No. 24-1744

Appeal from the United States District Court for the Eastern District of Michigan at Detroit. No. 2:22-cr-20309-1—Stephen Joseph Murphy III, District Judge.

Argued: July 30, 2025

Decided and Filed: August 19, 2025

Before: MOORE, GRIFFIN, and RITZ, Circuit Judges.

COUNSEL

ARGUED: Celeste Kinney, OFFICE OF THE FEDERAL COMMUNITY DEFENDER OFFICE, Detroit, Michigan, for Appellant. Henry Edward Moon, III, UNITED STATES ATTORNEY'S OFFICE, Detroit, Michigan, for Appellee. ON BRIEF: Celeste Kinney, Keshava Kirkland, Todd Shanker, OFFICE OF THE FEDERAL COMMUNITY DEFENDER OFFICE, Detroit, Michigan, for Appellant. Jeanine Brunson, UNITED STATES ATTORNEY’S OFFICE, Detroit, Michigan, for Appellee.

RITZ, Circuit Judge. Kyrrah Radaker-Carter appeals the denial of his motion to suppress pretrial identification evidence. We affirm.

On April 23, 2022, E.J. went to a store in Detroit to buy orange juice. As she left, two men carjacked her in the parking lot.

E.J. had noticed the men earlier, while shopping. Although the men were wearing masks, E.J. saw that one was “[l]ight-skinned” and the other was “dark-skinned.” RE 32, Suppression Hr’g Tr., PageID 190, 193, 198. The lighter-skinned man stood behind E.J. in the checkout line and E.J. heard him talking to another customer.

The darker-skinned man followed E.J. as she left the store. When E.J. got to her car, the man pointed a gun at her and demanded her fanny pack and car keys. Then, the lighter-skinned man came out of the store, also pointing a gun at E.J. E.J. was scared the men would shoot her, so she gave them her keys and fanny pack. The men got in E.J.’s rental car and drove away.

As E.J. waited for the police, she was approached by the customer she saw speaking with the lighter-skinned carjacker in the checkout line. The customer said she knew the man from school. E.J. and the customer exchanged phone numbers, and the customer later told E.J. that she would call her “in about twenty minutes with that name.” Bodycam Footage at 03:08.

When police officers showed up at the store, E.J. told them that the lighter-skinned man was 'mixed [race] or maybe . . . Mexican.' Id. at 04:46. Then, a few hours later, officers interviewed E.J. at her house. There, she described the man as “possibly mixed or Hispanic” and in his “late 20s or early 30s,” but because she was “focused on the gun,” E.J. could not remember everything. RE 57, Gov’t Ex. 4, Incident Rep., PageID 441 (citation modified).

After she met with police, E.J. started texting the customer. The customer told E.J. that the light-skinned carjacker was named 'Karaa'; he was from Inkster, Michigan; and he was about 30 or 31 years old.

The investigating officer was also texting the customer, having gotten her number from E.J. During their exchange, the customer sent the officer three photos of Kyrrah Radaker-Carter and screenshots from his Facebook profile containing his personal information.

The police scheduled a photo lineup with E.J. The lineup included pictures of six men, including Radaker-Carter. The pictures were arranged in a grid with two rows of three photos each, with Radaker-Carter’s photo in the center of the top row. The officer who assembled the lineup ordered the photos randomly.

In picking the photos, the officer looked for others of “similar sex, race, hairstyle, things [like] that to include as fillers.” RE 32, Suppression Hr’g Tr., PageID 161. The officer testified that he usually pulls lineup pictures from the Detroit Police Department's mugshot database because it has an effective search function. That's where he got the five filler photos. But because Radaker-Carter was not in that database, the officer had to get his photo from elsewhere. As a result, Radaker-Carter's photo had a darker background and different lighting than the others.

Before the lineup, E.J. asked the store customer-Radaker-Carter's schoolmate-for a picture of the man that the customer thought was the light-skinned carjacker. E.J. said she 'want[ed] to make sure [she] g[o]t it right' at the lineup. RE 57, Gov't Ex. 8, Text Msg. Screenshots, PageID 450. The customer sent E.J. a picture of Radaker-Carter from his Facebook page. But the customer said that the police “asked [her] not to show” E.J. the photo, so she told E.J., “don’t tell them you saw a picture.” Id. at PageID 451 . When E.J. got the customer's text, she responded, “[t]hat’s him, I remember [] his features.” Id.

The police conducted the photo lineup the day after the customer sent E.J. the photo, which was three days after the carjacking. E.J. '[i]nstantly' identified Radaker-Carter as the light-skinned carjacker. RE 32, Suppression Hr’g Tr., PageID 212; accord id. at 173. A day later, the police arrested Radaker-Carter while he was driving E.J.’s rental car. He was charged with carjacking, 18 U.S.C. § 2119, and brandishing a firearm in relation to a crime of violence, id. § 924(c).

Radaker-Carter moved to suppress E.J.’s identification. He argued that introducing the identification at trial would violate his due process right to a fair trial. See Neil v. Biggers , 409 U.S. 188, 198 (1972).

After a hearing, the district court denied Radaker-Carter’s motion. The court ruled that even if E.J.’s identification was affected because she saw a photo of Radaker-Carter before the lineup, due process does not require exclusion of an out-of-court identification where 'the suggestive circumstances were not arranged by law enforcement officers.' RE 36, Order, PageID 275 (quoting Perry v. New Hampshire , 565 U.S. 228, 232 (2012)). The court also concluded that the photo array was not unduly suggestive.

Radaker-Carter pled guilty to both counts. The district court sentenced him to 122 months’ imprisonment, and Radaker-Carter appealed.

Radaker-Carter argues that the district court should have suppressed the evidence that E.J. identified him as one of the carjackers. But E.J.'s identification was not the product of police-arranged suggestive circumstances, so we reject his argument.

We must first decide the applicable standard of review. Our case law on this point is less than clear. We have at times reviewed challenges to the admissibility of pretrial identification evidence using a mixed standard, reviewing the district court’s factual findings for clear error and its legal conclusions de novo. See United States v. Crozier , 259 F.3d 503, 510 (6th Cir. 2001); United States v. Marks , 209 F.3d 577, 585 (6th Cir. 2000). At other times, we have applied clear-error review to the entirety of such claims. See United States v. Beverly , 369 F.3d 516, 538 (6th Cir. 2004); accord United States v. Washington , 714 F.3d 962, 966 (6th Cir. 2013). For two reasons, the former approach is correct.

First, the mixed-standard cases predate the clear-error cases. “When a later decision from this court conflicts with its prior decisions, the earlier cases control.' Sowards v. Loudon Page 5County , 203 F.3d 426, 431 n.1 (6th Cir. 2000). Marks and Crozier were decided in 2000 and 2001, respectively. In those pretrial identification cases, we 'review[ed] [the] district court's . . . legal conclusions de novo.' Crozier , 259 F.3d at 510; accord Marks , 209 F.3d at 585. By contrast, Beverly , which reviewed the whole 'denial of a motion to suppress identification evidence for clear error,” was decided later, in 2004. 369 F.3d at 538. Per our practice of inter-panel accord, the earlier mixed-standard cases govern. See 6 Cir. R. 32.1(b); Helphenstine v. Lewis County , 60 F.4th 305, 317 (6th Cir. 2023).

To be sure, Beverly cited a 1982 pretrial identification caseUnited States v. Hamilton , 684 F.2d 380 (6th Cir. 1982)—to support its use of clear error. See Beverly , 369 F.3d at 538. But Hamilton did not establish a blanket clear-error-review rule. Rather, the Hamilton panel briefly commented that the court's 'findings [were] not clearly erroneous.' 684 F.2d at 383. That statement is ambiguous; it is not clear the panel was referring to the district court’s legal conclusions. Indeed, it makes more sense to read Hamilton as discussing the court's factual “findings.” Id. That reading would be consistent with our holding two years prior that '[t]he clearly erroneous standard of review' governs the 'factual finding[s] underlying [the district court’s] legal conclusion[s].” United States v. Coleman , 628 F.2d 961, 963 (6th Cir. 1980).

There is also no indication that Hamilton marked a departure from earlier pretrial identification cases, which appeared to use de novo review when analyzing legal issues. E.g. , United States v. Scott , 518 F.2d 261, 264-66 (6th Cir. 1975); United States v. Clark , 499 F.2d 889, 892 (6th Cir. 1974). As was typical at the time, those cases did not explicitly identify the relevant standard of review. See Amanda Peters, The Meaning, Measure, and Misuse of Standards of Review , 13 Lewis & Clark L. Rev. 233, 238 (2009) ('[I]t was not until the late 1980s and early 1990s that appellate courts routinely began to include a discussion on the applicable standard of review in most opinions.'). However, the analysis in Clark and Scott looks non-deferential. Clark , for example, 'examine[d] carefully the composition of the challenged photographic array' and agreed that the array was 'suggestive' but ultimately concluded that its admission was constitutional. 499 F.2d at 892; see also Scott , 518 F.2d at 26566 (independently “view[ing] the pictures in question” to decide the due process claim).

Page 6Moreover, as Marks and Crozier show, we did not initially treat Hamilton as requiring clear-error review of a trial court’s legal conclusions. Crozier , in fact, directly cited Hamilton for a different proposition, demonstrating that the panel was aware of Hamilton ’s analysis and still did not read the case to create a blanket clear-error rule. Crozier , 259 F.3d at 512. So Hamilton did not disturb our established non-deferential review of legal conclusions in pretrial identification cases.

Second, applying the clear-error standard to the issue presented here would arguably conflict with Supreme Court precedent. The Court has said that “[a]bsent a treaty or statutory prescription,” “questions of law are reviewed de novo .” Monasky v. Taglieri , 589 U.S. 68, 83 (2020); see also Highmark Inc. v. Allcare Health Mgmt. Sys., Inc. , 572 U.S. 559, 563 (2014) ('Traditionally, decisions on 'questions of law' are 'reviewable de novo '.' (quoting Pierce v. Underwood , 487 U.S. 552, 558 (1988))). And cases like this one present a legal question: whether the admission of identification evidence would violate a defendant's right to due process. See Perry , 565 U.S. at 237. True, that inquiry often “involves plunging into a factual record.” U.S. Bank Nat’l Ass’n ex rel. CWCapital Asset Mgmt. LLC v. Vill. at Lakeridge, LLC , 583 U.S. 387, 396 n.4 (2018). Even so, the Court has emphasized that “[i]n the constitutional realm, . . . the role of appellate courts 'in marking out the limits of [a] standard through the process of case-by-case adjudication’ favors de novo review' of even 'mixed question[s].' Id. (quoting Bose Corp. v. Consumers Union of U.S., Inc. , 466 U.S. 485 503 (1984)) (second alteration in original).

This circuit’s “history of appellate practice” also “indicat[es] the appropriate standard.” Monasky , 589 U.S. at 84 (quoting Pierce , 487 U.S. at 558); see also McLane Co. v. EEOC , 581 U.S. 72, 79-80 (2017) (deciding the standard of review for appeals from a motion to quash an EEOC subpoena by looking to other administrative subpoena cases). When assessing suppression claims in criminal cases premised on constitutional arguments, we generally “review findings of fact under the clear-error standard and review conclusions of law de novo.” United States v. Whitley , 34 F.4th 522, 528 (6th Cir. 2022) (motion to suppress evidence from an allegedly unconstitutional search); see also United States v. Jacobs , 63 F.4th 1055, 1058 (6th Cir. 2023) (motion to suppress an allegedly involuntary confession).

Page 7In sum, the mixed standard is consonant with our older pretrial identification cases and recent Supreme Court precedent. When reviewing a district court's decision on a motion to suppress pretrial identification evidence, we will depart from its factual findings only if they are clearly erroneous. But we apply de novo review to the court’s conclusions concerning whether the circumstances giving rise to the identification were unnecessarily suggestive and whether the identification was otherwise reliable.

Reviewing the district court's due process analysis de novo, we agree that E.J.'s identification was constitutionally admissible. A defendant's due process right to a fair trial prevents the introduction of out-of-court identification evidence when that evidence is 'so extremely unfair that its admission violates fundamental conceptions of justice.” Perry , 565 U.S. at 237 (quoting Dowling v. United States , 493 U.S. 342, 352 (1990)). We evaluate claims that a pretrial identification should be excluded in two steps. Salter v. City of Detroit , 133 F.4th 527, 538 (6th Cir. 2025). First, we ask whether “law enforcement officers use[d] an identification procedure that is both suggestive and unnecessary.” Perry , 565 U.S. at 238-39. “If so, we then consider whether the evidence was nevertheless reliable.” Haliym v. Mitchell , 492 F.3d 680, 704 (6th Cir. 2007).

Exclusion is a strong remedy reserved only for cases '[w]here the 'indicators of [a witness’] ability to make an accurate identification’ are ‘outweighed by the corrupting effect’ of law enforcement suggestion.” Perry , 565 U.S. at 239 (quoting Manson v. Brathwaite , 432 U.S. 98, 116 (1977)) (second alteration in original). Trial rights instead are the primary safeguards by which '[t]he Constitution . . . protects a defendant against a conviction based on evidence of questionable reliability.' Id. at 237. Absent exclusion of identification evidence, defendants maintain the right “to confront the eyewitness” and “to the effective assistance of an attorney, who can expose the flaws in the eyewitness’ testimony during cross-examination.” Id. at 245-46. Defendants can also use “[e]yewitness-specific jury instructions” to their advantage. Id. at 246. These protections help “caution juries against placing undue weight” on potentially “fallib[le]” eyewitness identifications. Id. at 245.

Page 8Because due process's restraint on the admissibility of pretrial identification evidence “turn[s] on the presence of state action,” such claims are cognizable only if government officials played some role in creating the suggestive circumstances giving rise to an identification. Id. at 232-33. We have accordingly declined to exclude identifications even when the identifying witness by chance saw the defendant’s photo in the newspaper, United States v. Peterson , 411 F. App'x 857, 865 (6th Cir. 2011), or was shown the defendant's social media profile by a coworker, United States v. Carson , 796 F. App’x 238, 248 (6th Cir. 2019), before participating in the lineup.

E.J.’s texts with the store customer, therefore, do not create constitutional concerns. The customer was not a state agent, and no evidence suggests that she was working with the investigating officers to influence E.J.'s selection from the lineup. In fact, the customer apparently disobeyed the police’s wishes by sharing the photo of Radaker-Carter with E.J.

By contrast, the photo lineup was police-arranged, so our normal due process framework applies. Under that standard, E.J.’s identification was admissible because the photo array did not “give rise to a very substantial likelihood of irreparable misidentification.” Sexton v. Beaudreaux , 585 U.S. 961, 965 (2018) (per curiam) (quoting Biggers , 409 U.S. at 197).

The photo array did not 'improperly single out' Radaker-Carter. United States v. Sullivan , 431 F.3d 976, 985 (6th Cir. 2005). Rather, like Radaker-Carter, '[t]he individuals pictured were all African-American males of similar build, color, complexion and hairstyle,” and the photos were “placed in random order.” United States v. Reamey , 132 F. App’x 613, 615 (6th Cir. 2005). Plus, the officer who assembled the lineup gave a “reasonable explanation” for any differences between the photos. See United States v. McComb , 249 F. App’x 429, 440 (6th Cir. 2007). He testified that Radaker-Carter's photo was not in the Detroit Police Department mugshot database. And the officer did not pull the other photos from the same database as Radaker-Carter's because Detroit's system was searchable 'generically . . . as opposed to searching by name.” RE 32, Suppression Hr’g Tr., PageID 159.

Page 9Still, Radaker-Carter argues that the cumulative effect of four of the lineup's features made the array suggestive. He points to the number of photos, their positioning, their backgrounds, and their shading.

None of these differences alone render E.J.’s identification inadmissible. For instance, a six-photo lineup is not per se suggestive. United States v. Stamper , 91 F. App’x 445, 460 (6th Cir. 2004); see also Wingate v. United States , 969 F.3d 251, 261 (6th Cir. 2020) (concluding that a six-photo lineup “does not look impermissibly suggestive”). And we have implied the same for a photograph's placement in a lineup-especially in the absence of evidence that the placement influenced the witness’s identification. Searcy v. Berghuis , 549 F. App'x 357, 365 (6th Cir. 2013).

The differently colored background in Radaker-Carter's photo also does not affect our conclusion. Generally, '[a] darker hue or different colored background does not 'in [itself] create an impermissible suggestion that the defendant is the offender.’” McComb , 249 F. App’x at 440 (second alteration in original) (quoting United States v. Burdeau , 168 F.3d 352, 358 (9th Cir. 1999)). And here, Radaker-Carter’s picture was not the only one with a darker background. Only three of the six photos had pure white backgrounds. The remaining pictures, including Radaker-Carter's, were set against darker grey backgrounds. To be sure, the background of Radaker-Carter’s photo was the darkest. However, E.J. “gave no indication that . . . the color of the background drew her attention to the picture.” Id.

Finally, we do not think the lineup was suggestive because Radaker-Carter’s complexion appeared lighter than that of some of the other men. True, “a glare from the camera flash” makes the relative color of the men’s skin appear slightly different. Washington , 714 F.3d at 967. But comparatively, 'there is not a drastic difference between the defendant's skin tone and the remaining photos.” Id.

Accordingly, none of the lineup’s features on their own raise due process concerns. And because the photos’ differences were themselves “minor,” their cumulative effect was weak. See Reamey , 132 F. App'x at 616-17. Reamey is instructive. There, we confronted a six-person photo lineup where the suspect 'was much older than the other individuals,' 'was the only Page 10individual . . . grimacing,' and was wearing 'clothing [that] was noticeably different from the clothing of the other individuals pictured.” Id. at 616. Despite those differences, we determined that the lineup was not overly suggestive. Id. We reasoned that, because all six photos depicted men of the same race with similar builds, complexions, and hairstyles, it was unlikely that the differences in age, clothing, and facial expression led the witness there to pick the defendant’s photo. Id. at 616-17.

So too here. As explained, the lineup photos all showed men with similar physical features who appeared to be approximately the same age. Additionally, the record does not suggest that E.J. identified Radaker-Carter because of any of the photos' differences. To the contrary, E.J. testified that she recognized Radaker-Carter’s nose. Given that E.J.’s “attention” was not “directed to a suspect” because of the array’s composition, the district court correctly ruled that admitting her identification was constitutional. Haliym , 492 F.3d at 704 (quoting Howard v. Bouchard , 405 F.3d 459, 469-70 (6th Cir. 2005)).

Our analysis ends there. Radaker-Carter 'failed to meet his burden to show that the photo array was unduly suggestive,” so it is “unnecessary” to decide “whether the identification was nonetheless reliable.” Washington , 714 F.3d at 968.

For the foregoing reasons, we affirm.