NOT PRECEDENTIAL

UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT _____________ No. 24-1538 _____________ UNITED STATES OF AMERICA v. BRIAN FOSTER, Appellant _____________ On Appeal from the United States District Court for the Middle District of Pennsylvania (D.C. No. 3:23-cr-00166-001) District Judge: Hon. Malachy E. Mannion _____________ Submitted Pursuant to Third Circuit L.A.R. 34.1(a) December 3, 2024 _____________ Before: SHWARTZ, MATEY, and McKEE, Circuit Judges (Filed: April 2, 2025) _____________ OPINION * _____________

This disposition is not an opinion of the full Court and, pursuant to I.O.P. 5.7, * does not constitute binding precedent.

MATEY, Circuit Judge.

Brian Foster pleaded guilty to possessing a weapon while incarcerated. He appeals the District Court’s denial of his request for discovery to support a selective enforcement claim. Because the District Court acted within its discretion to deny further discovery on this collateral issue, we will affirm.

I While serving his sentence for homicide at FCI Schuylkill, a corrections officer inspected Foster and found a piece of clear plexiglass sharpened to a point with a tape handle inside Foster’s pants pocket. Foster acknowledged that the plexiglass was a weapon, and he was indicted for possessing contraband in violation of 18 U.S.C. § 1791(a)(2), (b)(3).

Rather than defend against the charge directly, Foster alleged that FCI Schuylkill officials selectively enforce prison contraband laws based on race. In support of a motion for discovery on this theory, Foster claimed that over a three-year period the United States prosecuted thirty-nine FCI Schuylkill inmates for contraband-only offenses. Of those inmates, thirty were black, eight were hispanic, and one was white. He argued that when contrasted against the nationwide racial composition of inmates, his statistics permit an inference that contraband investigations at FCI Schuylkill are conducted on the basis of race. Foster asked the District Court to compel the United States to produce from the same three-year period: 1) all incident reports related to possession of a weapon or drugs; 2) all forms referring the same incidents to the United States Attorney’s Office; 3) all discipline reports related to the same incidents; 4) the race data related to the inmates involved in the same incidents; and 5) the prison policies on officials’ discretion to refer such incidents for prosecution.

The United States voluntarily produced several of the items Foster requested.

These materials clarified that only 18% of FCI Schuylkill’s 1,027 inmates were white.

The United States also confirmed that Foster identified all contraband-only prosecutions that occurred during his chosen period. And the United States provided FCI Schuylkill’s policy to refer all contraband offenses for prosecution. Based on this additional information, the United States opposed the rest of Foster’s discovery requests as overly broad and unreasonable.

The District Court denied Foster’s motion. The Court concluded that Foster’s proffer indicates “a racial disparity in the rate of prosecution” for contraband-only offenses. App. 145. But it found insignificant the sole prosecution of a white inmate based on the low population of white inmates housed at FCI Schuylkill, and the small sample of total prosecutions (some thirty-nine). The District Court also noted that prison- contraband prosecutions leave little room for bias as the facility limits investigations to inmates who have chosen to possess contraband. On balance, the District Court concluded it could not reasonably infer that prison officials declined to refer contraband offenses because the offending inmates were white.

Foster subsequently entered a conditional guilty plea, reserving the right to appeal the denial of his motion for discovery. 1 II The Fifth Amendment prohibits federal investigations of alleged criminal conduct “based on ‘an unjustifiable standard such as race, religion, or other arbitrary classification.’” United States v. Armstrong, 517 U.S. 456, 464 (1996) (quoting Oyler v. Boles, 368 U.S. 448, 456 (1962)); see United States v. Washington, 869 F.3d 193, 214 (3d Cir. 2017). A defendant may collaterally attack his indictment with “clear evidence” that investigations or prosecutions were not pursued against similarly situated offenders based on some impermissible consideration. Armstrong, 517 U.S. at 465 (quoting United States v. Chemical Found., Inc., 272 U.S. 1, 14 (1926)).

A district court may permit limited discovery on a selective enforcement claim if shown evidence of discriminatory effect through “reliable statistical evidence, or its equivalent,” that is “strong enough to support a reasonable inference of discriminatory intent.” Washington, 869 F.3d at 221. But it must also consider “judicial economy and the need to avoid protracted pretrial litigation.” Id. We review only whether the District Court’s weighing of Foster’s proffer was “arbitrary, fanciful[,] or clearly unreasonable.”

See United States v. Collins, 36 F.4th 487, 494 (3d Cir. 2022) (quoting Democratic Nat’l Comm. v. Republican Nat’l Comm., 673 F.3d 192, 201 (3d Cir. 2012)).

The District Court had jurisdiction under 18 U.S.C. § 3231. We have jurisdiction under 28 U.S.C. § 1291 and review the District Court’s discovery ruling for abuse of discretion. See United States v. Washington, 869 F.3d 193, 213 (3d Cir. 2017).

It was not. The District Court reasonably noted Foster’s proffer did not account for either the small sample size offered, see, e.g., Blunt v. Lower Merion Sch. Dist., 767 F.3d 247, 276 (3d Cir. 2014), or the self-selection of inmates who chose to possess contraband.

And unlike the Washington defendant, whose decision to engage in unconsummated criminal activity was not alone dispositive, Foster completed his crime before the investigation began. See Washington, 869 F.3d at 222. Additionally, the strength of Foster’s initial proffer diminished when the United States voluntarily produced data showing that white inmates are a substantial minority of FCI Schuylkill’s population and a policy indicating investigators lacked discretion. See id. at 221 (“[T]he end ‘goal’ of such a discovery motion is a valid claim of selective enforcement under the heightened substantive standards.”). The District Court therefore “operate[d] within its discretion to deny additional discovery.” Id. *** For these reasons, we will affirm the District Court’s order denying Foster’s discovery motion.

McKEE, Circuit Judge, concurring.

I agree that Foster has not produced sufficient evidence to support a finding of selective enforcement, but my reasoning differs somewhat from that of my colleagues.

I.

In United States v. Washington, we examined “the distinction between ‘selective prosecution’ and ‘selective enforcement’” claims. 1 We explained that selective prosecution claims focus on “the actions of prosecutors (in their capacity as prosecutors),” whereas selective enforcement claims focus on “the actions of law enforcement and those affiliated with law-enforcement personnel.” 2 Both “[s]ubstantive claims . . . are generally evaluated under the same-two part test”—“[a] defendant challenging a criminal prosecution at either the law enforcement or prosecution inflection points must provide ‘clear evidence’ of discriminatory effect and discriminatory intent.” 3 However, the standard for obtaining discovery to support each claim is distinct. 4 Under the Armstrong/Bass test 5 that applies to selective prosecution claims, a defendant must make a “credible” showing with “‘[s]ome evidence’ . . . that similarly situated persons were not prosecuted.” 6 In contrast, although “courts contemplating motions for discovery on selective enforcement claims must still be guided by the spirit

United States v. Washington, 869 F.3d 193, 214 (3d Cir. 2017). Id. at 214. Id. Id. at 220–21.

This test derives from United States v. Armstrong, 517 U.S. 456 (1996) and United States v. Bass, 536 U.S. 862 (2002). Id. at 197.

Id. at 214–15. of Armstrong/Bass,” that test does not apply to selective enforcement claims. 7 Rather, in selective enforcement claims: a district court retains the discretion to conduct a limited pretrial inquiry into the challenged law-enforcement practice on a proffer that shows “some evidence” of discriminatory effect. The proffer must contain reliable statistical evidence, or its equivalent, and may be based in part on patterns of prosecutorial decisions . . . . Distinct from what is required under Armstrong/Bass, a defendant need not, at the initial stage, provide “some evidence” of discriminatory intent, or show that (on the effect prong) similarly situated persons of a different race or equal protection classification were not arrested or investigated by law enforcement. However, the proffer must be strong enough to support a reasonable inference of discriminatory intent and non-enforcement. 8 Foster argues that the evidence he produced of a statistical difference between the racial/ethnic makeup of the thirty-nine contraband defendants and the racial/ethnic makeup of the population of FCI Schuylkill supports a reasonable inference of discriminatory intent and non-enforcement. Although the District Court accepted the evidence, it refused to conclude that it supported a reasonable inference of selective enforcement because the District Court believed that the sample size of thirty-nine defendants was too small to be reliable.

The sample size is obviously quite small, but that does not necessarily negate its ability to support a reasonable inference. Fundamentally, all statistical samples are quite small relative to the universe they attempt to represent. That is the beauty and significance of statistical sampling. It allows a reliable conclusion about a larger population based upon a small (but representative) sampling of that population.

Id. at 220.

Id. at 220–21 (emphasis added).

Although “[c]onsiderations such as small sample size may . . . detract from the value of [statistical] evidence,” 9 if the statistical sample here actually consisted of thirty-nine similarly situated defendants, I might be willing to assume that it was adequate to support Foster’s conclusion.

However, in Washington, we explained that “the proffer must contain reliable statistical evidence, or its equivalent.” 10 The government argues on appeal that Foster’s “limited proffer” did not establish a reasonable inference because it “included both questionable and inaccurate data.” 11 I agree. Foster’s evidence has the following flaws: A. Foster’s data collection methods were questionable.

Foster identified thirty-nine contraband defendants based on an ECF search but admitted there could have been additional defendants that he missed because their cases were either miscoded on ECF or filed under seal. Foster offered no evidence of the race of three defendants but assumed they were coded as White. He also assumed that seven of the eight White defendants were Hispanic based on their last names and a website of unverified accuracy called “namecensus.com.” 12 B. Foster’s choice of data to compare was questionable.

Foster initially compared the thirty-nine contraband defendants to the population of the BOP at large. Only after the government provided information about the

Int’l Bhd. of Teamsters v. United States, 431 U.S. 324, 340 n.20 (1977).

Washington, 869 F.3d at 221.

Answering Br. 16.

The government initially affirmed Foster’s assertion that there were thirty-nine contraband defendants of whom thirty were Black, eight were Hispanic, and one was White. Thus, it might be possible to overlook these particular flaws in this data. population of FCI Schuylkill did he change his comparison so that it was based on the relevant institution.

Although everyone in his dataset was charged with possessing contraband in violation of 18 U.S.C. § 1791(a)(2), the statute defines several different kinds of “prohibited objects,” including weapons, controlled substances, cellphones, and currency. 13 While some individuals in Foster’s dataset were charged with possessing weapons, many others were charged with possessing controlled substances or cellphones.

Foster failed to distinguish between these three categories of contraband defendants.

Foster was charged with weapons contraband. It is not only conceivable that prison officials might treat possession of each category of prohibited objects differently, it is understandable and perhaps even likely. For example, corrections officials could be granted discretion in how they enforce prohibitions against possessing currency but be required to refer weapons contraband for prosecution.

If we look at weapons only contraband, then one out of only sixteen defendants was White as opposed to one out of thirty-nine. This sixteen-person dataset is even less reliable and less statistically significant than Foster’s original dataset. Given the already small sample size, this reduction in relevant data further undermines the reliability of Foster’s conclusion.

C. Foster’s selected time period was questionable.

18 U.S.C. § 1791(d)(1)(A)–(D). The statute further distinguishes among numerous types of controlled substances such as methamphetamine, marijuana, and alcoholic beverages.

Foster did not explain why his statistical inquiry relied upon a three-year time period and there is no reason why we should conclude that even a statistically valid analysis of the three years relied upon would yield a “fair” statistical conclusion.

Although Foster was indicted on June 20, 2023, he continued the three-year dataset until July 19, 2023. There is no explanation of why the dataset continued after his suit began.

If we cannot consider cases after Foster’s indictment, the number of weapons contraband cases drops from sixteen to fourteen, and our dataset is further reduced to one White defendant out of the fourteen. There is absolutely no reason to conclude that a dataset that small can support a reliable inference of selective enforcement.

D. Foster’s number of contraband defendants was inaccurate.

Originally, Foster identified thirty-nine contraband defendants. In his reply brief, however, he noted that a superseding indictment had eliminated one of the thirty-nine but failed to identify the eliminated defendant’s name, race, or contraband category.

Considering this already very small dataset, I am not willing to assume that this change would have no bearing on the reliability of Foster’s conclusion about the likelihood of selective enforcement.

II.

As a result of the many flaws in the data, it is impossible to conclude that Foster proffered “reliable statistical evidence, or its equivalent.” 14 Foster could have strengthened his proffer by providing additional evidence, including expert analysis to explain and substantiate his argument about the validity of his dataset. 15 Absent more than appears on this record, we are in no position to hold that the District Court abused its discretion in rejecting Foster’s claims of selective enforcement. I therefore agree with the majority’s conclusion that we must affirm the District Court’s order.

Washington, 869 F.3d at 221.

Courts may also “consider the product of earlier investigations in deciding whether to conduct pretrial discovery on the individual claims they happen to confront.”

Washington, 869 F.3d at 221. If previous investigations or lawsuits involving FCI Schuylkill revealed selective enforcement or racial bias, this evidence could strengthen a motion for discovery. Moreover, an affidavit from Foster, or another inmate at the prison who witnessed selective enforcement, could also bolster his motion.