PUBLISHED

UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT

No. 19-4137

UNITED STATES OF AMERICA,

Plaintiff - Appellee,

v.

ALLEN H. LOUGHRY II,

Defendant - Appellant.

Appeal from the United States District Court for the Southern District of West Virginia, at Charleston. John T. Copenhaver, Jr., Senior District Judge. (2:18-cr-00134-1)

Argued: October 29, 2020 Decided: December 21, 2020

Before NIEMEYER, DIAZ, and QUATTLEBAUM, Circuit Judges.

Affirmed by published opinion. Judge Niemeyer wrote the opinion, in which Judge Quattlebaum joined. Judge Diaz wrote a separate opinion joining in Parts III and IV and dissenting from Part II.

ARGUED: Elbert Lin, HUNTON ANDREWS KURTH LLP, Richmond, Virginia, for Appellant. Richard Gregory McVey, OFFICE OF THE UNITED STATES ATTORNEY, Huntington, West Virginia, for Appellee. ON BRIEF: Nicholas D. Stellakis, Boston, Massachusetts, Katy Boatman, HUNTON ANDREWS KURTH LLP, Houston, Texas, for Appellant. Michael B. Stuart, United States Attorney, Philip H. Wright, Assistant United States Attorney, OFFICE OF THE UNITED STATES ATTORNEY, Charleston, West Virginia, for Appellee. NIEMEYER, Circuit Judge:

After the former Chief Justice of the Supreme Court of Appeals of West Virginia,

Allen H. Loughry II, was convicted of mail fraud and wire fraud for the misuse of public

assets, he filed a motion challenging the fairness of his trial on the grounds that a juror —

referred to by the district court and the parties as Juror A — allegedly engaged in

misconduct and was biased. He requested a new trial or at least a hearing on his motion.

The district court denied Loughry’s motion, concluding that the evidence Loughry

presented was insufficient to sustain his claims or even to justify a hearing.

The court thereafter sentenced Loughry to 24 months’ imprisonment, imposed a

$10,000 fine, and ordered restitution.

From the district court’s judgment dated February 25, 2019, Loughry filed this

appeal, alleging only that the district court abused its discretion in denying his request for

an evidentiary hearing to investigate Juror A’s potential misconduct and bias.

For the reasons that follow, we affirm.

I

In October 2017, the news media in Charleston, West Virginia, began investigating

and reporting about lavish spending of public funds by justices of the West Virginia

Supreme Court of Appeals for renovation and refurbishing of their offices, and shortly

thereafter a federal investigation ensued. The investigation led to evidence that Loughry

removed a historical desk from the court to his home; that he improperly used state vehicles

and gas credit cards for personal use; and that he obstructed justice during the course of the

2 investigation. The historical desk, which became prominent in the news coverage, was one

that was selected for use in the courthouse in the 1920s by Cass Gilbert, a prominent

architect who designed the West Virginia State Capitol, the United States Supreme Court

building, the Woolworth building in New York, and other well-known buildings. The desk

was thus referred to as the “Cass Gilbert desk.”

In June 2018, a grand jury returned a 25-count indictment charging Loughry with

mail fraud, wire fraud, and related crimes. During the same period, the West Virginia

Judicial Investigations Commission filed a complaint against Loughry, alleging numerous

violations of the state Judicial Code of Conduct, and the Judiciary Committee of the West

Virginia House of Delegates began impeachment proceedings against Loughry, as well as

three other sitting justices of the West Virginia Supreme Court.

The criminal trial against Loughry began on October 2, 2018, with voir dire of the

venire — the pool from which the jurors are selected. As is customary, the district court

summarized for the venire the charges made against Loughry in the indictment and inquired

whether any prospective juror knew Loughry; whether any knew the prospective witnesses;

whether any were related to law enforcement officers; and whether any had ever served on

a jury or as a witness in a criminal case.

With respect to the pending charges, the court inquired whether any of the

prospective jurors had any knowledge or exposure to “this case” or the “facts of this case,”

and whether they had discussed the case with anyone. The court asked similar questions

about the impeachment proceedings against Loughry and the other justices that were taking

place in the state legislature. In response to affirmative responses from prospective jurors,

3 the court inquired about whether those prospective jurors could set aside their knowledge

or experience and “listen to the evidence and base a verdict solely upon the evidence

received here in the courtroom.” Finally, the court followed up with general questions

about bias or preheld opinions about the guilt or innocence of the defendant. After these

and similarly general voir dire questions, the court allowed counsel for the parties to

conduct voir dire of individual prospective jurors who responded affirmatively to any of

the questions.

During this process, Juror A answered “no” to questions of whether she had

knowledge “of this case” or “facts of this case”; answered “yes” to questions of whether

she had knowledge of the impeachment proceedings; and answered “yes” to whether she

could set aside her knowledge and render a verdict based solely on the evidence presented

at trial. Loughry did not elect to conduct any further individual voir dire of Juror A, and

she was impaneled as a juror, as was another juror who had answered these questions

similarly to Juror A.

Following six days of trial and two days of deliberation, the jury returned a verdict

finding Loughry guilty of eleven counts — one count of mail fraud, seven counts of wire

fraud, one count of witness tampering, and two counts of making false statements to a

federal agent — and acquitting him of one count of mail fraud and nine counts of wire

fraud. The jury was unable to reach a verdict on one count of wire fraud. After the jury

returned its verdict, the district court entered a judgment of acquittal as to the witness-

tampering count for a lack of sufficient evidence. (The government had dismissed the three

other counts of the indictment before trial.) Of significance here, the jury acquitted

4 Loughry on the count charging him with mail fraud in connection with his removal of the

historical Cass Gilbert desk from the Supreme Court building to his home, which had been

the subject of extensive media coverage.

Shortly after trial, an individual on the street outside the Kanawha County

Courthouse approached counsel for Loughry and informed him that he should look at the

Twitter account of Juror A. Counsel did so and saw that Juror A had “liked” or “retweeted”

four tweets over the summer of 2018 related to the West Virginia Supreme Court scandal.

Twitter is a social networking platform that allows a person to post and read short

messages called “tweets.” Tweets can be up to 280 characters long and can include links

to websites and other resources. A Twitter user can also “follow” other Twitter users,

electing for those users’ tweets to appear on his or her “home timeline” or “feed.” The

Twitter user can reply to a tweet with a comment, indicate that the user “liked” a tweet by

tapping a heart icon, and republish a tweet to the user’s own followers by “retweeting” it

or quoting it. Twitter can thus be, and often is, used to receive news, to follow leaders and

celebrities, or simply to stay in touch with family and friends.

Loughry’s counsel found that Juror A had “liked” or retweeted some 11 tweets

during the four months before Loughry’s trial, and 4 of them related to comments about

the conduct being reported about the justices of the West Virginia Supreme Court, as

follows:

June 7, 2018: Juror A “liked” and retweeted a tweet by a state legislator, Delegate Mike Pushkin, stating: “When the soundness of the judiciary is questioned, coupled with the corrupt activities of other branches of government, how is the public ever to have any faith in State government?” The tweet contained a link to a West Virginia Gazette Mail article about the

5 civil complaint filed by the Judicial Investigations Commission charging ethics violations.

June 26, 2018: Juror A “liked” a tweet by another state legislator, Delegate Rodney Miller, stating: “Legis Special Session begins at noon today looking at Supreme Court impeachments; more state employees quitting/fired; DHHR $1 million overspending for nothing; RISE program dysfunctional until Gen. Hoyer gets involved. My goodness we’ve got issues to take care of!”

June 26, 2018: Juror A “liked” another tweet by Delegate Mike Pushkin, stating: “Justice Loughry should resign. The people of WV already paid for his couch, he should spare them the cost of his impeachment.” The tweet contained a link to a West Virginia Gazette Mail opinion piece entitled, “Ken Hall: WV Justices who take advantage of public funds should resign.”

August 7, 2018: Juror A “liked” a tweet by private citizen James Parker, stating: “Yes, it’s a sad day in WV to think these individuals who are supposed to be the pillars of what is right, just and truthful would be overcome with such an attitude of self importance that they thought the lavish spending was appropriate!”

Counsel also discovered that Juror A had accessed Twitter on at least two days during trial.

On October 3 (the day the government began presenting its case), Juror A “liked” a tweet,

and on October 6 (a Saturday on which the court was not in session), Juror A retweeted

one tweet and tweeted one of her own. That activity on both dates, however, was related

to football. Counsel also learned that Juror A was “following” two local journalists who

had reported on the trial but did not provide any evidence that Juror A “liked” or retweeted

those journalists’ tweets during trial.

Based on this Twitter activity, Loughry filed a motion for a new trial or,

alternatively, for an evidentiary hearing, contending that Juror A engaged in misconduct

and was biased. He argued that Juror A was biased against him based on her Twitter

activity during the four months before trial and because she had failed to indicate that she

6 had personal knowledge of the case or the facts of the case during voir dire. Loughry also

contended that Juror A engaged in misconduct during trial by accessing her Twitter account

on October 3 and October 6.

The district court denied Loughry’s motion, ruling that the evidence presented was

insufficient to show misconduct or bias. First, it concluded that “there is no reason to

believe that Juror A was anything but truthful in answering” the questions relating to “this

case” because Juror A’s Twitter activity shows only that she had knowledge of the

impeachment proceedings and ethics investigation — not “this case.” The court did

acknowledge that the facts relating to the impeachment proceedings and ethics

investigation overlapped with the “facts of this case” — something that also had to be

obvious to counsel at the time. But it concluded that an affirmative answer by Juror A

relating to the “facts of this case” would not have provided a valid basis for challenging

Juror A for cause in view of her other responses relating to her ability to consider only the

evidence presented at trial. Moreover, the court pointed out that the only facts discussed

in the tweets’ linked articles that overlapped with the “facts of this case” related to the Cass

Gilbert desk and the vehicle usage, but the jury acquitted Loughry of the desk-related count

and seven vehicle-related wire-fraud counts. The court explained further that Juror A’s

failure to affirmatively respond to the open-ended questions asking potential jurors if they

could “think of anything that might prevent them from rendering a fair and impartial

verdict” or if they had “anything further to add” was a “simple innocent failure to disclose

information that could have been elicited by questions counsel chose not to ask.”

7 The district court also rejected Loughry’s separate claim that Juror A engaged in

misconduct by using Twitter during trial. The court explained that it had never admonished

the jurors to make no use of social media during trial. “Rather, the jury was informed

repeatedly that the jurors were not to use social media to learn or discuss anything about

‘this case,’” and Juror A’s Twitter activity does not show that she read tweets about “this

case.” (Emphasis added).

The court concluded that “[w]ithout even a threshold showing of juror misconduct,”

it would not “expend its resources to allow the defendant to pry into a juror’s pretrial

conduct and fish for evidence of bias.”

After the district court sentenced Loughry and entered judgment, Loughry filed this

appeal, challenging only the court’s denial of an evidentiary hearing with respect to Juror

A’s alleged misconduct and bias.

II

Loughry contends first that Juror A’s use of social media during the trial

constituted misconduct, in violation of Remmer v. United States,

(1954),

entitling him to an evidentiary hearing. In Remmer, the Supreme Court held that any

outside contact “with a juror during a trial about the matter pending before the jury is . . .

presumptively prejudicial,” entitling the defendant to a hearing to determine if such contact

was in fact prejudicial.

. Loughry relies on evidence that Juror A used Twitter

on two days during the span of trial — once on October 3 and twice on October 6 — and

that she “followed” two local reporters, one of whom tweeted on October 9 (during trial)

8 that “[t]here seems to be quite a bit of evidence against the Justice.” Loughry presented no

evidence, however, that Juror A accessed Twitter on October 9 and was thereby exposed

to that tweet. Nonetheless, Loughry urges that because of the nature of social media,

any potential juror contact with social media during trial about a matter before the jury

triggers the Remmer presumption of prejudice, citing United States v. Harris,

(6th Cir. 2018). In Harris, the court held that a Remmer hearing was justified by evidence

that a juror’s live-in girlfriend had researched the defendant online and likely found

information about the defendant that had been excluded at trial.

. The court

reasoned that knowledge of this information was imputable to the juror because the

girlfriend did not know the defendant and had no reason to research the defendant except

for the juror’s participation in the case.

at 953–54.

Rejecting Loughry’s argument, the district court concluded:

There is no evidence or allegation that Juror A posted anything related to the case during [trial]. Although Juror A follows a number of West Virginia elected officials and members of the media — including Kennie Bass of WCHS-TV and Brad McElhinny of West Virginia MetroNews, who reported on the evidence admitted at trial — there is no evidence that Juror A was exposed to any content related to the case. [Moreover] . . . the court had instructed the jurors to refrain from using social media or the internet to obtain information on the case or communicate with anyone about the case, and Juror A has not been shown to violate that admonition.

The court added that Loughry failed to show that the reporter’s October 9 tweet “was of

such a character as to reasonably draw into question the integrity of the verdict.” (Quoting

Barnes v. Joyner,

(4th Cir. 2014)).

It is foundational to due process that a defendant in a criminal case be given the

right to a trial “by an impartial jury.” U.S. Const. amend. VI; see also United States v.

9 Small,

(4th Cir. 2019). And the Supreme Court has held that this

impartiality is presumptively compromised by “any private communication, contact, or

tampering directly or indirectly, with a juror during a trial about the matter pending before

the jury.” Remmer,

. A defendant seeking a hearing on this issue must

present “a credible allegation that an unauthorized contact was made, and that the contact

was of such a character as to reasonably draw into question the integrity of the trial

proceedings, constituting more than an innocuous intervention.” United States v. Johnson,

(4th Cir. 2020) (cleaned up). This requires “something more than mere

speculation.” United States v. Forde,

(4th Cir. 2011) (per curiam)

(quoting United States v. Barshov,

(11th Cir. 1984)); see also United

States v. Gravely,

(4th Cir. 1988) (“[T]he sixth amendment . . . [does

not] require[ ] an inquiry into possible external influence when a threshold showing of

external influence has not been made”); United States v. Wintermute,

(8th Cir. 2006) (rejecting a Remmer claim where the defendant had argued that a

juror’s comment suggested that she had “probably accessed the Internet” during trial on

the ground that such “hypothesis” was nothing but “[s]peculation”); United States v.

Robertson,

, 1294–95 (10th Cir. 2007) (rejecting a Remmer claim that was

based only on “innuendo”).

In this case, the record shows that on October 3 and October 6, Juror A accessed

Twitter, “liking” a tweet on the first of those days and retweeting a tweet and tweeting one

of her own on the second. All of this activity related to football, and none referred to any

facts about the case or, more broadly, the scandal at large. The record also shows that Juror

10 A followed reporters who were reporting on the Loughry trial, and one had tweeted on

October 9, “There seems to be quite a bit of evidence against the Justice.” But there is no

evidence that Juror A read that tweet. Indeed, there is no evidence that she accessed her

Twitter account on October 9. Loughry’s request for a Remmer hearing rests on the

argument that Juror A could have seen the reporter’s tweet on October 9 or other tweets by

the reporters because she had a Twitter account and used it. Such a standard is defined so

broadly as to reach not only Juror A’s activity but also the activity of any other juror who

had a social media account. In any event, the jurors were repeatedly instructed to avoid

social media “about this case,” and we presume that the jury followed these instructions.

See Stamathis v. Flying J, Inc.,

(4th Cir. 2004). The stubborn fact yet

remains in this case that Loughry did not make “a credible allegation that an unauthorized

contact was made.” Johnson,

(cleaned up).

To be sure, social media can facilitate improper contacts to a greater degree than

before the technology existed. Sites containing news and prohibited information can be

displayed instantly with but a touch of a finger. As a consequence of this new reality,

courts must become more circumspect in undertaking to guard jurors from inappropriate

contacts and communications during trial. Indeed, a committee of the Judicial Conference

— the Committee on Court Administration and Case Management — recently circulated

proposed model instructions designed to do just that. See Judicial Conference Committee

on Court Administration and Case Management, Proposed Model Jury Instructions, The

Use of Electronic Technology to Learn or Communicate about a Case, updated June 2020.

11 For example, the model instructions that are proposed for the beginning of trial read as

follows:

[D]uring the trial, you must not conduct any independent research about this case, or the matters, legal issues, individuals, or other entities involved in this case. Just as you must not search or review any traditional sources of information about this case . . . , you also must not search the internet or any other electronic resources for information about this case or the witnesses or parties involved in it.

Second, this means that you must not communicate about the case with anyone . . . . Most of us use smartphones, tablets, or computers in our daily lives to access the internet, for information, and to participate in social media platforms. To remain impartial jurors, however, you must not communicate with anyone about this case, whether in person, in writing, or through email, text messaging, blogs, or social media websites and apps (like Twitter, Facebook, Instagram, LinkedIn, YouTube, WhatsApp, and Snapchat).

* * * Finally . . . [m]any of the tools you use to access email, social media, and the internet display third-party notifications, pop-ups, or ads while you are using them. These communications may be intended to persuade you or your community on an issue, and could influence you in your service as a juror in this case. For example, while accessing your email, social media, or the internet, through no fault of your own, you might see popups containing information about this case or the matters, legal principles, individuals or other entities involved in this case. Please be aware of this possibility, ignore any pop-ups or ads that might be relevant to what we are doing here, and certainly do not click through to learn more if these notifications or ads appear. If this happens, you must let me know.

While these model instructions, and the other ones proposed, focus on problems raised

by social media, they do not recommend that jurors be told to avoid social media altogether

— only with respect to matters related to “the case” before the jurors.

We commend such expanded instructions and highlight also that they recognize that

accommodation should be made to allow jurors the use of electronic media for non-case-

related matters and to protect juror privacy. Of course, in seeking such a balance, an eye

12 must always be kept on ensuring that defendants in criminal cases have an impartial jury,

the touchstone of which is a jury “capable and willing to decide the case solely on the

evidence before it.” McDonough Power Equip., Inc. v. Greenwood,

(1984) (emphasis added) (quoting Smith v. Phillips,

(1982)).

Loughry contends that the increased risk of improper juror contact posed by social

media was realized in this case based on the possibility that Juror A saw the reporters’

tweets about the trial. Relying on the Sixth Circuit’s recent decision in Harris, he argues

that Remmer should be read to justify a hearing on improper contacts “merely [on] potential

juror contact with social media during trial.” But Harris does not, as we read it, reach so

broadly.

In Harris, it was discovered that a juror’s live-in girlfriend had accessed the

defendant’s LinkedIn page during trial.

. The circumstantial evidence of

improper contact arose because, other than her communication with the juror, the girlfriend

had no reason to research the defendant.

As the court explained, the girlfriend did not

know the defendant, and the trial “received little publicity.”

The only reasonable

explanation was that the juror “must have discussed the trial with his girlfriend.”

Worse, the court concluded that the girlfriend likely found the defendant’s LinkedIn page

by searching the defendant’s name on Google, a search method that would have also

exposed her to “prejudicial information that the government was precluded from

introducing at trial.”

. The Sixth Circuit concluded that this constituted “credible

evidence” that the juror discussed the case with his girlfriend, that the girlfriend then

searched the internet for the defendant, and that the girlfriend “potentially communicate[d]

13 her findings” to the juror.

at 953–54. The court therefore ordered a Remmer hearing.

.

The facts in Harris, however, are readily distinguishable from those before us, and

the Harris court did not relax the Remmer requirements for a hearing. Rather, it applied

them, finding that the defendant “presented a colorable claim of extraneous influence on a

juror.” Harris,

. This holding is entirely consistent with what we have

stated repeatedly — the standard for justifying a hearing under Remmer requires a

defendant to present “a credible allegation that an unauthorized contact was made,”

Johnson,

(cleaned up); see also Barnes,

, and the allegation

must be based on “more than mere speculation,” Forde,

.

Loughry also contends that whether or not Juror A actually viewed information

about the case on Twitter during trial, she nonetheless committed misconduct by violating

the district court’s instructions that jurors avoid all social media during trial. Specifically,

according to Loughry, the court instructed the jury to “avoid all social media” (on Day 2

of the trial) and “social networking exposure of any kind” (on Day 7 of the trial). Thus, he

argues, Juror A’s Twitter use on October 3 and 6 shows that she violated these instructions,

justifying a Remmer hearing to investigate the scope of this misconduct.

We conclude, however, that the statements on which Loughry relies were taken out

of context and that any reasonable juror receiving the district court’s instructions during

trial would have concluded that social media was prohibited only in connection with the

case. A review of the instructions demonstrates this.

Day 1 (October 2, at the outset of trial after the jury had been impaneled):

14 I want to mention to you one thing that is so very important at the outset, and that is, of course, as jurors, you must decide this case solely upon the evidence that you hear from the witness stand and the exhibits as they’re offered and introduced into evidence in the case.

This means that during the trial, you must not conduct any independent research about this case, the matters in this case, or the individuals involved in this case.

You must not consult dictionaries or reference materials; you must not search the Internet, websites, blogs, or use any other tools, electronics or otherwise, to obtain information about this case or to help you decide the case.

Do not try to find out information from any source outside the confines of this courtroom.

Until you retire and deliberate, you may not discuss this case with anyone, not even your fellow jurors.

You may not communicate with anyone about the case, on your cell phone, your iPhone, through e-mail, text messaging, Twitter, through any blog or website, including Facebook, Google, Myspace, LinkedIn, YouTube, anything imaginable. It’s all out. You must not use it in any sense.

Day 2 (October 3, at the end of the day before the jury was dismissed for the day):

You’re going to hear me say this more than once, but, continue to be guarded, that is, do not expose yourself to any media coverage of any kind; avoid all social media, as well, and avoid discussing this or letting anyone draw you into discussion about the case.

Day 3 (October 4, at the end of the day before the jury was dismissed):

Once again, I’ll remind you, it’s better for you to have someone else review the newspapers, and they can filter what you see. As you know, of course, when the newscasts come on television, because that’s pretty well fixed, you need to avoid that, of course. And radio is a little different, it gives the news at any moment, so you have to be very cautious about that. And if you happen to have it on and something is coming on about this case — and I’m not sure that that will happen, but it could very well happen — then click it off.

15 And continue to observe the Court’s direction that you not let anyone speak to you about this case nor [are] you to engage with anyone else, and avoid all social networking with respect to it as well.

Day 4 (October 5 (Friday), at the end of the day before the jury was dismissed for the weekend):

Avoid all social networking having to do with the case.

Day 5 (October 8, at the end of the day before the jury was dismissed):

It continues to be especially important that you observe the Court’s directive that you avoid all media coverage about this case, and that, of course, has to do with radio, television and newspapers, and all social networking, as well.

So continue to observe those same directions and avoid all contact. Don’t let anyone contact you about it, whether it is through social networking or otherwise, and you, of course, would not be contacting those as well.

Day 6 (October 6, at the end of the day before the jury was dismissed):

Avoid all news media and social networking having to do with this case.

Day 7 (October 10, at the end of the day before the jury was dismissed):

And I will just say briefly that, as you can understand, under no circumstances are you to discuss the case with anyone or let anyone discuss it with you. Continue to avoid all news media and social networking exposure of any kind until you’re back in here in the morning in the jury room.

Day 8 (October 11, at the end of the day after deliberations began and before the jury was dismissed):

And I am not going to go over all this with you again, but I want to impress upon you, continuing the necessity of your seeing to it that no one is in touch with you about this case, not even among yourselves, until all 12 of you are back in the jury room tomorrow morning.

Day 9 (October 12, no instructions given as deliberations concluded and the verdict was returned).

(Emphasis added throughout).

16 The instructions thus given throughout the trial unambiguously prohibited social-

media usage about the case — well anticipating the model instructions recently proposed

by the Judicial Conference Committee. The only instruction that did not use language

equivalent to “about this case” was a short-form, single-sentence instruction given on Day

7, after deliberations had begun. If, for some reason, Juror A understood the Day 7 short-

form instruction to prohibit all social media usage, it would nonetheless have been of no

moment, since that day was well after Juror A’s Twitter usage on Day 2 and on the Saturday

after Day 4.

At bottom, we conclude that the district court did not abuse its discretion in denying

Loughry’s motion for an evidentiary hearing under Remmer because Loughry failed to

make a credible allegation that an improper contact occurred.

III

Next, Loughry contends that, in view of the four tweets Juror A “liked” or retweeted

during the summer months before trial relating to the impeachment proceedings and ethics

investigation of the West Virginia Supreme Court justices, Juror A dishonestly answered

eight of the district court’s questions during voir dire about her knowledge of the case,

thereby indicating bias. See McDonough,

(holding that “to obtain a

new trial . . . a party must first demonstrate that a juror failed to answer honestly a material

question on voir dire, and then further show that a correct response would have provided a

valid basis for a challenge for cause”); see also Conaway v. Polk,

(4th Cir. 2006). More particularly, he argues that Juror A dishonestly answered five

17 questions when denying knowledge of or discussion about “this case” or the “facts of this

case” and three others when answering general questions of prior knowledge and bias. The

three general questions asked were: (1) whether prospective jurors had an opinion about or

had expressed an opinion about the “guilt or innocence of the defendant or the charges” in

the indictment; (2) whether there was anything that might “prevent [the prospective jurors]

from rendering a fair and impartial verdict based solely upon the evidence” at trial; and

(3) whether any prospective juror “wish[ed] to change or supplement” an answer.

Loughry’s argument that Juror A dishonestly answered the three general questions,

however, depends on his demonstrating that she dishonestly answered the five questions

about “this case” or the “facts of this case.”

The district court rejected Loughry’s claim that Juror A answered dishonestly during

voir dire and reasoned further that any arguably wrong answers would not have warranted

her dismissal for cause. We conclude that the evidence is insufficient to show that Juror A

answered any voir dire questions dishonestly.

The district court began the relevant portion of voir dire by asking whether jurors

had knowledge of “this case” or the “facts of this case.” And the transcript makes clear

that these questions related specifically, as the court explained, to the “case as set forth in

the indictment.” (Emphasis added). Juror A did not respond to these questions. One of

the prospective jurors who did, however, also indicated that he was “probably” getting the

criminal case “confused with the impeachment trials.” After concluding this exchange

with that juror, the court turned to the entire venire and asked, “Apart from what you’ve

told me about this case, let me ask who among you have heard anything about the

18 impeachment proceedings that are taking place in the state legislature?” (Emphasis added).

Juror A was among the prospective jurors who said “yes” to this question. The court then

asked a series of follow-up questions, all geared towards uncovering whether the

prospective jurors who answered “yes” would be able to ignore what they had learned

about the impeachment proceedings and base a verdict solely upon the evidence adduced

in court. Juror A and the other prospective jurors said that they could do so. Then, the

court posed some open-ended questions, asking “while we’re on the subject, is there

anything further that any of you would want to relate to the Court about your knowledge

of this case that goes beyond what we’ve already covered” and later, whether the

prospective jurors wanted “to change or supplement” any answers. Juror A did not answer

these questions. The court also asked the prospective jurors if “any of you now have an

opinion or have you at any time expressed an opinion as to the guilt or innocence of the

defendant of the charge or charges contained in this indictment in this case.” Juror A, like

the other prospective jurors, did not respond.

Given the series of questions the prospective jurors were actually asked, the four

pretrial tweets that Juror A “liked” or retweeted do not suggest that Juror A was dishonest

in answering, or failing to answer, any voir dire questions. The June 7 tweet linked to an

article discussing the civil complaint filed by the Judicial Investigations Commission for

ethics violations. One of the June 26 tweets mentioned impeachment, and the other argued

that Loughry should resign from the bench. And the August 7 tweet merely expressed how

“sad” it was that the justices “would be overcome with such an attitude of self importance

that they thought the lavish spending was appropriate!” So while Juror A’s Twitter activity

19 clearly shows that she was aware of the impeachment proceedings and the ethics

investigation, it reveals nothing about her awareness of the criminal case — which the

court carefully distinguished from the impeachment proceedings during voir dire. There

is thus no reason to conclude that Juror A was dishonest when she did not answer the

court’s questions about knowledge of “this case” or the “facts of this case.”

Crucially, when the court asked the venire about the impeachment proceedings,

Juror A was one of the 13 prospective jurors who admitted hearing about them. Juror A

was thus entirely forthcoming that she was aware of the impeachment proceedings. And

when asked by the court — repeatedly — whether she would be able to put aside her

knowledge of the impeachment and base a verdict solely upon the evidence presented in

court, she affirmed that she could do so.

As for the open-ended question asking if the prospective jurors had anything to add,

Juror A was not required to mention again her knowledge of the impeachment because she

had already told the court that she had knowledge of the impeachment proceedings. And

finally, with respect to the question asking whether any of the prospective jurors had “an

opinion as to the guilt or innocence of the defendant of the charge or charges contained in

this indictment in this case,” Juror A’s pretrial Twitter activity suggests only that she had

an opinion on the impeachment and the ethics investigation, not Loughry’s “guilt . . . of

the charge or charges contained in this indictment in this case.” Moreover, Juror A had by

this point in voir dire already affirmed to the court that she could put aside her knowledge

of the impeachment proceedings and decide the case based solely on the evidence

20 introduced at trial. Her pretrial Twitter activity therefore does not suggest that she

dishonestly hid her views regarding Loughry’s guilt or innocence.

Furthermore, even if we were to recognize that Juror A should have volunteered

more information about her views on the impeachment proceedings — a requirement we

do not impose — “a juror’s failure to elaborate on a response that is factually correct but

less than comprehensive” does not establish juror dishonestly “where no follow-up

question is asked.” Porter v. Zook,

(4th Cir. 2015). Here, Loughry’s

counsel was invited to inquire further into the jurors’ knowledge of the case or the

impeachment proceedings, and, in fact, he questioned other potential jurors who stated that

they had knowledge about those topics. He did not, however, question Juror A or two other

jurors who likewise stated that they had knowledge of the impeachment proceedings but

not the criminal case. Loughry cannot now use his failure to follow up with Juror A as

grounds for an evidentiary hearing. See Billings v. Polk,

(4th Cir. 2006)

(“Otherwise, defendants would be able to sandbag the courts by accepting jurors onto the

panel without exploring on voir dire their possible sources of bias and then, if their gambit

failed and they were convicted, challenging their convictions by means of post-trial

evidentiary hearings based on newly discovered evidence of possible juror bias”).

Loughry’s theory of dishonesty rests on parsing the transcript to isolate five stand-

alone questions that asked about “this case” or about the “facts of this case.” He then

contends that Juror A dishonestly answered those questions because the “facts of this case”

overlapped with the facts prompting the impeachment proceedings and that construing

“this case” to not include “the impeachment” is possible only by “hypertechnically

21 parsing” the court’s questions. But a fair reading of the voir dire transcript shows that

when the court referred to “this case” and the “facts of this case,” it was referring to the

criminal case — not the impeachment proceedings. As the court stated, “You’ve heard

what I told you about — what little I’ve told you about this case as set forth in the

indictment. Do any of you have personal knowledge of the facts of this case?” (Emphasis

added). And when asked about the impeachment proceedings — which Juror A’s pretrial

Twitter activity did reveal knowledge of — Juror A freely admitted to the court that she

had knowledge of those proceedings. Moreover, with Loughry’s counsel’s own

understanding that facts relating to impeachment overlapped with facts relating to the case,

counsel knew from Juror A’s answers that she had some knowledge of facts pertaining to

the case — i.e., the overlapping facts. Yet, he never inquired further and was satisfied to

let her sit on the jury.

In sum, we conclude that because Loughry has not made a colorable showing that

Juror A dishonestly answered material questions during voir dire, the district court did not

abuse its discretion in failing to hold an evidentiary hearing to investigate McDonough

bias.

IV

Finally, Loughry contends that the district court abused its discretion in failing to

hold an evidentiary hearing to investigate whether Juror A was actually biased against him.

Specifically, Loughry contends that Juror A’s failure “to answer numerous voir dire

questions honestly . . . is evidence of bias against the defendant, particularly if Juror A

22 acted deliberately for the purpose of getting onto the jury.” See Jones v. Cooper,

(4th Cir. 2002).

To demonstrate actual bias, a defendant “must prove that a juror, because of his or

her partiality or bias, was not ‘capable and willing to decide the case solely on the

evidence before it.’” Porter v. Zook,

(4th Cir. 2018) (quoting Phillips,

). Absent such proof, “[j]urors are ‘presumed to be impartial.’” United

States v. Powell,

(4th Cir. 2017) (quoting Wells v. Murray,

(4th Cir. 1987)).

Here, too, we conclude that the district court did not abuse its discretion in failing

to hold an evidentiary hearing to investigate Loughry’s allegations of actual bias. Juror

A’s pretrial Twitter activity reveals that she read tweets of Loughry’s impeachment

proceedings and ethics investigation and that she had seemingly approved of — “liked” —

what she read. But when asked directly by the district court whether she could put aside

what she had learned before trial about the impeachment proceedings and decide the case

“based solely on the evidence as we receive it here in the courtroom through the witnesses

and the exhibits that are admitted into evidence,” she said “yes.” It is difficult to see how

the same pretrial Twitter activity that likely led Juror A to tell the court that she had heard

about the impeachment proceedings now undermines that answer.

Insofar as Juror A’s Twitter activity reveals that she had some preexisting

knowledge of the case, “it is a long-settled proposition that mere knowledge of a case is

insufficient to support a finding of actual prejudice.” United States v. Higgs,

(4th Cir. 2003). Moreover, even if that activity suggests that she likely viewed

23 Loughry less than favorably, she nonetheless affirmed to the district court that she did not

have “an opinion” and has not “expressed an opinion as to the guilt or innocence of the

defendant of the charge or charges contained in this indictment in this case” and that she

could render a verdict based solely on the evidence at trial. (Emphasis added).

We conclude that the district court did not abuse its discretion in refusing to hold an

evidentiary hearing on actual bias.

* * *

The long and short of this case is that evidence indicates that Juror A had some

pretrial exposure to news of the investigations of the West Virginia Supreme Court justices

and participated modestly in the public dialogue via a few “likes” and retweets on Twitter.

But evidence further indicates that she engaged in no prohibited contacts or

communications during trial. As we have noted, social media does heighten the risk that

jurors will be exposed to external information about the case, but here Loughry has failed

to make a threshold showing that that risk was realized. In this case, all the evidence points

to a fair trial. The jury, including Juror A, assured the court that it was “capable and willing

to decide the case on the evidence before it.” McDonough,

(quoting

Phillips,

). And its verdict reflects just that, as the jury acquitted Loughry

on several charges.

At bottom, we conclude that the district court, which carefully scrutinized the

evidence advanced by Loughry in support of his motion, did not abuse its discretion in

denying Loughry’s request for an evidentiary hearing. The court’s judgment is

AFFIRMED.

24 DIAZ, Circuit Judge, dissenting in part:

My colleagues conclude that the district court acted within its discretion in denying

Loughry’s request for an evidentiary hearing to explore allegations of juror misconduct.

For the reasons explained by the majority, I agree that denying Loughry a McDonough 1

hearing wasn’t reversible error. On the present record, I also agree that Loughry failed to

make out a claim of actual bias. But because I would hold that Loughry is entitled to a

Remmer 2 hearing to ascertain the full extent of Juror A’s Twitter activity during the trial, I

respectfully dissent.

Under Remmer, a defendant is entitled to a presumption of prejudice and a hearing

when he “presents a credible allegation of communications or contact between a third party

and a juror concerning the matter pending before the jury” that could “reasonably draw

into question the integrity of the verdict.” Barnes v. Joyner,

(4th Cir.

2014). This is a “minimal standard.”

at 245 (quoting United States v. Cheek,

(4th Cir. 1996)).

As we explained in Barnes:

Extrajudicial communications or contact with a juror has been deemed to trigger Remmer in a variety of circumstances, including: a juror being offered a bribe during trial and subsequently being investigated by an FBI agent; a juror applying for a job at the prosecuting attorney’s office during the trial; a local restaurant owner suggesting to jurors in a capital case that “they ought to fry the son of a bitch,”; and allegations, if proven to be true during an evidentiary hearing, that a juror’s husband pressured her throughout the trial to vote for the death penalty.

1 McDonough Power Equip., Inc. v. Greenwood,

(1984). 2 Remmer v. United States,

(1954). 25 Id. (cleaned up). More recent examples include a juror asking her father, see Hurst v.

Joyner,

(4th Cir. 2014), and a pastor, see Barnes,

, what

the Bible says about the death penalty; and a juror researching the definition of an element

of a crime on Wikipedia, United States v. Lawson,

, 639–40 (4th Cir. 2012).

Here, Loughry’s request for a hearing was predicated on the claim that Juror A was

likely exposed to tweets published by two reporters during Loughry’s trial. Specifically,

Loughry alleged that after the trial, an individual approached Loughry’s counsel and told

him he should look into Juror A’s Twitter account. A review of the juror’s public account

revealed that, in the months leading up to trial, Juror A had “liked” and “retweeted”

comments criticizing Loughry and a news article detailing the judicial complaint filed

against Loughry by West Virginia’s Judicial Investigation Commission. The review also

revealed Twitter activity by Juror A on two dates during the trial: October 3 and October

6, 2018. 3 Loughry argued that because Juror A “followed” two reporters who covered the

trial, she would have seen their “near constant ‘tweets’ concerning the trial” because “such

posts appear on a user’s home timeline.” J.A. 834; cf. Oracle Am., Inc. v. Google Inc.,

(N.D. Cal. 2016) (“When one Twitter user ‘follows’ another

Twitter user, the latter’s posts appear in the former’s default real-time feed of tweets.”).

As Loughry details on appeal, the two reporters “tweeted” or “retweeted” about his

case a combined total of 73 times during the trial. Indeed, one of the reporters did so twelve

3 Loughry also alleged that Juror A used her Instagram account on October 7, 2018 and her Facebook account on October 8, 2018.

26 times on October 3 alone. And on October 9—the day before jury deliberations began—

the other reporter tweeted: “There seems to be quite a bit of evidence against the justice.”

Appellants’ Br. at 4–5. 4

While we haven’t previously addressed this precise issue, other courts have found

that social media use can trigger the Remmer presumption. See Ewing v. Horton,

(6th Cir. 2019) (juror looked up defendant’s Facebook profile); United

States v. Harris,

, 952–54 (6th Cir. 2018) (juror’s live-in girlfriend viewed

defendant’s LinkedIn page); State v. Smith,

, 48–49 (Tenn. 2013) (juror

exchanged Facebook messages with a government witness). And in a different context,

the Supreme Court has recognized “the extraordinarily high” risk of jurors being tainted

by “read[ing] reactions to a verdict on Twitter.” See Dietz v. Bouldin,

(2016).

In Dietz, the Court considered the extent of a district court’s authority to “rescind a

jury discharge order and recall a jury for further deliberations after identifying an error in

the jury’s verdict.”

. The Court held that a district court had such authority, but

cautioned that “[b]ecause the potential of tainting jurors and the jury process after discharge

is extraordinarily high, . . . this power is limited in duration and scope, and must be

exercised carefully to avoid any potential prejudice.”

The Court emphasized that, in

4 The government says that we should disregard these tweets because Loughry didn’t attach them as exhibits in the district court. Loughry invites us to take judicial notice of the tweets instead, as they remain publicly available to this day. Neither is necessary here, as the fact that the reporters tweeted throughout the trial wasn’t in dispute below, and the district court assumed as much in denying Loughry a hearing.

27 weighing its decision, a district court must consider the extent to which the dismissed jurors

may have accessed their smartphones or the internet after being dismissed.

. As

the Court explained,

It is a now-ingrained instinct to check our phones whenever possible. Immediately after discharge, a juror could text something about the case to a spouse, research an aspect of the evidence on Google, or read reactions to a verdict on Twitter. Prejudice can come through a whisper or a byte.

(emphasis added).

So too here. Loughry has made a credible allegation that Juror A was likely exposed

to tweets from reporters commenting about the trial. Indeed, it’s undisputed that (1) a

substantial percentage of Juror A’s Twitter activity in the months leading up to trial related

to the investigation of Loughry and the related impeachment proceedings; (2) Juror A

“followed” two reporters who covered the trial and tweeted regularly throughout it; and (3)

Juror A used Twitter on at least two days during the trial. And, in my view, Juror A’s

activity on other social media sites during additional trial days indicates that she likely

scrolled through her Twitter feed passively on at least some of the days when she didn’t

affirmatively interact with other accounts.

Loughry’s evidence is admittedly circumstantial. Nonetheless, it’s the most he

could possibly offer without the opportunity to conduct discovery or question Juror A. As

Loughry explains, Twitter “can be accessed by phone virtually anywhere and for any length

of time, and includes no visible record of whether a tweet has been seen or not.” Reply Br.

at 9. Thus, it’s impossible to obtain direct evidence of which tweets Juror A saw without

a hearing. Cf. Smith,

(“[T]echnology has made it easier for jurors to

28 communicate with third parties and has made these communications more difficult to

detect”); United States v. Fumo,

(3d Cir. 2011) (Nygaard, J., concurring

in part and dissenting in part) (“The Internet and social networking sites . . . have simply

made it quicker and easier to engage more privately in juror misconduct”). And the

evidence Loughry did offer is stronger than the evidence presented in Harris, where the

Sixth Circuit remanded for a Remmer hearing after determining that the defendant

presented “a colorable claim of extraneous influence.” Harris,

.

There, Harris offered evidence that a juror’s live-in girlfriend viewed Harris’s

LinkedIn 5 profile either during or shortly after the trial (the record was inconclusive

regarding the exact date).

. The Sixth Circuit agreed with Harris that the juror’s

girlfriend likely found the LinkedIn profile by searching for Harris on Google, where she

may have also seen prejudicial information that the government was precluded from

introducing at trial.

. And though the district court had admonished the jury not

to discuss the case with others, the Sixth Circuit reasoned it was “quite possible” that the

juror told his girlfriend about the trial, leading her to google Harris and potentially

communicate her findings to the juror.

at 953–54. The court determined that this mere

possibility of inappropriate communication with a juror was enough to warrant a Remmer

hearing and held that the district court abused its discretion in denying one.

.

5 “LinkedIn is a web-based social networking site that presents itself as an online community offering professionals ways to network.” Low v. LinkedIn Corp.,

(N.D. Cal. 2012). 29 There is much less need for speculation here. As the district court explained, when

a Twitter user publishes a tweet, that tweet appears on the homepage of all accountholders

who “follow” that user. Thus, when Juror A used Twitter during the trial, the reporters’

tweets were on her homepage, where she would have either read them or scrolled past them

to read other tweets. And since she was on the jury and interacted with other tweets about

Loughry in the months leading up to trial, it’s reasonable to assume that tweets about the

trial would have caught her eye.

The district court—and my friends in the majority—fault Loughry for failing to

prove with certainty that Juror A saw the reporters’ tweets. But again, there’s simply no

way Loughry could do so without being allowed, at minimum, to question Juror A about

her Twitter use during the trial. See, e.g., Harris,

(stating that although

Harris “did not establish that [a juror] was exposed to unauthorized communication, Harris

did present a colorable claim of extraneous influence, which necessitated investigation.”);

see also Fed. R. Evid. 606(b)(2)(A) (“A juror may testify about whether[ ] extraneous

prejudicial information was improperly brought to the jury’s attention.”).

My colleagues also express concern that granting Loughry’s request for a Remmer

hearing would open the floodgates to a hearing any time a defendant presents evidence that

a juror used social media during a trial. Not so. The mere fact that Juror A used Twitter

during the trial isn’t what warrants a hearing here. Rather, Loughry is entitled to hearing

because of Juror A’s past Twitter activity, coupled with who she follows (reporters) and

the fact that those reporters used Twitter repeatedly to report and comment on Loughry’s

trial.

30 * * *

On this record, I would remand for a Remmer hearing, where the government could

attempt to rebut the presumption of prejudice by showing either that Juror A didn’t see the

reporters’ tweets during the trial or that her exposure to them didn’t harm Loughry.

Because my colleagues disagree, I respectfully dissent.

31