United States v. McFadden

116 F.4th 1069

Opinion

Appellate Case: 23-1089 Document: 010111103175 Date Filed: 08/30/2024 Page: 1 FILED

United States Court of Appeals Tenth Circuit

PUBLISH

August 30, 2024

UNITED STATES COURT OF APPEALS Christopher M. Wolpert

FOR THE TENTH CIRCUIT Clerk of Court

_________________________________

UNITED STATES OF AMERICA, Plaintiff - Appellee, v. No. 23-1089 MICHAEL TRACY MCFADDEN,

Defendant - Appellant.

_________________________________

Appeal from the United States District Court for the District of Colorado (D.C. No. 1:19-CR-00243-GPG-1)

_________________________________

Ty Gee (Meredith O’Harris with him on the briefs) of Maddon, Morgan and Foreman, P.C., Denver, CO, for Defendant - Appellant.

J. Bishop Grewell, Assistant United States Attorney (Cole Finegan, United States Attorney, with him on the brief), Denver, CO, for Plaintiff - Appellee.

_________________________________

Before PHILLIPS, KELLY, and FEDERICO, Circuit Judges.

_________________________________

PHILLIPS, Circuit Judge.

_________________________________

A jury convicted Michael Tracey McFadden of five criminal counts related to McFadden’s sexual assault of two minors. The district court sentenced him to serve concurrent life sentences on each count. McFadden now appeals his conviction and sentence. He contends that various evidentiary

2

errors at trial, plus a mistake in his Guidelines calculation, require a new trial and resentencing. We disagree and so, exercising our jurisdiction under 28 U.S.C. § 1291 and 18 U.S.C. § 3742(a), we affirm.

BACKGROUND

I. Factual Background Two boys native to Grand Junction, Colorado were abused throughout their childhoods by the defendant, Michael Tracy McFadden. Both boys, known to us as J.W. and K.W., spent much of their youth playing and sleeping over at McFadden’s house. McFadden is distantly related to J.W. on J.W.’s mother’s side. McFadden and K.W. are unrelated, but McFadden was a close friend of the W. family. McFadden stepped up to help look after the boys because J.W.’s mother was battling addiction and an abusive relationship, while K.W.’s parents generally struggled to make ends meet. McFadden ingratiated himself with the boys and their families through his generosity. He provided necessities (clothing, food, rides to school, doctor appointments) and luxuries (BMX bikes, paintball guns, trampolines, videogames) that the boys’ families could not afford.

J.W. lived with McFadden for the better part of six years, from roughly ages six to twelve. When J.W. was about eleven and K.W. was about ten, K.W. began regularly going over to McFadden’s house to play with J.W. and J.W.’s siblings. K.W. began spending most weekends there and would frequently stay the night. During these overnights, both boys slept with McFadden in his bed.

3

Once the boys would fall asleep, McFadden would assault them. When McFadden moved to a different home in Grand Junction, J.W. went with him, and K.W.’s family moved to a neighboring property. K.W. continued to spend significant time playing with J.W. and sleeping over at McFadden’s house. The same pattern of abuse continued there. To facilitate his abuse, McFadden routinely gave the boys melatonin in high doses before bedtime.

Though McFadden’s Colorado homes were sites of repeated abuse, McFadden also assaulted both boys across state lines. McFadden was a truck driver, so he regularly traveled on interstate highways picking up and delivering loads. He often invited J.W. and K.W. to accompany him on these trips. On one such trip to Arizona, when J.W. was about nine, McFadden assaulted him during the night while J.W. slept next to McFadden on the mattress in the semi-truck’s sleeper cab. On another trip from Telluride, CO to Farmington, NM in December 2010, when J.W. was ten, McFadden once again assaulted J.W. while the two slept in the sleeper cab of McFadden’s semi-truck. Both times, McFadden penetrated J.W.’s anus while J.W. pretended to sleep. These assaults caused J.W. to feel pressure and a wet sensation in his rear.

In December 2012, McFadden took K.W. and K.W.’s brothers on a semi-

truck trip from Idaho to Nebraska. At the time, K.W. was eleven, his older brother (S.W.) was eighteen, and his younger brother (L.W.) was nine. S.W. slept across the driver and passenger seats, while K.W., L.W., and McFadden shared the mattress in the sleeper cab behind the front seats. During the night, McFadden

4

pulled down K.W.’s pants and inserted his penis into K.W.’s anus, which K.W. said “hurt a lot.” ROA vol. VI, at 183.

Around the time of the Nebraska trip, other children started coming forward with allegations that McFadden had abused them. From these accusations, Detective Edward Prescott with the Grand Junction Police Department obtained a warrant to arrest McFadden, who was still with the W. boys in Nebraska. The local Nebraska police executed the warrant and arrested McFadden on January 3, 2013. The W. boys’ mother picked them up at a Nebraska truck stop and drove them back to Colorado. On the ride home, K.W.’s mother asked K.W. if McFadden had ever “done anything” to him, and K.W. said no. Once they returned to Colorado, K.W.’s mother arranged for K.W. to meet with a child counselor. That meeting was the first time K.W. accused McFadden of sexually abusing him. The next day, January 16, 2013, Detective Prescott conducted a forensic interview with K.W., during which K.W. detailed McFadden’s history of abusing him. That interview was recorded on video.

The parties disagree about when J.W. first accused McFadden of sexual abuse. 1 McFadden insists that J.W. outcried before K.W. talked to Detective

1 The government has a motion pending before this court to supplement the record on appeal with an exhibit list from trial, which it contends will refute McFadden’s assertion that K.W.’s outcry was tainted by J.W.’s earlier admission of abuse. The government maintains that J.W. accused McFadden after K.W.’s interview on January 16, 2013, and so K.W.’s interview was (footnote continued)

5

Prescott, the government maintains that it was after. Regardless, J.W. came forward in early 2013 and told Detective Prescott that McFadden had been sexually abusing him for many years.

Years later, in 2018, J.W. and K.W. were interviewed again about McFadden’s assaults. The FBI conducted K.W.’s interview during his stint at a juvenile correctional facility. The audio of that interview was recorded. II. Procedural Background McFadden was convicted in Colorado state court on nineteen counts of child-sex abuse. 2 McFadden appealed that conviction and won. The Colorado Court of Appeals determined that McFadden’s speedy-trial rights had been violated, and so the court dismissed all charges. The Colorado Supreme Court denied certiorari, and McFadden was released. People v. McFadden, No. 17SC573, 2018 WL 827272 (Colo. Feb. 12, 2018).

untainted by his friend’s influence. McFadden opposes this motion. The government’s arguments do not compel us to exercise our power to supplement the appellate record. See United States v. Kennedy, 225 F.3d 1187, 1191 (10th Cir. 2000). Even if McFadden is correct that J.W.’s forensic interview occurred before K.W.’s, this does not impact our deferential review of the district court’s trustworthiness analysis under Rule 807(a)(1), see Discussion § I.A.1, infra. The exhibit list that the government seeks to admit would have no effect on our decision. The government’s motion is denied.

2 The Colorado Court of Appeals opinion and the entire state-court case record is sealed. See People v. McFadden, 2013-CR-27, 2013-CR-339, 2013- CR-342 (Mesa Cnty. Dist. Ct. 2015); People v. McFadden, No. 15CA1925 (Colo. Ct. App. June 22, 2017).

6

About a year after his release, a federal grand jury charged McFadden with five criminal counts related to his sexual abuse of J.W. and K.W. across state lines. Counts One and Three were charged under 18 U.S.C. § 2241(c) for crossing state lines with intent to engage in a sexual act with a minor under the age of twelve, and Counts Two, Four, and Five were charged under 18 U.S.C. § 2423(a) for transportation of a minor with intent to engage in sexual activity.

Before trial, the government filed a notice of intent to introduce the video recording from K.W.’s 2013 forensic interview with Detective Prescott under Federal Rule of Evidence 807, the residual exception to hearsay. In its notice, the government identified three recorded statements for admission at trial: (1) K.W.’s statement regarding McFadden’s assaults; (2) K.W.’s statement about the melatonin that McFadden gave him; and (3) K.W.’s statement that, one time, he saw McFadden touch J.W. under a blanket. The government argued that these statements met Rule 807’s requirements because the statements were “supported by sufficient guarantees of trustworthiness” and were “more probative,” given that K.W.’s 2013 statements were made closer in time to the alleged assaults. ROA vol. I, at 157–59. McFadden objected to the government’s notice. The district court then held a pretrial evidentiary hearing, in part to assess the 2013 video’s admissibility under Rule 807. After hearing arguments from both sides, the court reserved its ruling on the Rule 807 issue until trial.

7

At trial, the government examined K.W. about the alleged assaults.

Asking whether McFadden had penetrated K.W. with his penis, K.W. initially said that McFadden had “tried to.” ROA vol. VI, at 182–83. The government zeroed in: “When you say he tried to put his penis in you, did he put his penis into your butt?” Id. at 183. K.W. then answered, “Yes.” K.W. further confirmed that, “Yes,” he felt pressure against his anus from McFadden and that it “hurt a lot” “because [he] was little at the time.” Id. In testifying about the Nebraska semi-truck trip, K.W. said that he “slept right next to Mike” and identified the government’s exhibit (a photo of the semi-truck’s sleeper cab mattress) as the location where McFadden penetrated him. Id. at 187.

After this testimony, the government moved under Rule 807 to admit the 2013 forensic-interview video recording, in which K.W. makes more direct statements to Detective Prescott about McFadden’s penetrative assault. The district court admitted the evidence under Rule 807. The court agreed with the government that Rule 807’s admissibility requirements were satisfied and that “[i]t was apparent that [K.W.] only answered with prompting and he was equivocal about the penetration, which he was not in the video.” Id. at 195. Concluding that “the video evidence is more probative on the point for which it is offered than any other evidence that the Government can obtain through reasonable efforts,” the court admitted the video recording. Id. McFadden objected. Over McFadden’s objection, the jury was shown the 2013 video recording of K.W.’s forensic interview, but the video was not given to the jury

8

to use during deliberations. A transcript of the interview was distributed to the jury as an aid while they watched the video, but the transcript was not given to the jury to reference during deliberations.

After the jury watched the 2013 video recording, K.W. was cross-

examined by defense counsel. During cross, defense counsel used the transcript from K.W.’s 2018 FBI interview to impeach him. Defense counsel then moved to admit the audio recording of K.W.’s 2018 interview with the FBI. The court denied the defense’s motion because, unlike the government, defense counsel had given no notice of his intent to introduce Rule 807 evidence. So the court rejected the 2018 audio recording as inadmissible hearsay.

Over the course of McFadden’s five-day trial, the jury heard testimony from the following witnesses: J.W.; J.W.’s mother; K.W.; K.W.’s mother; S.W. (K.W.’s older brother); two of McFadden’s former bosses from the construction company and trucking company where he worked; the Nebraska police officer who arrested McFadden; Detective Prescott; Sue Goebel, the nurse who performed a SANE (sexual assault nurse examiner) exam on J.W. in March 2013; and a child therapist who offered expert testimony about the psychological impact of sexual abuse on children. Most of this testimony is unchallenged on appeal, except for two instances of alleged vouching, which we discuss later in this opinion.

The jury found McFadden guilty on all counts, and the district court sentenced McFadden to serve concurrent life sentences on each count. The

9

presentence report calculated McFadden’s sentence by splitting the five convicted counts into two groups, one for each child victim. Counts One and Two, based on the abuse of K.W. in Nebraska comprised the first group, and Counts Three, Four, and Five, based on the abuse of J.W. on the trips to Arizona and New Mexico comprised the second group. The base level for each group was 28, according to U.S.S.G. § 2G1.3(a)(3), the guideline for violations of 18 U.S.C. § 2423(a). The base level was then increased by 14 levels under various specific offense characteristics. McFadden challenges one of these increases on appeal, the two-level enhancement imposed under § 2G1.3(b)(2)(B) for “unduly influenc[ing] a minor to engage in prohibited sexual conduct.” McFadden timely appealed his conviction and sentence.

DISCUSSION

McFadden contends that multiple errors committed at trial demand reversal of his conviction and that an error in his Guidelines calculation requires resentencing. We agree with McFadden that the district court erred in admitting the 2013 forensic-interview video recording under Rule 807, but this error was harmless. Because the district court made no other errors, we affirm. I. Rule 807 The residual-hearsay exception under Rule 807 is a catchall to the recognized hearsay exceptions housed in Rules 803 and 804—present sense impressions, excited utterances, and the like. See Idaho v. Wright, 497 U.S. 805, 817 (1990) (explaining that the residual rule “accommodates ad hoc

10

Appellate Case: 23-1089 Document: 010111103175 Date Filed: 08/30/2024 Page: 10

instances” where a statement may not “otherwise fall[] within a recognized hearsay exception” but is “nevertheless . . . sufficiently reliable to be admissible at trial”).

Rule 807 states:

(a) In General. Under the following conditions, a hearsay statement is not excluded by the rule against hearsay even if the statement is not admissible under a hearsay exception in Rule 803 or 804:

(1) the statement is supported by sufficient guarantees of trustworthiness—after considering the totality of circumstances under which it was made and evidence, if any, corroborating the statement; and

(2) it is more probative on the point for which it is offered than any other evidence that the proponent can obtain through reasonable efforts.

(b) Notice. The statement is admissible only if the proponent gives an adverse party reasonable notice of the intent to offer the statement—including its substance and the declarant’s name—so that the party has a fair opportunity to meet it. The notice must be provided in writing before the trial or hearing—

or in any form during the trial or hearing if the court, for good cause, excuses a lack of earlier notice.

Fed. R. Evid. 807 (emphasis added).

“The residual exception ‘should be used only in extraordinary circumstances.’” United States v. Burgess, 99 F.4th 1175, 1183 (10th Cir. 2024) (quoting United States v. Dalton, 918 F.3d 1117, 1133 (10th Cir. 2019)). Such circumstances exist when the court is “satisfied that the evidence offers guarantees of trustworthiness and is material, probative and necessary in the

11

interest of justice.” Id. (quoting Dalton, 918 F.3d at 1133). We interpret the residual exception with “caution” so that it does not “swallow the entirety of the hearsay rule.” United States v. Hammers, 942 F.3d 1001, 1011 (10th Cir. 2019).

McFadden argues that the district court abused its discretion in applying Rule 807 to two pieces of evidence offered at trial: (1) a video recording of K.W.’s 2013 forensic interview with Detective Prescott, introduced by the government; and (2) an audio recording of K.W.’s 2018 interview with the FBI, offered by the defense. Under Rule 807, the district court admitted the 2013 video recording but excluded the 2018 audio recording. McFadden contends both rulings were error.

We review such evidentiary decisions for an abuse of discretion. United States v. Hay, 95 F.4th 1304, 1318 (10th Cir. 2024). We do so recognizing that a legal error constitutes an abuse of discretion per se. United States v. Geddes, 71 F.4th 1206, 1214 (10th Cir. 2023). “[L]egal conclusions about the Federal Rules of Evidence” are subject to de novo review. Hay, 95 F.4th at 1318.

A. The 2013 Video Recording To be admissible under the residual-hearsay exception, the offered evidence must contain “sufficient guarantees of trustworthiness” and be “more probative on the point for which it is offered than any other [reasonably attainable] evidence.” Fed. R. Evid. 807(a). “[T]he party offering the evidence bears” the “heavy burden” of showing that both prongs are met. United States

12

v. Trujillo, 136 F.3d 1388, 1396 (10th Cir. 1998). The government offered the 2013 video recording, so it bore the burden of satisfying Rule 807’s two- pronged admissibility standard. See id.

1. The district court’s ruling on trustworthiness was not error.

In child-sex-abuse cases, there are various factors that courts consider in deciding whether a child victim’s hearsay statement is admissible under Rule 807. See Wright, 497 U.S. at 821. These include spontaneity, consistency between repeated allegations, mental state of the child, use of age-appropriate terminology, and lack of motive to fabricate. United States v. Tome, 61 F.3d 1446, 1452–53 (10th Cir. 1995) (citing Wright, 497 U.S. at 821). Rule 807 also instructs courts to evaluate “the totality of circumstances under which [the statement] was made.” Fed. R. Evid. 807(a)(1). A totality analysis might implicate the training and expertise of the interviewer, the interviewer’s use of open-ended rather than leading questions, the time between the alleged abuse and the hearsay statements, and the declarant child’s age. 3 See Tome, 61 F.3d at

3 After the 2019 amendments to Rule 807 (formerly Rule 803(24)) we should also consider corroborating evidence that bolsters the trustworthiness of a hearsay statement. Fed. R. Evid. 807(a)(1); see Burgess, 99 F.4th at 1184 n.5 (supplanting Tome’s rule “that corroborating evidence d[oes] not bear on the trustworthiness of a hearsay statement” because “the 2019 amendments to Rule 807 . . . specifically require the court to consider corroborating evidence in the trustworthiness enquiry” (cleaned up)). But in this case there is no such evidence. The government argues that S.W.’s testimony “corroborat[ed] [K.W.’s] account of the sleeping arrangements” in the semi-truck and K.W.’s recollection that “(S.W.) did not wake up during the night.” ROA vol. I, at 158.

(footnote continued)

13

1453; see also United States v. Farley, 992 F.2d 1122, 1126 (10th Cir. 1993) (affirming that the hearsay testimony was admissible under Rule 807 because the child spoke to her mother about the alleged assault within a day of the incident, the child was “still suffering pain and distress from the assault” when she spoke about it, and the child’s “youth” favored reliability (citation omitted)). Together, these factors guide the determination for whether the child was “particularly likely to be telling the truth when the statement was made.” Tome, 61 F.3d at 1453 (quoting Wright, 497 U.S. at 822). The goal is to ensure that a child’s hearsay statement admitted under the residual-hearsay exception carries the same “particularized guarantees of trustworthiness” that would attend any other hearsay statement admissible under a “firmly rooted” hearsay exception. Wright, 497 U.S. at 821.

The government’s notice of intent gave several reasons why K.W.’s 2013 video statements were sufficiently trustworthy: (1) the consistency of K.W.’s statements about the assaults; (2) the interview’s proximity in time to the alleged Nebraska assault (about two weeks); (3) K.W.’s use of childlike terminology (i.e., describing penises and anuses as “no-no’s”); (4) K.W.’s

But that testimony doesn’t corroborate the assault, and it doesn’t lend any greater reliability to K.W.’s story. In fact, McFadden argued at trial that S.W.’s testimony credited his version of events because, in McFadden’s view, it’s unrealistic that S.W. could have been sleeping mere feet away from McFadden and the younger brothers and yet not have heard or seen any of signs of abuse. For these reasons, we don’t view S.W.’s testimony as corroborative evidence that must be considered in the trustworthiness inquiry.

14

particular and detailed descriptions of the assaults; (5) K.W.’s lack of motivation to fabricate; and (6) the interview being recorded on video, which allowed the jury to assess K.W.’s credibility. 4 And at the pretrial evidentiary hearing, the government presented testimony from Detective Prescott to explain the trustworthy circumstances under which he conducted K.W.’s forensic interview.

The district court was particularly persuaded by Detective Prescott’s testimony. In issuing its ruling at trial, the court observed that “Detective Prescott went through his experience, his training on forensic interview techniques,” and his “attempt[] to determine that the witness [knew] the difference between a truth and a lie.” ROA vol. VI, at 195. From those observations, the district court concluded that the 2013 video “weighs in favor of a finding of trustworthiness.” Id. The court added that it had “carefully reviewed the video in light of relevant facts . . . under Rule 807 and Idaho v. Wright, 497 U.S. 805.” Id. The court then ruled that the 2013 video was sufficiently trustworthy under Rule 807(a)(1).

4 In its response brief, the government again makes the point that K.W.’s 2013 forensic interview is more trustworthy because it was recorded. Neither Wright nor Tome list the recording of a statement as a factor that bears on its trustworthiness. We agree with other courts that have found this detail unpersuasive. See, e.g., United States v. Bruguier, 961 F.3d 1031, 1033 (8th Cir. 2020) (“Although a recording ensures a declarant’s statement is faithfully reproduced, it provides little assurance that the statement was truthful and reliable when spoken.”).

15

Properly admitting a child’s hearsay statement under the catchall exception requires the district court “not . . . merely to find an absence of evidence that the statement was unreliable,” but to find that the child “was particularly likely to be telling the truth.” Burgess, 99 F.4th at 1184 (citation omitted). Though the court has “leeway” in considering the “appropriate factors” in its Rule 807(a)(1) analysis, it must consider those “factors [that] relate to whether the child declarant was particularly likely to be telling the truth.” Wright, 497 U.S. at 822. Once the court identifies the relevant factors, given the facts and arguments before it in a particular case, the court must then decide whether those factors unequivocally demonstrate the statement’s trustworthiness. See Tome, 61 F.3d at 1453. If the “circumstances surrounding [a child’s] statement . . . are equivocal” as to trustworthiness, then the “statement [is] . . . inadmissible hearsay.” Id.

“[H]earsay determinations are particularly fact and case specific,” so “we afford heightened deference” to the district court’s evidentiary rulings to admit hearsay evidence. United States v. Lovato, 950 F.3d 1337, 1341 (10th Cir. 2020) (quoting Trujillo, 136 F.3d at 1395). Because the district court stated that it considered the “relevant facts . . . under Rule 807 and Idaho v. Wright,” ROA vol. VI, at 195, before it admitted the 2013 video, we take the court at its word, see Lovato, 950 F.3d at 1341. Under this highly deferential review, the district court did not err in determining that the 2013 video carried sufficient guarantees of trustworthiness under Rule 807(a)(1).

16

But we would be remiss not to point out that the district court’s reasoning in ruling on the government’s Rule 807 motion was wanting. Before admitting the video, the court should have made findings on the record for each of the relevant factors under Wright/Tome that it considered in evaluating the video statements’ trustworthiness. The district court is not typically “required to make a finding on the record as to each of the Rule 807 [factors],” but only so “long as the record demonstrates that the district court considered the relevant factors.” United States v. Smith, 591 F.3d 974, 980 (8th Cir. 2010) (emphasis added). And especially “[w]hen a statement is admitted under the catchall, the court should make an on-the-record finding that the requirements [of Rule 807] have been satisfied.” 5 Christopher B. Mueller & Laird C. Kirkpatrick, Federal Evidence § 8:140 (4th ed.) (August 2023 update); see United States v. Palacios, 556 F.2d 1359, 1363 n.7 (5th Cir. 1977) (“The [Senate Judiciary] Committee . . . stated that the special facts and circumstances which lead the trial judge to allow an exception under Rule 803(24) [(now 807)] should be stated in the record.”).

At the pretrial hearing, the court enunciated the cluster of factors pertinent to Rule 807(a)(1)’s trustworthiness inquiry, including spontaneity and use of age-appropriate language, but then declined to address these factors in its ruling on the record. This omission would be more acceptable if the record absolutely supported the court’s decision to admit the video. See Smith, 591 F.3d at 980; see, e.g., Burgess, 99 F.4th at 1184 (recognizing that, even beyond

17

the “relevant” factors that the district court considered in assessing trustworthiness, “other guarantees of trustworthiness” were apparent from the record that additionally supported the court’s ruling). But here, those factors weigh against the video’s trustworthiness, and yet the court offered no analysis or findings on these points. It instead confined its discussion to Detective Prescott’s experience, the nature of his interview techniques, and his efforts “to determine that the witness did know the difference between a truth and a lie.” ROA vol. VI, at 195. Based on that, the court found that “K.W.’s statement [wa]s supported by sufficient indicia of reliability such as to be admissible with respect to the first prong of the Rule 807 inquiry.” Id. at 195–96. Our concern is that most forensic interviews will likely be conducted by qualified law- enforcement agents, trained to ask the right questions in the right way so as to render every forensic interview inherently trustworthy by the district court’s standards. The court focused almost singularly on these elements of K.W.’s 2013 interview, without paying equal mind to the factors that diluted its trustworthiness. See, e.g., Tome, 61 F.3d at 1453 (weighing equally the factors that supported and the factors that undermined the hearsay statement’s trustworthiness). From our perspective, several factors potentially subverted the trustworthiness of the 2013 video statements, which the district court never addressed on the record.

First, K.W.’s statements to Detective Prescott were not spontaneous.

K.W. knew that the purpose of the interview was to talk about McFadden’s

18

inappropriate behavior with children. And unlike the child victim in Burgess, K.W. never made any spontaneous allegations about McFadden’s abuse before his interview. See 99 F.4th at 1184 (affirming the admissibility of Rule 807 evidence partly because the child’s recorded statements were consistent with spontaneous admissions she had made before the forensic interview). When a child is brought into a forensic interview knowing that the purpose of the interview is to discuss alleged abuse, this court has concluded that those interview statements lack spontaneity. See Tome, 61 F.3d at 1453 (noting that the child’s statement to a caseworker wasn’t spontaneous because the child knew the purpose of the meeting was to talk about “what defendant had done to her”). And the government doesn’t contest that K.W.’s 2013 statements lacked spontaneity.

Second, though K.W. used some age-appropriate terminology in his descriptions of McFadden’s assaults, he also used sophisticated language beyond his years—for instance, stating that McFadden has “a disease and it makes him like little children” and that McFadden was “overdosing” the children with melatonin. ROA vol. I, at 171, 193. Not only does some of K.W.’s language suggest adult intervention, but the record substantiates that K.W. spoke with several adults about McFadden before his interview with Detective Prescott, including his mother and a child counselor. This type of intervention could undermine the trustworthiness of a child’s out-of-court statements. See United States v. Barrett, 8 F.3d 1296, 1300 (8th Cir. 1993)

19

(instructing the district court to consider on remand whether evidence of “prior interrogation, prompting, or manipulation by adults” of the child victim impacted the spontaneity of the child’s admission of abuse, making spontaneity “an inaccurate indicator of trustworthiness” (citation omitted)). More still, the taint of adult influence was McFadden’s principal argument against the video’s admission in his objection to the government’s notice of intent and at the pretrial evidentiary hearing, yet the court didn’t address this point in ruling on the motion at trial. We can’t know for sure which aspects of K.W.’s video statements were genuine and which, if any, were the product of outside adult influence. But it was the district court’s job to grapple with that possibility on the record as it “relate[d] to whether [K.W.] was particularly likely to be telling the truth.” Wright, 497 U.S. at 822.

Third, the government stated in its notice of intent that the consistency between K.W.’s 2013 interview statements and his state-trial testimony make the video statements more reliable. But that analysis mistakes how we apply the consistency factor under Wright/Tome. In United States v. Harrison, we found that a child’s accusations about her abusive step-father were “consistent” because the child’s three separate allegations (two to law enforcement officers and one to a doctor) after the incident all conveyed the same story. 296 F.3d 994, 996–99 (10th Cir. 2002). So too in Burgess, we deemed the child’s statements “consistent” because her spontaneous admissions about the abuse to a trusted adult matched the statements she made to her mother the next day, as

20

well as the statements she made in a forensic interview conducted three days later. 99 F.4th at 1179, 1184. In both cases, we limited our consideration of the child’s consistency to the time “when the [hearsay] statement was made.” Id. (quoting Tome, 61 F.3d at 1453).

Contrary to the government’s theory, K.W.’s consistency between his 2013 forensic interview and his 2015 state-trial testimony is irrelevant. Testimony that K.W. gave two years after the interview has no bearing on the circumstances that existed when K.W. made his accusations in 2013. See Tome, 61 F.3d at 1453. For that matter, the record contains no other accusatory statements from K.W. in the days around his 2013 interview from which we might gauge the consistency of his accusations against McFadden. K.W. and his mother both testified that he did not accuse McFadden of assault on the ride back from Nebraska, and any statements K.W. made to the child counselor the day before his forensic interview are not in the record. So we cannot say whether K.W. was consistent or inconsistent when he accused McFadden in January 2013. And without any reasoning from the court explicitly addressing Tome’s consistency factor, it’s unclear whether the court accepted the government’s erroneous arguments when it decided to admit the video. If it did, then that acceptance would have been error.

We recognize that many factors favor the video’s trustworthiness—

K.W.’s lack of motive to fabricate; Detective Prescott’s training and interview style; K.W.’s detailed statements about the abuse; and the short time gap (two

21

weeks) between the last alleged abuse and the forensic interview—though the district court failed to mention most of these in its ruling. This balance of factors allows us to imagine how the district court arrived at its conclusion to find the video sufficiently trustworthy. But the absence of a record weighing the “appropriate factors” in this case hampers our ability on appeal to determine whether Rule 807 was correctly applied. Wright, 497 U.S. at 822.

Tome doesn’t command merely that a majority of the factors signal trustworthiness. Its standard is even higher: Tome instructs that the circumstances surrounding the offered hearsay statement must be “[un]equivocal” such that its trustworthiness is “guarantee[d].” 61 F.3d at 1453. “[A] suggestion of trustworthiness cannot suffice.” Harrison, 296 F.3d at 1006. Our intolerance for equivocality in this arena echoes the extreme caution with which we apply Rule 807. The district court’s perfunctory statement about having considered the “relevant facts” under Rule 807 and Wright leaves us doubtful that the appropriate level of caution was exercised here. ROA vol. VI, at 195. Going forward, district courts assessing child hearsay statements in sex- abuse cases would do well to articulate their reasoning on the record for each of the pertinent trustworthiness factors before admitting out-of-court statements under Rule 807’s catchall exception to the rule against hearsay.

Regardless, we resolve next that the 2013 video was not more probative on the point for which it was offered compared to other reasonably attainable

22

evidence. So we conclude that the court’s admission of the hearsay evidence was error on that ground.

2. The 2013 video was not more probative on the point for which it was offered.

Rule 807’s second prong requires that an admissible residual-hearsay statement be “more probative on the point for which it is offered than any other evidence that the proponent can obtain through reasonable efforts.” Fed. R. Evid. 807(a)(2). We begin with the point for which the government offered the 2013 video evidence.

In its notice of intent, the government asserted that the 2013 video was more probative on three points: (1) McFadden’s penetrative assault of K.W.; (2) McFadden’s administering melatonin to K.W.; and (3) K.W.’s observing McFadden assault J.W. under a blanket. Then, at the pretrial evidentiary hearing, the government argued that because K.W. had “professed either a lack of memory of the incident or . . . a very strong desire to not talk about the incident” when he was interviewed in 2018, the government anticipated that K.W. would not be “present at trial” to testify. Suppl. ROA vol. I, at 36. The suggestion being that the 2013 video would provide the most probative, if not the only, available evidence as to all the points the government had raised in its notice of intent.

But against the government’s expectations, K.W. did testify at trial and did recall the pertinent events. So when the government moved at trial to admit

23

the video recording, the court asked: “In light of the fact that [K.W.] testified that he did remember the truck assault, what is the purpose of the video?” ROA vol. VI, at 193. The government answered that the video “remains the most probative evidence on that point”—“that point” being the “truck assault”— because “K.W., while he did discuss the events in question, did so reluctantly and often after . . . prompting.” Id. Most concerning to the government was K.W.’s being “pretty equivocal about penetration,” which he was not in the video. Id. at 194. The government added that the video “is closer in time to the events in question.” Id. at 193. Defense counsel objected because K.W. “was able to recall the events” and therefore that there was no purpose for admitting the video. Id. The district court agreed with the government that K.W.’s trial testimony was “reluctant” and “equivocal about the penetration, which [K.W.] was not in the video.” Id. at 195. On that basis, the court ruled that the 2013 video was “more probative on the point for which it is offered than any other evidence that the Government can obtain through reasonable efforts.” Id.

The government maintains on appeal that the 2013 video was the more probative, reasonably available evidence to prove two points: anal penetration and K.W. observing McFadden jerk his hand away from J.W.’s penis under a blanket. 5

5 We decline to address the government’s third argument from the notice of intent—that the video was more probative proof of McFadden’s administering melatonin to K.W. After the notice of intent was filed, the (footnote continued)

24

We can quickly dispense with the second point. The government argues that the video was more probative of whether K.W. witnessed McFadden touching J.W. inappropriately because, when the prosecutor asked K.W. at trial “whether K.W. saw ‘McFadden do anything to anyone else that made [him] uncomfortable,’ K.W. answered: ‘I did not, no.’” Resp. Br. at 30 (quoting ROA vol. VI, at 189). A vague question about K.W.’s discomfort was a feeble attempt to elicit testimony about one specific incident of alleged assault from a man who experienced years of abuse throughout his childhood. Rule 807(a)(2) requires the proponent of the hearsay evidence to exercise “reasonable efforts” to procure alternative evidence before resorting to the catchall rule. K.W. was an available witness, and yet the government did not ask him directly about the incident to which it now refers. This is not the sort of reasonable effort that we expect from the government to obtain nonhearsay evidence. Cf. United States v. Nucera, 67 F.4th 146, 171 (3d Cir. 2023) (affirming the district court’s decision to exclude an out-of-court statement, in part, because the declarant “was available to testify” and yet not called as a witness, “and his in-court testimony

government never again raised that ground for probativeness before the district court, neither at the pretrial evidentiary hearing nor at trial, and the government has explicitly abandoned that point in its briefing to this court on appeal. Regardless, we would not find that argument persuasive of the video’s admissibility under Rule 807(a)(2) because K.W. testified at trial that McFadden regularly supplied him and the other children with melatonin. His live testimony about the melatonin is more probative than the out-of-court video statements. See United States v. W.B., 452 F.3d 1002, 1005–06 (8th Cir. 2006) (establishing that a child victim’s in-court testimony is “generally more probative” than hearsay statements).

25

. . . would have been more probative on th[e] point”); see also Mueller & Kirkpatrick, Federal Evidence § 8:142 (explaining that “more effort is expected of the government than of the accused” in demonstrating reasonable efforts under Rule 807(a)(2) and “more effort is expected if other evidence”—like live testimony—“would be superior to a statement offered under the catchall”).

As to the point about anal penetration, K.W. testified sufficiently to this at trial. On direct examination, the government asked K.W. whether McFadden penetrated him, and K.W. responded that McFadden had “tried to.” ROA vol. VI, at 183. The government clarified whether “tried to” meant that “[McFadden] tried to put his penis in [K.W.],” specifically whether McFadden “put his penis into [K.W.’s] butt.” Id. To that, K.W. answered, “Yes.” Id. K.W. added that he felt pressure against his anus from McFadden’s penetration which “hurt a lot” “because [he] was little at the time.” Id.

It is axiomatic that the Federal Rules of Evidence favor live testimony over statements made outside the courtroom. See Harrison, 296 F.3d at 1007. This is especially true “[w]hen the ‘key factual issues’ at trial turn on the ‘credibility’ and ‘demeanor’” of the witnesses. Garcia-Martinez v. City & Cnty. of Denver, 392 F.3d 1187, 1191 (10th Cir. 2004). Even in child-sex-abuse cases, where “exceptional circumstances generally exist” to admit residual- hearsay evidence, “a child’s in-court statements are generally more probative than a child’s out-of-court statements.” United States v. W.B., 452 F.3d 1002, 1005–06 (8th Cir. 2006). Simply because a case involves child-sex abuse does

26

not make Rule 807’s probativeness prong a guarantee, particularly when the victim has testified at trial. Cf. United States v. Balfany, 965 F.2d 575, 582 (8th Cir. 1992) (determining that, even though an adult’s hearsay testimony “included some facts that the testimony of [the child] and the other witnesses did not,” the hearsay statements were still less probative than the child’s live trial testimony which was “very comprehensive”).

K.W. and J.W.’s credibility shoulders the government’s case because no one else bore witness to the alleged assaults. In this he said/he said case, the jury’s ability to assess K.W.’s credibility from live testimony was paramount. See Garcia-Martinez, 392 F.3d at 1191–92. Because the 2013 video statements are not superior to K.W.’s live testimony, there was no justification to override our preference for in-court testimony. See 2 McCormick on Evidence § 324 n.39 (8th ed.) (July 2022 update) (explaining circumstances when an “out-of- court statement will be superior” to live testimony, including when a witness’s trial testimony is “incomplete[]”).

Our most recent dispatch on Rule 807 and child hearsay statements offers a helpful contrast. In Burgess, we ruled that the video recording of a child’s forensic interview was admissible under Rule 807 because the recorded statements were more probative than the child’s trial testimony on a determinative issue of fact. 99 F.4th at 1185–87. Burgess contained nearly identical facts to the ones we confront now: a child accused a trusted adult of illegal sexual contact, the child testified to this contact at the accused’s

27

criminal trial, and the government additionally sought to introduce the video recording of the child’s forensic interview under Rule 807. See id. 1179–82. But Burgess contained one key difference—the child’s testimony differed from the original accusation that she made during her forensic interview. See id. at 1182. In the interview, the child alleged that the defendant committed different sexual acts than those she testified to at trial. Id. at 1186. Given this inconsistency, we affirmed the trial judge’s “discretion to permit hearsay testimony” to “determin[e] that the recorded interview was ‘more probative on the point for which it [wa]s offered’—what sex acts [the defendant] committed against [the child].” Id. at 1187. We found support for this decision in several other circuit cases where courts have deemed out-of-court statements to be admissible residual hearsay, even though the child testified at trial, because the child’s testimony was inconsistent or unclear compared to the hearsay evidence. See id. at 1186–87 (citing Harrison, 296 F.3d at 995–96, 1000, 1003– 07; W.B., 452 F.3d at 1004; United States v. Peneaux, 432 F.3d 882, 887, 891– 93 (8th Cir. 2005); United States v. Wandahsega, 924 F.3d 868, 874, 881–82 (6th Cir. 2019)).

In contrast, K.W.’s trial testimony reflects his original allegation from the 2013 video accusing McFadden of anal penetration. When asked directly at trial whether McFadden inserted his penis into K.W.’s anus, K.W. responded, “Yes.” ROA vol. VI, at 183. And he even elaborated that the act “hurt a lot” due to his age and size. Id. K.W. also identified the government’s exhibit—a

28

photo of the sleeper cab’s mattress—as the location where McFadden penetrated him. Though we generally afford the district court “a deferential standard of review” due to the court’s ability “to watch and listen to the witnesses as they testified,” we disagree that the 2013 video was more probative than K.W.’s live testimony about the alleged incident in the semi- truck. Burgess, 99 F.4th at 1187 (quoting Peneaux, 432 F.3d at 893).

The government insists that K.W.’s trial testimony was ambivalent, making the 2013 interview more probative and necessary to prove penetration. McFadden responds that, even if the trial evidence weren’t probative enough to show penetration—though he believes it was—a conviction under 18 U.S.C. § 2241(c) does not necessarily require proof of penetration. The government acknowledges this but adds that, without penetration, the government would have “had to prove the additional element,” Resp. Br. at 30, of intention to “abuse, humiliate, harass, degrade, or arouse or gratify the sexual desire” of K.W, 18 U.S.C. § 2246(2) (providing four definitions for a “sexual act” punishable under 18 U.S.C. § 2241(c)). The government maintains that it was not obliged to add to its burden in this way. Even if we accept that the government needed to prove penetration as an element of the charged crime, the forensic-interview video was not more probative for that purpose. We disagree with the government that K.W.’s initial statement that McFadden had tried to penetrate him made his testimony unclear as to whether penetration occurred. In a follow-up answer, K.W. immediately explained what he meant.

29

A victim’s tepid trial testimony does not justify admitting more zealous, yet consistent, statements under the residual-hearsay exception. Were that true, Rule 807’s purportedly stringent admissibility standard would wilt. See Dalton, 918 F.3d at 1133 (reserving Rule 807 hearsay admissions for “extraordinary circumstances”). The government’s theory that K.W.’s ambivalence, alone, warranted the 2013 video’s admission would allow the residual exception to “swallow . . . the hearsay rule.” Hammers, 942 F.3d at 1011. To admit forensic- interview recordings even when victims testify at trial and offer testimony consistent with their original allegations would be to risk creating a per se rule that such recordings are always admissible in child-sex-abuse cases. We cannot condone such a potentially far-reaching result.

Because the government failed to satisfy Rule 807’s admissibility standard for residual-hearsay evidence, the district court’s decision to admit the 2013 video recording constituted legal error and therefore an abuse of discretion per se. Geddes, 71 F.4th at 1214.

B. The 2018 Audio Recording K.W. was interviewed by the FBI in 2018 during his detention at a juvenile correction center. The audio of that interview was recorded and a transcript issued. At trial, defense counsel impeached K.W. with statements he had made during the 2018 interview. Defense counsel reminded K.W. of his 2018 statement that he didn’t remember the incident with “Mike in the truck.” ROA vol. VI, at 207. K.W. explained that he “did say that, but the reason why

30

[he] said that is [he] d[id]n’t like talking about this.” Id. at 207–08. Defense counsel recited several more statements from the 2018 interview, which K.W. confirmed having made. In going through this line of impeachment questioning, defense counsel touched on some inconsistencies between K.W.’s 2018 interview and his trial testimony. For example, defense counsel noted K.W.’s 2018 statement that, at the time, he and J.W. were still close friends, compared to K.W.’s trial testimony that he “had stopped talking to J.W. after this incident” with McFadden. Id. at 209. And there were some other 2018 statements that K.W. testified he could not recall having made.

After finishing his cross-examination of K.W., defense counsel moved to admit the audio recording of K.W.’s 2018 interview. Defense counsel sought to admit the audio recording partly because he believed that the government had introduced the 2013 forensic-interview video based on K.W.’s 2018 statements that he couldn’t remember the Nebraska trip. So, “[i]n fairness,” defense counsel advanced, “the [audio] recording that gave rise to the video should be admitted.” Id. at 212. The government replied that “the [2018 audio] recording [wa]s not the reason that [the 2013 forensic-interview video] was admitted.” Id. The government posited that defense counsel was trying to “impeach [K.W.] based on that [2018] transcript, which [defense counsel] had already done and c[ould] continue to do if he fe[lt] he ha[d]n’t done it adequately.” Id. at 213. But the government maintained that the defense’s impeachment of K.W. did not require admitting the 2018 audio recording.

31

The district court construed McFadden’s submission of the 2018 audio recording as a Rule 807 motion, and so it excluded the recording on the ground that defense counsel had failed to give the requisite notice under Rule 807(b). Defense counsel explained that he was unable to give notice because he “did not plan on asking for the admission of [the audio recording]” based on his “assum[ption] that the [2013] video was not going to be admitted.” Id. The defense’s position remained that K.W.’s lack of memory made him an unavailable witness, so “under the circumstances” the audio recording should be admissible. The court dismissed this argument and rebuked the defense for failing to “follow the steps that needed to be followed in order to have an 807.” Id. at 214. Because defense counsel failed to follow the notice standard under Rule 807(b), the court excluded the audio recording.

Before us, McFadden argues that two rules of evidence foreclose the district court’s ruling: Rule 807(b) and Rule 613(b). McFadden also maintains his general appeal to fairness advancing that the district court’s evidentiary ruling deprived him of his constitutional right to “present a complete defense.” Op. Br. at 28 (quoting Crane v. Kentucky, 476 U.S. 683, 690 (1986)).

“[W]e review [the district court’s] legal interpretation of the Federal Rules of Evidence de novo and its application of the rules for abuse of discretion.” United States v. Armajo, 38 F.4th 80, 84 (10th Cir. 2022). Under this standard, we discern no error in the district court’s exclusion of the 2018 audio recording.

32

1. The district court did not abuse its discretion by enforcing Rule 807(b)’s notice requirement.

McFadden did not move expressly under Rule 807 to admit the 2018 audio recording, but the district court nevertheless ruled on 807 grounds. On appeal, McFadden contends that even “[a]ssuming arguendo that [Rule] 807 governed [his] request to admit K.W.’s 2018 interview,” the district court erred by excluding the recording under the rule’s notice requirement. 6 Op. Br. at 33.

Rule 807(b) provides that evidence offered under the residual exception to hearsay must be introduced with “notice” to the opposing party, either before or during trial, unless the court chooses to excuse lack of notice for “good cause.” McFadden disputes the district court’s “unyielding adherence” to Rule 807(b)’s notice requirement and offers four reasons why “good cause” existed to excuse his lack of notice: (1) the “2018 recording was created by the government”; (2) “K.W. could not remember making the inconsistent statements”; (3) the “government asked K.W. rehabilitating questions after defense counsel questioned him about the recording”; and (4) the “2018 recording became especially relevant once the court admitted K.W.’s forensic interview.” Op. Br. at 34.

None of these reasons demonstrate good cause or persuade us that the district court clearly erred in finding no good cause for McFadden’s lack of

6 The government agrees that the Rule 807 notice issue for the 2018 audio recording is preserved.

33

notice. The committee notes to Rule 807 advise that the good-cause exception might apply when “the proponent may not become aware of the existence of the hearsay statement until after the trial begins,” or “the proponent may plan to call a witness who without warning becomes unavailable during trial.” Fed. R. Evid. 807 advisory committee’s note (2019 Amend.). McFadden certainly knew about the existence of the 2018 recording before trial. Indeed, he argued before the district court that he did not make a pretrial motion for its admission because he “assumed that the [2013] video was not going to be admitted into evidence since [K.W.] was going to be able to testify.” ROA vol. VI, at 213. That assumption was unreasonable. Caselaw predating McFadden’s trial should have alerted the defense that hearsay statements are sometimes admissible even when a witness is available to testify at trial, especially in child-sex-abuse cases. See, e.g., Harrison, 296 F.3d at 1007.

McFadden further states that the district court was “permitted,” and so not required, to waive Rule 807’s notice requirement. Op. Br. at 34. Plenty of courts opt to strictly enforce the notice provision. See, e.g., Burgess v. Goldstein, 997 F.3d 541, 561 (4th Cir. 2021) (reversing the district court’s admission of hearsay evidence because the movant did not satisfy the rule’s requirement to give notice including the declarant’s name and address); Rotolo v. Digital Equip. Corp., 150 F.3d 223, 224–25 (2d Cir. 1998) (reasoning a videotape that the plaintiff attempted to admit without notice because it was publicly available failed Rule 807’s notice requirement and thus was

34

inadmissible under the rule). We see no reason to intrude on the district court’s well-exercised discretion not to excuse McFadden’s failure to give the government notice of his intent to introduce the 2018 audio recording, as Rule 807(b) requires. See Burgess, 99 F.4th at 1183 (noting this court’s deference to the district court on Rule 807 rulings). And even though McFadden’s motion to admit the 2018 recording was not made directly under Rule 807, when the court openly treated it as a Rule 807 motion, the defense made no effort to correct the record. We cannot fault the court for ruling on Rule 807 notice grounds when the defense’s motion might reasonably have been construed under Rule 807 and counsel never supplied an alternate rule of evidence to support the recording’s admissibility. For all these reasons, the district court’s exclusion of the 2018 audio recording under Rule 807(b) was proper.

2. The arguments presented under Rule 613(b) are waived.

Rule 613(b) allows for the admission of “[e]xtrinsic evidence of a witness’s prior inconsistent statement . . . if the witness is given an opportunity to explain or deny the statement and an adverse party is given an opportunity to examine the witness about it, or if justice so requires.” McFadden argues that this rule entitled him to admit the 2018 audio recording as extrinsic evidence of the inconsistency between K.W.’s 2018 statements and his trial testimony. The government responds that this argument is unpreserved for appellate review. We agree with the government.

35

Appellate Case: 23-1089 Document: 010111103175 Date Filed: 08/30/2024 Page: 35

At trial, defense counsel never cited Rule 613(b) as grounds to admit the 2018 recording. And the colloquy between the court and counsel reveals that the court understood McFadden’s motion as one under Rule 807. McFadden never corrected this misunderstanding—if, indeed, that’s what it was—and so any alternative argument that McFadden might have had under Rule 613(b) was not “apparent from the context.” Burke v. Regalado, 935 F.3d 960, 1014 (10th Cir. 2019); see United States v. Roach, 896 F.3d 1185, 1191 (10th Cir. 2018) (“To preserve an objection to the exclusion of evidence for appeal, the proponent must make an offer of proof at trial, first, describing the evidence and what it tends to show and, second, identifying the grounds for admitting the evidence.” (cleaned up)). When the district court denied the defense’s motion on notice grounds under Rule 807(b), counsel responded that K.W. was “unavailable” because he “was unable to remember” certain statements from 2018. ROA vol. VI, at 213. Citing a witness’s purported unavailability as justification for introducing hearsay evidence is not an obvious invocation of Rule 613(b)—if anything, this is more suggestive of Rule 804(b). See Fed. R. Evid. 804(b) (providing an exception to the rule against hearsay for former testimony given by an unavailable witness). And when the district court asked defense counsel his purpose for offering the audio recording, counsel responded that “[i]n fairness, . . . the recording that gave rise to the 2013 [forensic- interview] video should be admitted.” ROA vol. VI, at 212. Nowhere in this

36

explanation does defense counsel reference impeachment, which must be the purpose of evidence offered under Rule 613(b). 7 McFadden attempts to connect his “fairness” assertion with Rule 613(b)’s provision to allow extrinsic evidence of prior inconsistent statements when “justice so requires.” Reply Br. at 14. This connection is too attenuated, and it comes far too late. We cannot see how McFadden’s broad appeal to “fairness” would have alerted the district court that the defense was making a Rule 613(b) motion under the rule’s justice-so-requires prong, without any mention of the rule or the defense’s intent to use the audio recording for impeachment purposes. See Burke, 935 F.3d at 1014; United States v. Cates, 73 F.4th 795, 809 (10th Cir. 2023) (stating that, though “a party need not ‘use any particular language . . . to properly preserve an issue for appeal,’” the party must “br[ing] [it] to the court’s attention” (quoting Holguin-Hernandez v. United States, 589

7 McFadden’s failure to cite impeachment as the purpose for offering the 2018 audio recording undermines his reliance on United States v. White, where we affirmed the district court’s ruling to admit testimony for impeachment purposes under Rule 613(b). 68 F. App’x 870, 874 (10th Cir. 2003) (unpublished). In White, the district court initially denied the government’s offered testimony on hearsay grounds, but then reversed that ruling when the government explained its intent to use the testimony purely for impeachment purposes. Id. at 873. McFadden gave the district court no similar explanation. We consider United States v. Mitchell, 113 F.3d 1528 (10th Cir. 1997), abrogated on other grounds by United States v. Shipp, 589 F.3d 1084 (10th Cir. 2009), inapposite for the same reason. Mitchell is the one instance where this court has concluded that a district court’s failure to admit testimony under Rule 613(b) constituted an abuse of discretion. Id. at 1532. But in Mitchell, both parties and the court all understood that the evidentiary rule at issue was Rule 613(b), which was not the case here. Id.

37

U.S. 169, 174 (2020))). Because McFadden did not preserve his Rule 613(b) arguments at trial and he has not argued for plain-error review on appeal, his claim of error under Rule 613(b) necessarily fails. United States v. Martinez, 92 F.4th 1213, 1238 n.6 (10th Cir. 2024) (recognizing that an argument “not first presented to the district court” hits “the end of the road” when the defendant “fail[s] to argue for plain error” on appeal).

3. The 2018 audio recording’s exclusion did not violate McFadden’s constitutional right to present a defense.

McFadden squeezes his fairness argument for all the juice he can: he contends next that this argument also raised a constitutional claim that the district court’s ruling denied him his due process right to “a meaningful opportunity to present a complete defense.” Op. Br. at 28 (quoting Crane, 476 U.S. at 690). The government concedes that McFadden’s fairness argument is preserved for our review. So we accept that McFadden’s constitutional arguments are properly before us. Even so, they are meritless.

McFadden insists that by excluding the 2018 audio recording the district court “violated [his] right to present a defense and confront the witnesses against him.” Reply Br. at 22. The Fifth and Sixth Amendments guarantee a defendant “the right to present a defense,” which includes “the right to testify, present witnesses in his own defense, and cross-examine witnesses against him.” United States v. Tapaha, 891 F.3d 900, 905 (10th Cir. 2018) (cleaned up). This right, though “a keystone of our legal system,” is “not absolute”

38

because sometimes it “must bow to accommodate legitimate, competing interests in the trial process.” United States v. Rivas-Macias, 537 F.3d 1271, 1277–78 (10th Cir. 2008) (citation omitted). One such competing interest is adherence to the Federal Rules of Evidence. United States v. Williams, 934 F.3d 1122, 1131 (10th Cir. 2019). “Thus, an evidentiary ruling infringes a defendant’s due process rights” to present a defense “only if the district court violates the Federal Rules of Evidence.” United States v. Oldbear, 568 F.3d 814, 820 (10th Cir. 2009). If the court does violate the rules, we then consider whether “the excluded evidence was of such an exculpatory nature that its exclusion affected the trial’s outcome.” Williams, 934 F.3d at 1131 (quoting Tapaha, 891 F.3d at 905).

We have already established that the district court acted within its discretion by excluding the 2018 audio recording under Rule 807(b). And even if McFadden’s Rule 613(b) arguments were preserved, he contends that the recording was admissible under the rule’s “if justice so requires” prong. Fed. R. Evid. 613(b). The decision to admit extrinsic evidence in the interest of justice is a discretionary determination for the district court to make, and here it was never given the opportunity to do so. See id. advisory committee’s note (2024 Amend.). The district court did not abuse its discretion by declining to admit evidence on grounds that McFadden never brought to its attention. See Cates, 73 F.4th at 809. The district court is capable but not clairvoyant.

39

Our analysis of McFadden’s constitutional claim stops there. See id.;

United States v. Serrano, 406 F.3d 1208, 1215 (10th Cir. 2005) (affirming that “the accused does not have an unfettered right to offer [evidence] . . . otherwise inadmissible under [the] standard rules” (cleaned up)). McFadden was not deprived of his right to present a defense at trial because the district court’s exclusion of the 2018 recording was not an abuse of discretion under the Federal Rules of Evidence. Williams, 934 F.3d at 1131.

C. Harmless Error Because the district court erred in admitting the 2013 forensic-interview video under the residual-hearsay exception, we next consider whether that error was harmless. United States v. Chavez, 976 F.3d 1178, 1204 (10th Cir. 2020). McFadden challenges the video’s admissibility under Rule 807, so we apply the nonconstitutional harmless-error standard. 8 See United States v. Blechman, 657 F.3d 1052, 1067 (10th Cir. 2011) (“[W]hen a defendant objects to a district court’s admission of hearsay based solely on the Federal Rules of Evidence, we apply the nonconstitutional harmless error standard.” (cleaned up)). Under this standard, an error is harmless “unless a substantial right of a party is affected.”

8 McFadden contends that the stricter constitutional harmless-error standard applies to his Rule 807 objection. He thinks so because he claims that the video’s admission violated his due process right to receive a “fair trial.” Op. Br. at 14. But the constitutional harmless-error standard applies to hearsay objections when the defendant claims that the admission of the out-of-court statement violated his rights under the Confrontation Clause. See United States v. Blechman, 657 F.3d 1052, 1067 n.15 (10th Cir. 2011). McFadden makes no such claim.

40

Chavez, 976 F.3d at 1204 (alterations omitted) (quoting United States v. Charley, 189 F.3d 1251, 1270 (10th Cir. 1999)). The government bears the burden to prove harmlessness. United States v. Ledford, 443 F.3d 702, 712 (10th Cir. 2005).

Yet here, the government’s response brief fails to assert or even address harmlessness as it relates to the 2013 video recording. When the government waives harmless error, “this court may in its discretion ‘initiate harmless error review in an appropriate case.’” United States v. Samaniego, 187 F.3d 1222, 1224 (10th Cir. 1999) (quoting United States v. Torrez-Ortega, 184 F.3d 1128, 1136 (10th Cir. 1999)). In deciding whether to proceed with a discretionary harmless-error review, we consider “(1) the length and complexity of the record; (2) whether the harmlessness of the errors is certain or debatable; and (3) whether a reversal would result in protracted, costly, and futile proceedings in the district court.” 9 United States v. Holly, 488 F.3d 1298, 1308 (10th Cir. 2007). These factors favor exercising our discretion to reach harmless-error review in this case.

9 This court has remarked that it’s “unclear what the third factor contributes to a court’s decision in determining whether” to “address the issue of harmlessness” “sua sponte.” Samaniego, 187 F.3d at 1225 n.2. This court has declined to apply it before. See Torrez-Ortega, 184 F.3d at 1136–37. We agree that the third factor adds little to the discretionary-review calculus for harmlessness because “cost and potentially protracted proceedings cannot preclude reversal if an error was not harmless.” Id. So we base our decision on the first two factors, only.

41

First, the record here is not especially complicated—the trial lasted less than a week and involved one defendant and two key witnesses. Compare United States v. Doe, 572 F.3d 1162, 1175 (10th Cir. 2009) (proceeding to harmless-error review because “[t]he record in this case is not lengthy or complex” and both defendants’ trials “lasted less than one week”), with Torrez- Ortega, 184 F.3d at 1136 (declining to exercise discretion to address harmless error when the record was “extensive and complex,” because it included “twenty-five volumes cover[ing] a two-week, multi-defendant trial”). Plus, all the criminal counts in McFadden’s conviction are at issue on appeal, thus we are not left with a “difficult task of determining which portions of the transcript” are most relevant. Holly, 488 F.3d at 1308.

Second, and most importantly, the question of harmlessness in this case is certain. See United States v. Little, 829 F.3d 1177, 1189 (10th Cir. 2016) (Holmes, J., dissenting) (noting that “the certainty of harmlessness”—the second factor—is “[t]he most important” and “can be dispositive under certain circumstances”). Exercising our discretion to conduct harmless-error review, even when the government has failed to brief the issue, is appropriate “where the certainty of the harmlessness is readily apparent.” Holly, 488 F.3d at 1308; see United States v. Silver, 954 F.3d 455, 459 (2d Cir. 2020) (explaining that, by opting to address harmless error “without the benefit of briefing,” courts of appeal may avoid becoming “locked into interminable cycles of remand or

42

requests for additional briefing from the parties”). We are convinced that is the case here.

The video recording is harmless partly for the same reason that it is not “more probative” under Rule 807(a)(2): K.W.’s video statements from 2013 mirror his trial testimony. See Torrez-Ortega, 184 F.3d at 1135 (considering under a harmless-error analysis “whether the testimony was cumulative” (citation omitted)); 21 Charles A. Wright & Arthur R. Miller, Fed. Prac. & Proc. Evid. § 5035.2 (2d ed.) (June 2024 update) (noting that “improperly admitted” evidence “will usually be harmless” “[i]f the . . . evidence was simply cumulative of properly admitted evidence”); see also Untied States v. Otuonye, 995 F.3d 1191, 1208 (10th Cir. 2021) (“Cumulative evidence is defined as evidence which goes to prove what has already been established by other evidence.” (citation omitted)). Everything that eleven-year-old K.W. says on the 2013 video recording was elicited at trial through properly admitted testimony: the sleeping arrangements in the semi-truck during the Nebraska trip, McFadden’s habit of giving the boys melatonin, McFadden’s previous inappropriate touching of K.W. at his Colorado homes, and McFadden’s efforts to make his home an appealing place for the boys to come play and sleep over.

Additionally convincing are that J.W.’s testimony corroborated K.W.’s regarding McFadden’s patterns of abuse, that K.W. was old enough to understand “the true nature of [his] account[],” Charley, 189 F.3d at 1271, and that McFadden had an opportunity to thoroughly cross-examine K.W. using

43

statements from his 2018 FBI interview and his 2015 state-trial testimony to impeach K.W.’s credibility, see United States v. Begay, 937 F.2d 515, 524–25 (10th Cir. 1991) (determining that defendant’s inability to cross-examine child victim in a sex-abuse case was not harmless because that limitation “shut off a line of vital defense evidence” that, had it been allowed, might have produced defendant’s “best evidence of disproving penetration”). The jury also viewed several exhibits displaying photos of K.W. when he was eleven-years old, which the government introduced with no objection from McFadden. So to the extent the video might have “provoke[d] an emotional response in the jury” due to K.W.’s young age, the jury already had a mental image of K.W. as a child to connect with the allegations in his trial testimony. Otuonye, 995 F.3d at 1207. Most of all, when the government introduced the 2013 video at trial, the defense affirmatively declined the opportunity to request any redactions from the thirty-minute-long recording. This, and the above, resolve for us that the district court’s error in admitting the 2013 forensic-interview video under Rule 807 did not “substantial[ly] influence . . . the outcome of the trial.” Blechman, 657 F.3d at 1067. II. Vouching McFadden alleges that two witnesses (Detective Prescott and Nurse Goebel) vouched for K.W.’s and J.W.’s credibility, which deprived him of a fair trial in violation of the Due Process Clause. McFadden concedes that his challenge to Goebel’s testimony is unpreserved. So we review his vouching

44

claim as to Goebel for plain error and his claim as to Detective Prescott for an abuse of discretion. United States v. Willis, 826 F.3d 1265, 1279 & n.6 (10th Cir. 2016).

This court uses the terms “vouching” and “bolstering” somewhat interchangeably when referring to expert testimony or prosecutorial statements that “express[] a belief or opinion regarding a witness’s credibility.” United States v. Walker, 85 F.4th 973, 985 n.9 (10th Cir. 2023); see also United States v. Coulter, 57 F.4th 1168, 1186–87 (10th Cir. 2023) (noting that we apply “the prohibition on bolstering to statements by prosecutors and individuals affiliated with the prosecution,” including law-enforcement officers). Vouching might involve “blunt comments” affirming the witness’s credibility—as in, “I think the witness is honest”—or “comments that place the prestige of the government behind a witness.” 75A Am. Jur. 2d Trial § 577 (May 2024 update). At bottom, testimony that “usurp[s] the exclusive function of the jury to weigh the evidence and determine credibility” should be excluded as improper vouching. United States v. Hill, 749 F.3d 1250, 1260 (10th Cir. 2014) (quoting United States v. Samara, 643 F.2d 701, 705 (10th Cir. 1981)).

A. Detective Prescott McFadden contends that Detective Prescott vouched for the truth of K.W.’s sexual-assault allegations. Throughout the trial, McFadden disputed K.W.’s credibility because K.W. had initially told his mother on the car ride back from Nebraska that McFadden never touched him, and then mere weeks

45

later K.W. changed his tune during the forensic interview with Detective Prescott, after he had met with a child counselor and talked to his mother. According to McFadden, this “exposure to outside information” before the forensic interview makes K.W.’s interview statements unreliable. Op. Br. at 40.

McFadden probed Detective Prescott on cross-examination about this potential contamination of K.W.’s forensic interview. So on redirect, the government asked the detective: “In a forensic interview do you do your best to elicit the truth regardless of whether or not the child has been interviewed or talked to before you have a chance to talk to them?” ROA vol. VI, at 295. Prescott responded, “I do.” Id. McFadden objected to this testimony on relevance grounds, which objection the district court swiftly overruled. McFadden reasserts that objection to this court under Federal Rule of Evidence Rule 401, and he argues that Detective Prescott’s response of “I do” vouched for K.W.’s credibility, depriving him of a fair trial.

Detective Prescott’s testimony was not vouching. 10 On cross-examination, defense counsel attempted to undermine the forensic interview as a reliable

10 McFadden’s Rule 401 objection and his corresponding arguments on appeal are rooted in his supposition that Detective Prescott’s testimony constituted vouching. He alleges that vouching testimony is generally inadmissible under Rule 401 because the rule prohibits testimony that “encroaches upon the jury’s vital and exclusive function to make credibility determinations.” Op. Br. at 39 (quoting Charley, 189 F.3d at 1267). Because we conclude that Detective Prescott’s testimony was proper, McFadden’s relevancy objection fails. Even so, this court’s deference to the district court to make relevancy determinations is high, and “[t]he bar for admission under Rule 401 (footnote continued)

46

process for extracting truthful allegations from children. This line of questioning opened the door to the integrity of forensic interviews, generally. So naturally, on redirect, the government sought to shore up the forensic- interview process as a reliable means for ferreting out truthful statements from often-reticent child victims. Detective Prescott’s response spoke generally to the truth-seeking function of forensic interviews. Such general statements from government agents or law enforcement about the reliability of a particular investigatory or fact-finding process is not vouching. See, e.g., United States v. Brooks, 736 F.3d 921, 934–35 (10th Cir. 2013) (concluding that a government agent did not vouch for the witnesses’ credibility by testifying that the proffer process was designed “to gauge whether potential cooperators were inclined to be truthful”); United States v. Jones, 468 F.3d 704, 708 (10th Cir. 2006) (determining that testimony stating that, “proffers are used to ‘gauge truthfulness’” “do not meet our standard for vouching” because the testimony “did not amount to guarantees concerning the veracity of the witnesses”); United States v. Magallanez, 408 F.3d 672, 679 (10th Cir. 2005) (rejecting defendant’s argument that government agents’ “[g]eneral testimony” about “how . . . wire records, phone records, and hotel records help investigators to determine contact and money transfers” was “improper vouching,” especially

is very low.” United States v. Wells, 38 F.4th 1246, 1260 (10th Cir. 2022) (quoting United States v. Jordan, 485 F.3d 1214, 1218 (10th Cir. 2007)). Nothing in McFadden’s brief convinces us that the district court erred so egregiously.

47

when defendant “opened the door for the government to explain the purpose of [those] documents”); United States v. Bowie, 892 F.2d 1494, 1498 (10th Cir. 1990) (clarifying that “[u]se of the ‘truthfulness’ portions” of plea agreements “becomes impermissible vouching only when the prosecutors explicitly or implicitly indicate that they can monitor and accurately verify the truthfulness of the witness’ testimony”). Nothing Detective Prescott said during his testimony advocated for the specific truthfulness of K.W.’s forensic-interview statements, so his testimony did not cross the line into vouching.

McFadden’s argument leans heavily on United States v. Jones, a recently decided case in which this court determined that a mother’s testimony about her daughters’ truthfulness constituted plainly erroneous vouching—an error that ultimately led us to vacate the defendant’s conviction and grant a new trial. 74 F.4th 1065, 1069–70, 1073 (10th Cir. 2023). But Jones doesn’t support McFadden’s argument. In Jones, the defense had not yet attacked the daughters’ credibility when the mother testified, so “the door had never been opened for testimony vouching for their truthfulness.” Id. at 1069. McFadden assailed K.W.’s credibility from the beginning of trial during his opening statements by stating that K.W.’s story “gr[ew] over time” between earlier and later law-enforcement interviews because “people other than the children” were spreading lies about McFadden. These statements invited countervailing evidence on the reliability of K.W.’s forensic interview. Jones also differs because the mother’s testimony was separately inadmissible under the Federal

48

Rules of Evidence. Id. (rejecting the mother’s testimony under Rule 608(a) as impermissible character evidence). McFadden identifies no other evidentiary rule that would bar Detective Prescott’s testimony. 11 Because Detective Prescott’s testimony arose in response to McFadden’s attack on K.W.’s credibility and his testimony was otherwise admissible, the district court’s decision overruling McFadden’s objection to Detective Prescott’s testimony was not an abuse of discretion. See Willis, 826 F.3d at 1279 & n.6.

B. Nurse Goebel Nurse Goebel evaluated J.W. for signs of sexual abuse in March 2013 and testified at trial to her findings from that exam. McFadden argues that Goebel vouched for J.W.’s credibility by testifying to “her ultimate ‘assessment’ and ‘conclusion’” that “there was sexual assault” in J.W.’s case. Op. Br. at 41 (quoting ROA vol. VI, at 134). McFadden concedes that he did not object to this testimony when it was offered, and so we review its admissibility for plain

11 McFadden asserts for the first time on appeal that the testimony was “unfairly prejudicial” and so “inadmissible under [Rule] 403.” Op. Br. at 39. McFadden made no such argument before the district court; regardless, it has no merit. For this proposition McFadden looks to Charley, where we determined Rule 403 was violated when health counselors testified to their conclusions that the child victim was sexually abused because “these statements . . . were manifestly outside the counselors’ direct knowledge,” “had minimal probative value,” and “were unquestionably prejudicial.” 189 F.3d at 1270. Detective Prescott’s statements do not resemble the health counselors’ from Charley. Detective Prescott’s statements were probative as to the reliability of forensic interviews, an issue that McFadden raised, his statements about the forensic-interview process were well within the scope of his knowledge, and his statements were not unquestionably prejudicial because they did not assert as “a statement of fact,” id., that K.W. was sexually assaulted.

49

error. Walker, 85 F.4th at 983. This requires McFadden to establish (1) an error (2) that is plain (3) that affected his substantial rights and (4) that undermined the fairness, integrity, and public reputation of the judicial proceeding. United States v. Booker, 63 F.4th 1254, 1258 (10th Cir. 2023).

The government concedes the first two prongs of plain error. 12 Even so, McFadden hasn’t shown a “reasonable probability” that Goebel’s testimony affected his substantial rights, so his claim fails on the third prong of the plain- error test. United States v. Benford, 875 F.3d 1007, 1017 (10th Cir. 2017). Under the third prong, the defendant “bears the burden of establishing” that “the error impacted [his] substantial rights.” United States v. Harlow, 444 F.3d 1255, 1261 (10th Cir. 2006). “[W]hen reviewing vouching for plain error, we weigh the seriousness of the vouching in light of the context of the entire proceeding.” Id. This analysis includes consideration of whether the vouching was “slight or confined to a single instance” or “pronounced and persistent.” Carter v. Bigelow, 787 F.3d 1269, 1292 (10th Cir. 2015) (citation omitted).

In cases “where the outcome boils down to a believability contest[,] . . .

testimony vouching for the credibility of the victim is often prejudicial,” but even then, prejudice is not a given. Jones, 74 F.4th at 1072 (emphasis added). In Jones, the mother’s testimony improperly vouched for her daughters’ tendency toward truthfulness under Rule 608(a) because, rather than testify to

12

We decline to decide whether Goebel’s testimony constituted a plain error.

50

their character of truthfulness, she testified that they were truthful “on a specific occasion.” Id. at 1069. We concluded that this testimony affected the defendant’s substantial rights and reversed his conviction on plain-error review. Id. at 1072–73. Because the daughters’ credibility was central to the jury’s verdict, we reasoned that the effect of the mother’s vouching was “amplified” and so created a “reasonable probability” of changing the trial’s outcome. Id. at 1072. This made Jones different from Charley, where we “held it was harmless error to admit the inadmissible testimony because there was overwhelming evidence implicating the defendant.” Id. at 1072 n.4.

Though McFadden’s trial was a “believability contest” between J.W., K.W., and McFadden, Goebel’s testimony lacked the same outsized impact that the mother’s testimony had in Jones. The mother in Jones testified about her daughters’ specific truthfulness, so both of their allegations were tainted by her improper vouching. Id. at 1069. At best, Goebel’s testimony taints J.W.’s credibility, only. K.W. testified extensively and compellingly to McFadden’s habit of giving K.W. large doses melatonin, McFadden being on top of him in the middle of the night with an erect penis, McFadden groping him, and most of all, McFadden taking K.W. to Nebraska and anally penetrating him during the night in the back of the semi-truck. So it’s not true in this case that “the only other evidence cited by the Government . . . is testimony that [the witness] vouched for.” Id. at 1072. The conduct described in K.W.’s unvouched-for testimony accounted for two of the five criminal counts on which the jury

51

convicted McFadden and two of the five concurrent life sentences to which the district court sentenced him. So even if J.W.’s credibility was tainted, as McFadden alleges, McFadden would still have been convicted on two counts and sentenced to life imprisonment.

But we are unconvinced that any vouching in Goebel’s testimony was so “pronounced and persistent” as to have affected McFadden’s trial. Carter, 787 F.3d at 1292 (citation omitted). In a single comment, Goebel stated her impression from J.W.’s exam that he had been sexually assaulted. See, e.g., United States v. Garrett, 648 F.3d 618, 624–25 (8th Cir. 2011) (concluding that “an isolated comment from one witness in the midst of testimony from twelve other witnesses,” though objectionable, did not merit a mistrial). The bulk of her testimony covered typical symptoms that she observes in child-sex-abuse cases, particularly cases involving anal penetration, and her extensive explanations that “[i]n the vast majority of cases” there is “no physical finding of anal penetration of the child.” ROA vol. VI, at 145. Goebel never claimed to have any expertise in ascertaining the truth from her child patients. And she admitted that her exam “isn’t the forensic interview.” Id. at 142. So the jury had all the information it needed to weigh the credibility of Goebel’s statements. Cf. Magallanez, 408 F.3d at 682 (“It is not the role of an appellate court to consider the credibility of the witnesses or weigh the conflicting evidence, as these matters are within the exclusive province of the jury.”). Besides, Goebel’s conclusion was not the lynchpin of the government’s theory

52

of guilt. Cf. United States v. Whitted, 11 F.3d 782, 787 (8th Cir. 1993) (concluding that admission of expert-vouching testimony required reversal in part because “[t]he Government heavily relied on [the doctor’s] testimony” in prosecuting the case). The prosecution never mentioned Goebel’s impression that J.W. had been sexually assaulted during its closing argument, and it did not return to the topic when examining her on redirect. See Garrett, 648 F.3d at 624–25 (considering that “the government did not mention [a witness’s] comment during the opening statement or closing argument” in affirming the district court’s denial of defendant’s motion for a mistrial).

McFadden additionally argues that the lack of eyewitnesses in this case makes Goebel’s vouching substantially prejudicial. Though it is true that no one witnessed the charged assaults, other testimony corroborated J.W.’s general account of McFadden’s years-long abuse. McFadden’s boss at the trucking company verified that McFadden indeed transported a truck delivery from Telluride, CO, to Farmington, NM in December 2010, when J.W. was about ten-years old. McFadden’s supervisor at the construction company and landlord testified that there were “normally” “three or four” young boys “around [McFadden],” even though there was no business reason for young boys to accompany McFadden on work trips. ROA vol. VI, at 461. And it’s not as though McFadden was accused of one, isolated incident of assault to which there were no eyewitnesses. See United States v. Velarde, 214 F.3d 1204, 1212 (10th Cir. 2000) (concluding that expert testimony vouched for the victim’s

53

credibility and affected defendant’s trial in part because “in this case, a young girl testified to a single instance of alleged sexual abuse” with “relatively little other evidence” (emphasis added)). J.W. testified to an intimate relationship with McFadden that formed throughout his childhood, which then gave McFadden the opportunity to assault J.W. on the out-of-state truck trips. J.W.’s mother corroborated this dynamic. She testified that McFadden “took the lead” in taking care of J.W., that she struggled with a methamphetamine addiction during J.W.’s youth, and that she knew J.W. slept in McFadden’s bedroom when he stayed at McFadden’s house. She also confirmed that J.W. frequently accompanied McFadden on semi-truck trips, sometimes out of state, and that due to her drug addiction there were times when she “didn’t even realize that [J.W.] was gone.” ROA vol. VI, at 448. So even though there were no eyewitnesses to McFadden’s sexual abuse, multiple witnesses substantiated J.W.’s testimony about McFadden’s pattern of traveling across state lines with young boys and his unfettered access to J.W. for large portions of his childhood. This evidence establishes that McFadden had “the opportunity to commit the crimes.” Charley, 189 F.3d at 1271 (considering favorably under a harmless-error analysis that defendant “was on supervised release at the time of the events in question” and that the victims were home alone).

Based on the evidence in this case, McFadden has not carried his burden to show “a reasonable probability that but for” Goebel’s limited testimony

54

about J.W.’s report of abuse, “he would not have been convicted.” Hill, 749 F.3d at 1266. III. Cumulative Error “In a cumulative-error analysis, we aggregate all the separate nonreversible trial errors to assess whether, together, they created reversible error.” United States v. Guinn, 89 F.4th 838, 850 (10th Cir. 2023). When preserved and unpreserved errors occurred at trial, we consider whether the preserved errors—even if, together, they are harmless—plus the unpreserved errors, collectively create prejudice that demands reversal. United States v. Caraway, 534 F.3d 1290, 1302 (10th Cir. 2008).

McFadden argues that a combination of preserved and unpreserved errors were committed during his trial. He asserts that the admission of the 2013 forensic-interview video recording, the exclusion of the 2018 audio recording, and the vouching testimony from Detective Prescott and Nurse Goebel cumulatively “misled the jury about the quantity and quality of evidence against Mr. McFadden,” creating “collective prejudice” that now requires reversal and a new trial. Op. Br. at 46. Among these, we have recognized one nonreversible error: the admission of the 2013 video recording. But McFadden must demonstrate “at least two errors before we may find cumulative error.” Willis, 826 F.3d at 1280. Because we lack multiple errors to aggregate in the cumulative-error analysis, McFadden’s argument fails.

55

IV. Undue-Influence Enhancement The district court applied a two-level undue-influence enhancement to McFadden’s total offense level under U.S.S.G. § 2G1.3(b)(2)(B). McFadden objected to the enhancement’s application during sentencing and challenges it again on appeal. When we evaluate the district court’s calculation of a Guidelines sentence, we review the court’s factual findings for clear error and its legal conclusions de novo. United States v. Jackson, 82 F.4th 943, 949 (10th Cir. 2023). In this case, the district court properly applied the § 2G1.3(b)(2)(B) enhancement.

The undue-influence enhancement authorizes two levels for defendants who “unduly influenced a minor to engage in prohibited sexual conduct.” § 2G1.3(b)(2)(B). The Guidelines Commentary explains,

In determining whether subsection (b)(2)(B) applies, the court should closely consider the facts of the case to determine whether a participant’s influence over the minor compromised the voluntariness of the minor’s behavior. The voluntariness of the minor’s behavior may be compromised without prohibited sexual conduct occurring.

§ 2G1.3(b)(2)(B), cmt. 3(B). If the defendant is at least ten years older than the minor, then the Commentary implements a “rebuttable presumption that subsection (b)(2)(B) applies.” Id. The rebuttable presumption “shifts the burden of producing evidence to rebut the presumption to [the defendant],” though the government always bears the ultimate burden of proving that a sentencing

56

enhancement applies. United States v. Castellon, 213 F. App’x 732, 737 (10th Cir. 2007) (unpublished).

At the sentencing hearing, the district court found that McFadden submitted no evidence to rebut the presumption. McFadden’s “sole argument,” the court noted, was that K.W. and J.W. “were either asleep or so sick” when McFadden abused them and therefore that “there could be no undue influence to get them to engage in sexual activity with him.” ROA vol. VI, at 546. But the court discerned that by instilling himself as a trusted figure in the boys’ lives, McFadden “induced” both boys to “sleep in the bed with him and go on overnight trips,” where he was then “able to rape them.” Id. Because McFadden failed to rebut the presumption, the district court overruled his objection and applied the enhancement.

McFadden concedes that the rebuttable presumption applies because he is over ten years older than J.W. and K.W. McFadden argues instead that he couldn’t have exercised undue influence over the boys because “there is no allegation that J.W. or K.W. ‘voluntarily’ or ‘willingly’ engaged with [him] during their alleged sexual assaults.” Op Br. at 50 (quoting Castellon, 213 F. App’x at 736).

This argument interprets § 2G1.3(b)(2)(B) too narrowly. The enhancement is not limited to circumstances where a victim voluntarily engaged in sex—it also “reaches ‘manipulating’ and ‘preying upon’ a vulnerable victim” in order to procure sex. United States v. Reid, 751 F.3d 763,

57

768 (6th Cir. 2014); see, e.g., Castellon, 213 F. App’x at 738 (affirming the district court’s application of § 2G1.3(b)(2)(B) partly because the court had correctly considered “the fact that [defendant] picked [victim] up in the middle of the night for sexual liaisons at hotels he paid for, . . . drove her all the way from Albuquerque to Mexico for further such liaisons, all while she had no money and no identification”). Even if the boys never consented to the specific sex acts that McFadden performed, his efforts to groom them for sex are enough for the enhancement to apply. See, e.g., United States v. Kempter, 29 F.4th 960, 966 (8th Cir. 2022) (applying the § 2G1.3(b)(2)(B) enhancement “based on a manipulative adult’s building a relationship with a minor for the purpose of eventual sexual activity” (cleaned up)); United States v. Brooks, 610 F.3d 1186, 1199 (9th Cir. 2010) (affirming the § 2G1.3(b)(2)(B) enhancement on appeal because victims “had no money, no job and, as runaways, nowhere to live” when defendants enticed them into prostitution).

Even setting the presumption aside, ample evidence in the record supports the district court’s conclusion that McFadden leveraged his position in J.W.’s and K.W.’s lives to lure them into scenarios that facilitated his abuse. J.W. testified that he saw McFadden as a father-figure. J.W. explained how his mother struggled to provide for him and how McFadden stepped in to give J.W. clothing, food, housing, toys, and a fun environment away from his mother’s abusive boyfriend. K.W. similarly testified that he loved going to McFadden’s house to play with J.W. and the other children. K.W. said that his family didn’t

58

have much money, and so he got clothes, toys, and necessities from McFadden. After cultivating a close relationship with these two similarly disadvantaged boys, McFadden arranged for them to sleep in his bed with him. McFadden then took both boys out of state on long trucking routes, where the only true place to sleep was a shared mattress in the semi-truck’s sleeper cab. Once McFadden got the boys in his bed and in his truck, he sexually assaulted them. This evidence justifies the district court’s application of the § 2G1.3(b)(2)(B) enhancement.

We affirm McFadden’s sentence as imposed by the district court.

CONCLUSION

Affirmed.

No. 23-1089, United States v. McFadden FEDERICO, Circuit Judge, dissenting in part

The majority rightly concludes that the admission into evidence of the 2013 video recording of a forensic interview with a then eleven-year-old K.W. was error. But I do not agree this error was harmless. Rather, I would accept the government’s concession that the error was not harmless and thus decline to sua sponte do the government’s work for them (which they are not even arguing should be done). The exercise of this court’s discretion to engage in a sua sponte harmless error review is not warranted when, as here, the harmlessness is easily debatable. Nonetheless, engaging in a harmless error review, I reach a different conclusion than the majority.

Because the admission of the 2013 video affected McFadden’s substantial rights, I would set aside the convictions and sentence for Counts 1 and 2 (the two charges naming K.W. as the victim), vacate the judgment on these counts, and remand for further proceedings.1

I

Not every erroneous admission of evidence merits relief to a defendant.

An error is harmless if it “does not affect substantial rights.” Fed. R. Crim. P. 52(a). The government bears the burden to prove harmlessness. United States

1 I otherwise agree with the majority on the remainder of the issues before us in this appeal.

v. Ledford, 443 F.3d 702, 712 (10th Cir. 2005). By assigning a burden to a party, the law thus requires that party to either meet their burden or suffer defeat on that matter before the court. Here, the government surrendered this point, declined to argue harmlessness, and conceded the video was prejudicial. Aple. Br. at 25–40; Oral Arg. at 19:00 – 19:40.

Pause to reflect on this point. On appeal, the government’s duty is to defend the convictions it secures at trial after the expenditure of significant time, labor, and resources.2 The government is very capable of doing this and, rightfully, is usually not timid about arguing all potential legal paths to affirmance.3 But here, the government concluded the video was so prejudicial that if this court were to find that its admission into evidence was erroneous (as we have now done), it would not be a plausible argument or strategy to fall back on harmless error.

2 Consider the government’s determination and effort to admit the 2013

video into evidence. To meet its burden of admissibility under the Rule 807 residual hearsay rule, the government provided pretrial notice to the defendant and then presented the evidence and its argument to the district court in a pretrial evidentiary hearing. At trial, it had a transcript produced and ready to provide to the jury when it moved the video into evidence. Upon admission, the jury was given the transcript as an aid to assist in understanding what was said during the video, which the jury watched in open court.

3 In fact, the government argued harmless error regarding two of the

issues raised in this appeal – the 2018 audio recording and the alleged vouching testimony of Nurse Goebel. Aple. Br. at 47, 64.

Even when the government does not argue an error was harmless under Rule 52(a), we retain the discretion to independently assess harmlessness “in an appropriate case.”4 United States v. Samaniego, 187 F.3d 1222, 1224 (10th Cir. 1999) (quoting United States v. Torrez-Ortega, 184 F.3d 1128, 1136 (10th Cir. 1999). But determining what constitutes “an appropriate case” is not a straightforward task.

The Rule 52 harmless error rule was adopted in 1944 as “a restatement of existing law.” Fed. R. Crim. P. 52 1944 Advisory Committee Notes. However, it was not until 1992 that we affirmatively stated that our court may exercise its discretion to initiate a harmless error review when the government did not

4 Rule 52(b) states, “a plain error that affects substantial rights may be

considered even though it was not brought to the court’s attention.” Fed. R. Crim. P. 52(b). It is the flip side of the coin to the Rule 52(a) harmless error standard. Plain error is argued by defendants seeking appellate relief when they have not adequately preserved their objections before the district court. However, unlike harmless error, our cases hold firm that “it is the obligation of the [defendant] to identify, and argue for” plain error, and “[t]he failure to argue for plain error and its application on appeal . . . surely marks the end of the road for an argument for reversal not first presented to the district court.” United States v. McBride, 94 F.4th 1036, 1044 (10th Cir. 2024) (quoting Richison v. Ernest Grp., Inc., 634 F.3d 1123, 1131 (10th Cir. 2011)). So, when it comes to applying Rule 52, what is good for the government goose is apparently not also good for the defendant gander.

make that argument independently.5 United States v. Langston, 970 F.2d 692, 704 n.9 (10th Cir. 1992).

In Torrez-Ortega, we adopted the factors established by the Seventh Circuit in United States v. Giovannetti, 928 F.2d 225, 227 (7th Cir. 1991), as relevant for deciding whether to exercise our discretion to conduct an independent harmless error analysis in the absence of government presentation on the inquiry. 184 F.3d at 1136. The three factors are:

(1) the length and complexity of the record, (2) whether the harmlessness of the error or errors found is certain or debatable, and (3) whether a reversal will result in protracted, costly, and ultimately futile proceedings in the district court.

Id. Although we purposefully did not decide in Torrez-Ortega whether the three Giovannetti factors are exhaustive, we have effectively treated them as such in subsequent cases. See, e.g., Samaniego, 187 F.3d at 1224–25; United States v. Holly, 488 F.3d 1298, 1308 (10th Cir. 2007); United States v. Doe, 572 F.3d 1162, 1175 (10th Cir. 2009); United States v. Spence, 721 F.3d 1224, 1230 n.6

5 Our cases that established and discussed this discretionary authority

do not mention the Federal Rules of Appellate Procedure mandate that an appellee’s brief (which is typically the government in criminal appeals) “must contain . . . the argument, which must contain . . . appellant’s contentions and the reasons for them” and “for each issue, a concise statement of the applicable standard of review.” Fed. R. App. P. 28(a)(8), (b) (emphases added).

(10th Cir. 2013); United States v. Little, 829 F.3d 1177, 1189 (10th Cir. 2016) (Holmes, J., dissenting).

Generally, the three enumerated factors are problematic in their application. Ditching the third factor and applying only the first two, the majority concludes that we should engage in a harmless error analysis without the benefit of the parties’ arguments about it. I respectfully disagree.

The first factor – length and complexity of the record – is phrased in the conjunctive, recognizing that these are not necessarily harmonious terms. Length does not necessarily convey complexity, nor does complexity mean a record is of substantial length. Nevertheless, our cases usually apply this factor based on the number of pages in the record or the number of days of the trial. Again, it is not obvious why those metrics – page numbers and trial days – matter at all as to whether we can or should conduct our own harmless error inquiry.

Rather, the better question is whether the record is complete and adequate for us to do a harmless error review because it contains all that we need to review to do it, not whether it is hard or labor intensive because the record is lengthy and documents that a lot happened at the trial court. Samaniego, 187 F.3d at 1225 (noting the “record itself is abysmally inadequate for a harmless-error review”). We have the video, all admitted exhibits, and the entire trial transcript before us in the record on appeal. In applying this factor,

I agree with the majority that we have all we need in this record for a harmless error review, so this factor weighs in favor of us exercising our discretion.

The second factor – whether the error’s harmlessness is certain or debatable – has been deemed “the most important” factor and was correctly emphasized as so by the majority. Little, 829 F.3d at 1189 (Holmes, J., dissenting). But certainty seems not only elusive in this context but non- existent. How can we conclude the harmlessness of the error is certain when the government, represented by lawyers at trial and on appeal who have the professional obligation and wherewithal to save convictions from reversal in the face of erroneous trial rulings, not only failed to argue harmless error but conceded it should not apply? To answer my own question – we cannot and, indeed, should not.

Reaching this conclusion is not ceding our own judicial obligations to the government. Rather, it recognizes that “harmlessness could be vigorously debated here.” Id. That debate should include the government raising harmlessness and the defendant being afforded the opportunity to respond to the arguments first made by the government. Indeed, as I will elaborate in the next section, a vigorous debate on harmlessness is warranted and, in my view, leads to a conclusion different from that reached by the majority.

Finally, the majority is correct that the third factor – whether reversal would result in protracted, costly, and futile proceedings in district court – adds

7

little to the discretionary review calculus. See also Mollett v. Mullin, 348 F.3d 902, 920 (10th Cir. 2003) (noting that “confusion about what the third factor contributes to the analysis has caused this and other courts to merely reference [the third factor] but not apply it” (quoting Samaniego, 187 F.3d at 1225 n. 2)). Rather than just casually abandoning this factor, we should permanently drop it from our case law altogether.

In sum, only one of the three Torrez-Ortega factors weigh in favor of a sua sponte harmless error review. Because the question of harmlessness is far from certain, we should not exercise our discretion to engage in this inquiry on our own behalf and without the benefit of party presentation.

II

A

Although we should not conduct a sua sponte harmless error review, I nevertheless do so to track the majority opinion. I conclude that the admission of the video was not harmless.

The error here was a violation of a rule of evidence, so it was not a constitutional error. For non-constitutional errors, the government must prove the error did not have a “substantial and injurious effect or influence in determining the jury’s verdict.” Kotteakos v. United States, 328 U.S. 750, 776 (1946). Specifically:

[I]f one cannot say, with fair assurance, after pondering all that happened without stripping the erroneous action from the whole, that the judgment was not substantially swayed by the error, it is impossible to conclude that substantial rights were not affected. The inquiry cannot be merely whether there was enough to support the result, apart from the phase affected by the error. It is rather, even so, whether the error itself had substantial influence. If so, or if one is left in grave doubt, the conviction cannot stand.

Id. at 765. To conduct this review, we “review[ ] the record as a whole de novo to evaluate whether the error [was] harmless, examining the context, timing, and use of the erroneously admitted evidence at trial and how it compares to properly admitted evidence.” United States v. Blechman, 657 F.3d 1052, 1067–68 (10th Cir. 2011) (quoting United States v. Hanzlicek, 187 F.3d 1228, 1237 (10th Cir. 1999)) (alterations in original).

In other words, we must scrutinize the entire trial record, years after the trial occurred in another time and place, to determine whether the jury’s verdicts on Counts 1 and 2 were influenced by viewing a video they should not have seen. And, to again beat the drum of the preceding section, for us to conclude that this factor weighs in the government's favor, we must be certain that the video did not substantially sway the jury. With this framework in mind, I explain why I reach the opposite conclusion of the majority – that the erroneous admission of the video was harmful, not harmless.

B

K.W. is the victim in Counts 1 and 2 of the indictment. Both counts arise from a trucking trip that began in Colorado, stopped in Idaho, and ended in Nebraska, when the defendant was arrested on January 2, 2013, at a Love’s truck stop. K.W., who was eleven years old at the time of the trip, was accompanied by his two brothers, one older and one younger. J.W., the victim in Counts 3, 4, and 5, was not on the trip. During the trip, the defendant sexually assaulted K.W., who disclosed the assault to Detective Prescott during a video-recorded forensic interview on January 16, 2013, two weeks after the defendant’s arrest.

The trial occurred nine years later, in November 2022.6 K.W. was twenty years old when he testified at trial. K.W. testified about the sexual assault that occurred during the trip as follows:

GOVERNMENT: Did Mr. McFadden do anything on that trip specifically that made you uncomfortable?

K.W.: Yes.

GOVERNMENT: What did he do?

K.W.: He tried to stick his thing in me.

GOVERNMENT: By stick his thing in you, did Mr. McFadden put his penis in your butt?

6 As the majority explains, McFadden was first tried and convicted in

state court. The federal charges were not filed until after his state conviction and sentence were overturned on appeal.

K.W.: He tried to, yes.

GOVERNMENT: What do you mean he tried to?

K.W.: It just wouldn’t go in necessarily, so he just tried to, and then after a while he just gave up.

* * *

GOVERNMENT: When you say he tried to put his penis in you, did he put his penis into your butt?

K.W.: Yes.

GOVERNMENT: How did that feel?

K.W.: It hurt, because I was little at the time. It hurt a lot. Again, I don’t remember a lot. Again, I’ve tried to suppress most of these memories, just try to fade them out.

R.V at 182–83. At the end of K.W.’s direct examination, the government moved to admit his 2013 forensic interview with Detective Prescott under Federal Rule of Evidence 807.

The district court asked: “In light of the fact that the witness testified that he did remember the truck assault, what is the purpose of the video?” Id. at 193. The government asserted the video was the “most probative evidence on that point” because “K.W., while he did discuss the events in question, did so reluctantly and often after my prompting and after I used the words that he seemingly cannot.” Id. The district court then admitted the video, finding that

K.W. was “reluctant” and “only answered with prompting and he was equivocal about the penetration.” Id. at 195.

Context and a description of the video are also important to our harmless error review. The video (Government Exhibit 6) is 32 minutes and 40 seconds. The camera is set in a corner, so the viewpoint of the recording is looking down upon a room, which has a small blue sofa, a blue chair, and a white dry-erase board between the sofa and chair. Here is an image from the video (K.W. is blocked in this image by Detective Prescott while he closes the door):

K.W. walked into the room and sat on the blue sofa wrapped in a blanket. He portends the nervous energy of a child. His first comments to

Detective Prescott were, “There is a computer in that other room,” as he pointed towards the door, adding, “I want to go play on it.” Govt. Ex. 6 00:00 – 00:13. During the interview, K.W. acts his age: he throws himself around the couch and floor, lies on the couch, and mumbles a lot of his words.

The majority concludes that the video’s admission into evidence is harmless for the same reason it is not more probative under Rule 807(a)(2) because, it says, K.W.’s video statements from 2013 mirror his trial testimony. I respectfully disagree.

First, his testimony at trial and his statements in the video are not mirrors of each other. The substance is similar, but they are not the same. For example, in both his testimony and the video he said the words to describe a “sexual act,” an element of Count 1 and defined for the jury in the court’s instructions. R.I at 908. But although some of the words and language used in both are the same or similar, they are not “mirrored” because these types of evidence are qualitatively different.

Evidence is not just the words or text later put to paper on a transcript. Trials are live events, where jurors watch how witnesses testify as much as they listen to their words. The consumption of information through live testimony is qualitatively different than watching a video of

13

an interview that was conducted in a controlled environment.7 The eleven- year-old K.W. talking to a detective in a comfortable room, while draped in a blanket, and lying on a small blue sofa, is not the same as a twenty-year- old K.W. testifying in person before the jury, eight years later, recounting what happened in the past.

This is not to say that the differences in the types of evidence make the video “more probative” than his live testimony under the residual hearsay rule, which requires a court to weigh the proffered evidence against “any other evidence that the proponent can obtain through reasonable efforts.” Fed. R. Evid. 807(a)(2). In balancing between the two, the majority is correct that live testimony was superior and more probative evidence.8

7 This is also why I do not hang my harmless error hat on a cumulative

evidence analysis. United States v. Otuonye, 995 F.3d 1191, 1208 (10th Cir. 2021).

8 The majority notes there is one key difference between this case and

United States v. Burgess, 99 F.4th 1175, 1182 (10th Cir. 2024), in which this court affirmed the admission of a forensic interview video of a child victim under the residual hearsay rule. That difference being that the child’s testimony in Burgess diverged from the original accusation made during the video-recorded forensic interview.

I emphasize another key difference – the child in Burgess was seven years old when she was abused, seven when she sat for a forensic interview, and nine when she testified at trial. In other words, she remained a young child throughout the entirety of the case, from allegation through trial. Not so here, where K.W. grew from a young boy of eleven in 2013 when the video was recorded to a young man of twenty when he testified at trial.

United States v. Harrison, 296 F.3d 994, 1007 (10th Cir. 2002); United States v. W.B., 452 F.3d 1002, 1005–06 (8th Cir. 2006).

Second, the video is not harmless because it improperly bolstered K.W.’s credibility. See Fed. R. Evid. 608(a)(2); United States v. Bowie, 892 F.2d 1494, 1499 (10th Cir. 1990). The jurors heard K.W.’s tepid testimony about the abuse. Although other facts and circumstantial evidence generally corroborated K.W.’s testimony, his testimony was the only evidence regarding the element that a “sexual act” occurred. As the majority states, K.W.’s credibility shoulders the government’s case because no one else testified that they bore witness to the alleged assault.9 That is why the 32-minute video, which had the purpose and effect only to improperly bolster K.W.’s credibility, was not harmless. The jurors10 heard his testimony and then watched him on video, in a controlled setting, give similar answers to similar questions posed by a trained detective.

9 The jurors were instructed that each count on the indictment “involve

separate additional factual determinations that you must make. . . . Each crime or offense as charged and the evidence applicable thereto should be considered separately as to each count.” R.I at 907.

10 The fact that this was a jury trial is key to the bolstering analysis. Had

this been a bench trial, witness bolstering would not have the same effect. See, e.g., United States v. W.B., 452 F.3d 1002, 1006 (8th Cir. 2006) (“The district court conducted a bench trial and in such a situation, we find little prejudice in the admission of cumulative evidence or testimony which improperly bolsters a prior witness.”).

The majority also concludes that any harm done by the jurors watching the video and seeing K.W. as a child was offset by the fact that the government also admitted into evidence photographs depicting K.W. at eleven years old. They conclude that this mitigated the “emotional response in the jury,” having already seen him as a child in the photographs. United States v. Otuonye, 995 F.3d 1191, 1207 (10th Cir. 2021). But again, seeing photographs of a child compared to watching that child for 32-plus minutes on video is not the same. People are more than their images. When it comes to how jurors will consume and absorb evidence, other personal characteristics also matter, such as: voice, mannerisms, temperament, conduct, inflection, etc. This is one reason why our judicial system prefers jurors to observe live testimony of witnesses. Garcia-Martinez v. City & Cnty of Denver, 392 F.3d 1187, 1191 (10th Cir. 2004).

Finally, the majority concludes that “most of all” McFadden’s failure to request redactions from the video renders its admission harmless. The defense objected to the admission of the video, so I do not see how the failure to request redactions to the video after it was admitted into evidence then, ipso facto, converts this error to being harmless.

I simply cannot conclude that the erroneous admission of the video did not have a substantial effect or influence on the jury’s verdicts for Counts 1 and 2. Kotteakos, 328 U.S. at 765. Because I both disagree with

16

the decision to conduct a sua sponte harmless error analysis and conclude that McFadden’s substantial rights were violated, I respectfully dissent from that part of the majority opinion.