dissenting:
Were the standard of review for factual determinations in this case de novo, I might join my colleagues. However, while the standard of review for a district court’s “in custody” determination is a mixed question of law and fact warranting de novo review, we review the factual findings *888underlying the district court’s decision for dear error. United States v. Kim, 292 F.3d 969, 973 (9th Cir.2002). After setting forth the five prong test for determining if a person was “in custody” when he made incriminating statements, the Kim court itself reviewed the district court’s findings in that case on each of the five prongs for clear error, not de novo:
The district court’s factual findings are not clearly erroneous, as they are supported by testimony in the record that the judge determined was credible. After reviewing the factual findings under all of the circumstances ... we conclude that Kim was “in custody” for Miranda purposes because a reasonable person in Kim’s circumstances would not have felt free to leave.
Kim, 292 F.3d at 974.
A police officer is required to read a person his Miranda rights before conducting an interrogation if the person is “in custody.” “To determine whether an individual was in custody, a court must, after examining all of the circumstances surrounding the interrogation, decide whether there[was] a formal arrest or restraint on' freedom of movement of the degree associated with a formal arrest.” Kim, 292 F.3d at 973 (internal quotation marks and citation omitted) (alteration in original). Here, the State concedes it did not read Bassignani the Miranda warnings before he confessed. Thus, we must determine whether the district court committed clear error when it determined Bassignani was “in custody.”
“A finding of fact is clearly erroneous when the evidence in the record supports the finding but ‘the reviewing court is left with a definite and firm conviction that a mistake has been committed.’ ” Burlington N., Inc. v. Weyerhaeuser, 719 F.2d 304, 307 (9th Cir.1983).
Here, both the majority and the district court employ the analysis in Kim. The district court analyzed and weighed each of the five factors to determine whether Bassignani was in custody for purposes of Miranda v. Arizona, 384 U.S. 436, 86 S.Ct. 1602, 16 L.Ed.2d 694 (1966). United States v. Bassignani, 2007 WL 2406868, *4-6 (N.D.Cal.). The district court also correctly acknowledged the Kim factors are not exhaustive and that the “in custody” determination rests upon an objective inquiry of the totality of the circumstances. Id. at *4.
After reviewing the parties’ declarations and listening to the entire interrogation of Bassignani (as did I), the district court determined that Bassignani was in custody for Miranda purposes. Bassignani, at *5-6. The district court based its conclusion on the following factual findings applied to the Kim factors:
First, the district court found Bassignani did not voluntarily approach or accompany the officers. On this we all agree. Ms. King, the Tellabs Human Resource manager, stated in her declaration that Detective Williams and another officer approached Bassignani’s workspace and instructed Bassignani to follow Ms. King and the police officers to an enclosed conference room. Thus, this factor weighs in favor of Bassignani’s claim he was in custody.
Second, the district court found Bassignani was confronted with evidence of his guilt. The majority disagrees. The evidence supports the district court’s finding that this confrontation took place: The taped interrogation showed that during questioning, Detective Williams repeatedly confronted Bassignani with evidence of his guilt, including, inter alia, stating, “it’s your laptop1 ... I’m not going to lie to *889you ... we’ve got your email connected to the [child pornography] images, it’s a done thing.” The police so told Bassignani after their search of his residence had uncovered a laptop computer with child pornography on the hard drive. Bassignani therefore knew the police had found incriminating evidence at his house. The majority finds this evidence insufficient to turn an “otherwise cordial interview” into a confrontation. Maj. Op. at 885 n. 7. Here, the district court’s finding that Bassignani was confronted by the police with evidence of his guilt is plausible. Regardless the “tone” of voice used to confront Bassignani with this evidence — which the district court also heard — the content of the conversation makes this case much closer to United States v. Beraun-Panez, 812 F.2d 578, 579 (9th Cir.1987); it provided sufficient evidence for the district court’s finding.
Further, I disagree with the majority’s assessment that because Bassignani was trying to assess what the police knew, it somehow turned this into a friendly chat. The majority relies on the following remarks by Bassignani: “I’m trying to save my own ass here. I mean, let’s be honest. I’m trying to get minimal impact of this as possible.” The majority may conclude this made it a cordial conversation, but the district court surely was not clearly erroneous in finding these statements indicated Bassignani knew he was in trouble, he was trying to limit the damage he was about to suffer, and there was nothing friendly about the conversation. Rather, it was plausible for the district court to find the conversation was a police interrogation and the police were looking for evidence to convict him, not someone else.
Because there is evidence in the record to support the district court’s finding, we cannot say the district court committed clear error. This finding required the district court to examine exactly what the police said to Bassignani and how a reasonable person would have interpreted such statements, which is a factual determination. When reviewing for clear error, “this court will not reverse if the district court’s findings are plausible in light of the record viewed in its entirety ... even if it is convinced it would have found differently.” Katie A., ex rel. Ludin v. Los Angeles County, 481 F.3d 1150, 1155 (9th Cir.2007) (citation omitted).
Third, the district court found that the setting was a neutral factor in the analysis. The majority disagreed and found that because Bassignani was familiar with the conference room and because his employment superior, King, left the room, the surroundings were familiar and this factor should weigh against Bassignani’s claim he was in custody. The evidence, however, showed that the location of the interrogation limited Bassignani’s freedom. Bassignani was instructed to go to the conference room by the officers and his employer. See first factor, supra. It was at his place of work, a place where his freedom of movement was subject to his superior’s orders and the orders of the police. I agree there is no evidence in the record that the officers locked the door, as they did in Kim, but it was not made clear to Bassignani he could leave “at any time” or that he was free to have anyone else accompany him. Accordingly, the record does not allow us to hold the district court’s finding that the setting weighed neither for nor against a finding Bassignani was in custody was clearly erroneous.
Fourth, the district court found the tape recording of the interrogation showed the questioning lasted approximately two and a half hours, which is at the “high end” of this court’s precedent for what constitutes an “in custody” determination, see Kim, 292 F.3d at 977 (finding fifty minutes of questioning was not a “brief inquiry” but a *890“full-fledged interrogation”). On this point, we are all agreed. Thus, the district court was not clearly erroneous in finding this factor weighed toward Bassignani’s claim he was in custody.
Finally, the district court found that under all the circumstances a reasonable would not have believed he could freely walk away from the interrogation at any time, but would instead have thought he had to stay until the police were finished questioning him. Kim, 292 F.3d at 974. The majority disagreed. Here again, the record contains evidence that supports the district court’s finding. Detective Williams’s statement to Bassignani at the beginning of the interview when he said, “[yjou’ll walk out of here when we’re done,” implied Bassignani was not free to leave at any time, but had to stay until the officers had finished their work. Bassignani, at *4-6 (emphasis added). The majority looks to Detective Williams’s statement that Bassignani was not under arrest as a basis upon which to disagree with the district court’s finding.
This was a factual finding by the district court based on the evidence in this particular record. I fail to see how we can overturn the district court’s finding a reasonable person would not have thought he was free to leave at any time when he was instructed to accompany three police officers to the conference room, told he would be free to leave “when we’re done,” and then confronted more than once with evidence of his guilt. Again, even though we might have found differently were we sitting as a trial court, the district court’s factual findings here are plausible in light of the record viewed in its entirety. Thus, we cannot say the district court has committed clear error. Katie, 481 F.3d at 1155.
The majority concedes this is a close case. Maj. Op. at 886-87. Factual determinations in close cases are the function of the trial court, not the court of appeals, and especially not where our standard of review is for clear error. Clear error means it is not a close case. Because there is evidence in the record supporting each of the district court’s factual findings, I see no way we can say the district court committed clear error, thus I respectfully dissent.
. Bassignani has denied he owned or possessed a laptop computer.