21-1901-cr United States of America v. Nieves

UNITED STATES COURT OF APPEALS FOR THE SECOND CIRCUIT

August Term, 2022

Argued: October 28, 2022 Decided: January 26, 2023

Docket No. 21-1901-cr

UNITED STATES OF AMERICA,

Appellee,

— v. —

CHRISTIAN NIEVES, also known as Sealed Defendant 1,

Defendant-Appellant,

ELIAS POLANCO, also known as Sealed Defendant 2,

Defendant.*

* The Clerk of Court is directed to modify the caption to conform to the above. B e f o r e:

LYNCH, LEE, and ROBINSON, Circuit Judges.

Christian Nieves appeals from a judgment of the United States District Court for the Southern District of New York (Rakoff, J.) sentencing him to 36 months in prison following his conviction by a jury of witness retaliation. Nieves challenges the manner in which the district court conducted jury selection, arguing primarily that the district court neglected to adequately screen prospective jurors for bias against gang members and that the voir dire process was too abbreviated to allow for informed peremptory and for-cause challenges. We agree that, under these circumstances, the district court exceeded its discretion by failing to sufficiently account for the risk of gang-related bias among prospective jurors, and we therefore VACATE the judgment of the district court and REMAND for further proceedings.

JUN XIANG, Assistant United States Attorney (Allison Nichols, Karl Metzner, Assistant United States Attorneys, on the brief), for Damian Williams, United States Attorney for the Southern District of New York, New York, NY, for Appellee.

EDWARD S. ZAS, Federal Defenders of New York, Inc., Appeals Bureau, New York, NY, for Defendant-Appellant.

GERARD E. LYNCH, Circuit Judge:

“[W]e have never reversed a conviction for the failure to ask a particular

question of prospective jurors.” United States v. Bright, No. 20-3792,

2 53621, at *1 (2d Cir. Jan. 6, 2022) (summary order). That is no great surprise:

district judges are afforded broad discretion in conducting voir dire. That

discretion, however, is not boundless.

Christian Nieves (“Nieves”) challenges his conviction, by jury trial, on one

count of witness retaliation. The essence of his appeal is that the district court’s

(Jed S. Rakoff, J.) abbreviated voir dire left him, and the district court, unable to

meaningfully screen prospective jurors for bias against gang members, rendering

Nieves’s trial fundamentally unfair. While we disagree with Nieves’s more

ambitious arguments on appeal that various individual district court decisions in

connection with voir dire were, on their own, per se reversible – including his

contention that the district court was required specifically to ask prospective

jurors about gang-related bias – we agree that under these circumstances, the

district court abused its discretion by failing to take any of several possible steps

that could have effectively screened prospective jurors for such bias, or to take

other steps to counter any bias that may in fact have existed among the venire. We

therefore VACATE Nieves’s conviction and REMAND to the district court for a

new trial.

3 BACKGROUND

I. The Government’s Case against Nieves

A. The Theory of the Case

In May 2019, Nieves – who also goes by the name Eric Rosario and the

moniker “White Boy” – was indicted on four counts: witness retaliation in

violation of

(b)(1) (Count One); conspiring with codefendant Elias

Polanco to commit witness retaliation in violation of

(f) (Count

Two); witness tampering in violation of

(b)(3) (Count Three); and,

again with Polanco, conspiring to commit witness tampering in violation of

(k) (Count Four).

Those charges stemmed from a February 2019 clash between Nieves and

Miguel Carela. Both men were members of the Trinitarios, a “violent gang” of

Dominican-American origin. App’x 126. The government’s theory was that

Carela’s decision to testify for the government in a separate 2018 criminal trial ran

afoul of the group’s strict “code of silence,” which includes a pledge not to

cooperate with law enforcement. App’x 630. Carela, again taking the stand in this

case, testified that after his 2018 testimony and subsequent release from prison, he

4 confided in two Trinitario peers that he “didn’t want to be part of that [Trinitarios]

life anymore.” App’x 255.

According to the government, Carela’s perceived disloyalty made him a

marked man. On February 5, 2019, after encountering Carela on a Bronx sidewalk,

Nieves pursued and physically confronted Carela, slashing Carela’s face with a

razor blade and running off with his wallet and sneakers. Although the parties

disputed some of these details at trial, including whether there was actually a

slashing or whether this was merely a fistfight, the basic fact that the two men

fought was not in dispute.

Rather, the trial centered on why they fought. While Nieves characterized

the incident as an unremarkable “Bronx street fight,” App’x 1004, the government

construed Nieves’s motivations as both retaliatory and proactive – i.e., as witness

retaliation (for Carela’s past cooperation) and witness tampering (to discourage

future cooperation). In support of that theory, the government elicited testimony

from Carela that Nieves told him in the midst of the fight that “[t]his is happening

to you for being a snitch.” App’x 265-66. According to Carela, he received phone

calls the next day from codefendant Polanco warning Carela to “take my name

out of your mouth” and threatening to “come over to where you are”; at one

5 point, Nieves joined the call and told Carela that “[i]t was a good thing that that

had happened to [Carela], for being a snitch.” App’x 284, 288-90.

B. The Evidence at Trial

Trial began on April 14, 2021. The government’s case-in-chief centered

largely on the above testimony from Carela, along with general insight into the

Trinitarios from expert witness Detective Paul Jeselson. Jeselson walked the jury

through the Trinitarios’ written constitution, which lays out the group’s “code of

silence” and provides that “traitor[s]” who flout that rule will be “severely

punished.” App’x 630-32. Jeselson also discussed his general experience

investigating “hits . . . put out” by the Trinitarios, as well as “other violence,”

including “kidnappings, robberies, [and] shootings” by members of the group.

App’x 626. This dovetailed with Carela’s testimony concerning his direct

participation in Trinitario missions that involved shootings, assaults, murder, and

other crimes. Those themes resurfaced at closing, where the government

contended that Carela, himself a “violent gang member,” App’x 982, had been

targeted by the defendants because he was a “snitch” – as “corroborated” by

Jeselson’s testimony regarding the “nature and organization and rules of the

Trinitarios gang.” App’x 986-90.

6 For his part, Nieves’s defense consisted of brief testimony from two

witnesses: a probation officer who spoke about Carela’s post-incarceration

association with Trinitarios members and continued cooperation with law

enforcement, and a detective who had investigated Nieves’s assault and could not

recall Carela reporting any other assailants or subsequent threats. In closing,

Nieves’s counsel argued that Carela had lied about Nieves calling him a “rat and

a snitch,” and in so doing had “transformed a Bronx street fight into a federal

case.” App’x 1004-05.

C. Verdict and Sentence

Only two counts reached the jury. Prior to trial, the government abandoned

Count Two (conspiracy to commit witness retaliation). After the government

rested, the district court entered a judgment of acquittal on Count Three (witness

tampering). Ultimately, the jury found Nieves guilty on Count One (witness

retaliation), but acquitted both Nieves and Polanco on Count Four (conspiracy to

commit witness tampering).

In July 2021, Nieves was sentenced to 36 months’ imprisonment to be

followed by three years’ supervised release.

7 II. Jury Selection

The jury selection process at the heart of this appeal took place in the

opening moments of trial proceedings.

A. The Parties’ Proposed Voir Dire Questions

Prior to trial, both sides proposed gang-related voir dire questions. The

government suggested asking whether prospective jurors or their families had

ever “been involved . . . with a case involving a gang or other racketeering

activity,” and whether the fact that this case “involves racketeering activity,

specifically, a gang” would affect their judgment. App’x 33. The defendants

recommended asking, among other similar inquiries, whether jurors had any

“particular interest” in “[s]treet gangs” or “street violence,” or any “personal

experiences or feelings arising from any acts of violence . . . that would affect your

impartiality in a case involving alleged gang violence.” App’x 50, 52. The

defendants also requested that the district court query whether any prospective

jurors “ha[d] any negative feelings, stereotypes, or opinions about Latino

Americans,” and whether they believed that “Latino American immigrants are

more likely to commit crimes than immigrants of [other] backgrounds,” along

with related questions about immigration. App’x 51.

8 The parties also suggested that the district court question jurors about a

suite of other topics, including jurors’ feelings about law enforcement, their

educational and employment backgrounds, what they like to read and watch,

their families, and their hobbies. More generally, the defendants urged the district

court to individually question each prospective juror in a manner that “explores

specific concerns related to the circumstances of this case, and potential biases.”

App’x 43. A perfunctory individualized voir dire, they insisted, would be

“insufficient to ensure the defendants an impartial jury and a fair trial.”

The district court instead opted for a far swifter approach than either side

had advocated.

B. Voir Dire

At the outset, the basic format of voir dire was nothing out of the ordinary.

The court seated 12 initial prospective jurors to be questioned as a group and

directed the balance of the pool to “listen carefully to my questions, because some

of those people may be excused and then you may be called up in their place and I

don't want to have to repeat all of my questions.” App’x 59-60. As prospective

jurors were dismissed, the court confirmed that their replacements had heard its

initial questions and asked whether those questions raised “[a]ny other issues you

9 need to respond to.” E.g., App’x 68.

The court then presented a high-level overview of the charges, without

referencing gangs:

The two defendants are named Christian Nieves and Elias Polanco. They have been charged with conspiring to threaten a government witness, named Miguel [Carela]. And also, Mr. Nieves has been charged with assaulting Mr. [Carela] in retaliation for Mr. [Carela’s] testimony at a prior trial.

App’x 60. It next explained several foundational trial concepts, including the role

of a criminal jury, the presumption of innocence, and the reasonable doubt

standard, and asked whether any of the first 12 prospective jurors had doubts

about their ability to apply those basic principles. None answered. The court

continued: “Every once in a while there is a case – not usually this kind of case –

involving some other kinds of crimes that people have such strong feeling[s]

about that they don't think they could be fair.” App’x 60-61. Having explained “a

little bit about what this case is about,” the court then asked whether any jurors

had any doubts about their ability to be fair. App’x 61. Again, none answered.

Finally, the court asked whether prospective jurors or their loved ones had

any connection to the defendants, lawyers, or witnesses who would feature in the

10 trial, any connection to various law enforcement agencies, any eyesight or

language impediments, any prior criminal charges, or any past experience as a

party to a civil lawsuit or as a juror in a criminal trial. When prospective jurors

raised their hands to answer in the affirmative, the court briefly followed up with

additional queries. That was the extent of the court’s generalized questioning.

The court’s three- or four-question individualized voir dire was similarly

brisk. It solicited each prospective juror’s (1) “county or borough” of residence; (2)

occupation; (3) “whether or not you have a significant other”; and (4) if so, the

significant other’s occupation. App’x 69. At a few points, the court asked

follow-up questions concerning specific law firms where jurors or their significant

others worked, the nature of a prospective juror’s company, and a college

student’s major. That was the extent of the court’s individualized questioning.

It was at this point – after the court announced to the venire that counsel

would now begin peremptory challenges – that defense counsel first objected. At

a sidebar, Polanco’s counsel urged the district court to “do some further inquiry of

these jurors” in line with the parties’ proposed questions, arguing that the

questioning thus far was not “enough . . . to make any cause challenges.” App’x

72. The court rebuffed the request: “after careful study,” it had “rejected” the

11 suggested questions as either “inappropriate or otherwise not useful.” App’x 73. It

assured counsel that the “objection is preserved for appeal.”

When Nieves’s counsel specifically broached the topic of “gangs and the

Trinitarios” as a voir dire subject, the court responded that it had “thought about

that question, but I’m sorry, I’m not giving that question.” App’x 73. Nieves’s

counsel insisted that the questioning thus far was not “sufficient to allow us to

meaningfully exercise any peremptory challenges.”

The court responded that it

had a “long standing practice of not asking what might be called attitudinal

questions,” explaining that “I think that is an invasion of privacy and I also think

it is counterproductive and leads to colloquies.”

That ended the discussion.

The parties then proceeded to exercise their peremptory challenges based

on the information that had surfaced to that point. Throughout that process, the

court also dismissed several more prospective jurors for a range of reasons,

including language difficulties, personal ties to law enforcement or to incarcerated

relatives, familiarity with a witness, recent service as a government witness, and

recent criminal charges. Upon seating new prospective jurors at this stage, the

district court asked if they had any responses to its general questions and its four

individualized questions. In short order, a panel of twelve jurors had been

12 finalized.

There was one last snag. One prospective alternate juror requested to be

excused “[d]ue to the nature of the charges.” App’x 88. At a sidebar, he explained:

“I just feel uncomfortable. My own biases. . . . I don’t know if they might be gang

members or . . . . I don’t feel comfortable being involved in such a case.”

The

court granted his request to be excused, but did not inquire if any other

prospective jurors had similar concerns. It would later explain that “part of my

concern was that jurors not feel any sense, during this trial, of intimidation or

potential fear or anything like that. It’s another reason why I didn’t put the

question that defense counsel had asked for about gangs.” App’x 217.

After excusing the juror who had expressed concerns about gangs, the court

separately dismissed three other prospective alternates based, respectively, on

religious concerns, professional ties to the U.S. Attorney’s Office, and personal ties

to incarcerated relatives. Shortly thereafter, twelve jurors and three alternates

were empaneled.

C. Nieves’s Mistrial Motion

After the jury was sworn but prior to opening statements, Nieves moved for

a mistrial, arguing that voir dire had been inadequate for several reasons,

13 including: (1) that prospective jurors had been masked, as required under the

court’s COVID-19 protocols at the time, rendering counsel unable to assess their

demeanor; (2) that counsel had not had an opportunity to object to the court

dismissing jurors for cause because “things were moving extremely quickly”; and

(3) that thanks to the district court’s scant questioning, defense counsel had heard

many jurors say “maybe five or ten words” at most. App’x 117-18.

The district court quickly disposed of all three arguments. First, as to masks,

it remarked that a challenge based on a juror’s (hypothetically maskless)

appearance “would have been overruled by increasingly clear precedent as to

how not only irrelevant it is, but how it is just a cover for discrimination.” App’x

116-17. Next, as to counsel’s purported lack of opportunity to object to the court’s

dismissals for cause, it explained to counsel: “[Y]ou were right there. . . [a]nd you

sat there silently.” App’x 117.

Finally, as to the brevity of its questioning, the district court said: “Show me

where in the constitution or the laws of the United States that says you're entitled

to 30 words as opposed to 10.” App’x 118. It continued: “[T]his Court’s view is

that it needs to question the jurors sufficiently to make sure that any who need to

be excused for cause are excused,” and that all its questions “are directed towards

14 that.” App’x 120. By contrast, its questions were intentionally not directed towards

“go[ing] beyond that and provid[ing] more than very basic details about the

persons who are being questioned.”

The court volunteered several justifications for that approach. Questions

asking for an “indication of the jurors’ attitudes” are not only an “invasion of

privacy,” it explained, but they also pose a “source of potential confusion” and

“most important[ly]” invite follow-up questions that “inevitably get the Court

into a description of the case in some way or another, which would be totally

improper.” App’x 120. Thus, the court explained, “I never give those questions.”

Defense counsel clarified that he was not advocating “attitudinal questions,”

but rather “even a few questions that would elicit something about the jurors’

background,” such as “their education” or the “number of children they have.”

App’x 121. The court again brushed the notion aside, explaining that these sorts of

questions – “Do you read the New York Times? What is your favorite

newspaper?” – are not only a “clear invasion of privacy” but also “an exercise of

stereotyping of people.”

With respect to the specific gang-related questions that counsel had

requested, the district court further opined:

15 [Y]ou can’t describe gangs in the abstract. There are gangs and gangs. In order to ascertain whether a juror had a meaningfully improper view or is biased because the case involved gangs, which is the Court’s job when it is looking to excuse someone for cause, the Court would . . . inevitably have to get into the description of the facts of the particular case. You could make it a little abstract but it would still be that and that would be totally improper. For all those reasons, but especially the latter, I think those are not an appropriate role for the Court.

App’x 120-21.

Defense counsel made one last-ditch appeal for questions that would

provide some “basis for exercising . . . peremptory challenges.” App’x 122. The

court replied: “Duly noted; denied.”

Trial then began.

DISCUSSION

The Supreme Court has recognized that the voir dire process “plays a

critical function in assuring the criminal defendant that his Sixth Amendment

right to an impartial jury will be honored.” Rosales-Lopez v. United States,

(1981) (plurality opinion). That is because an inadequate voir dire

compromises the trial court’s “responsibility to remove prospective jurors who

will not be able impartially to follow the court’s instructions and evaluate the

evidence.”

Therefore, notwithstanding the Constitution’s literal silence on the

16 matter, it is uncontroversial that “part of the guarantee of a defendant’s right to an

impartial jury is an adequate voir dire to identify unqualified jurors.” Morgan v.

Illinois,

(1992).

Still, “[d]espite its importance, the adequacy of voir dire is not easily subject

to appellate review.” Rosales-Lopez,

. That is because the trial court,

“not unlike . . . jurors later on in the trial,” is best positioned to “reach conclusions

as to impartiality and credibility by relying on [its] own evaluations of demeanor

evidence and of responses to questions.”

To that end, district judges have long

been “accorded ample discretion in determining how best to conduct . . . voir dire”

pursuant to Rule 24 of the Federal Rules of Criminal Procedure. United States v.

Quinones,

(2d Cir. 2007), quoting Rosales-Lopez,

.

Rule 24, in turn, sets forth broad, discretionary parameters. It permits the

district court to examine prospective jurors directly or to allow counsel to do so.

Fed. R. Crim. P. 24(a)(1). Where the district court itself examines the prospective

jurors, it requires that counsel be permitted either to “ask,” or “submit” for the

court to ask, “further questions that the court considers proper.” Fed. R. Crim. P.

24(a)(2).

17 I. Standard of Review

“Accordingly, we will not reverse unless we determine” that the district

court’s ample discretion “has been abused.” United States v. Lawes,

(2d Cir. 2002). A district court abuses its discretion when its decision

incorporates an “error of law,” rests on a “clearly erroneous factual finding,” or

“cannot be located within the range of permissible decisions.” Zervos v. Verizon

New York, Inc.,

(2d Cir. 2001).1 In this context, that deferential

standard applies to both the general “manner in which [voir dire] has been

conducted,” United States v. Salameh,

(2d Cir. 1998), and the

specific questions a district court elects to ask, or not to ask. See United States v.

Kyles,

(2d Cir. 1994); see also United States v. Tsarnaev,

, 1034 (2022) (“A trial court’s broad discretion in this area includes deciding

what questions to ask prospective jurors.”).

1 Consequently, we have frequently pointed out that the term “abuse” of discretion is an unfortunate misnomer. Despite its harsh connotations, it is substantively “a nonpejorative term of art” that “implies no misconduct on the part of the district court.” United States v. Bove,

, 607 n.1 (2d Cir. 2018).

18 II. The Court’s Duty to Screen for Bias

But that discretion is not boundless. “The standard set by the [Supreme]

Court, which remains the standard today, is that the trial court’s discretion must

be exercised consistent with ‘the essential demands of fairness.’” United States v.

Barnes,

(2d Cir. 1979) (footnote omitted), quoting Aldridge v.

United States,

(1931); see also United States v. Bright, No. 20-3792,

, at *1 (2d Cir. Jan. 6, 2022) (summary order) (reiterating this

standard). Even in that forgiving, deferential light, “the defense deserves ‘a full

and fair opportunity to expose bias or prejudice on the part of veniremen.’” United

States v. Colombo,

(2d Cir. 1989), quoting Barnes,

.

Where the court finds that such bias or prejudice exists, a juror is subject to

dismissal for cause. United States v. Torres,

(2d Cir. 1997)

(noting that dismissal is mandatory in some circumstances and discretionary in

others).

For-cause challenges are generally based on one of three species of bias: (1)

actual bias, or “bias in fact”; (2) implied bias, or bias that is “presumed as a matter

of law” where a typical person in the juror’s position would be biased,

irrespective of whether actual bias exists; and (3) inferable bias, which arises

19 “when a juror discloses a fact that bespeaks a risk of partiality sufficiently

significant to warrant granting the trial judge discretion to excuse the juror for

cause, but not so great as to make mandatory a presumption of bias.” United States

v. Greer,

(2d Cir. 2002) (internal quotation marks omitted).

Crucially, all three forms of bias must be “grounded in facts developed at

voir dire.” Torres,

. A finding of actual bias arises “upon either the

juror’s own admission or the judge’s evaluation of the juror’s demeanor and

credibility following voir dire questioning as to bias.”

A finding of implied bias

arises from a “disclosed fact” that “establish[es] the kind of relationship between

the juror and the parties or issues in the case that mandates the juror’s excusal for

cause.”

Finally, a finding of inferable bias arises from “responses from the juror

that permit an inference that the juror in question would not be able to decide the

matter objectively.”

To that end, even as voir dire challenges have historically not fared well in

this Court, we have often reiterated the basic principle that “[t]here must be

sufficient information elicited on voir dire to permit a defendant to intelligently

exercise” both for-cause and peremptory challenges. Colombo,

,

quoting Barnes,

. In other words, there must be sufficient

20 factfinding at voir dire to allow for facts probative of any of these forms of bias to

reveal themselves. Otherwise, “[f]undamental unfairness arises if voir dire is not

‘adequate . . . to identify unqualified jurors.’” United States v. Whitten,

(2d Cir. 2010) (second alteration in original), quoting Morgan,

.

In United States v. Lawes, we sketched out three basic archetypes of voir dire

that falls short in this regard:

(i) a voir dire so demonstrably brief and lacking in substance as to afford counsel too little information even to draw any conclusions about a potential juror’s general outlook, experience, communication skills, intelligence, or life-style;

(ii) a failure to inquire about, or warn against, a systematic or pervasive bias, including one that may be short-lived but existent at the time of trial, in the community that would have been cured by asking a question posed by a party; or

(iii) a record viewed in its entirety suggesting a substantial possibility that a jury misunderstood its duty to weigh certain evidence fairly that would have been clarified by asking a requested voir dire question.

21

(internal citations omitted). Conspicuously, however, we have

never overturned a conviction based on a finding that a particular voir dire

actually fell within any of the Lawes categories.

III. Analysis

At the heart of this appeal is Nieves’s contention that his is the rare case that

does just that. He argues that the government’s case against him – which revolved

around his Trinitarios affiliation – implicated the second Lawes category: a

“systematic or pervasive bias . . . in the community” against people associated

with gangs.

We agree. We hold that (1) prejudice against people associated with gangs

represented a pervasive bias relevant to a key dynamic likely to arise at trial, and

(2) the district court neglected to “inquire about, or warn against,” that bias.

We address each point in turn.

A. Systemic or Pervasive Bias

In this case, voir dire took place on April 14, 2021, in Manhattan. At that

time, New York City was only beginning to emerge from a frenetically-publicized

22 reported spike in violent crime that local2 and federal3 law enforcement had

broadly attributed to gang violence, including in large part to violence committed

by members of traditionally Latin-American gangs like MS-134 and the

Trinitarios.5 To be clear, the issue here is not whether the district court should

2 See, e.g., Ali Watkins, Violent Year in New York and Across U.S. as Pandemic Fuels Crime Spike, N.Y. TIMES (Dec. 29, 2020), https://www.nytimes.com/2020/12/29/ nyregion/nyc-2020-crime-covid.html (“Police officials in New York have pointed to gang disputes as a key driver of the violence over the summer . . . .”). 3 See, e.g., Acting United States Attorney Seth D. DuCharme Announces Expansion of Strategies For Rapid Federal Response to Spikes in Gun-Related Violence in the Eastern District of New York, U.S. ATTORNEY’S OFFICE FOR THE EASTERN DISTRICT OF NEW YORK (Oct. 14, 2020), https://www.justice.gov/usao-edny/pr/acting-united-states- attorney-seth-d-ducharme-announces-expansion-strategies-rapid (announcing a cross-departmental initiative in response to the spike in gun violence, attributing the “local epidemic of shootings in our city” in significant part to “violent gang members” turning portions of the city into “battle zones” (internal quotation marks omitted)). 4 See, e.g., MS-13's Highest-Ranking Leaders Charged with Terrorism Offenses in the United States, U.S. ATTORNEY’S OFFICE FOR THE EASTERN DISTRICT OF NEW YORK (Jan. 14, 2021), https://www.justice.gov/usao-edny/pr/ms-13-s-highest-ranking- leaders-charged-terrorism-offenses-united-states (“MS-13 is responsible for a wave of death and violence that has terrorized communities, leaving neighborhoods on Long Island and throughout the Eastern District of New York awash in bloodshed.” (internal quotation marks omitted)); Liz Robbins and Michael D. Shear, Trump, Visiting Epicenter of MS-13 Killings, Demands Tougher Immigration Laws, N.Y. TIMES (May 23, 2018), https://www.nytimes.com/2018/05/ 23/us/politics/trump-immigration-gangs.html (“Mr. Trump again described gang members as ‘animals . . . .’”). 5 See, e.g., Edgar Sandoval, Crime Wave Hits Bodegas, Threatening a Lifeline in the

23 have inquired into this news coverage during voir dire. Rather, we cite this

coverage to illustrate the broader risk of prejudice against gang members (and

especially members of Latin-American gangs) among prospective jurors who may

have been affected by a reported surge in violence broadly attributed to the

precise kinds of organizations – and, notably, the precise gang – at the center of

the government’s case against Nieves. See Staehr v. Hartford Fin. Servs. Grp., Inc.,

(2d Cir. 2008) (recognizing that courts may take judicial notice of

Pandemic, N.Y. TIMES (Dec. 10, 2020), https://www.nytimes.com/2020/12/10 /nyregion/nyc-crime-bodegas.html (discussing a “steep rise” in New York City violence that “shows no signs of slowing down,” and highlighting a Bronx bodega murder committed by Trinitarios members as typical of that surge); Clayton Guse and Rocco Parascandola, Teen Gang Member Busted for Threatening Brooklyn Subway Riders with Machete, N.Y. DAILY NEWS (Mar. 25, 2021), https://www.nydailynews.com/new-york/nyc-crime/ny-nyc-machete-subway- teen-arrest-20210325-cba6hxk4jvh57lhz6ewdhv66e4-story.html (“[The suspect] is a member of the Trinitarios gang, police said.”); Jorge Fitz-Gibbon, FBI Offers $5K Reward for Arrest of Trinitarios Gang Member, N.Y. POST (Oct. 15, 2020), https://nypost.com/2020/10/15/fbi-offers-5k-reward-for-arrest-of-trinitarios-gang- member/ (“Eric Rosario, 28, a reputed member of the Trinitarios street gang known as ‘White Boy,’ is sought for slashing a man in the face at Grand Concourse and 196th Street on Feb. 3, 2019 . . . .”); Robert Gearty, The Gang Just as Vicious as MS-13 but without the Notoriety: Savage Trinitarios Return to the Spotlight after Years in the Shadows, FOX NEWS (July 28, 2018), https://www.foxnews.com/us/ the-gang-just-as-vicious-as-ms-13-but-without-the-notoriety-savage-trinitarios-re turn-to-the-spotlight-after-years-in-the-shadows (equating the “[s]avage” Trinitarios’ “brutality” with that of MS-13).

24 the fact that news articles and their contents exist, without respect to the truth of

those contents).

Of course, we have never required district judges to sua sponte curate an

exhaustive catalog of every bias that a prospective juror could conceivably track

into court. And our holding today does not suggest otherwise. That is because –

even setting aside the government’s concession at oral argument that it would be

reasonable to expect a court to be aware of this context without the parties

specifically presenting it – it is undisputed that the district court was in fact well

aware of the risk of gang-related bias in this case.

For one thing, it was told of the risk. As a preliminary matter, the district

court had long been on notice that the government’s theory of the case was that

Nieves and Polanco had targeted Carela specifically because of the Trinitarios’

“‘zero-tolerance’ policy towards members of their gang who cooperate with law

enforcement.” Gov’t Opp’n Br. 1, SDNY Dkt. No. 50. It bears reiterating that

Nieves was not charged with simple assault, and that it was undisputed that the

run-in with Carela in fact occurred. Instead, Nieves was charged specifically with

witness retaliation and intimidation, both of which hinge on why the fight (and

25 surrounding events) occurred. As the district court knew well, the government’s

answer to that question was simple: gang affiliation. See id. at 1-4.

Accordingly, prior to trial, both the defense and the government requested

questions specifically concerning gangs in their proposed voir dire submissions.

Two of the eleven questions in the portion of the government’s proposed voir dire

entitled “Ability to Render a Fair and Impartial Verdict” asked, respectively,

about any past involvement with “case[s] involving a gang” and whether the fact

that this case involved “a gang” would “affect your ability to be fair and impartial

in this case.” App’x 32-33. Twice as many of the defendants’ proposed voir dire

questions explicitly referenced gangs, including whether prospective jurors had

“any personal experiences or feelings arising from any acts of violence that might

impact the way you think and feel about this case, or that would affect your

impartiality in a case involving alleged gang violence.” App’x 52. While a request

from counsel – “even all counsel” – that a particular subject be explored during

voir dire does not obligate the district court to do so, Lawes,

, a

court can hardly be considered ignorant of a risk that both sides have presented as

worthy of its attention.

26 Moreover, the district court was reminded of that risk by defense counsel’s

repeated objections prior to trial. And its on-the-record responses are telling. At

the outset, it explained that the decision not to ask about gangs was neither an

oversight nor an unconsidered rejection: “I thought about that question, but I'm

sorry, I'm not giving that question.” App’x 73.

Its subsequent elaborations make clear that it was perfectly aware how

fraught the topic of gangs could be for prospective jurors. It recognized, for

example, that jurors might feel “intimidation or potential fear” based on the fact

that the case involved gangs and gang members. App’x 217. And it not only noted

the possibility that a juror could harbor “a meaningfully improper view” or

indeed “bias[] because the case involved gangs,” but it acknowledged that it “is

the Court's job when it is looking to excuse someone for cause” during voir dire to

“ascertain” whether that is in fact the case for any particular juror. App’x 120-21.

The court’s rationale for how it approached that job was not to discount the risk of

bias against gang members, but rather to insist that “you can’t describe gangs in

the abstract” and that it would be “totally improper” to “get into the description

of the facts of the particular case.”

27 Importantly, then, no one – not the district court, not the government, and

certainly not Nieves – has disputed that the prospect of bias against gang

members was sufficiently “pervasive” in the community at the time of trial to

implicate the second Lawes category. See

. The government itself

proposed questions to ward off that risk, and on appeal does not retreat from its

stance below that such questions would have been prudent. At best, it now argues

that Nieves “did not make a record below” of that risk, and therefore that the

district court’s approach should be reviewed for plain error. Appellee’s Br. 21.

But, to start, that argument overstates Nieves’s obligations under Rule 51 of

the Federal Rules of Criminal Procedure; the defendants preserved the issue for

appellate review by requesting gang-related questions and objecting to their

omission on the grounds that “gangs and the Trinitarios” would be key concepts

at trial, and that the district court’s approach compromised defense counsel’s

ability to make for-cause challenges. App’x 72-73; see Holguin-Hernandez v. United

States,

(2020) (“[A] party may preserve a claim of error by

informing the court . . . of [1] the action the party wishes the court to take, or [2]

the party’s objection to the court’s action and the grounds for that objection.”

(second alteration in original) (quoting Fed. R. Crim. P. 51(b)); see also United States

28 v. Garcia,

(2d Cir. 2002) (“We presume that the objection of a

co-defendant is an objection for all defendants, and it is sufficient to preserve the

issue for appeal.”).

And more fundamentally, although it is true that Nieves did not specifically

document below the presence of gang-related bias in the community, there was

simply no need to. The parties already agreed that the venire should be

questioned about their feelings on gangs. The district court, in turn, quickly made

clear that it understood the risks posed by possible juror reactions to gangs – and

that its decision not to mention gangs during voir dire was attributable neither to

ignorance of those risks in general, nor to a factual finding that the risks did not

exist in this case, but rather to other justifications that we discuss and reject below.

The absence of dissent as to the prospect of bias against gang members is

entirely unsurprising. In our view, it would have been obvious to anyone living in

the New York City area at the time that there was a serious risk that jurors tasked

with impassively assessing the credibility of the anticipated testimony from

Carela and Detective Jeselson concerning the Trinitarios, and impartially applying

that testimony to defendants associated with that organization, would have some

difficulty maintaining an open mind while doing so. It would have been obvious

29 that prospective jurors might be predisposed to condemn Nieves by association

because the specific thing he was associated with had long (and particularly in the

months preceding trial) been the persistent target of an onslaught of local and

national media reports reciting the violent practices of groups like the Trinitarios –

much of it stimulated by pronouncements from members of the very executive

branch conducting Nieves’s prosecution.

In short, as other Circuits have acknowledged in a variety of contexts, it

would have been obvious here that there was a significant “possibility that a jury

will attach a propensity for committing crimes to defendants who are affiliated

with gangs or that a jury’s negative feelings toward gangs will influence its

verdict.” United States v. Irvin,

(7th Cir. 1996) (adding that

“[g]uilt by association is a genuine concern whenever gang evidence is admitted,”

and concluding that the district court had erred by admitting certain gang-

affiliation evidence); see United States v. Easter,

(9th Cir. 1995)

(affirming a conviction where the district court “took several steps to minimize

possible prejudice” including “prob[ing] potential jurors, during voir dire, on the

issue of bias due to gang affiliation” and “dismiss[ing] three potential jurors who

indicated that such evidence might influence their deliberations”); United States v.

F.3d 825, 831, 834-35 (4th Cir. 2019) (affirming the denial of a mistrial

where the district court made “appropriate inquiries into juror bias” after

members of the jury had expressed mid-trial anxiety in response to evidence of a

gang’s “violent nature”).

Accordingly, we hold that under the circumstances presented in this case,

Nieves’s alleged gang association was a “material issue” likely – indeed, certain –

to arise at trial as the cornerstone of the government’s theory of motive, and that

there was a significant risk that potential “strong feelings” concerning gangs

would “skew deliberations in fact.” Barnes,

(internal quotation

marks omitted). Therefore, because the risk of juror prejudice arising from bias

against gang members was both significant and apparent to the district court, it

was outside of the court’s discretion to altogether decline to protect against that

risk.

B. The District Court’s Failure to Adequately Screen for Bias

Having established that much, we must next decide whether the manner in

which the district court conducted voir dire sufficiently accounted for that risk.

We hold that it did not.

31 We start with what a district court is required to do. A district court must

“permit at least some questioning” – with nearly limitless discretion as to what

that questioning may look like – “with respect to any material issue that may

arise, actually or potentially, in the trial.” Barnes,

(emphasis

added).

Our decisions illustrate the range of options available to the district court.

In these decisions, as is often the case in voir dire appeals, the defendants have

targeted specific choices made by the district court during voir dire, arguing that

those choices failed to protect against a specific risk of bias. In affirming, we have

typically relied upon the fact that each district court accounted for that risk in

some other permissible way.

For example, in United States v. Treacy, which also concerned Judge Rakoff’s

voir dire practices, we held that his “categorical” policy never to use written

questionnaires during voir dire was “well within his bailiwick as a trial judge so

long as he conducts adequate voir dire by some other means.”

(2d

Cir. 2011). We concluded that the judge had sufficiently accounted for concerns

about jurors’ “potential prejudice against people associated with Wall Street” by

asking (1) whether jurors “could be fair and impartial in a securities or stock fraud

32 case involving a former [executive]”; (2) whether they or their loved ones “had

ever received any stock options”; (3) whether they or their loved ones “owned

stock in or were associated with” the defendant executive’s company; and (4)

whether any prospective jurors “had been exposed to any pretrial publicity about

the case.”

. Thus, we concluded, even after eschewing a written

questionnaire containing inquiries that “pertained specifically to the jurors’

experiences with and views of the financial sector,” the district court had

permissibly found another way to screen for the relevant bias.

.

In United States v. Rahman, we affirmed a variety of terrorism-related

convictions against primarily Muslim and Arab defendants accused, among other

things, of conspiring to bomb New York City landmarks.

, 104 (2d Cir.

1999). On appeal, we rejected the defendants’ argument that the district court’s

voir dire had not accounted for a range of potential biases among the venire; we

noted that the district court had asked not just whether each defendant’s religion

or ethnicity would compromise jurors’ impartiality, but also “more subtle,

detailed questions” about jurors’ “personal experiences” such as “whether they or

their loved ones regularly use the Holland and Lincoln Tunnels and the George

Washington Bridge.” Id. at 121.

33 Three of our recent summary orders have conformed to that same template.

In United States v. Miller, we acknowledged that it might have been “helpful” to a

defendant on trial for visa fraud for the district court to have inquired during voir

dire about jurors’ occupations, relationships to law enforcement personnel, and

“views on immigration or visa fraud generally.”

, 52-54 (2d Cir.

2018) (summary order). Nonetheless, we held that the decision not to ask those

questions was permissible because the district court had instead screened for

relevant biases in other ways, including by: (1) explaining to the venire that the

defendant was “charged with fraudulently obtaining a non-immigrant visa that

he then used to enter the United States from Jamaica”; (2) posing several questions

on that subject, including questions about jurors’ “experience with the visa

process [and] relationships with law enforcement and immigration authorities”;

and (3) asking whether the nature of the case, which the court had explained in

some detail, “would affect [jurors’] ability to impartially assess the evidence.”

.

Similarly, in United States v. Diaz, we rejected an appeal centered on the

district court’s decision not to ask a “proffered voir dire question concerning

implicit racial bias,” emphasizing that the district court “did question the potential

34 jurors on the issue of racial prejudice” in other ways, including by warning

prospective jurors that “it would be improper for any of you to consider any

personal feelings you may have about a defendant’s race, religion, national origin,

sex, or age” and then immediately asking whether any of the venire would “have

any difficulty putting aside any feelings, positive or negative, about the

defendant . . . based on these considerations.”

, 388-89 (2d Cir.

2021) (summary order) (emphasis in original) (internal quotation marks omitted).

Most recently, in United States v. Bright, we rejected a defendant’s argument

that the district court had impermissibly declined to inquire about jurors’

“possible prejudices against people who pursue nonconventional sexual practices

and fetishes, including age-based role play.” No. 20-3792,

, at *1 (2d

Cir. Jan. 6, 2022) (summary order). Again, we underscored that “even assuming

arguendo that the district court was required to question jurors about the

potential biases that [the defendant] identifies, the record reflects that the court

did, in fact, do so” in other ways, including by “describ[ing] the non-conventional

sexual practice that [defendant] was concerned about – ‘age-based role play’ –

nearly verbatim from his definition of the term,” and inquiring whether “evidence

involving age play” would prevent jurors from being fair and impartial.

35 Together, these cases animate the unremarkable but critical proposition that

the district court must provide some opportunity for prospective jurors to be

meaningfully screened for biases relevant to a particular defendant or the charges

against that defendant. Or, as the Rahman panel put it, a district court’s decision

not to ask certain questions probative of bias “will not be grounds for reversal” so

long as voir dire otherwise “‘covers the subjects’ that may arise in the case to

ensure that jurors will be impartial.” 189 F.3d at 121 (alterations omitted), quoting

Aldridge,

.

Deciding precisely how to do so remains within the district court’s

expansive voir dire discretion. It could, of course, inquire directly about relevant

biases. See, e.g., Bright,

, at *1. But we have never reversed for

failure to take that specific path, and again refrain from doing so today.

There are many other tacks a district court might justifiably deem more

prudent under the particular circumstances of a case. For example, it could opt

instead to ask the kind of “subtle[r]” questions that we endorsed in Rahman,

sussing out relevant bias without explicitly invoking it. See, e.g., 189 F.3d at 121. It

could ask generalized questions about jurors’ ability to serve impartially,

provided that it also presents sufficient context about the case for jurors’ answers

36 to that general question to actually convey information pertinent to the relevant

potential biases in that case. See, e.g., Miller, 752 F. App’x at 52-54. It could warn

jurors of their general duty of impartiality, again assuming it also provides

sufficient context for a juror to be able to self-identify the relevant biases before

jury selection is complete. See, e.g., Diaz, 854 F. App’x at 388-89; see also Treacy,

(“[A] district court may find that warning a jury against an improper

bias may be more effective in some cases than inquiring about that bias.”). And,

although we decline Nieves’s invitation to so declare in this case,6 even if the

district court does none of these things, it may even be the case that it can

6 Nieves argues that this case triggers not just the second Lawes category, but also the first: “a voir dire so demonstrably brief and lacking in substance as to afford counsel too little information even to draw any conclusions about a potential juror’s general outlook, experience, communication skills, intelligence, or life-style.” Lawes,

. We disagree. Although we hold that the district court’s brief, three- or four-question individualized voir dire did nothing to affirmatively cure its larger failure to adequately screen for bias against gang members, we do not conclude that the limited scope of questioning concerning prospective jurors’ backgrounds in itself constituted reversible error. Judicial efficiency is an important value in busy district courts, and district judges are best positioned to balance that value against the benefits of extensive voir dire. While many – perhaps most – judges would perhaps strike a different balance, and some might lean heavily in the direction of extensive voir dire, we can discern no quantitative standard governing precisely how many questions are required to give the parties a sense of jurors’ possible biases.

37 accomplish this end – or at least render harmless7 the failure to do so – by asking

7 The government does not argue that the failure to inquire on voir dire about venire members’ attitudes towards gang members, if error, was harmless. But even assuming arguendo that the district court’s voir dire errors could be rendered harmless by other steps taken outside of voir dire – a hypothetical question that we do not decide today, cf. Tankleff v. Senkowski,

(2d Cir. 1998) (“[A] Batson/Powers claim is a structural error that is not subject to harmless error review.”) – it did not take any such steps here. It did not, for example, give any limiting instruction with respect to Jeselson’s expert gang testimony, nor did its pre-deliberation jury instructions deal with gang-related members in any way (or even refer to gangs specifically). Cf. Tsarnaev, 142 S. Ct. at 1035 (holding that the district court’s voir dire was sufficient, highlighting jury instructions “the court gave during voir dire and repeated during the trial”). Separately, without expressly invoking the concept of harmless error, the government argues that even if jurors biased against gang members were selected, both sides would have been equally affected. That is because, it contends, any biased juror’s perception of the credibility of victim-witness Carela, himself a Trinitario, would suffer as much as that juror’s perception of Nieves. That argument is both speculative and ill-conceived. Gang membership might affect perceptions of credibility, or it might predispose a jury to believe that a gang member charged with a violent crime is a dangerous person who deserves to be locked up notwithstanding doubts about whether the government has actually proved each element of the charged crime beyond a reasonable doubt. There is no reason to believe that those biases would cancel each other out, or that jurors biased against a defendant should be seated so long as they are equally biased against a victim or a witness. In any event, the government’s reliance on Rosales-Lopez,

, to support this otherwise speculative argument is misplaced. In that case, the Supreme Court held that there was no “reasonable possibility that racial or ethnic prejudice” would have influenced the jury’s treatment of a defendant of Mexican descent charged with “aiding members of his own ethnic group to gain illegal entry into the United States.”

But the Court’s justification for that conclusion was not that all of the people involved shared the same potentially bias-inducing characteristic. Rather, it

38 sufficiently detailed questioning to allow defendants to indirectly “ferret out more

subtle biases” through peremptory challenges based on circumstantial indicators of

bias. Torres,

n.4.

Here, the district court’s deliberate decision to jettison any of those options

exceeded the bounds of its discretion. It did not ask – overtly or otherwise – about

gang-related bias, notwithstanding that the government’s central theory of the

case was that the assault was motivated by gang rules, and its evidence in support

of that theory consisted of gang-member testimony and gang-related expert

testimony. The court in fact deliberately declined to mention that the case

concerned gangs at all, or at least to caution the jurors that any feelings they might

have about gangs or gang members must be set aside in service of their duty to

decide, based on the evidence presented, whether the government established all

elements of the charged crimes beyond a reasonable doubt. Even after a

prospective juror volunteered his own discomfort and “biases” stemming from

the mere possibility that the case could involve gangs (and was thereafter

emphasized that the district court had not only asked a question about “aliens” in general, which “jurors would have understood . . . to at least include Mexican aliens,” but had also dismissed two jurors for cause based on their answers to that question. Id. at 193. There is no analog to that question in this case.

39 dismissed), the district court neglected to allow for the possibility that others’

impartiality could be compromised by the fact that all the central players in the

case belonged to allegedly violent gangs, and for that matter, that the defendants

were accused of retaliation and intimidation specifically related to court

proceedings. Cf. Smith, 919 F.3d at 834 (rejecting defendants’ challenge to “how

[the district court] addressed the jurors’ fears of gang retaliation,” highlighting

that “[a]fter learning of Juror No. 5’s fears, the district judge here questioned each

juror individually . . . and, where necessary, followed up”). Given those choices,

there was no way for the district court’s generic questions and vague warnings to

actually root out, or otherwise guard against, the specific bias against gang

members potentially at play in this case.

Nor, as a last resort, did the district court attempt to counterbalance any of

this by asking additional biographical or attitudinal questions that might have

circumstantially unearthed inferable bias among potential jurors, permitting

either the court to dismiss them for cause, or Nieves to exercise his peremptory

challenges to the same effect. For example, the district court might have asked not

simply what borough or county each juror was from, but what specific areas in

each borough or county jurors hailed from, which could have permitted the

40 defendants to use their peremptory challenges to dismiss jurors who lived in

places known to be Trinitario strongholds. Or, in lieu of the open-ended

questioning on media consumption that it considered invasive, the district court

might have asked whether prospective jurors had read news stories concerning

the reported crime surge in New York City. A question like that would have

permitted follow-up that might have surfaced gang concerns, in turn occasioning

the court to caution jurors that they must set aside preconceived notions about

gangs when assessing the evidence in this case, and to dismiss those jurors unable

to heed that warning; at the very least, it would have permitted defense counsel to

peremptorily weed out jurors they deemed most problematic. We, again, do not

hold that the district court was required to ask these specific questions, nor that its

failure to do so was error on its own, nor even that asking these questions would

have been sufficiently curative in this case. We offer these examples merely to

illustrate some of the tools available to district courts in these situations. Here, the

district court availed itself of none of those tools.

That was no accident. The district court reiterated multiple times that it

would be “improper” not just to ask gang-related questions, but also to offer the

venire any description of the case other than an abstract definition of the statutory

41 charges. It dismissed the notion that it was obligated to solicit any information

beyond what was necessary to ensure that “any [jurors] who need to be excused

for cause are excused.” App’x 120. And while Nieves argues that this statement

disregarded the court’s accompanying obligation to also facilitate informed

peremptory challenges, see Mu’Min v. Virginia,

(1991), the more

problematic fault was the court’s exceedingly narrow take on its obligation to

probe for information relevant to for-cause challenges. Even as the court

acknowledged its for-cause factfinding role, it insisted that its hands were tied:

“In order to ascertain whether a juror had a meaningfully improper view or is

biased because the case involved gangs, which is the Court’s job when it is looking to

excuse someone for cause, the Court would . . . inevitably have to get into the

description of the facts of the particular case . . . and that would be totally

improper.” App’x 120-21 (emphasis added).

As a proposition of law, if not of the district court’s own preference, that

statement is simply inaccurate. See Zervos,

(a district court abuses

its discretion when its decision relies upon an “error of law”); see also Fed. R.

Crim. P. 24(a)(2)(B) (“If the court examines the jurors, it must permit the attorneys

for the parties to submit further questions that the court may ask if it considers

42 them proper.”). District courts routinely give such descriptions without appellate

rebuke, and this Court has relied upon that practice in rejecting defendants’

arguments concerning the adequacy of voir dire. See, e.g., Miller, 752 F. App’x at

53. We cannot endorse the district court’s apparent belief that it would be “totally

improper” to describe the essence of a case to the jury. App’x 120. That view is

inconsistent not only with our frequent approval of just that sort of description,

but also with our requirement in Lawes that the district court’s voir dire “inquire

about, or warn against,” any widespread and relevant bias.

. In

most cases, any meaningful inquiry or warning will be impossible without some

identification of the facts likely to evoke that bias.8

Judge Rakoff’s past voir dire choices reflect as much. Despite Nieves’s

attempts to attribute Judge Rakoff’s suggestion that a question or description

concerning gangs would have been improper to a larger pattern of purportedly

“inflexible” voir dire practices, Appellant’s Br. at 27, Judge Rakoff himself, as

discussed above, has asked precisely the kinds of questions he declined to ask

here, and this Court has approved his doing so. In Treacy, Judge Rakoff did not

8 While we are of course mindful not to over-parse a district court’s spontaneous, on-the-record remarks, we highlight this one because it is emblematic of the court’s larger error in this case.

43 simply advise potential jurors that the case to be tried generally involved

allegations of fraud; to the contrary, he specifically asked jurors “whether they

could be fair and impartial in a securities or stock fraud case involving a former

chief operating officer and president of [a public company].”

. That

question was in no way improper, and is indistinguishable from asking whether

venire members could be fair and impartial in a case involving an alleged gang

member.

The district court’s other rationales were no more persuasive. For instance,

it later explained that, in its view, invoking gangs before trial could contribute to

jurors feeling “intimidation or potential fear” during trial.9 App’x 217. At no point,

9 The government suggested at oral argument that although the district court’s overview of the charges did not expressly mention gangs, any prospective juror who already harbored strong feelings regarding gang members would have understood the court’s description of the defendants’ “conspiring to threaten,” and Nieves’s “assault in retaliation” against, a government witness as a description of gang behavior, or at least as behavior just as likely to incite those same strong feelings. But that theory cuts against the government’s briefed argument – that the district court reasonably felt that broaching the topic of gangs in any form “might do more harm than good.” Appellee’s Br. 36. The government cannot have it both ways. Either the district court ducked the subject or it did not. In any event, we are not at all convinced that the district court’s recitation of the statutory charges against the defendants somehow signaled to biased jurors that this was a gang case, or some other kind of case that would similarly test their impartiality. That certainly was not the district court’s intent: for better or worse, the district court insisted that any allusion whatsoever to

44 however, did it explain how that dynamic would be assuaged by stifling any

mention of gangs until, as it turned out, the second sentence of the government’s

opening statement – early enough to color jurors’ perception of everything they

would hear afterwards, but too late to filter any fear-driven bias from their ranks.

Nor did the court offer any justification for its decision to affirmatively

represent to the venire that this was not the kind of case that would typically

provoke prejudice or bias. “Every once in a while,” it intimated, “there is a case –

not usually this kind of case – involving some other kinds of crimes that people have

such strong feeling[s] about that they don’t think they could be fair.” App’x 60-61

(emphasis added). As we have already discussed, that also was simply incorrect.

Shortly after the district court made those remarks, one juror spontaneously

voiced concerns about serving in a case potentially involving gangs. That can only

have vindicated the district court’s own belief that, as it expressed later, fear or

bias surrounding gang members could be a factor in the case. But avoiding

mention of gangs until the subject inevitably and expectedly headlined the

government’s opening, and much of its key evidence, was never going to

somehow suppress the kinds of emotional reactions that could have been detected

gangs risked spooking the venire. App’x 217.

45 by appropriate voir dire questions, and thereafter mitigated either through

cautionary instructions or disqualification of jurors unable to fairly assess the

evidence.

Relatedly, the district court’s reluctance to ask either “attitudinal” questions

or other questions that it considered invasive cannot justify its refusal to inquire

about attitudes towards gang members. It is within a judge’s discretion to decide

whether or not to ask questions that elicit jurors’ general attitudes about political

or social questions, permitting counsel to speculate about jurors’ reactions to the

evidence in a case when allocating their peremptory challenges. But where the

“attitude” in question is a bias or predisposition for or against one side or the

other, and particularly where key evidence targets the subject of that bias or

predisposition, it is the judge’s responsibility to make some inquiry. Prospective

jurors have no right to keep those attitudes private. Here, the jury would be

tasked, among other things, with assessing the credibility of Carela’s and

Detective Jeselson’s insights into the rules and practices of the Trinitarios, and the

effect those rules and practices may have had on Nieves’s (disputed) motivation

for his (largely undisputed) conduct in this case. Any predisposition, based on

46 extra-judicial information or prejudice, to believe or disbelieve that sort of

testimony was therefore entirely relevant to jurors’ ability to do their job.

In short, in our view, a capable and experienced district judge faced with an

admittedly tricky task made the wrong call for the wrong reasons. As Judge

Pooler wrote in dissent in Lawes, “[i]f any set of facts presents an abuse of

discretion in the administration of voir dire, it is this case.”

(Pooler,

J., dissenting). That perspective did not prevail under the circumstances presented

in Lawes. It does under those presented here. Although we do not doubt the

sincerity of the district court’s belief that it was charting the most appropriate

course, it is difficult for us, sitting in review, to imagine what less it could have

done to guard against potential bias against gang members, even as the court

itself acknowledged that the issue warranted close attention. Like Judge Pooler,

we “do not advocate that the trial judge was required to ask any particular

questions”; rather, the error was the district court’s resolve not to “ask any

question” that might uncover potential predispositions against gang members, or

to provide sufficient context to allow its broader questions and warnings to

encompass concerns about such bias, therefore leaving that potential prejudice

“completely unexplored.”

(emphasis in original).

47 Because the district court’s failure on voir dire to explore, or to take other

steps specifically to counter, such potential prejudice unfairly deprived Nieves of

the opportunity to unearth a pervasive bias relevant to an issue pivotal to the

government’s case against him, we hold that the district court abused its

discretion.

CONCLUSION

For the reasons set forth above, we therefore VACATE the district court’s

judgment and REMAND for a new trial.

48