded men. The price tag for this unequal treatment amounts to an estimated $25 million a year in traffic fines and higher insurance premiums. For a decade, the national debate over racial profiling has focused on raw accounts of indignity: stops for “driving while black,” unjustified searches, videotaped beatings. But in Massachusetts, the most routine use of discretion by police - writing a ticket or a warning - points to a subtler side effect: the economic impact on minorities and men, who are hit with traffic fines and insurance surcharges that others avoid.

That there is a pattern of differential treatment is clear.

Statewide, when local police cited drivers for speeding 45 m.p.h. in a 30 m.p.h. zone, the most common offense, whites drove away with a ticket 31 percent of the time, while 49 percent of minorities received a ticket. The figures are based on
an analysis of 166,000 tickets and warnings from every police department in the state in a two-month period, April and May of 2001. These were the only months for which the Registry of Motor Vehicles collected all warnings as part of a state-sponsored test for profiling.

<snip>

When factors of race and sex are considered together, the records reveal a tiered system of ticketing. Local police allow white women to drive faster without penalty, while reserving the harshest treatment for minority men. When drivers went 45 m.p.h. in a 30 m.p.h. zone, white women were ticketed 28 percent of the time in the two sample months; white men, 34 percent; minority women, 44 percent; and minority men, 52 percent.

The striking exception to this pattern was the Massachusetts State Police. The records show that troopers gave almost exactly equal treatment to all drivers, regardless of race, sex, or age. No local police department of any size was as fair as the State Police.

(Boston Globe, 20 July 2003, Metro/Region p. A1)

So, while raw data on the number of tickets issued to different groups may not provide much to go on, i think it’s considerably more troubling when minorities get more tickets than whites when stopped for the same offence. Actually, it’s also troubling that men get more tickets than women when stopped for the same offence.

I’m sorry i can’t provide links, but i found this information using a Lexis/Nexis search, and you need a subscription to gain access to the databate.

If anyone has Lexis/Nexis, you can find this easily by searching in “General News,” specifying the Boston Globe, and using search terms “profiling,” “speeding,” and “traffic.”

[IzzyR]

mhendo,

Thanks a lot - excellent post and valid point.

In the interests of clarification though, I should point out that Jack McDevitt and Lisa Bailey did not, as you state, write “a piece in the Boston Globe in which they expanded on the resuts of their study” - their study was not completed at the time (8/03) - as noted in the OP, it was released yesterday. What they did was write an Op-Ed piece in support of the Boston Globe’s study that you refer to in your second quote.

Here are the links:

Op-Ed: Looking Deeper at Racial Profiling

Race, sex, and age drive ticketing

So I don’t think your (apparent) complaint that the linked OP article misrepresented the study is a valid one - unless you can show that this study too looked at tickets vs. warnings.

In sum, the Boston Globe’s study appears to be a valid and meaningful one (they themselves note some qualifiers, but it is certainly pretty strongly suggestive of a disparity), while the one released yesterday remains an invalid one.

Cheesesteak, I’d be interested if you could find a link to the study that you refer to - I looked around a bit and couldn’t find any reference to it.

[theR]

@mhendo

Actually, it’s also troubling that men get more tickets than women when stopped for the same offence.

I would be interested to see whether and how a higher percentage of female police officers would change the gender disparity with regard to getting tickets. I believe most police departments are still by far predominantly male.

[mhendo]

@IzzyR

mhendo,

In the interests of clarification though, I should point out that Jack McDevitt and Lisa Bailey did not, as you state, write “a piece in the Boston Globe in which they expanded on the resuts of their study” - their study was not completed at the time (8/03) - as noted in the OP, it was released yesterday. What they did was write an Op-Ed piece in support of the Boston Globe’s study that you refer to in your second quote.

You’re right, of course. What i should have written was that they were commenting on the Globe’s study, not on their own which, as you say, had not even been completed.

And i didn’t mean to suggest that the article linked in the OP misrepresented the study–only that it didn’t tell the full story of racial profiling possibilities.

[mhendo]

@IzzyR

mhendo,
In sum, the Boston Globe’s study appears to be a valid and meaningful one (they themselves note some qualifiers, but it is certainly pretty strongly suggestive of a disparity), while the one released yesterday remains an invalid one.

I can’t agree with this at all–i think you’re making way too big a leap.

Sure, the study might be inconclusive, but that’s a far cry from being “invalid.” Even the article you linked to in the OP makes no claim that anyone is using this study (yet) to accuse police of racial profiling, and that article also contains quotes from various officials stating that the study simply provides a basis for discussion of the issue.

Furthermore, the IRJ Website announcing the release of the report specifically states that the figures being released are a “preliminary set of tabulations,” and are

being released to provide information to communities about traffic citations that have occurred in their jurisdiction between April 1, 2001, and June 30, 2003. For many communities this will be the first time they have seen an official report concerning their traffic citation data and been able to review the information to see if it corresponds to the traffic stop goals of their law enforcement departments. We are releasing these preliminary tabulations so that police departments and community members can offer comment about the report and can receive clarification about what questions traffic citation data can answer.

They are quite open about the fact that there are certain questions that the data cannot answer. That the IRJ is aware of this was already apparent in the earlier piece written by McDevitt and Baile, and the current summary on the website restates the case, saying:

From the outset it is important to note that aggregate data, such as the data presented in this preliminary report, can indicate patterns of disparate traffic citation activity in a department but cannot identify the motives involved in individual traffic stop, citation, or other enforcement decisions. Therefore, this preliminary report should not be read as an indication of racial profiling by any Massachusetts law enforcement agency.

To call the study “invalid” is to presume that the data will be used for illegitimate purposes. The fact that the data collected need to be interpreted correctly in order to form a basis for conclusions and future actions does not in any way invalidate the study.

[Cheesesteak]

@IzzyR

Cheesesteak, I’d be interested if you could find a link to the study that you refer to - I looked around a bit and couldn’t find any reference to it.

No luck, may have been lost when the boards went down. I’ll check to see if i can find the original study.

That said, mhendo’s post is certainly food for thought. Certainly seems to indicate bias, not just against minorities, but against men as well. Better argument than just a higher number of tickets vs. population %, it’s higher chance of getting a ticket vs. a warning, big difference.

[IzzyR]

@mhendo

Sure, the study might be inconclusive, but that’s a far cry from being “invalid.” Even the article you linked to in the OP makes no claim that anyone is using this study (yet) to accuse police of racial profiling, and that article also contains quotes from various officials stating that the study simply provides a basis for discussion of the issue.

To call the study “invalid” is to presume that the data will be used for illegitimate purposes. The fact that the data collected need to be interpreted correctly in order to form a basis for conclusions and future actions does not in any way invalidate the study.

Well I think its pretty clear that that is what the study intends to support - I can’t think of any other reason for such a study. The AP and the people quoted in the article also interpreted this as being the (inconclusive) implication of the study.

I am aware that they loaded on some qualifiers. Nonetheless, even a report filled with qualifiers can have the effect of making the apparent implication to be sort of a baseline hypothesis, with the burden of proof shifting to the other side to explain it away. (As an example, the cop that I quoted in the OP who suggested that the drivers in his community did not reflect the ethnic makeup of the population at large). My point here is that there is no need to explain anything, since there is no basis for a null hypothesis that assumes that ticket rates should be equal. There are cultural, socio-economic, and demographic differences between the different groups, and looking at the numbers of tickets per capita is not merely inconclusive as regards the issue of profiling, but it is not even suggestive of anything, absent reason to believe that the numbers should be similar.

@Cheesesteak

That said, mhendo’s post is certainly food for thought. Certainly seems to indicate bias, not just against minorities, but against men as well. Better argument than just a higher number of tickets vs. population %, it’s higher chance of getting a ticket vs. a warning, big difference.

I agree, as noted earlier.

That said, it is possible for this too to be accounted for by other factors. In particular, it is possible that due to the history of enmity and distrust between minority communities and the police, that minorities are more likely to adopt an argumentative and confrontational attitude when pulled over. And if there is one sure way to make sure you get a ticket and not a warning it is adopting an argumentative and confrontational attitude.

But this is rather speculative. So I agree in this case that the results are strongly suggestive of actual bias, absent support for some alternative rationale.