Profiles in Distortion: Misusing Data to Justify Racism and Privilege in Policing

An Analysis of Traffic Stops, Race and Media Deception in Nashville, Tennessee

April 2004


On April 3, 2004, Nashville’s daily paper, The Tennessean, ran as its lead story an article entitled, “White women more likely ticketed when pulled over.” Therein, the author reported that based on the paper’s analysis of local police data, it is white women, and not men and women of color who are most likely to be ticketed in the event of a traffic stop. According to the paper, this fact flies in the face of conventional wisdom, and essentially debunks the notion that Nashville law enforcement officers engage in racial profiling or biased policing towards people of color.

Yet a careful examination of the Tennessean data indicates that the conclusions reached by its reporter are terribly flawed. In fact, the data presented in the paper’s own story actually does much to prove the existence of the very racial profiling it seeks to refute.

The primary findings of this report are:

  • Although the rates at which various racial groups are ticketed are roughly equal, and whites are slightly more likely to be ticketed than blacks when stopped by police, this fact has absolutely no bearing on whether or not racial bias in policing exists. Even if all motorists pulled over were ticketed at the same rate of 100%, racial bias could continue to skew who gets stopped and who doesn’t in the first place;
  • The Tennessean data actually indicates substantial racial disparity in rates of being stopped by police. Of all drivers stopped, blacks were 34 percent of the total, despite being 25 percent of the driving-age population in Nashville, and 23 percent of Nashville drivers;
  • Since national studies indicate no significant racial differences in violating traffic laws (and thus, blacks, at 23 percent of drivers should also represent roughly 23 percent of driving infractions), the fact that African Americans are 34 percent of all drivers stopped means that blacks are almost 50 percent more likely to be stopped than would be expected based on random chance and under conditions of equal treatment;
  • White women, though slightly more likely to receive a ticket after being stopped by police, represent only 21 percent of all motorists pulled over, which is far less than their 35 percent share of Nashville drivers. This means that white women are 40 percent less likely to be stopped than would be expected based on random chance. In all, whites are roughly 71 percent of Nashville drivers but only 59 percent of persons stopped by police for traffic infractions: a rate that is 17 percent below what would be expected;
  • Amazingly, black potential drivers are stopped at a per capita rate of 36 percent in Nashvilleã80 percent higher than the 20 percent per capita rate for potential white drivers;
  • Even these estimates of disparity are conservative. National studies find that whites, on average, take about 183 more car trips per person than blacks each year, meaning that Nashville whites would literally have millions more opportunities annually to commit a traffic violation, be caught by police, stopped and ticketed;
  • Although law enforcement tends to be deployed more heavily in communities of color due to higher crime rates there, and thus, the lawbreaking of persons in those places will more readily come to the attention of police, there is little reason to believe that such an excuse would explain traffic stop disproportions of this size. National studies have found that even after controlling for local population demographics and crime rates, blacks and Latinos are still far more likely to be stopped by police than could be expected by random chance or under conditions of equal treatment;
  • Furthermore, justifying disparity in traffic stops because of saturation policing still contributes to a cycle of profiling: namely, police saturate certain communities, thereby stumbling upon illegality in those places; then the persons living in such neighborhoods become marked as less law-abiding, which then serves to justify more saturation policing, and in effect maintains high levels of ticketing, arrest and incarceration (for more serious offenses) for blacks and Latinos, both locally and nationwide.


There’s an old adage that numbers can be made to say anything one chooses; and while this is not exactly true–after all, one can’t make 2+2=5–it is certainly the case that data often gets manipulated in order to advance a particular argument.

Or in the words of the famous quip: “Liars always figure and figures always lie.”

Nowhere is this more apparent than in the efforts of those who try valiantly to deny the persistence of racism in American life.

Whenever the issue is raised, skeptics have a ready excuse for what appears, to most, to be clear evidence of racial discrimination. Whether the area of interest is income inequality, wealth accumulation, differential outcomes for blacks and whites seeking mortgages, or disparities in the justice system, there will be those who insist the gaps are caused by anything but racial bias.

Case in point, a recent report based on police data in the Nashville Tennessean, which suggests that there is no racial bias against people of color in terms of ticketing them for traffic violations. Contrary to popular belief, the story claims, white women are slightly more likely than anyone else to be ticketed after a traffic stop, and this, they imply, disproves the existence of racial profiling by Metro Nashville police (1)

For those untrained in the issue of racial profiling, or in statistical interpretation, the “investigative” piece on the front page of the April 3, 2004 edition might seem the ultimate “I told you so” to the civil rights community. To conservative whites, the data presented will cement their belief that people of color are merely bellyaching about racism, when there is little or no evidence to justify their concerns.

Yet a close examination of the Tennessean data indicates that, as with other denials of racism, this one tends to say more about those trumpeting it than it says about the data itself. As this analysis will demonstrate, data from Nashville police not only fails to debunk the notion of racial profiling and bias, it actually supports claims of both. That white women are slightly more likely to be ticketed than others has no bearing on either of these issues, as will be explained below.



On the one hand, it is true that there seems to be very little difference in terms of the rates at which various demographic groups receive tickets (versus verbal warnings) upon being stopped by Nashville police for traffic offenses. For example, white females received tickets 93 percent of the time when stopped, as did 92 percent of white men, compared to only 91 percent for black women, 90 percent for Latinos, and 88 percent for black men (2).

Yet contrary to what the police themselves claim, and counter to the theme of the Tennessean story, these numbers don’t even come close to disproving the existence of racial profiling and racial bias in policing.

To begin with, whenever a police department operates under a policy of essentially ticketing everyone they stop, and only stopping those they intend to ticket–as is the case in Nashville, which according to the story has a much higher ticketing rate, per stop, than in other cities–we would naturally expect there to be very little difference in the rates at which various sub-groups received tickets. But this would in no way foreclose the possibility of biased policing and profiling.

Even with a 100 percent ticketing policy (i.e., everyone stopped was given a ticket), racism could be present because police could be stopping blacks in the first place at a disproportionate rate, relative to their share of drivers breaking various laws. The fact that all drivers, once stopped, were ultimately ticketed, would merely obscure the differential rates at which drivers were pulled over to begin with. So, for example, blacks could be stopped at a rate that was 2, 3, 4, or even 10 times higher than their share of all drivers, or their share of speeders, and yet the fact that all persons stopped for speeding, of whatever race, were ticketed would make it appear as though there were no bias involved.

As an extreme example, and using the logic of the Nashville Police, if officers in a 90 percent white community were to stop 1000 people in a year and give them all tickets, and they stopped blacks 900 times and whites only 100 times, the fact that each group had 100 percent ticket rates when stopped would “prove” that no bias existed. Yet this would be absurd, since the disproportionate stoppage of blacks alone would signal that racial discrimination was occurring before the point of issuing a ticket. This is especially true since, as will be noted below, there are no significant racial differences in the rates at which various racial group members violate traffic laws.

Beyond hypotheticals, real world examples also make the point. Some of the most egregious cases of proven profiling in recent years have involved disproportionate stops of blacks and Latinos, without those stopped being ticketed. So, in New Jersey, for example, one African American male–a dentist who drove a BMW–noted that he had been stopped fifty times by state troopers on the New Jersey Turnpike, but never ticketed. Instead he was repeatedly pulled over and asked if he had drugs in the car (3). Does the fact that this individual wasn’t ticketed–or that many blacks stopped aren’t–mean that police are operating fairly? As this example amply demonstrates, the answer is obviously no.

Likewise in Florida, one notorious case of profiling (in an attempt to find drugs), involved stops along a stretch of I-95 in Volusia County, in which blacks and Latinos comprised 70 percent of all drivers stopped, while being only 5 percent of drivers on that stretch of road. Of 1100 people stopped, only nine received any ticket at all, meaning that there was a similar rate of ticketing for all racial groups–virtually nil in all cases–but clearly a pattern of selective stoppage, effecting minority drivers (4).

Or consider a recent incident filmed by the NBC newsmagazine, Dateline, in which a Cincinnati police officer stopped and searched a young black man on the street, merely because he “looked away” from the officer and “seemed to be trying to avoid” contact with him. The stop, which was hardly based on serious reasons for suspicion, led to no arrest or citation of any kind, because the young man had done nothing wrong (5). But should the fact that he was not cited for any illegality be seen as evidence of fair treatment, or rather highly unfair and unjust harassment, and racially-biased suspicion on the part of the officer?


Since traffic laws are a mix of well-known and obvious rules, as well as more arcane, little-known and obscure regulations, virtually all of us will violate several traffic laws every time we pull out of our driveways. Yet police can’t stop everyone who fails to signal a lane change, or doesn’t signal for a great enough distance before a turn, or who comes to a rolling stop at a stop sign. Law enforcement in such an area is, by necessity, selective, thereby inviting potential bias. The question is not whether people are ticketed upon being stopped, but rather, who is stopped and why; and who is not stopped as often, and why not?


Indeed, as regards this question, the same data splashed on the Tennessean’s front page indicates what could be evidence of a very significant problem with racial bias in stoppage rates. For example, as the provided data table indicates, there were 121,838 traffic stops last year where the race of the driver was known to police. Of these, stops of African Americans represented 34 percent of the total, stops of Latinos represented 5.5 percent, and people of color overall were 41 percent of all drivers stopped for a traffic offense, in those cases where the race of the offender was known to police. White women, though receiving tickets slightly more often when stopped, represented only 21 percent of all persons stopped in 2003 (6).

A few things stand out from these numbers.

First, that 34 percent of all persons stopped by Nashville police in 2003 were black is troubling, not only because this is well above the share of Nashville’s driving-age population that is black (about 25 percent) (7), and well above the black share of drivers (which is around 23 percent since blacks are far more likely to be without a car than whites), (8) but also because there is no evidence to indicate that blacks violate basic traffic laws any more often than anyone else. In fact, as David Harris documents in his groundbreaking book Profiles in Injustice, studies have almost uniformly found little if any difference in driving behaviors based on race (9).

Thus, if we assume that roughly 23 percent of drivers in Nashville are black, and that about 23 percent of traffic violations are committed by blacks, this would mean that blacks are 48 percent more likely to be stopped by police in the first place, than would be expected based on random chance, and under conditions of equal treatment.

Likewise, Latinos appear to be stopped at a disproportionate rate as well. Although Latinos represent only 3.8 percent of all Nashvillians with access to a vehicle, they made up 5.5 percent of all drivers stopped by police in 2003 (see Appendix). This would suggest an over-stoppage rate of approximately 44 percent, relative to what would be expected under conditions of equal treatment and random chance.


It is also important to note that while traffic stops seem to disproportionately target blacks and Latinos, there is a flipside to that disparate treatment that is often unremarked upon: namely, that white drivers are thereby receiving preference and privilege relative to drivers of color.

For example, white women are obviously far more than 21 percent of the drivers on Nashville roads, and unless one believes they are uniquely law abiding, far above that percentage of persons breaking traffic laws. Indeed, whites are about two-thirds of Nashville’s driving age population, and around 71 percent of its drivers (10) meaning that white women are approximately 35 percent of all drivers in the city at any given time. To be 35 percent of drivers, and likely the same rough share of rule violators, means that for white women to be only 21 percent of persons stopped is to be stopped at a rate that is 40 percent below what would be expected.

By comparison, white men are roughly 35 percent of all drivers in Nashville, and comprise 38 percent of all persons stopped–a very small gap, which suggests that white men are stopped about as often as one would expect. Overall, this means that although whites are about 7 in 10 drivers in Nashville, they are slightly less than 6 in 10 persons stopped by police, for an overall rate that is 17 percent below what would be expected by random chance.

In other words, the data that the Nashville police use to disprove preferential treatment for whites and bias against blacks, not only fails to disprove it; rather, it may demonstrate it conclusively. In fact, when we compare the racial stop numbers with the numbers of black or white drivers in Nashville, we discover a disturbing picture; one left completely unexamined by the police or the press. Namely, that white potential drivers (those with access to a vehicle who could be driving at a given moment) are stopped by police at a per capita rate of 20 percent, while black potential drivers are stopped at a per capita rate of 36 percent: meaning that any given potential black driver was 80 percent more likely to be stopped than any given potential white driver (11).

In order to justify the disparities present in the data, police would have to demonstrate that black drivers violated traffic laws at a rate that was almost double the rate of white drivers: a claim for which there is no evidence presented, implied, or even logically intuited from any local or national study ever done.


In fact, even these estimates of bias are conservative, since they are based on stoppage rates relative to car ownership rates, while the more relevant comparison figure would be how many people on the road at any given time are likely to be black or white: a figure that will not simply mirror the rates at which blacks and whites own cars.

To begin with, data suggests that white drivers average more trips per day in their cars than blacks–roughly 4.4 as compared to 3.9 (12). Figured over the course of a year, this amounts to about 183 more individual car trips for white drivers each year compared to blacks, or roughly 183 more chances annually to break a law, be caught, and be ticketed for each white driver in Nashville. Multiplied by the city’s white driving population, this amounts to literally millions more opportunities for whites collectively to be stopped or ticketed relative to blacks.

Controlling for this factor–as well as the fact that whites are more likely to commute longer distances, thereby necessitating more time on the road and making themselves even more disproportionately available for a stop or ticket–there is every reason to believe that blacks, instead of representing 23 percent of potential targets for a stop (based on ownership data), would represent well below 20 percent of all drivers on Nashville roads at any given time. If so, the level of disparate treatment between the share of drivers on the road who are black and the share of persons stopped who are black would be substantially higher than the previously claimed 48 percent. Likewise, the under-representation of whites in the stop data would be larger, since, instead of 71 percent of drivers (the number based on ownership figures), whites would likely represent at least 75 percent or more of drivers on the road at a given time.

Furthermore, one recent study in Cincinnati suggests that rates of disparity like those found in Nashville between white and black stops may be somewhat obscured by a failure to disaggregate moving violations from less serious types of infractions (busted taillights, missing rearview mirrors, expired tags, windows that are too darkly tinted, being illegally parked, etc.) According to the study, in which researchers at Dateline NBC examined over 100,000 police stops over a three-year period, blacks were three times more likely to be stopped than whites for non-moving violations. This compared to a previous study at the University of Cincinnati, which examined a smaller sample of both moving and non-moving violations and found a much smaller disparity–on the order of 35 percent or so (13).

Essentially, the NBC study suggests that while serious infractions are treated more similarly by police, less serious ones–which allow for more discretion on the part of police as to whether to enforce the law–are met with substantial unequal treatment. If a similar tendency exists in Nashville, it may be that racial disparities for certain infractions are far larger than the aggregate figures provided here thus far. In order to determine the extent of such disparities, police should provide disaggregated information on moving versus non-moving violations. While it could be argued that people of color are more likely to be in violation of the less serious types of infractions–since busted equipment or expired tags are likely to be more common among lower-income folks and since people of color are more likely than whites to be low-income–it is doubtful that the rates at which blacks are in violation of such things would amount to 2-3 times the rate at which whites violate such rules


There is one other possible explanation for racial disparities in traffic stops, other than bias, and other than the possibility (largely debunked by national studies) that people of color simply violate traffic laws more often. Indeed, it is this explanation to which law enforcement most often turns in an attempt to explain away disparities such as those found in Nashville. The argument goes roughly like this:

Crime rates are higher in communities of color, especially low-income African American communities. Thus, police will be deployed to such areas in larger numbers than in areas that are mostly white. Since police are more likely to saturate black neighborhoods (ostensibly for legitimate reasons of fighting more serious crime), they will naturally be more likely to observe all kinds of legal violations in such places: from serious crime, to drug crime, to traffic violations. As such, blacks will be stopped, searched, arrested, etc. more often than whites, not because of bias, and not because they necessarily are less law-abiding than anyone else in the area of traffic offenses, for example, but simply because police will see their violations more often than they will those of whites, due to their heavier presence in communities of color.

While there is obviously some truth to this argument (of course police will be more likely to see violations in places where they patrol than in places where they don’t patrol) there are several problems with this argument as it relates to justifying significant racial disparities in traffic stops.

First, if saturation policing explained the disparity, Nashville police should be able to point to data that would illustrate that reality. In other words, they should be able to indicate that, controlling for local demographics in a neighborhood (which obviously mean that blacks will be stopped in areas that are almost all black) and crime rates (which serve as a good proxy for police presence), there is no real disparity in rates of traffic stops. Yet they have provided no such data, nor do they claim to even collect data that would allow for such a comparison. Until Nashville police provide precinct-by-precinct and neighborhood-by-neighborhood data on who is stopped and for what kind of violation, it will be hard to know if the saturation excuse is legitimate or not.

What we do know, from national studies that have examined this issue, is that saturation policing, though occasionally explaining some of the racial disparities in stoppage rates, cannot come close to explaining all of the disparity. For example, in New York City, one study found that even after local demographics and crime rates were considered, police were still about twice as likely to stop blacks and Latinos for bodily searches, as they were to stop whites. This was true in mostly white neighborhoods as well as mostly black and brown neighborhoods (14).

Additionally, although it is true that police will often saturate certain neighborhoods, it is also true, as mentioned previously, that whites saturate the roadways, and actually make millions more trips in their cars annually than blacks and Latinos. Thus, even though there may be fewer police, per capita, patrolling or driving around in white neighborhoods, there are so many more opportunities for whites to break a law (and thus be observed, even by a smaller net police presence) that one would expect the two forms of saturation (police in black neighborhoods and whites on the roads generally) to balance out at some point.

This is especially true when you consider that it is in low income black and brown communities where people of color would be even less likely to have access to a car in the first place, due to economic constraints. So in the places where police “saturate,” there is statistically speaking, less opportunity for them to witness traffic violations than many other kinds of violations that don’t require a vehicle.

Also, the reason for which police “saturate” low-income communities is, presumably, to fight serious crime. As such, giving out a large number of relatively insignificant traffic tickets, or merely stopping people heavily for minor infractions can hardly be explained by saturation policing. After all, if the reason for which police are heavily present in certain areas, and the kinds of law enforcement they are dispensing are totally different, it makes sense to question whether or not they are really doing the jobs for which they were dispatched. One would assume that police in such places had more serious concerns than stopping motorists for busted break lights, failure to signal a lane change, or other minor infractions. If so, then stopping people for such minor issues merely detracts from actual crime-fighting (in which case, what is the point of saturation policing?), and if not, then there was no need for saturation policing to begin with.

Finally, the claim that otherwise legitimate saturation policing explains and justifies disparate traffic stops ignores the cycle that such an excuse sets up. In effect, if we are to justify disparities of this sort because of saturation policing (which itself is “justified” by higher crime rates in certain communities), then we establish a never-ending cycle, in which persons of color will always find themselves trapped. Since black and brown “law breaking” will come to the attention of authorities more often (whether or not it is actually more prevalent), higher stoppage rates will then be used to justify still more disproportionate stops of black and brown persons, which then will re-justify more saturation policing. Meanwhile, lawbreakers in the white communities will go undetected and un-stopped, no matter their tendency to violate certain laws.

While this pattern may be abstractly justifiable under such an analysis, it still begs the question as to how much unequal treatment should be tolerated in the justice system, especially when that unequal treatment then contributes directly to the appearance of greater criminality in communities of color, which then “justifies” more unequal treatment.


Even putting aside the issue of disparate stops, there are a number of reasons for roughly equivalent ticketing rates (among those stopped by police), none of which would necessarily, or even logically indicate the absence of racial bias.


For example, slightly lower rates of ticketing for blacks–as was evidenced in the Nashville data–could signify that African Americans were being stopped on less legitimate legal grounds. As a result, police didn’t ticket them as often, merely because the legal basis for the ticket would have been more questionable. As such, the lower ticketing rate could actually be evidence of bias (in the form of unjustified harassment of black motorists), rather than a refutation of it. Cops, according to this analysis may be stopping African Americans out of suspicion that the driver in question has committed some unspecified crime, and yet there is insufficient evidence to justify a ticket at the end of the stop.

Much in this fashion, a study in New York (15), as well as national data available from the Justice Department (16), indicates that although blacks and Latinos are disproportionately likely to be stopped and searched for drugs, whites are actually more likely, when stopped and searched, to be found with contraband on them. So, as a result, whites stopped will have higher rates of arrest than blacks and Latinos who are stopped; yet would anyone suggest this is evidence of anti-white bias, or proof that people of color are treated fairly? Of course not. If anything, it proves the opposite to be true, as with the recent Dateline study in Cincinnati, mentioned previously.

Had black men, for example, been ticketed at the same rate as white men after being pulled over, there would have been an additional 961 black men ticketed in 2003; likewise, had black women been ticketed at the same rate as white women after being stopped, an additional 369 black women would have been ticketed in 2003 (17). While at first, this may seem to suggest that blacks in Nashville are being treated better than whites, in fact it could merely signify that more than a thousand African Americans in the city are being stopped annually for offenses that are too minor to justify ticketing, and for which they never would have been stopped at all had they simply been white


Similarly, that whites have slightly higher rates of being ticketed, once stopped, in Nashville (or in any other jurisdiction) for traffic offenses could be evidence of substantial preference for white drivers.

For example, imagine a community where police routinely looked the other way when whites committed minor infractions, and only stopped them for more serious offenses. In such a scenario, those whites stopped would be almost guaranteed to receive a citation (since police are more concerned about serious infractions than minor ones), while people of color might be let off more often simply because they had committed relatively inconsequential offenses, which were less likely to merit a ticket in a given situation. So the rate of ticketing for whites would be higher than for blacks, but only because they were ignored for the kinds of minor offenses that resulted in a stop for people of color.


Contrary to the claims of the Nashville Tennessean, police data do not indicate a lack of racial profiling against people of color in the metropolitan area. While it is possible that police treat drivers equitably, without regard to race and ethnicity, such a conclusion cannot be sustained merely on the basis of the data presented thus far. In fact, a careful reading of the numbers suggests that racial bias may heavily influence who gets stopped in the first place on Nashville roads, even if motorists of all races are roughly equally likely to be ticketed after a stop.

If Nashville police wish to be exonerated from the charge of racial profiling, they must compile and make public all data related to the race of persons stopped, on surface streets and interstates; the reasons for each stop; the rates of ticketing for each offense, for each racial group; and the rates at which searches of vehicles are conducted, for each racial group. Furthermore, the police and/or a news organization such as the Tennessean should conduct random videotaped surveillance of driver behavior in and around the city, so as to determine the relative racial percentages of drivers breaking various traffic laws. These numbers could then be directly compared to the percentages of each group being stopped, so as to demonstrate either the existence of disparity or the lack thereof.

That the Nashville media, to this point, has put forth such a flawed report, while failing to examine the many reasons for the results they obtained, only indicates how un-investigative journalism tends to be on issues of race. Having little background in issues of discrimination or statistical analysis, media merely ends up parroting the line of officials, be they political or law enforcement types.

That said, for those of us who seek to eradicate racial inequality, and its various manifestations like racial profiling, it is vitally important that we learn to properly analyze and break down the fuzzy math of those who deny what all other evidence and personal experience tells us is true: namely, that racism is still very real, unequal treatment still very much the norm, and that claims to the contrary, though laden with propaganda value, are without the least amount of merit.



According to the Census Bureau, the population of Metro Nashville in 2001 was approximately 585,000 people. Of these, approximately 381,783 (or 65 percent) were white, non-Hispanic, while 147,696 (or 26 percent) were black. Additionally, approximately 4.6 percent of the population identifies themselves as Latino, while a little less than 1 percent identify as Asian. This leaves the remaining 3.4 percent undetermined, or persons identifying as mixed race on Census forms.

(Source: United States Department of Commerce, Bureau of the Census, State and County Quick Facts,


Note: the aggregate figures for black/white/other percentages of the population must be adjusted when figuring out the percentage of each group among the driving population. This is so, because different racial groups are more or less likely to include larger numbers of young people (under 15 or 16 who don’t drive), and over 75 (and also not as often likely to drive). In fact, African Americans are more likely than whites to be below the age of 15, and less likely to be found among older drivers, because of lower life expectancy. As such, a disproportionate share of blacks will be in the non-driving age category of youth and a disproportionate share of the older drivers will be white as well. As such, it would be reasonable to adjust the black population numbers to reflect those within the driving-age range of 15-75, which would reduce the black share of the relevant population, at least to 25 percent, down from 26 percent in the aggregate. Indeed, this is likely a conservative estimate.


To determine the approximate share of drivers in Nashville who are white, black, etc., one need only do the following:

1. Start with the population percentage of blacks who are driving age (roughly 15-75). According to Census data, this would be approximately 25 percent of the population;

2. Multiply that number (.25) by the percentage of black households (and thus, likely individuals) who do not have a car (.21, see footnote).

.25x.21=. 0525

3. Subtract the result (.0525) from the black population percentage (.25)

.25-.0525=. 1975 (19.75%)

This then offers a number representing the share of the Nashville population that is black with a vehicle.

4. Take the white (non-Hispanic) share of the population (65 percent according to the Census Bureau), and multiply that by the percentage of white households (and thus, likely individuals) not having a car. (For this number, I use 7 percent, which is likely a high estimate, since the national average is 8 percent, and the rates for people of color are so large. By estimating high, I actually likely underestimate the share of white drivers and thus the amount of racial privilege received by whites in terms of their share of traffic stops)

.65x.07=. 0455 (4.6%)

5. Subtract the result from the white population percentage

.65-.0455=. 6045 (60.45%)

This then represents the share of the Nashville population that is white with a vehicle.

6. Take the share of Latinos in the population (4.6 percent according to the Census), and multiply that by the percentage of Latinos without a vehicle. Although this data is not as readily available as the data for blacks and whites, there are rough estimates one can use to make the calculation. According to Census data found in Cheryl Russell’s 1998 volume, Racial and Ethnic Diversity: Asians, Blacks, Hispanics, Native Americans and Whites, Ithaca, NY: New Strategist, Inc. (page 252), roughly 17 percent of Hispanic/Latino households (defined here as occupied housing units, which is a pretty good proxy for households), have no motor vehicle available for their use.

So, if we assume that about 17 percent of Latino households (and thus persons of driving age) are without access to a car:

.046x.17=. 00782 (0.78%)

7. Subtract this figure from the Latino population percentage:

.046-.0078=. 0382 (3.82 percent)

This then represents the share of the Nashville population that is Latino with a vehicle.

8. Take the percentage of the Nashville population that is Asian (roughly 1 percent according to the Census), and multiply that by the share of Asians not having a car. Although this data is not available in any database known to the author, social indicators related to economic status for Asians tend to closely resemble those for whites, meaning it is reasonable to estimate that about 7 percent of Asians have no access to a car, as with whites.

.01x.07=. 0007

9. Subtract this figure from the Asian population percentage: .01-.0007=. 0093 (0.93 percent)

This then represents the share of the Nashville population that is Asian with a vehicle.

10. Now, rounding off each figure to the nearest half of a percentage point, add together the percentages of each group with car access:

60.5+20+4+1=85.5 percent of Nashville’s population has a vehicle

11. Now divide each sub-group number in step 10 by the sum of all sub-group figures (85.5) to determine the share of drivers represented by each racial subgroup

Thus, for whites: 60.5/85.5=71 percent of all drivers are white Thus, for blacks: 20/85.5=23 percent of all drivers are black Thus, for Latinos: 4/85.5=5 percent of all drivers are Latino Thus, for Asians: 1/85.5=1 percent of all drivers are Asian

This means that overall, people of color are 29 percent of Nashville drivers, though they are 41 percent of all persons stopped by police. Whites are 71 percent of drivers but only 59 percent of persons stopped by police. This means that people of color, overall are 41 percent more likely to be stopped than would be expected based on random chance, while whites are 17 percent less likely to be stopped than would be expected.


To determine the per capita stop rates for blacks and whites who are potential drivers, one merely divides the number of whites and blacks stopped in 2003 by the number of white or black potential drivers (the latter of which can be determined by the computations in the previous section of this appendix, and which refers to those persons in households with access to a vehicle).

Whites stopped = 72,492
Blacks stopped = 41,612
White potential drivers = 355,058 Black potential drivers = 116,680

72,492/355,058 = .20 (the share of white potential drivers stopped by police in 2003) 41,612/116,680 = .357 (the share of black potential drivers stopped by police in 2003)

In other words, on a per capita basis, 20 percent of whites and 36 percent of blacks with access to a vehicle in Nashville were stopped in 2003. Thus, potential black drivers are 80 percent more likely to be stopped than potential white drivers in Nashville, according to the data for 2003.

Note: Obviously this number does not represent the percentage of actual black and white drivers stopped in 2003, since the denominator in the above equation includes all persons in households with access to a car, including those who don’t drive, who are under legal driving age, etc. So in fact, since the denominator would be smaller in a more restricted analysis, the share of black and white drivers stopped in a given year would actually be higher, per capita, than claimed here, and if anything the disparity ratio would grow, since a larger share of blacks would have to be removed from the actual driving pool, given lower economic status, lower average age (including more youth under legal driving age), etc. In other words, the estimates here are conservative, but still provide a fairly clear glimpse of the extent of racial bias likely operating in the system.


  1. Demsky, Ian. 2004. “White women more likely ticketed when pulled over,” The Tennessean. April 4: 1A .
  2. Demsky, 2004: 1A, Table
  3. Harris, David A. 2002. Profiles in Injustice: Why Racial Profiling Can’t Work. NY: New Press: 53.
  4. Ibid., 63-4.
  5. Bird, Rick. 2004. “NBC: City Cops Profile,” Cincinnati Post. April 9. The Dateline study also found that blacks in Cincinnati were three times more likely than whites to be stopped for minor, non-moving infractions by police.
  6. Demsky, 2004: 1A, Table, and calculations by the author. White women were stopped 26,120 times in 2003. As a share of 121,838 stops where the offender’s race was known, this represents 21 percent of the total.
  7. United States Department of Commerce, Bureau of the Census, 2004. State and County Quick Facts, at:, and calculations by the author (see A.2 of appendix at end of report.
  8. According to data collected by profiling expert David Harris, of the University of Toledo School of Law, roughly 21 percent of black households have no car. This can be compared to a national average of only 8 percent of all households without a car. (See, U.S. Department of Transportation, Bureau of Transportation Statistics, 2002. “Highlights of the 2001 National Household Travel Survey: Household, Individual and Vehicle Characteristics,” available at: In order for the national average of carless households to reach 8 percent, and given that the black rate of carlessness is 21 percent, this would mean that the white rate of household carlessness would have to be well below 8 percent, actually. Even a white rate of 7 percent would mean that black households are three times more likely to be carless than white households
  9. Harris, 2002: 54, 61.
  10. United States Bureau of the Census, 2004; also, Harris, David. 1999. “The Stories, the Statistics, and the Law: Why Driving While Black Matters,” 84 Minnesota Law Review 265-326, and additional calculations by the author (see A.1 and A.3. of Appendix below).
  11. Demsky, op.cit., and U.S. Bureau of the Census, op.cit., and additional calculations by the author, (see A.4 of appendix.)
  12. Harris, 2002: 68; also, Harris, 1999.
  13. Bird, 2004.
  14. Fagan, Jeffrey and Garth Davies, 2000. “Street Stops and Broken Windows: Terry, Race and Disorder in New York City,” 28 Fordham Urban Law Journal.
  15. Fagan and Davies, 2000: 457, 479. The study found that blacks and Latinos are approximately 2-2.5 times more likely than whites to be stopped and frisked by police, even after controlling for local neighborhood demographics and relative racial crime rates, both of which might “explain” disproportionate rates of stoppage. Yet, when stopped, whites were more likely to be found to have violated a law such as drug or weapons possession.
  16. Matthew R. Durose, Erica L. Schmitt and Patrick A. Langan, Contacts Between Police and the Public: Findings from the 2002 National Survey. U.S. Department of Justice, (Bureau of Justice Statistics), April 2005. This data indicates that blacks are three times more likely than whites to be stopped in their vehicles and searched for drugs or other illegal contraband, even though whites are nearly four-and-a-half times more likely to actually be found in possession of such items, on those occasions when they are stopped.
  17. To calculate this number is simple, based on the data provided by the Tennessean. There were 121,838 traffic stops, where the race of the driver was known to police, and black men comprised 25,643 of those stopped. They were ticketed at a rate of 88.2. White men were stopped 46,372 times, and ticketed at a rate of 91.95. Had black men been ticketed at a rate of 91.95, the additional number of persons ticketed in this group would have been 961. For black women, the difference between the 91.01 rate at which they were ticketed after a stop, and the 93.32 percentage rate at which white women were ticketed, suggests that had black women been stopped at the rate of white women, an additional 369 persons in this group would have received a ticket.

Leave a Reply