Opinion
Criminal Action No. 09–10352–LTS.
04-16-2015
UNITED STATES of America, v. Foster L. STARKS, Jr.
David G. Tobin, United States Attorneys Office, Boston, MA, for United States of America.
David G. Tobin, United States Attorneys Office, Boston, MA, for United States of America.
ORDER DENYING DEFENDANT'S MOTION TO SUPPRESS
LEO T. SOROKIN, District Judge.
Defendant Foster Starks, Jr.'s motion to suppress touches upon a serious issue that has been the subject of much discussion by, among others, Courts and law enforcement leaders: whether African Americans are stopped and searched by police because of their race. See, e.g., United States v. Leviner, 31 F.Supp.2d 23, 33 (D.Mass.1998) (“Motor vehicle offenses, in particular, raise deep concerns about racial disparity. Studies from a number of scholars, and articles in the popular literature have focused on the fact that African American motorists are stopped and prosecuted for traffic stops, more than any other citizens.”); James B. Comey, Director, Fed. Bureau of Investigation, Hard Truths: Law Enforcement and Race (Feb. 12, 2015), available at http://www.fbi.gov/news/speeches/hard-truths-law-enforcement-and-race (“Serious debates are taking place about how law enforcement personnel relate to the communities they serve, about the appropriate use of force, and about real and perceived biases, both within and outside of law enforcement.”). The pending motion, though, concerns a specific and narrow question: whether, as Starks asserts, Massachusetts State Trooper Jason Vital followed and then stopped the car Starks was driving because Starks is black.
Originally, another judge of this court denied Starks's motion to suppress on standing grounds. Doc. No. 312 at 2. The First Circuit reversed and remanded for an evidentiary hearing, United States v. Starks, 769 F.3d 83 (1st Cir.2014), which the Court conducted on March 9 and 18, 2015. At the conclusion of the hearing on March 9, 2015, the Court denied Starks's Motion to Suppress insofar as it raised a Fourth Amendment challenge to the stop, finding that there was reasonable suspicion for the stop based on a registration discrepancy with respect to the color of the car. Doc. No. 354. Subsequently, the Court issued a written decision finding that the marked-lanes violation also provided probable cause to stop Starks's car. Doc. No. 352 at 1. The Court now proceeds to consider the second theory advanced in Starks's Motion to Suppress—that this Court should suppress the evidence gathered as a result of the stop because Vital stopped Starks due to race in violation of the Equal Protection Clause of the Fourteenth Amendment. Doc. No. 344 at 5, 13.
The basic facts are clear. On May 24, 2009, Starks was driving a car rented by a friend. He was alone. At approximately 11:05 p.m., Starks pulled over into the breakdown lane of Route 24. Vital saw this and pulled over to see if everything was all right. Before getting out of his cruiser, and pursuant to his customary practice, Vital “ran” the license plate for the car Starks was driving. While awaiting a response regarding the plate, Vital approached the rear of Starks's car and stopped near the rear driver's side wheel. Starks was standing just outside the driver's door. They spoke briefly. Starks explained he had dropped his cigarette on the floor, and he retrieved it as the two spoke. Vital confirmed that Starks did not need assistance. During this brief discussion, Vital observed that Starks was black. Nothing unusual or untoward occurred during the interaction, after which both men returned to their cars. Thereafter, Starks pulled back into the travel lanes, and Vital followed Starks's car for a relatively short distance. As he followed behind Starks, and perhaps as soon as he returned to his cruiser, Vital learned that the information listed on the car's registration differed in color (but not make or model) from the car Vital observed. Also, while following Starks, Vital observed Starks's car partially moving into a different lane several times. Citing both the registration discrepancy and the lane shifting, Vital initiated a traffic stop. Upon stopping the car, Vital obtained Starks's license, learned that it was revoked, and arrested him for this offense. Then, during an inventory search pursuant to established state police policy, Vital discovered a firearm and ammunition in a bag in the front seat of the car.
Nearly twenty years ago, the Supreme Court unanimously held that “ulterior motives,” such as race-based discrimination, only impact the Fourth Amendment analysis “in the absence of probable cause.” Whren v. United States, 517 U.S. 806, 811–12, 116 S.Ct. 1769, 135 L.Ed.2d 89 (1996) (emphasis in original). Although “the Constitution prohibits selective enforcement of the law based on considerations such as race[,] ... the constitutional basis for objecting to intentionally discriminatory application of laws is the Equal Protection Clause, not the Fourth Amendment.” Id. at 813, 116 S.Ct. 1769. Therefore, once the Court found that Vital possessed probable cause to pull over Starks, Supreme Court precedent compelled the Court's finding that Vital did not violate Starks's Fourth Amendment rights. See Doc. Nos. 352, 354.
Whether suppression is the appropriate remedy for Equal Protection violations remains an open issue in the First Circuit. Starks, 769 F.3d at 90 n. 6. Assuming for present purposes, without deciding, that suppression is a permissible remedy, Starks must prove not only that Vital's actions had a discriminatory effect, but also that Vital was motivated by a discriminatory purpose. Wayte v. United States, 470 U.S. 598, 608, 105 S.Ct. 1524, 84 L.Ed.2d 547 (1985). In his motion to suppress, Starks sought to establish discriminatory effect and discriminatory purpose through the use of statistical evidence. Starks obtained two sets of data regarding troopers in Area D–4 (Vital's area at all relevant times): the first set encompassed the period October 2008 to May 2009, and the second set encompassed the period June 2008 to May 2009. Both sets of data contained breakouts by trooper and, for each trooper, the number of stops of white drivers and black drivers. While the first set of data made no distinction between stops during daytime and nighttime, the second set of data divided the data for each trooper between day and night stops. Starks retained an academic expert to analyze the data and submitted both an affidavit from the expert and live testimony. The government made no challenge to the expert's qualifications or credibility (in the sense of honesty or truthfulness), though it disputes the expert's analysis and conclusions. Neither side presented any other evidence.
Testimony at the hearing clarified that this was the true time period, rather than the time period listed in the expert's affidavit. Doc. No. 361 at 37:15.
There was miscommunication between Starks's attorney and the expert regarding whether the data that the expert relied upon related to only stops of black and white drivers or of black, white, and Hispanic drivers. See, e.g., Doc. No. 361 at 50:23–54:10. These discrepancies only buttress the Court's finding that the evidence is weak.
--------
The defense expert determined that from October 2008 to May 2009 (the first data set), 19.8% of drivers that Vital stopped were black, whereas 14.3% of drivers that all other troopers in Area D–4 stopped were black. Doc. No. 357 at 1. The expert opined that there was a 3% probability that the difference between Vital's stop rate and the other troopers' stop rate was due to chance. Id. The expert also analyzed stop data from May 2008 to May 2009 (the second data set) and found that during this period, 17.5% of all drivers that Vital stopped were black, whereas 15.3% of drivers that all other troopers stopped were black. Id. at 2. The expert opined that there was a 7% probability that the difference between Vital's stop rate and the other troopers' stop rate was due to chance. Id. The expert acknowledged that if the longer data set provided a more accurate picture of Vital's stop patterns, it weakened his findings in the first data set. Id. However, the expert hypothesized that perhaps Vital and/or the other members of his troop changed their stop behavior over time. Id. To explore this possibility, the expert split the second data set (May 2008–May 2009) into a 2008 subset and a 2009 subset. The expert chose calendar year as the break point a priori (without regard to month by month differences), simply based on the normal convention of looking at data in different years separately. Id. In 2008, 14.5% of all drivers Vital stopped where black, whereas in 2009, 21% of all drivers Vital stopped were black. Doc. No. 357 at 2.
Finally, the expert analyzed data specific to nighttime stops (the second data set, spanning May 2008–May 2009). Id.; Doc. No. 361 at 32:10–21. Twenty-point-eight percent (20.8%) of the drivers Vital stopped at night were black, compared with 20.7% for the rest of the troop. Doc. No. 357 at 2. Given the differences that the expert found comparing 2008 with 2009 data for the overall stops, the expert decided to break up the nighttime data by calendar year as well. Id. In 2009, 23.7% of drivers Vital stopped at night were black, whereas 17.5% of the drivers the rest of the troop stopped were black. Id. In May 2009 (the month Vital stopped Starks), 25.5% of drivers Vital stopped at night were black, whereas 17.8% of drivers the troop stopped at night were black. Id. The expert opined that “the probability that randomness would have produced [a] difference this large or larger is 9.1%.” Id.
The record presents no direct or circumstantial evidence that Vital stopped Starks because of his race other than the statistical evidence. For the reasons explained below, the Court finds that the expert's opinions fail to support the conclusion that Vital stopped Starks because of Starks's race.
First, the expert's own data defeats Starks's contention that Vital engages in racial profiling. Vital testified that he typically patrols at night. Doc. 98 at 85:23–86:7. This suggests, and I find, that the nighttime data provides the proper basis for comparing Vital's stops to those of other members of his troop. The nighttime data is a natural close fit for comparison purposes, while the daytime data distorts the analysis because of the differences between day and night traffic on highways. Some of these differences (such as volume) are self-evident, while others (such as whether the racial composition of drivers between the two time periods differ) are less clear, especially in the absence of any evidence on the point.
Vital's nighttime stop rate from May 2008 to May 2009 is virtually identical to the nighttime stop rate of the troop: 20.8% and 20.7%, respectively. The defense expert did not suggest that this slight difference had any statistical significance. In other words, Vital stopped black drivers at the same rate as the other members of the troop during that period. The expert also examined the nighttime data by comparing 2008 to 2009. While the expert found that, in 2009, Vital had a nighttime stop rate of black drivers higher than that of the troop (23.7% versus 17.5% percent). The Court does not find this analysis reliable. Neither the expert nor any other evidence explains the increased percentage, and the expert made no effort to determine, for example, whether the difference arose from a continuous trend, resulted from a one-month spike, or occurred because the data set was too small to permit meaningful conclusions.
Second, “[statistical] results are considered significant if the probability that the observed effect occurred by chance is lower than a certain percentage level, usually 5%. ” Fudge v. City of Providence Fire Dep't, 766 F.2d 650, 658 (1st Cir.1985) (citing D. Baldus & J. Cole, Statistical Proof of Discrimination 291 (1980)) (emphasis added); U.S. S.E.C. v. Selden, 632 F.Supp.2d 91, 94 (D.Mass.2009) (“A p value of 0.05 or less (indicating a 95% level of certainty that the observed effect was not randomly induced) is generally accepted as showing a statistically-significant effect.”). During the hearing, the Court asked the expert whether he thought there was enough data to draw conclusions about the racial profiling trend that the expert testified he had identified. Doc. No. 361 at 24:14–18. The expert did not respond to this question directly, but stated that “the way to understand the numbers in the affidavit, if I say something is significant at the 7 percent level, it says that there was only a probability of 7 percent that there would be this big or a bigger disparity between Officer Vital and other officers.” Id. at 25:8–12. However, in his report, the expert stated: “For the month of May 2009, in which Foster Starks was stopped, the ratios are 25.5% to 17.8%; the probability that randomness would have produced difference (sic) this large or larger is 9.1%. This relatively small probability is due to small sample size. ” Doc. No. 357 at 2 (emphasis added). This statement reveals the answer to the Court's question—the expert presented data with a lower level of statistical significance (which correlates to a 9.1% likelihood of chance) than is typically accepted (which correlates to a 5% likelihood of chance), due to small sample size. The expert never explained, in his affidavit or his testimony, the basis for assigning meaning to data with a lower than ordinary level of statistical significance, nor did he acknowledge the fact that he did not apply the generally accepted standard. Therefore, the Court declines to rely upon the expert's testimony in this regard.
Third, Starks points out that the percentage of blacks in the population of the towns surrounding the area of the car stop, according to the Census, is 11%, and he contrasts that figure with the fact that 20.8% of the drivers Vital stopped at night were black. The government suggests that the Census undercounts African Americans, especially in urban areas like Brockton, which is within the Census area referred to by the parties here. Neither side presents any data or evidence regarding the actual or predicted racial composition of drivers on Route 24 (other than the Census and the aforementioned stop data).
Finally, the finding requested is significant, the relief sought is substantial, and the general issue of great concern, but the evidence offered in support of the motion is weak. The most salient evidence establishes that Officer Vital stops black drivers no more or less than the other troopers in Area D–4. As to whether the Area D–4 troopers generally stop a disproportionate number of black drivers on the road, there is only the evidence from the Census that black people constitute eleven percent of the overall population in the towns patrolled by Area D–4, almost all of whom live in Brockton. In this case, that is really not helpful evidence at all. The troopers patrol two limited access highways, Interstate 495 and Route 24, along with Route 140. The universe of drivers on these roads is not merely a mirror image of the nearby towns falling within the purview of Vital's barracks. Accordingly, the Court finds that Starks has not met his burden to prove a prima facie case of discriminatory purpose, let alone demonstrated that Vital acted with a discriminatory purpose in violation of the Equal Protection Clause. See United States v. Nichols, 512 F.3d 789, 795 (6th Cir.2008) (stating that although the burden of proof shifts to the government after the defendant establishes a prima facie case of discrimination, “[t]he defendant retains the ultimate burden of proving discrimination”). Starks's motion to suppress, Doc. Nos. 71, 344, is DENIED.
SO ORDERED.