specially concurring:
Because I agree with the majority that we should reverse and remand this case for a new trial, I concur in the judgment.1 I also agree with the majority that, to decide the voluntariness of a confession resulting from a non-custodial police interview, we consider the totality of the circumstances in which the suspect makes the confession, including the relevant characteristics of the suspect. Doody v. Ryan, 649 F.3d 986, 1008 (9th Cir.2011) (en banc). But I view the totality of the circumstances differently than the majority. In particular, I think that the officers’ false promises about the nature of the interview, coupled with Preston’s intellectual disability, coerced Preston into confessing, but most of the tactics employed by the officers were not coercive.
*1030I.
Many aspects of the interview were not coercive.
First, Preston was not in custody at the time of the interview. Two law enforcement officers, wearing plain clothes and arriving in unmarked cars, approached Preston outside his home. One was not armed, and the other did not display a weapon at any time. The entire conversation took place outside Preston’s home; when asked whether he would prefer to talk privately in one of the officers’ cars, Preston said “no” and the interview continued at the same outdoor location. The officers told Preston that he was not under arrest, that he was free to stop answering their questions at any time, and that he was free to leave at any time. (Because the discussion took place outdoors, Preston had only to walk through his own front door to terminate the interview.) The officers were not outsiders to the Navajo community; one was a tribal officer. Their tones of voice were soft, and they asked questions in a slow and low-key manner, with many pauses and silences in order to give Preston time to think and respond. The entire interview lasted only about 38 minutes.2
Second, apart from the false promises of leniency discussed below, the tactics used by the officers (and the instances of their mistakes) in this case are benign: Over the course of a casual 38-minute interview, the officers asked almost entirely open-ended questions and discussed highly charged topics in broad terms. The officers also, clearly by mistake, referred to the wrong day of the previous week during portions of the interview. In my view, the method of asking questions and the mistake were not coercive.
The officers asked Preston only a few either-or questions, and many were followup questions to clarify steps in the narrative that were necessarily binary — for example, an officer asked, “Did [the child] put the condom on or did [Preston]?” and “[W]as [the child] standing up or was [the child] sitting down?”3 Although either-or questions could be overly leading in certain *1031contexts, they are not necessarily leading and can, at times, serve to clarify a narrative.
Contrary to the majority, maj. op. at 1024-25, I do not think that the district court erred — clearly or otherwise — in finding that Preston’s admissions were not tied to a particular day of the week. Over the course of the interview, the discussion moved from questions about a particular weekday to questions about an event unmoored from any particular day. I would not find Preston’s confession involuntary, even in part, by reason of such a slight misstep.
As the Supreme Court has recognized, there is “no talismanic definition of ‘voluntariness,’ mechanically applicable to the host of situations where the question has arisen.” Schneckloth v. Bustamonte, 412 U.S. 218, 224, 93 S.Ct. 2041, 36 L.Ed.2d 854 (1973). Rather, voluntariness is viewed as a spectrum: At one end is “the acknowledged need for police questioning as a tool for the effective enforcement of criminal laws,” while at the other is “the set of values reflecting society’s deeply felt belief that the criminal law cannot be used as an instrument of unfairness.” Id. at 225, 93 S.Ct. 2041. But such a rule is also inherently ambiguous and, as the Court acknowledged in crafting the bright-line Miranda rule, the totality-of-the-circumstances test is challenging for law enforcement officers. Dickerson v. United States, 530 U.S. 428, 444, 120 S.Ct. 2326, 147 L.Ed.2d 405 (2000).
In order to respect the “acknowledged need for police questioning as a tool for the effective enforcement of criminal laws,” we must craft clear rules that do not place officers in constant fear of condemnation for deploying the mildest of tactics. In the course of investigating reported crimes, officers necessarily have to approach most suspects without any information regarding their intellectual capabilities and necessarily have to employ tactics similar to those that the officers used here. Those tactics may, in certain circumstances, coerce a highly sensitive suspect, but that determination must be made on a case-by-case basis by looking at the totality of the circumstances at hand. See Doody, 649 F.3d at 1008 (“The Supreme Court has observed that, ‘[t]he application of these principles involves close scrutiny of the facts of individual cases.’ ” (alteration in original) (emphasis omitted) (quoting Gallegos v. Colorado, 370 U.S. 49, 52, 82 S.Ct. 1209, 8 L.Ed.2d 325 (1962))).
II.
Our disagreements aside, the majority and I reach the same conclusion: that Preston confessed involuntarily. But contrary to the majority, I do not think that we must condemn every interview tactic that the officers used or embark on an exploration of empirical literature on intellectual disabilities. Rather, I think that this case is resolved squarely by our earlier precedent that prohibits false promises of confidence in exchange for confessions.
The Supreme Court has held that the test for voluntariness of a confession “is whether the confession was extracted by any sort of threats or violence, or obtained *1032by any direct or implied promises, however slight, or by the exertion of any improper influence.” Hutto v. Ross, 429 U.S. 28, 30, 97 S.Ct. 202, 50 L.Ed.2d 194 (1976) (per curiam) (internal quotation marks and brackets omitted). Although we have held that general allusions to leniency are insufficient to render a confession involuntary, United States v. Coleman, 208 F.3d 786, 791 (9th Cir.2000); United States v. Leon Guerrero, 847 F.2d 1363, 1366 (9th Cir. 1988), we also have held that statements by officers conveying that the interview is in complete confidence render any resulting confession coerced, Henry v. Keman, 197 F.3d 1021, 1027 (9th Cir.1999).
The officers’ suggestions that any incriminating statement would be kept in the United States Attorney’s file and that the purpose of the interview was to extract an apology are troubling. Indeed, those suggestions and that appeal to culturally appropriate closure, in my view, overbore Preston’s will. As the record reveals, Preston remained incredibly reserved and circumscribed in the facts that he chose to share with the officers until after their promise of confidence. Immediately following the promise of confidence, Preston agreed with the officers’ version of the incident:
[Officer Greg]: I mean we’re not going to make a judgment; are you — on you or — you know. We’re just here. We do this day in and day out. We talk to people left and right, and we don’t say anything about it.
We don’t tell this to anybody. It stays with the folder, and it stays with the U.S. Attorney’s Office and that’s it. So what happened Friday?
[Officer James]: So you were inside in there. Did the kids come inside then?
[Preston]: No, nobody came inside. It’s just like what you guys said, that guy came in, but I didn’t do nothing.
(Emphasis added.)
Aso problematic is the fact that the officers perpetuated the misimpression that the interview was confidential by describing the form as a mere apology note to the child (who is Preston’s second cousin and a member of a neighboring family that was feuding with his own) and by securing Preston’s signature under that misimpression:
[Officer Greg]: Uh-huh. Do you feel sorry?
[Preston]: I guess.
[Officer Greg]: No. I mean, it’s either yes or no, one — I mean, there’s nothing in between.
[Preston]: Yeah.
[Officer Greg]: You’re sorry? And, you know, if — do you have — you want to say anything to the kid?
[Preston]: Who?
[Officer Greg]: To [the child]?
[Preston]: No.
[Officer Greg]: No?
[Officer James]: Do you want to write any- — -usually what we do is we write a statement and like if you wanted to say you’re sorry or something like that, you could — you could definitely do that, and we can provide that to him.
“Such misleading comments were intended to convey the impression that anything said by [Preston] would not be used against him for any purposes.” Henry, 197 F.3d at 1027-28. Those comments, coupled with Preston’s intellectual disability, his youth, and the particular cultural and jurisdictional4 context in which the *1033officers made the statements, overcame Preston’s will and tricked him into confessing. “Because the police tactics and trickery produced a confession which was neither rational nor the product of an essentially free and unconstrained choice,” Henry, 197 F.3d at 1028, the confession was rendered involuntary.
For these reasons, I conclude that the district court erred in admitting Preston’s confession at trial. I therefore concur in the judgment reversing the conviction and remanding for a new trial.
. I also concur fully in Part V of the majority opinion and with the majority's holding in Part II.E. that we must overrule Derrick v. Peterson, 924 F.2d 813, 818 (9th Cir.1991).
. The majority makes much of an officer's comment to Preston that, "[ajfter the interview, you’re free to go.” Maj. op. at 1013, 1026. Not only does the majority's broad interpretation of that phrase disregard the officers' two earlier statements articulating clearly that Preston was not under arrest and was free to leave and free to stop talking at any time, but the majority also reads the phrase out of context. Immediately following the quoted comment, the officers reaffirmed their meaning: "We're not here to put the handcuffs on you today....” Rather than communicating to Preston that he was currently not free to leave, as the majority suggests, the statements that Preston was not going to be arrested after the interview and that he would be free to leave after the interview merely served to affirm that he would continue to be free to leave.
. A close analysis of the interview transcript reveals that the officers asked Preston approximately 17 either-or questions over the course of a 38-minute interview that included dozens of non-either-or questions. Moreover, those 17 questions, contrary to the majority’s summary treatment of them as homogenous, fall into two distinct categories: (1) two either-or questions regarding repeat offender status; and (2) a series of either-or questions to clarify steps in Preston’s narrative of the incident.
The officers asked only two of the 17 questions before Preston began volunteering information regarding the incident. Those two either-or questions involved inquiring into whether Preston was a routine assaulter or whether the alleged event was a one-time occurrence. Contrary to the majority, I did not interpret Preston’s response to those questions as confused acquiescence. Preston responded repeatedly in the negative or with "I don't know,” and at no point did he acquiesce in response to those questions. Rather, what I find troublesome about those questions is not that they trapped Preston into a binary answer — an answer that he notably did not provide — but that the officers coupled their *1031description of a one-time offender with an implication that such a person would not be prosecuted. I will discuss this issue below.
Second, the 15 either-or questions that the officers asked after Preston began to volunteer information regarding the incident were either open-ended or were necessarily binary. The majority speculates that answering "neither one” to such questions "could well have exceeded [Preston's] intellectual abilities.” Maj. op. at 1024. The record does not support the majority’s speculations. For example, in response to at least three of those either-or questions, and arguably more, Preston responded in the negative to both options *1032offered him, or answered that he did not know.
. See, e.g., United States Government Accountability Office, U.S. Department of Justice Declinations of Indian Country Criminal Matters, at 3 (Dec. 13, 2010) (finding that United States Attorney offices had declined to prosecute "67 percent of sexual abuse and related matters” in Indian Country for fiscal *1033years 2005 through 2009). Because of those very low rates of prosecution, which are widely known, Preston may have been more in-dined to believe the officers’ promises of confidentiality.