OMNIBUS ORDER DENYING DEFENDANT’S MOTIONS FOR DISCOVERY, DISMISSAL AND SUPPRESSION
Defendant Gerardo Duque-Nava has filed three pretrial motions in this case: a
Facts
The Traffic Stop
On January 11, 2003, at approximately 12:44 p.m., Russell County Deputy Sheriff Sergeant Kelly Schneider was on routine patrol in Russell County, Kansas, on Interstate 70 near mile marker 185. As a blue 1997 Ford F-250 pick-up truck approached him traveling eastbound on 1-70, Deputy Schneider observed that the windshield of the pick-up had a “substantial” crack. Deputy Schneider was aware that Kansas law 1 prohibits driving a vehicle with a damaged front windshield that substantially obstructs a clear view of the highway, and initiated a traffic stop of the pick-up. Deputy Schneider testified that he decided to stop the pick-up to determine if the cracked windshield obstructed the driver’s view, because sometimes it is not obvious without further investigation. He further testified that he has stopped vehicles with cracked windshields and further investigation over 30 times.
Defendant is a Hispanic male of Mexican descent. Deputy Schneider testified that at the time he decided to stop the pick-up for having a cracked windshield, he could not see the occupants, and thus did not know their gender, race or ethnicity. He testified that he did not perceive that the driver or passenger were dark-skinned or Hispanic until he pulled up next to the truck. Deputy Schneider testified that he pulled alongside the pick-up truck as a standard safety practice, so that he could determine the number of occupants of the vehicle. He testified that in his experience, it was possible for a passenger to lie behind the driver’s seat in an extended cab pick-up such as the one stopped in this case. When he pulled alongside, he looked at the crack through the driver’s window and determined that it could have obstructed the driver’s vision of a merging vehicle or at an intersection. He then pulled behind the pick-up and activated his lights.
After stopping the vehicle Deputy Schneider approached from the passenger side, where defendant was riding. Schneider observed that the crack in the windshield was approximately 18 inches long and curved. He advised the driver that the windshield was broken and needed to be fixed. The driver was identified through his Arizona driver’s license as Isidro Duque of Phoenix, Arizona, defendant’s father. Deputy Schneider testified that both the driver and defendant were extremely nervous, and that the driver’s hands shook as he handed him his license. He further testified that defendant appeared overly friendly and provided him
Deputy Schneider returned to his patrol vehicle to run computer checks. After determining that everything was in order, he wrote a warning citation and returned to the truck. He returned the license and paperwork to the driver, along with the warning citation. Deputy Schneider testified that he noticed the driver was still shaking as he returned the documents to him. He told the driver and defendant to be careful, turned to go back to his patrol car, then stopped and asked them if he could ask them another question. Defendant stated that he could, and Deputy Schneider advised them that there were a lot of illegal narcotics being transported over the interstate, and asked if they had any of these items. Defendant replied that they did not. Deputy Schneider then asked for consent to search the truck and defendant consented. In searching two small duffel bags located in the truck, Deputy Schneider found approximately 2.3 pounds of methamphetamine. Defendant was immediately placed under arrest. After interrogation that resulted in incriminatory admissions by defendant, he was booked into the Russell County jail.
Racial and Ethnic Proñling
In addition to the evidence concerning the traffic stop, search and seizure, the parties presented stipulated evidence concerning defendant’s claim that Deputy Schneider stopped him solely on the basis of his race or ethnicity, or in today’s parlance, because of racial profiling. This evidence was previously presented to this Court in the case of United States v. Mesa-Roche, 2 and consists of the following: transcript of motions hearing in Mesa-Roche-, Racial Profiling Study and Services (“Lamberth study”); Russell County Sheriff Racial Profiling Statistics for 2001-2002; Russell County Sheriff Racial Profiling Statistics for Deputy Schneider; Russell County Racial Profiling Statistics for seven remaining sheriffs officers; Russell County Sheriff Racial Profiling Statistical Analysis-Monthly Statistics on Traffic Stops; curriculum vitae of George A. Milliken; Russell County Tables of identification by ethnicity; Comparison of Kansas Highway Patrol stops to Deputy Schneider stops; Russell County Sheriffs Department Book-in sheets; affidavits of Kelly Schneider.
Defendant contends that the statistical data and evidence and testimony introduced and analyzed in Mesa-Roche are applicable to this case as defendant was stopped in the same time period, in the same county and by the same sheriffs officer. The Court agrees, and further finds it appropriate to reiterate its summary and analysis of the evidence as detailed in the order entered in the Mesa-Roche case. 3
The Russell County Sheriffs office produced records indicating the race and ethnicity of motorists in traffic stops initiated by its officers in 2001 through 2002. The records indicate that Schneider initiated traffic stops that were on average: 56.59% white motorists; 33.69% Hispanic motorists; and 9.72% black motorists. Schneider patrols the stretch of 1-70 passing through Russell County exclusively; he does not patrol any other roads in the county. Other officers in his department, who patrol other parts of the county but not 1-70, initiated traffic stops that were on average: 90.27% white motorists;
As further evidence of racial and ethnic profiling, defendant offers the Lamberth 4 study, commissioned by then Governor Bill Graves. In Mesa-Roche, the defendant and the government presented expert statistical and research testimony from Dr. George Milliken 5 and Dr. Brian Withrow, 6 respectively. Both Dr. Milliken and Dr. Withrow testified at length about the validity of the Lamberth study and other studies that have attempted to record and compare the incidence of law enforcement action against discrete racial or ethnic groups, for purposes of evaluating claims of racial profiling or selective enforcement.
The Lamberth study surveyed traffic stops during four months in 2002, by seven law enforcement agencies in Kansas. 7 The participating law enforcement agencies were instructed to compile stop data on actual traffic stops conducted by the law enforcement officers during the period of the study. The officers were instructed to record race or ethnicity as determined by the officer during the course of the stop. This could be determined either through the officer’s observation, or by the officer asking the motorist to identify his or her race or ethnicity. This study recorded, among other things, the race and ethnicity of motorists stopped by the Kansas Highway Patrol in 2002, on 1-70 between Colby, Kansas and the Kansas-Colorado border. The Court takes judicial notice that Colby is about 130 miles west of Russell County.
Based on the actual stop data recorded in the Lamberth study, on the Colby-Colorado border stretch of 1-70, the Kansas Highway Patrol initiates traffic stops that are on average 5.7% black motorists and 6.8% Hispanic motorists, a significantly lower percentage than Schneider’s stops that are 9.72% black motorists and 33.69% Hispanic motorists. In this case, the defendant and government have stipulated to the Court’s consideration of the data in the Lamberth study, but not its conclusion. Because the author of the Lamberth study was not called as a witness, the Court does not admit the conclusions of the study, including the conclusion that officers were stopping black and Hispanic motorists for suspect reasons. 8
All studies that seek to evaluate law enforcement action must compare the data of actual law enforcement action with some type of “benchmark data.” In measuring traffic stops, studies have compared actual stop data with essentially three types of benchmarks: the census population; the violator population; or the transient motorist population.
Some studies use a violator benchmark, comparing data of actual law enforcement action against data of actual violators. Both experts agreed that the violator benchmark is in theory, the superior benchmark. But, as Dr. Milliken testified, if the violator benchmark is based on data of actual law enforcement action, the benchmark may be flawed. Dr. Withrow agreed, testifying that a violator benchmark ignores the sociological phenomena of a higher incidence of law enforcement activity in urban, high crime areas, where there happens to be a larger population of minorities. The violator rate is thus affected by the higher incidence of police activity in locations with a higher incidence of minorities.
It may be possible to establish a violator benchmark that is not reliant on, nor biased by actual law enforcement activity. The State of New Jersey tried to determine whether its troopers were stopping minority drivers with greater or lesser frequency than warranted by the actual number of violations committed by these drivers. The New Jersey Study 9 used a violator benchmark that was not based on actual traffic stops, speeding tickets or other actual law enforcement activity. Instead, the study attempted to determine the violator benchmark without the taint or questionable effect of police bias in defining or affecting who the violators are. In this study, teams of observers recorded the race or ethnicity of turnpike motorists by studying high-quality digital photographs taken as the motorists entered the turnpike. The speed of the motorists’ vehicles was recorded by radar gun. Thus, the teams of observers could compare the incidence of speeding by the race or ethnicity of the motorists. Their statistical analysis of the incidence of speeding was used to create a violator benchmark. 10
The
Lamberth
study, as well as some other studies, have used the transient motorist population as a benchmark for com
As in Mesar-Roche, defendant relies on the transient motorist benchmark developed in the Lamberth study, although the Lamberth study did not measure the transient population on the stretch of 1-70 that Schneider patrols in Russell County. Nevertheless, Dr. Milliken testified that one could expect the transient Hispanic population in Russell County to be about the same as such population on the Colby-Colorado stretch, because their census populations were about the same, and no other factor would significantly change the transient motorist population between the Colby-Colorado stretch and the Russell County stretch of 1-70.
Some of the Hispanic transient motorist population on 1-70 might be diverted to state Highway 83, which intersects 1-70 at Colby and meanders south to Garden City and Dodge City, where there is a sizeable Hispanic population. But, the Court cannot conclude that Highway 83 would affect only that transient motorist population between Colby and Russell. While westbound 1-70 travelers might exit 1-70 at Colby to take Highway 83, one could expect that eastbound 1-70 travelers might exit 1-70 at Colby to take Highway 83 as well. Nothing suggests that Highway 83 significantly increases or decreases the transient motorist population of Hispanics on 1-70 between Russell and the Colorado border.
Thus, the Lamberth study’s transient motorist population benchmark of 2.6% black and 1.8% Hispanic motorists on 1-70 can be extrapolated to the stretch of 1-70 patrolled by Schneider. Assuming the transient motorist population benchmark established in the Lamberth study was based on an accurate measure of the race and ethnicity of transient motorists, 11 Schneider’s incidence of 33.69% stops involving Hispanic motorists and 9.72% involving black motorists far exceeds the transient motorist population of those two groups.
Analysis
A. Fourth Amendment
Legitimacy of Traffic Stop
In order to make a traffic stop, Deputy Schneider needed reasonable suspicion that a traffic violation had or was occurring. Tenth Circuit cases “establish that a traffic stop is reasonable under the Fourth Amendment at its inception if the officer has either (1) probable cause to believe a traffic violation has occurred ... or (2) a reasonable articulable suspicion
The Tenth Circuit has examined K.S.A. 8-1741 under similar facts. In affirming the district court’s refusal to suppress evidence, the court stated:
[Defendant’s] windshield had a crack about 12 inches across and 6 inches high, large enough that Officer Voigt could view it from behind the car. This gave Officer Voigt reasonable articulable suspicion-“a particularized and objective basis”-to believe that the crack substantially obstructed [defendant’s] view of the street.... It is irrelevant whether the observed crack was, in fact, large enough to constitute a violation of the law.... 14
Deputy Schneider’s testimony at the evi-dentiary hearing was credible and undisputed. He testified that not only could he see the cracked windshield as defendant’s vehicle approached him, but also that when he pulled alongside the truck, he looked through the driver’s window and believed the crack could obstruct the driver’s vision of merging vehicles or from an intersection. The Court finds that the crack was not so small that it could not have been seen by the officer, but was of sufficient size and placement that it gave Deputy Schneider reasonable articulable suspicion to believe that the crack substantially obstructed the driver’s clear view of the highway. Accordingly, defendant’s challenge to the initial traffic stop lacks merit.
Defendant further contends that Deputy Schneider’s decision to stop was motivated by defendant’s Hispanic ethnicity, and is therefore unreasonable under the Fourth Amendment. Schneider has testified that he was unaware of defendant’s ethnicity prior to stopping him. For the reasons set out in this order infra denying defendant’s motions alleging violation of the Equal Protection Clause of the Fifth and Fourteenth Amendments, the Court concludes that defendant’s evidence on racial profiling does not show that Schneider’s stop was premised on racial or ethnic discrimination. Furthermore, since Schneider’s search of the vehicle was with the consent of the defendant, the evidence seized is not subject to suppression. Defendant’s Fourth Amendment challenge fails.
B. Selective Prosecution, Selective Enforcement and Equal Protection
Defendant moves for discovery and to dismiss on the basis of an equal protection violation based on Deputy Schneider’s selective enforcement of traffic laws, which led to the prosecution of defendant. Although defendant’s motions are styled as motions regarding selective prosecution, the Court finds that they are in fact, motions regarding selective enforcement. Defendant challenges the decision and actions of the law enforcement officer, not the exercise of discretion by the prosecu
The Equal Protection Clause precludes selective enforcement of the law based on race or ethnicity. 16 In its recent decision in Marshall v. Columbia Lea Regional Hospital, 17 the Tenth Circuit noted that such claims are troubling in that:
Racially selective law enforcement violates this nation’s constitutional values at the most fundamental level; indeed, unequal application of criminal law to white and black persons was one of the central evils addressed by the framers of the Fourteenth Amendment. 18
A defendant claiming unequal enforcement of a facially neutral statute must show both that the enforcement had a discriminatory effect, and that the enforcement was motivated by a discriminatory intent.
19
In order for a defendant to obtain discovery on these issues, the defendant need not establish a prima facie case of selective enforcement.
20
Nevertheless, given the heavy burden that discovery can impose on the government,
21
the showing necessary to obtain discovery for a selective prosecution defense must “itself be a significant barrier to the litigation of insubstantial claims.”
22
A defendant who claims he was targeted for enforcement of traffic laws because of race or ethnicity is entitled to discovery on that claim only if he presents
“some evidence tending to show
the existence of the essential elements of the [selective enforcement] defense, discriminatory effect and discriminatory intent.”
23
A defendant cannot obtain discovery based on a bald, unsupported
Discriminatory Effect
As a matter of law, long-standing equal protection jurisprudence has recognized that some measure of selectivity in the law enforcement arena is constitutionally permissible. 24 With that recognition, courts have required a defendant to make a credible showing that a similarly situated individual of another race or ethnicity could have been subjected to the same law enforcement action as the defendant, but was not. 25 This element of a “similarly situated individual” has been applied in claims of selective prosecution, to require a defendant to show not only that his racial or ethnic group is prosecuted more than another group, but that a similarly situated individual in another group was not prosecuted for the same offense. 26
In this case, the government urges the Court to find that Defendant failed to make a credible showing of discriminatory effect because he has not shown that a similarly situated individual of another race was treated differently than he was. In Armstrong, a selective prosecution ease, the Supreme Court stated that it is not enough to show a discriminatory effect, “a claimant must show that similarly situated individuals of a different race were not prosecuted.” 27 While recognizing that such a showing presented some burden to the claimant, the Supreme Court balanced the burden discovery places on the government with the burden placed on the defendant, and specifically found that “[t]he similarly situated requirement does not make a selective-prosecution claim impossible to prove.” 28 The fact that the similarly situated requirement did not foreclose a defendant from a showing of discriminatory effect, was critical to the Court’s adoption of this balance, as addressed in Armstrong.
In selective prosecution claims, making a showing of a similarly situated individual is possible. The claimant is dealing with a defined universe of individuals with whom to compare; individuals who were arrested and prosecuted for the same offenses in that jurisdiction or some other jurisdiction. In Armstrong, the defendant raised a claim of selective prosecution in federal court of black defendants charged with distributing crack cocaine. Defendant could potentially make the showing that a similarly situated white defendant charged with distributing crack cocaine, was charged in state court, where the penalties were less severe.
In the seminal cases that the Supreme Court reviewed in its
Armstrong
decision, claimants could also rely on a defined universe of individuals with which to compare and show different treatment of a similarly situated individual. In
Ah Sin,
29
a habeas
But, as the Seventh Circuit’s well reasoned opinion in Chavez v. Illinois State Police 31 concludes, Armstrong is distinguishable from a selective enforcement claim based on a traffic stop. Armstrong and other selective prosecution and enforcement eases in which there is a limited and identifiable class of individuals pose no insurmountable burden to a claimant required to show a similarly situated individual. 32 But to a claimant raising a selective enforcement claim that challenges a traffic stop, such requirement imposes an insurmountable burden on the claimant. In striking the balance between the burden on the defendant of requiring a showing, with the burden that discovery places on the governmental agency, the Supreme Court in Armstrong relied on the fact that the burden on the claimant was not insurmountable, stating “[t]he similarly situated requirement does not make a selective-prosecution claim impossible to prove.” 33
In the context of a challenged traffic stop, however, imposing the similarly situated requirement makes the selective enforcement claim impossible to prove. It would require the defendant to make a credible showing that a similarly situated individual was
not stopped
by the law en
More importantly, how could a defendant even ivith discovery, show a similarly situated individual who was not stopped? Law enforcement agencies keep records of law enforcement activity. Like many police agencies, the Russell County Sheriff does not record every traffic stop an officer makes. Rather, the only recorded stops are those that result in a written ticket or warning citation. Stops that result in a verbal warning are not recorded. To the extent that there is a record, it often doesn’t include pertinent data such as race or ethnicity.
In fact, a law enforcement agency cannot record any, much less all pertinent information on the dozens, hundreds, or even thousands of motorists who are not stopped within a patrolling officer’s given space and time. If a law enforcement agency engaged in this type of activity, a nation of our democratic principles and constitutional liberties would find abhorrent such a practice of recording information on every person the officer encounters while on patrol, even those who are not arrested, stopped or otherwise engaged by the officer.
It is virtually impossible to identify a “similarly situated” individual who was not stopped. The person cannot be identified at all, nor is there any recorded information from which one can compare whether the motorists presented similar factors to an observing officer, such that there had been disparate treatment or not. Because law enforcement agencies do not make or keep records on individuals they do not stop, and certainly not on “similarly situated” individuals they do not stop, imposing such a requirement on this defendant or any defendant who challenges a traffic stop as selective enforcement, effectively denies them any ability to discover or prove such a claim. Thus, the defendant challenging a traffic stop for selective enforcement, must be allowed to show discriminatory effect in some other way.
The Tenth Circuit has indicated that a defendant can demonstrate discriminatory effect by either showing a similarly situated individual,
or
by relying on statistical evidence. In its recent
Marshall
decision,
34
the Tenth Circuit stated that claims of racially discriminatory traffic stops and arrests should be held to a “similarly high standard,” comparing them to the standard required for claims of selective prosecution. Yet, the Tenth Circuit did not foreclose the use of statistical evidence to meet the requisite showing of discriminatory effect. In
Chavez,
the Seventh Circuit stated that statistical evidence of discriminatory effect should be accepted as proof of a selective enforcement claim based on a traffic stop.
35
The Supreme
The Supreme Court has certainly considered statistical evidence as sufficient evidence of discriminatory effect in some cases. In Yick Wo, for example, discriminatory effect or impact was proved by showing that all 200 applications by Chinese launderers were denied; while only 1 of 90 applications by white launderers was denied. 37 And, in Hunter v. Underwood, 38 the Court was satisfied with statistical evidence of discriminatory effect, to wit: the challenged state law made disenfranchisement of blacks at least 1.7 times more likely than disenfranchisement of whites. While helpful, purely statistical evidence is rarely sufficient to support an equal protection claim, but can be sufficient to establish discriminatory effect. 39
Defendant’s statistical evidence
Defendant presents statistical evidence of the race and ethnicity of traffic stops by Deputy Schneider in Russell County, compared with stops by other officers in the Russell County Sheriffs Office, and compared with stops by Kansas Highway Patrol troopers patrolling a stretch of 1-70 west of the stretch that Deputy Schneider patrols. These statistics show that Deputy Schneider is stopping Hispanic motorists at a significantly higher rate than other officers in his department. Deputy Schneider’s traffic stops are on average 56.59% white motorists; 33.69% Hispanic motorists; and 9.72% black motorists. In comparison, his fellow officers’ traffic stops are on average: 90.27% white motorists; 5.57% Hispanic motorists; and 4.6% black motorists. Yet, this statistical difference may be explainable in whole or in part by the demographic differences in the motorists that Deputy Schneider encounters while on patrol with the motorists that his fellow officers encounter while on patrol. Because Deputy Schneider patrols 1-70 almost exclusively, the transient motorist population he encounters is quite different from the population encountered by his fellow officers who patrol in the city and county of Russell, but not on 1-70. In fact, there is no dispute that the transient motorist population of Hispanics along I-70 in Russell County is greater than the census population of Hispanic residents of Russell .County.
A better comparison is Deputy Schneider’s traffic stops on 1-70 with the
Lam-berth
study’s findings concerning the traffic stops by Kansas Highway Patrol on I-70. Deputy Schneider’s traffic stops of Hispanic motorists are significantly higher than the traffic stops of Hispanic motorists by Kansas Highway Patrol troopers who patrol a stretch of 1-70 some 130 miles to the west of Deputy Schneider’s area of
Clearly, “statistics are not irrefutable; they come in infinite variety and, like any other kind of evidence, they may be rebutted. In short, their usefulness depends on all of the surrounding facts and circumstances.” 40 And, this Court’s analysis of the methodology of the Lamberth study convinces the Court that there are flaws in the Lamberth statistics. 41 The Lamberth study measured the actual stop data of participating law enforcement agencies; and it measured the transient motorist population that the participating law enforcement agency was exposed to. In this ease, defendant compares Russell County’s stop data on Deputy Schneider’s traffic stops with the Lamberth study’s data on traffic stops by Kansas Highway Patrol troopers on 1-70 from Colby to the Colorado line. To the extent the Lamberth stop data is flawed, it affects the validity of a comparison with the data on Deputy Schneider.
Based on the testimony of the experts and analysis of the
Lamberth
study methodology and design, the Court is not firmly convinced that the stop data is completely accurate and reliable. The stop data was self reported by the agencies, and as Dr. Withrow noted, this presents the potential for injection of bias into the sample. Indeed, three of the agencies originally included in the study were excluded because of irregularities that might evidence some bias.
42
As Dr. Withrow testified, when a person knows their activity is being observed or evaluated, they may change their activity, a phenomena referred to as “reactivity.” The researchers did nothing
The Lamberth study breached an inviolate rule of the scientific method: the agencies surveyed were allowed to pick or suggest the sites to be studied. Whether the agencies selected the sites they perceived to be least problematic, or most problematic, or whether the agencies used some other criteria, the study is certainly not based on a random sample.
The Court finds no reason to impugn or question the data self-reported by the Kansas Highway Patrol. The Kansas Highway Patrol admittedly selected areas in which it was concerned there might be a high incidence of traffic stops of minorities. In effect, the Kansas Highway Patrol may have selected areas that exhibit a higher than average incidence of traffic stops of minorities. If that is what happened here, then the Kansas Highway Patrol’s 6.8% incidence of Hispanic motorist stops, makes even more suspect, Deputy Schneider’s 33.69% incidence of Hispanic motorist stops. Even if the Kansas Highway Patrol’s data is not overstated, or even if the Lamberth stop data on the Highway Patrol is not entirely accurate or reliable, Deputy Schneider’s incidence of stops of Hispanic motorists still establishes discriminatory effect.
There are also questions about the reliability and accuracy of the Lamberth study’s measurement of the transient motorist population. The methodology of collecting the transient motorist data was flawed. It is telling that in Mesa-Roche, Dr. Milliken testified that to obtain a good sample of data in Russell. County, one would need to collect data on 3000 to 4000 stops, over a period of two years. Yet, the Lamberth study collected data on fewer than 3000 stops, and those stops were not isolated in one city or area, but were the total stops associated with the areas surveyed for the seven participating law enforcement agencies. Moreover, the Lam-berth study collected data over a four month period. Dr. Withrow noted that there was an inadequate number of samples collected, over an inadequate period of time. He further noted that the study failed to account for seasonal changes in the demographics of drivers. This study collected data over four winter months, a period in which older drivers are less likely to drive.
Furthermore, the accuracy and reliability of those who observed and recorded the race or ethnicity of the motorists, is suspect. This data was collected by using “spotters” who observed and recorded the race and ethnicity of the motorist (as well as age, gender and the state in which the license tag was issued). An obvious concern is the accuracy of the spotters’ recorded data, in light of the short period of time they have to record this information as a car passes them by. According to Dr. Milliken, the overall reliability generally ranged from .84 for tests done in dusk/ dark conditions to .93 for tests done in daylight. But, as Dr. Withrow pointed out, there was no verification of the accuracy of the observer, by spot check, by multiple observations of the same vehicle, or otherwise. For example, the study does not detail the individual observer’s rate of completion, i.e., how many cars in which they were able to evaluate the driver, and how many they were not.
The study did not sufficiently test the accuracy of the recorded racial and ethnic data, by using multiple spotters per each vehicle. And although the study did use an “interrater reliability” test for quality control, it used this test only twice out of the many incidences of spotter tests conducted for multiple police departments at multiple places and times. Thus, there is some question as to whether the spotters accurately recorded the race or ethnicity of those they observed.
The Court also rejects the government’s criticism that the entire Lamberth study should be disregarded because it rests on an invalid presumption, that Hispanic, black and white motorists violate traffic laws at the same rate, such that any difference in traffic stops for violations suggests an improper reason for the stop. To be sure, the courts have rejected a number of Equal Protection and Fourth Amendment challenges that were based on assumptions that discrete racial or ethnic groups violate particular laws at the same rate as other groups. 43 But it is important to distinguish this case from Armstrong, 44 where the Supreme Court rejected a Court of Appeals analysis that was based on a presumption that all races commit all types of crimes. The Court noted that statistics of the United States Sentencing Commission showed distinct differences in the incidence of persons sentenced in 1994 for certain offenses: more than 90% of those sentenced for crack cocaine trafficking were black; 93.4% of those sentenced for LSD trafficking were white; and 91% of those sentenced for pornography or prostitution were white 45 The Lamberth study does not rest on a presumption that a particular offense is violated with equal frequency by members of all racial or ethnic groups. Surely those defendants who have asked the courts to accept such assumptions were misdirected, given the anecdotal, as well as statistical evidence suggesting otherwise, particularly with respect to certain offenses. Rather, in this case, Defendant is not claiming selective prosecution or selective enforcement of a particular offense.
The Lamberth study presumes that there is' no appreciable difference in the incidence of racial and ethnic groups violating one or more of the many traffic laws that justify a traffic stop. At first blush, this may seem an unsupported and unsupportable assumption, but it is not. Law enforcement officers are charged with policing a large universe of traffic violations. Deputy Schneider testified that there are dozens of reasons why he has stopped people everything from equipment problems, to signs of fatigue, to speeding, to reckless driving. In Kansas, there are well over fifty traffic violations, running the gamut from speeding, to driving under the influence, to equipment violations. 46
Officers are faced with multiple violations by multiple motorists, requiring a choice of whom to stop. Officers know that effecting a stop of one motorist may prevent them from stopping a more serious violation they observe while they are in the middle of processing that stop. Thus, they make decisions about when and when not to stop. This endowed discretion is necessary and appropriate for the effective enforcement of traffic laws and for the effective protection of public safety.
In Mesa-Roche, the government touted the New Jersey Study, as refuting any presumption that traffic laws are violated at the same rate by all groups. But, the New Jersey Study considered only one traffic violation, speeding; it did not attempt to measure the violation rate of all traffic violations that officers use to justify traffic stops. Moreover, the New Jersey Study is itself seriously flawed. 47
Thus, the proper consideration here is whether there are any differences in the incidence of traffic violations by different racial and ethnic groups when one considers all such violations, and when one considers the host of reasons officers rely on in effecting a traffic stop. Based on the testimony of the officers, that they stop vehicles for many reasons and that they stop motorists of all races and ethnic groups, but not with regard to their race or ethnicity, the Court concludes that the premise of the Lamberth study is valid. There is simply no evidence suggesting that as a whole, a particular racial or ethnic group presents the facts and circumstances justifying a traffic stop, with any greater frequency than any other racial or ethnic group.
Discriminatory Intent
Having satisfied the discriminatory effect prong, defendant must also make a credible showing of discriminatory intent, in order to obtain discovery or dismissal of his selective enforcement claim. In order to prevail, a defendant must prove that the government’s enforcement of traffic laws against him “was invidious or in bad faith and was based on impermissible considerations such as ... the desire to prevent the exercise of constitutional rights.”
48
Discriminatory intent requires more than a showing of an intentional act done with knowledge of the racial impact: “It implies that the decisionmaker ... selected or reaffirmed a particular course of action at least in part ‘because of,’ not merely ‘in
Defendant offers scant direct evidence specific to his case that would support an inference that racial considerations played a part in his stop. There is no evidence that Deputy Schneider treated, spoke to, or otherwise exhibited discriminatory behavior towards defendant, the driver or any other persons previously stopped by Schneider for a traffic violation. Schneider denies that he stopped the defendant because he was Hispanic. In fact, Schneider testified that at the moment he decided to stop the pick-up truck because of the cracked windshield, he had “no idea” of the driver’s or defendant’s race or ethnicity. Defendant notes that Deputy Schneider pulled next to the truck before signaling the driver to stop, implying that Schneider observed the driver’s or defendant’s appearance before making the stop. But Schneider testified that this is a standard practice that allows the officer to determine the number of occupants of the car for purposes of officer safety. The Court finds no reason to believe that Schneider’s decision to stop defendant came only after observing defendant’s race or ethnicity.
Direct evidence of discriminatory intent was sufficient to avoid summary judgment on a § 1983 claim of selective enforcement in the Tenth Circuit’s recent Marshall decision. 51 There the defendant was able to present evidence of the officer’s behavior during the events in question as well as the officer’s alleged record of racially selective stops and arrests in drug cases under similar circumstances. 52 The defendant offered evidence that he did not commit the alleged traffic violation and that the officer made eye contact with him prior to activating his emergency lights. As soon as the officer approached the defendant, he accused him of being on crack, an accusation the officer repeated several times during their encounter. When the officer filled out the citation form, he noted the defendant’s race, although the form called for no such designation. Most eompellingly, this officer had an extensive recorded history of similar misconduct during his prior employment as a police officer in another jurisdiction. 53
The Tenth Circuit stated that:
A challenge to the specific acts of a particular police officer bears some resemblance to a claim of racial discrimination in the use of peremptory jury challenges, which also involves the acts of a single state actor (the prosecutor) in the course of a single incident (the selection of the jury). In such cases, the Supreme Court has instructed that the court should consider all relevant circumstances, including the prosecutor’s pattern of strikes against black jurors, and the prosecutor’s questions and statements, which may support or refute an inference of discriminatory purposes.Similarly, a police officer’s pattern of traffic stops and arrests, his questions and statements to the person involved, and other relevant circumstances may support an inference of discriminatory purpose in this context. 54
In this case, however, defendant has no direct evidence, nor any circumstantial evidence showing discriminatory intent or motive, and asks the Court to do that which higher courts have already refused to do, infer discriminatory intent from the statistical evidence. 55 Despite the acknowledgment in criminal cases, and even in much of civil jurisprudence, that state of mind is not something usually subject to proof by direct evidence, in the context of the federal discrimination statutes, or in the context of constitutional claims under the Equal Protection Clause, the proponent cannot rely solely on statistical evidence to show discriminatory effect or impact.
Rather, only in “rare cases [has] a statistical pattern of discriminatory impact demonstrated a constitutional violation,” 56 though “the [Supreme] Court has accepted statistics as proof of intent to discriminate in certain limited contexts.” 57 Specifically, “[t]he Court has accepted statistical disparities as proof of an equal protection violation in the selection of a jury venire in a particular district” and “has accepted statistics in the form of multiple regressions analysis to prove statutory violations under Title VII ....” 58 It is possible that the Supreme Court would also accept statistics as sole proof of intent in the context of challenges to legislative redistricting. 59 None of these situations, however, is present in this case. Instead, the defendant asks the Court to rely on his statistics, which tend to show discriminatory effect, to conclude that Deputy Schneider is intentionally discriminating against Hispanic motorists by stopping them on the basis of their ethnicity.
Dr. Withrow testified that the
Lamberth
study, like the other studies to date, is designed to measure effect. Moreover, the agencies were instructed to have the officers record the motorists’ race and ethnicity, as determined or perceived by the officer during the course of the stop. One police department’s data was excluded because the officers recorded their perception of race or ethnicity before they made the stop. The
Lamberth
study, by design, only measures the actual race or ethnicity of the motorist perceived or determined by the officer after the stop. The
Lamberth
study does not measure the officer’s perception of race or ethnicity before making the stop. This is an important distinction,
There has been no study designed to measure an officer’s intent in making traffic stops or taking other law enforcement action. Although statistics alone are generally viewed as insufficient evidence of intent, certainly a comparison of an officer’s stops with similarly situated officers in his own police department might be evidence of an officer’s particular pattern of discriminatory intent or motive. But in this case, the other officers in the Russell County Sheriffs department were not similarly situated because Deputy Schneider patrols 1-70 exclusively, while his fellow officers patrol the town and county exclusively. The transient population is not the same as the census population. To be sure, a comparison of Deputy Schneider’s stops with the stops of Kansas Highway Patrol on the stretch of 1-70 some 130 miles west of Russell County supplies a basis to compare the activity of similarly situated officers. But the disparity that establishes discriminatory effect, does not establish that the discriminatory effect of Deputy Schneider’s conduct was itself the product of discriminatory intent.
To show intent, one must show causality. As Dr. Withrow testified, causality has three components: temporal order; correlation; and lack of alternative plausible explanations. The Lamberth study was designed to show a correlation, that is, some association (positive or negative) between the alleged cause (race or ethnicity of motorist) and effect (traffic stop).
The Lamberth study also attempted to demonstrate a lack of alternative plausible explanations. The seven agencies in the study were given an opportunity to explain why the data showed a statistically significant higher incidence of traffic stops of blacks and/or Hispanics. Most of the agencies offered no explanation, or explanations that were discredited by the agencies’ own records. 60 The government counters that there are a host of factors that go into an officer’s decision to make a traffic stop. Notably, although the government argued that the genesis of the license tag was a factor, for there are states as well as countries that are viewed as sources of drug contraband, the officers surveyed denied stopping motorists on the basis of this factor.
Dr. Withrow testified that temporal order, that is, that cause precedes effect, is difficult to show in connection with traffic stops. One would have to show that the officer perceived the motorist’s race or ethnicity before the stop and then made the stop. The Lamberth study instructed officers to record their perception or determination of race after the stop, not their perception of race before the stop. As Dr. Withrow testified, no study to date has shown temporal order. Defendant has presented no evidence that Deputy Schneider knew or perceived defendant’s race or ethnicity before he decided to make the stop. Nor is there any evidence that in any or all of his other recorded traffic stops, Schneider perceived the motorists’ race before deciding to make the stop. To the contrary, Schneider, as well as Dr. Withrow (a former law enforcement officer) testified in Mesa-Roche that officers often do not or cannot perceive the race or ethnicity of the motorist before making a decision to stop.
In the context of racial profiling, a study would be well on the road towards an identification of racial motive or intent using a valid scientific methodology, if the evaluative process of the law enforcement officer is observed in process, and data is collected on the various factors used with respect to each stop. Relevant factors would include pertinent demographic information about the motorist, and the “success” of the stop, that is, did it result in something other than apprehension for a traffic violation, such as the interdiction of drugs, money, or other contraband. Such a study might definitively determine whether law enforcement officers consider such factors as deployment strategies, departmental enforcement emphasis, resources allocation, crime rates, traffic enforcement patterns and other factors that are relatively race-neutral, as the government has posited. The identification of appropriate, race-neutral, sound and effective factors could then be used to evaluate an individual officer’s decision making, through logistic regression analysis. Of course, such a study would require candid explanations of the decision maker as he or she processed through their discretionary decision making. Perhaps there are other inherent obstacles in such a social science study of decision making. But, if an effective study of racially discriminatory motive or intent will ever be accomplished, it will probably be based on or allegiant to this model of study.
In this context, statistics may not be the sole proof of a constitutional violation and the defendant has not presented sufficient non-statistical evidence to demonstrate discriminatory intent. 61 Defendant has thus not met his burden of showing that Deputy Schneider intentionally discriminated against him. Defendant’s Motion for Discovery is denied.
Needless to say, since defendant has not satisfied the threshold showing required for discovery relating to his claims, he has not made a prima facie showing of selective enforcement or prosecution. Accordingly, defendant’s Motion to Dismiss is also denied.
IT IS THEREFORE ORDERED BY THE COURT that the Defendant’s Motion for Discovery Regarding Selective Prosecution (Doc. 20) is DENIED; Motion to Dismiss (Doc. 18) is DENIED; and Motion to Suppress (Doc. 19) is DENIED.
IT IS SO ORDERED.
Notes
. K.S.A. 8 — 1741 (b) forbids a person to "drive any motor vehicle with a damaged front windshield or side or rear windows which substantially obstructs the driver's clear view of the highway or any intersecting highway."
.
. Id. Defendant did not appeal the Court’s decision to the Tenth Circuit in that case. The Court declines, however, to go so far as to extend the doctrine of collateral estoppel to the issue of discriminatory impact, as urged by defendant.
. John C. Lamberth, Ph.D., Racial Profiling Study and Services: A Multijurisdictional Assessment of Traffic Enforcement and Data Collection in Kansas (Police Foundation, 2003).
. Dr. Milliken is a professor of statistics at Kansas State University.
. Dr. Withrow is an assistant professor of criminal justice at Wichita State University.
. The study attempted to survey ten law enforcement agencies in Kansas, including large, medium and small agencies. But, only seven agencies were effectively surveyed: Overland Park; Wichita; Emporia; Olathe; Osage County; Park City; and Kansas Highway Patrol. Three agencies submitted inconclusive or incomplete data and thus were not included in the analysis: Kansas City, Hutchinson and Marysville.
.Fed.R.Evid. 803(8)(C).
. James E. Lange, Ph.D., et al., Speed Violation Survey of the New Jersey Turnpike: Final Report (2001). This study was commissioned after the State of New Jersey entered into a Consent Decree with the United States to promote non-discriminatoiy law enforcement by New Jersey troopers under the supervision of a federal district court.
. Highly summarized, the New Jersey Study revealed that very few drivers were traveling at speeds which the study defined as speeding (i.e. 15 mph over the speed limit in 65 mph zones). Of those drivers who were speeding, white drivers were found to be under-represented and black drivers were found to be over-represented. Black motorists comprised 16% of the turnpike drivers, comprised 25% of the speeders and comprised 23% of the stops by New Jersey troopers.
. See discussion infra regarding the questionable reliability and accuracy of this measurement.
.
United States v. Ozbirn,
.
United States v. Cortez,
.
United States v. Callarman,
.
United States v. Barlow,
.
Whren v. United States,
.
. Id. at 1167. The court further noted that while it is appropriate that law enforcement identify any problem of racial profiling and undertake administrative steps to eliminate its improper use in law enforcement, it is quite another thing to "craft judicially manageable standards for determining liability,” because of the broad discretion vested in executive branch officials to determine when to prosecute and by analogy, when to conduct a traffic stop or initiate an arrest. Id.
.
United States v. Armstrong,
.
United States v. James,
.
See Armstrong,
. at 464,
.
Id.
at 468,
.
See Oyler v. Boles,
.
James,
.
See, e.g., Armstrong,
.
Armstrong,
. Id. at 466,
.
Ah Sin
v.
Wittman,
.
Yick Wo v. Hopkins,
.
. There is another notable distinction between
Armstrong
and the seminal cases it discusses, and the selective enforcement claim at bar. In
Armstrong,
the Court was mindful of the great deference due the Executive Branch in its exercise of its special province to enforce the laws, thus recognizing that long standing “presumption of regularity” accorded prosecutorial decisions. The Court was also mindful that to invalidate the San Francisco ordinances at issue in
Ah Sin
and
Yick Wo,
would constitute a serious interference by the Federal Court, with the "course of criminal justice” of a State.
Armstrong,
.Armstrong,
.
.
Chavez,
.Id.
at 637 (quoting
Int’l Bhd. of Teamsters v. United States,
.
.
.
Chavez,
.
Teamsters,
. Indeed, the Honorable Sam A. Crow has found the
Lamberth
study to be so flawed that it could not be relied upon to establish discriminatory impact.
See United States
v.
Alcaraz-Arellano,
. The data submitted by Kansas City, Kansas, a large city police department, was inconclusive because the officers recorded the race of the motorist based on their perception "before” the stop, not based on their perception after the stop, as the study required. Because the officers were trained to record "unknown” if they weren't sure, the data includes a high number of unknowns. Although this problem was corrected, the recorded stops at the benchmark locations dropped by two-thirds, suggesting that the officers were under-reporting’ the stops. Likewise, the data from Hutchinson was inconclusive because the number of reported stops at the benchmark locations dramatically decreased once the study was underway. The data from Marysville was inconclusive for another reason; during the study, the police force decreased by one-third, such that there were too few stops to create a sufficient sample.
.See, e.g., Armstrong,
.
. Id.
. The Uniform Act Regulating Traffic; Rules of the Road is found at K.S.A. Article 15, 8-1501 et. seq. This Act includes, by the Court's count, over 50 rules of the road and thus, over 50 potential traffic violations.
. In the New Jersey Study, great measures were taken to assure the accuracy of the identification of race. The race or ethnicity of the motorists was determined by teams of three persons who studied each photograph. If two of the team of three could not agree on the race or ethnicity, the data was discarded. Thus, persons whose race or ethnicity was not readily determinable were not included in the study. But persons whose race was more readily perceivable, such as dark or brown skinned blacks, might therefore be over represented in the study. For this reason alone, the New Jersey Study’s conclusion that black motorists have a statistically significant higher incidence of speeding is suspect.
.
United States v. Salazar,
.
McCleskey,
.
Armstrong,
.
Marshall,
. Id.
. Id. at 1170-71.
.
Id.
at 1171 (internal quotations omitted) (quoting
Batson v. Kentucky,
.
Cf. United States v. Jones,
.
McCleskey,
.
Id.
at 293,
.
Id.
at 293-94;
see, e.g., Teamsters,
.
Chavez,
. For example, Osage County explained that their dramatically higher incidence of stops of Hispanics might be attributed to a high incidence of Hispanic motorists in the county, who drive old cars back and forth from Mexico, resulting in traffic stops for equipment violations. But this explanation proved to be without merit, for the county's records did not show that Hispanics were being stopped for equipment violations.
. See id. at 648.