One of the hallmarks of the American criminal justice system is the importance it attaches to the protection of the rights of criminal defendants. Just because someone is charged as a defendant does not mean that he committed the crime in question; that simple truth lies behind the presumption of innocence and the many procedural rules that combine to assure that due process is observed. In this direct criminal appeal, we confront one of the most disturbing claims that can be brought: Larry Hall asserts that the government got the wrong man when it convicted him for the kidnapping (and murder) of Jessica Roach, which occurred in late September 1993. He tried to show this through the use of certain expert testimony, that would have put before the jury his claim that his "confession" to the crime was actually false. Because we believe that the district court erred in excluding this testimony, we reverse and remand for a new trial.

* Fifteen-year-old Jessica Roach was last seen in the afternoon of Monday, September 20, 1993, riding her new bicycle near her home in rural Georgetown, Vermilion County, Illinois. About one-half hour later, her sister saw Jessica's bicycle lying in the roadway. Concerned, she returned home, and her father called the police. About six weeks later, on November 8, 1993, Jessica's decomposed body was found across the state line, near Perrysville, Indiana, mutilated by a farmer's combine. Due to the extensive damage caused by both the machine and the passage of time, it was impossible for investigators to be sure of the cause of death. The police were hampered in their efforts to solve the crime because of the total absence of physical evidence either at the scene of her disappearance or where her body was found.

Months later, on October 22, 1994, two girls in Georgetown reported that they had been followed by a man in a van. He asked them questions as he drove by, but he did not attempt to get out of the vehicle. Nevertheless, they found him suspicious, and they retreated to a house where they could call for help. They also obtained his license plate number. The man turned out to be Larry Hall. Following up on the girls' report, Vermilion County Sheriff's Investigator Gary Miller went to Hall's home in Wabash, Indiana, on November 2, 1994, to question him about the incident. Hall confirmed that he had stopped and spoken to the girls, but he denied doing anything more. The police realized that Hall was the man who had also followed two other teenaged girls on May 29, 1994, in Georgetown. Those two were able to elude Hall by riding their bicycles through an alley that was blocked to motor traffic by a truck. It became apparent that Hall had a habit of stalking, or following, teenaged girls. Other incidents (some of which Hall admitted, others of which he contested) occurred on April 6, 1993, in Marion, Indiana; on March 30, 1994, in Gas City, Indiana; on May 29, 1994, in Perrysville, Indiana; and on July 24, 1994, in Logansport, Indiana.

Hall's November 2, 1994, interrogation was not the first time he had encountered the police in conjunction with his stalking activities. Back in March 1994, when he was stopped in Gas City, Indiana, police officers had questioned him and searched his van. They found a number of suspicious objects in it, including a knife, a can of starter fluid, a mask, a bundle of rope, and a flyer about the disappearance of a young woman named Tricia Reitler. This caused them to turn Hall over to Sgt. Bruce Bender of Marion, Indiana, who was responsible for the Reitler case. Reitler had been a student at nearby Indiana Wesleyan University who had vanished from the campus on March 29, 1993. The Reitler case had received extensive publicity in the Marion newspapers. On that occasion, however, the police learned nothing from Hall that led them to believe he was involved in the Reitler disappearance.

Between March 1994, and November 15, 1994, a Detective Phil Amones of the Wabash Police Department had several conversations with Hall. Apparently he realized that Hall had mental health problems, because at his recommendation Hall agreed to see a therapist at the Otis R. Bowen Center, a mental health facility in Wabash. Amones kept in touch with the treating therapist and provided him with information about the accusations concerning Hall's propensity to bother young girls. The therapist, in turn, shared his assessment of Hall's condition with Amones, and Amones kept other local law enforcement agencies informed about Hall's treatment. While it is unclear whether every person involved in the Roach case knew about his counseling at Bowen, many did.

It was Sgt. Amones who asked Hall to come to the Wabash Police Department on November 2, 1994. He told Hall that someone from Illinois wanted to speak with him, and that person turned out to be Officer Miller. Miller began by questioning Hall about the October incident in Georgetown, but he soon turned to the Jessica Roach case. (It is worth noting that Miller had, by this time, taken a statement from another individual who had confessed to abducting Roach and dumping her body in the cornfield. Transcript at 455-56. This fact was not developed at the trial.) Hall initially denied ever seeing Jessica. Some of the evidence indicated that Miller became upset with Hall's responses, moved closer to Hall, and started suggesting the "right" answers as the questioning progressed. The government does not directly dispute this, although it claims that no one physically abused Hall. Hall began crying at some point, and he asked Amones during a break what was expected of him. After two and a half hours or so, the police allowed Hall to leave.

On November 15, the police told Hall that they needed to question him again. This session lasted from around 10:00 a.m. that morning until about 3:20 a.m. the next morning. By now, the FBI was involved in the investigation. Hall refused to take a polygraph test that FBI Agent Randolph was prepared to administer. Randolph interrogated him alone for about two hours. About that time, Hall began to talk about Jessica Roach; another 20 or 30 minutes later, he began making admissions about his involvement in the Roach case. There were no notes, tape recordings, or video recordings of the session. Instead, Randolph wrote out a statement in narrative format and asked Hall to sign it. Miller was present for at least part of this questioning, and continued to talk with Hall for an hour or so after the statement was signed. It is a little difficult to make the hours here add up, but the record shows that Hall was booked into the Grant County Jail at 3:25 a.m. on November 16, 1994.

On December 21, 1994, Hall was charged in a one-count indictment with the offense of kidnapping Jessica Roach for purposes of sexual gratification and transporting her from Illinois to Indiana, in violation of 18 U.S.C. § 1201(a)(1). The district court denied Hall's motion to suppress his confession. After an eight-day trial, the jury convicted him. The court denied his voluminous motion for new trial (alleging 77 errors by the trial court!) and sentenced him to a term of life imprisonment. On appeal, Hall makes three arguments: first, the trial court should have suppressed his confession; second, the court erroneously admitted evidence of other crimes under Rule 404(b) (in particular, the Reitler case); and third, the court erred in refusing to permit Hall's experts to testify about false confessions and his susceptibility to coercion.

At the trial, Hall's entire theory of defense boiled down to a simple proposition: due to a personality disorder that makes him susceptible to suggestion and pathologically eager to please, he "confessed" to a crime that he did not really commit, in order to gain approval from the law enforcement officers who were interrogating him. He was, in the words of the Wabash police, a "wannabe." At the trial, one psychologist (Greg Helgesen) and one psychiatrist (Arthur Traugott) testified about his mental and emotional problems. Social worker Timothy Myers, who had treated Hall at the Bowen Center, testified about the fact of his treatment, but the district court did not permit him to offer expert testimony about Hall's condition. Most importantly, the district court significantly limited the scope of Dr. Traugott's testimony, and it excluded altogether another expert. These two decisions are the focus of Hall's appeal.

Hall tendered Dr. Richard Ofshe as a social psychologist expert in the field of coercive police interrogation techniques and the phenomenon of false or coerced confessions. Ofshe had been on the faculty of the University of California at Berkeley since 1962, had a Ph.D. in psychology from Stanford University, and had published widely. He had also worked extensively both with law enforcement officials and with defense counsel. He indicated that he would have testified about the fact that experts in his field agree that false confessions exist, that individuals can be coerced into giving false confessions, and that certain indicia can be identified to show when they are likely to occur. He described his methodology in general terms, and what factors experts in the field rely upon to distinguish between reliable and unreliable confessions. The district court rejected the proffer of Dr. Ofshe's testimony in its entirety, on two grounds: (1) Dr. Ofshe would need to judge the credibility of Randolph's and Miller's testimony about what happened during the interrogation of Hall, and (2) in the final analysis, Dr. Ofshe's testimony would add nothing to what the jury would know from common experience.

Hall also proffered testimony from Dr. Traugott, a psychiatrist who had examined him. Like Dr. Ofshe, Dr. Traugott's credentials as a psychiatrist were not in question. The district court allowed Dr. Traugott to testify about Hall's mental condition (e.g., his attention-seeking behavior and his high level of suggestibility) and to opine that one of the problems for someone interrogating Hall (including himself, as a professional) was that Hall could easily be led to give the type of response he believed the questioner was seeking. Hall's appeal focuses on other testimony that he wanted Dr. Traugott to offer, which the district court refused to allow. The proffer shows that Dr. Traugott would have testified about Hall's susceptibility to various interrogation techniques, the propriety of suggesting answers to Hall, and Hall's capability of confessing to a crime that he did not commit. The court found that the jury could appreciate whether police interrogation techniques were suggestive by themselves and that Dr. Traugott's testimony would invade the prerogative of the jury to assess Randolph's and Miller's credibility.

In Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993), the Supreme Court established the approach a trial judge must take when faced with a Rule 104(a) proffer of expert scientific testimony, for possible admission under Rule 702. The judge must determine "whether the expert is proposing to testify to (1) scientific knowledge that (2) will assist the trier of fact to understand or determine a fact in issue." Id. at 592, 113 S.Ct. at 2796. With respect to the first question, the Court noted that a number of factors would bear on the scientific validity of the testimony, but it disclaimed any intention of creating a rigid or exclusive list. The inquiries that the Court singled out for special attention were (1) whether the theory or technique can be and has been tested, (2) whether it has been subjected to some form of peer review or scrutiny within the scientific community, and (3) the reliability of the technique, including where applicable the potential rate of error and the general acceptance it enjoyed within the relevant scientific community. Id. at 593-94, 113 S.Ct. at 2797. With respect to the inquiry as a whole, the Court stressed that other applicable parts of the Federal Rules of Evidence had to be considered, including Rules 703, 706, and 403. Id. at 595, 113 S.Ct. at 2797-98.

This court considered Daubert at length in Porter v. Whitehall Laboratories, Inc., 9 F.3d 607 (7th Cir.1993). After reviewing the first part of Daubert, which essentially relates to the scientific validity of the expert's methodology, it explained that the second part is essentially a relevance inquiry. Id. at 613. An expert's credentials and methodology may be impeccable, but if the proffered testimony fails the general test of relevance under Rule 402, or if, in the words of Rule 702 it is not likely to "assist the trier of fact to understand the evidence or to determine a fact in issue," then the district court should reject the proffer. When a question about the acceptance or rejection of expert testimony arises on appeal, this court first undertakes a de novo review of whether the district court properly followed the framework set forth in Daubert. See Bradley v. Brown, 42 F.3d 434, 436 (7th Cir.1994). If we conclude that the court did so, we will not disturb its findings unless they are clearly erroneous. See id. at 436-37, citing Cella v. United States, 998 F.2d 418, 423 (7th Cir.1993).

In this case, we cannot be confident that the district court applied the Daubert framework. The judge never mentioned Daubert specifically, and thus he never focussed on the individual questions that must be answered. The only thing that is clear is his conclusion that the testimony would not assist the jury in its task. (More precisely, he concluded that their testimony would in some respects usurp the jury's role and thus that it would not properly assist the jury in its task.) This conclusion, as we explain below, appears to have been based on a misunderstanding of the "helpfulness" aspect of Rule 702.

In evaluating a Rule 104(a) proffer of expert testimony, the ultimate test is whether the testimony would assist the trier of fact in understanding the evidence. See, e.g., United States v. Solis, 923 F.2d 548 (7th Cir.1991). If the expert testimony would be helpful and relevant with respect to an issue in the case, the trial court is not compelled to exclude the expert just because the testimony may, to a greater or lesser degree, cover matters that are within the average juror's comprehension. See Weinstein's Evidence, para. 702 at 702-20 to 702-21 (1996); Michael H. Graham, Federal Practice and Procedure--Evidence, § 6644 at 265-66 (Interim edition 1992). As Graham summarizes it, the court makes two preliminary determinations when it evaluates a Rule 104(a) proffer of expert testimony:

First, the court must decide whether expert testimony could assist the trier of fact in understanding the evidence or determining a fact in issue. The court may be required as an aspect of this inquiry to determine whether a sufficiently reliable body of scientific, technical, or other specialized knowledge has been developed. Second, the court must also determine whether the witness called is properly qualified to give the testimony sought.

Graham, § 6641 at 242-44.

The first question the court should address is whether the proffer demonstrated that a sufficiently reliable body of specialized knowledge existed. Social science in general, and psychological evidence in particular, have posed both analytical and practical difficulties for courts attempting to apply Rule 702 and Daubert. See, e.g., C. Robert Showalter, "Distinguishing Science from Pseudo-Science in Psychiatry: Expert Testimony in the Post-Daubert Era," 2 Va. J. Soc. Pol'y & L. 211 (1995); David L. Faigman, "The Evidentiary Status of Social Science Under Daubert: Is It 'Scientific,' 'Technical,' or 'Other' Knowledge?," 1 Psychol. Pub. Pol'y & L. 960 (1995). Notwithstanding these difficulties, however, social science testimony is an integral part of many cases, ranging from employment discrimination actions, to family law matters, to criminal proceedings. As such, whether it is hard to do or not, courts must apply the rules of evidence to these experts as faithfully as they can.

Because the fields of psychology and psychiatry deal with human behavior and mental disorders, it may be more difficult at times to distinguish between testimony that reflects genuine expertise--a reliable body of genuine specialized knowledge--and something that is nothing more than fancy phrases for common sense. It is nevertheless true that disorders exist, and the very fact that a layperson will not always be aware of the disorder, its symptoms, or its consequences, means that expert testimony may be particularly important when the facts suggest a person is suffering from a psychological disorder. Suppose, for example, it were relevant for a jury to decide whether a person's use of foul or abusive language was intended to harm another person. Most of the time, the jury would be able to assess the circumstances without the need for expert testimony, since foul language is an unfortunate part of everyday life. In some cases, however, the individual might be suffering from Gilles de la Tourette's syndrome, which is a rare disorder manifested by grimaces, grunts, and in about half of all cases, episodes of the use of foul language. AMA Encyclopedia of Medicine at 487 (1989). See also The Encyclopedia of Mental Health, Ada P. Kahn & Jan Fawcett, eds., at 375-76 (1993). A defendant wishing to explain his behavior by showing that he had Tourette's syndrome would need expert testimony both on the condition itself and his own affliction. In other cases, the question whether a person has voluntarily joined certain activity may be central. If a possible explanation is that the person is suffering from post-traumatic stress disorder, the jury would need expert testimony to allow it to take this possibility into account. See, e.g., United States v. Winters, 729 F.2d 602 (9th Cir.1984); Encyclopedia of Mental Health at 300-01.

Even if it is clear that the field in general qualifies for expert testimony, the proffered testimony may or may not be based upon the expert's special skills. In United States v. Benson, 941 F.2d 598 (7th Cir.1991), this court made the straightforward point that "[a]n expert's opinion is helpful only to the extent the expert draws on some special skill, knowledge, or experience to formulate [his] opinion; the opinion must be an expert opinion (that is, an opinion informed by the witness's expertise) rather than simply an opinion broached by a purported expert." Id. at 604. Unless the expertise adds something, the expert at best is offering a gratuitous opinion, and at worst is exerting undue influence on the jury that would be subject to control under Rule 403. In the Benson case itself, IRS Agent Cantzler had testified that certain payments were fees for investigative services (hence taxable), not money received in a settlement. The court found that this amounted to nothing more than Cantzler's personal opinion about whose account of the deal was believable. He did not, for example, need to unravel a complex transaction so that the jury could understand it, or to explain an arcane part of tax law to them. Concluding that the jury was just as capable of deciding whom to believe as Cantzler, and that the inclusion of his testimony might have unduly influenced them, the court reversed Benson's conviction. Id. at 605-06.

In the proper circumstances, experts in psychiatry and psychology can meet both these hurdles: real science, and testimony based thereon. In United States v. Shay, 57 F.3d 126 (1st Cir.1995), for example, the First Circuit addressed a district court's decision to exclude expert testimony on Munchausen's Disease, a mental disorder (known formally as "pseudologia fantastica") characterized as an extreme form of pathological lying. 57 F.3d at 129 & n. 1. The defendant, Shay, Jr., had been accused of helping to build a bomb that had been placed under his father's car and that had detonated when police officers called by Shay, Sr. were attempting to dismantle it, killing one officer and wounding the other. Shay, Jr. wanted to argue to the jury that his statements after the fact, taking responsibility (in many ways, inaccurately) for the incident, were the result of his mental disorder. The district court excluded the testimony primarily because it believed that the jury could consider the reliability of Shay, Jr.'s statements without it.

The First Circuit reversed. It interpreted this court's decision in Benson to stand for the limited proposition that "an expert's opinion that another witness is lying or telling the truth is ordinarily inadmissible pursuant to Rule 702 because the opinion exceeds the scope of the expert's specialized knowledge and therefore merely informs the jury that it should reach a particular conclusion." 57 F.3d at 131. It found no categorical reason to exclude expert testimony that bears on truthfulness in the Federal Rules of Evidence, noting instead that Rules 608(a) and 405(a) contemplate that truthful or untruthful character may be proven by expert testimony. The expert in Shay was prepared to testify that Shay, Jr. suffered from a mental disorder that caused him to make grandiose statements similar to those the government was trying to use against him. The court of appeals found that the jury was "plainly unqualified to determine without assistance the particular issue of whether Shay, Jr. may have made false statements against his own interests because he suffered from a mental disorder." 57 F.3d at 133 (emphasis in original). The expert, in short, would have given them a reason to reject the common sense evaluation of the facts that they would otherwise be entitled to use.

As noted earlier, the district court is not compelled to exclude all expert testimony that may in some way overlap with matters within the jury's experience. The test of Rule 702 is whether the testimony will assist the jury. If it meets that test, the district court may still use the normal controls on scope of testimony and relevance that are available to it. If the proffered testimony duplicates the jury's knowledge, Rule 403 might counsel exclusion of the expert testimony to avoid the risk of unduly influencing the jury. Or the court might conclude that certain matters go beyond the individual's expertise. In United States v. Whitted, 11 F.3d 782 (8th Cir.1993), a sexual abuse case, the court found that the expert doctor could testify about the alleged victim's bedwetting, nightmares, and inability to sleep at night. Id. at 787. On the other hand, his diagnosis of "repeated sexual abuse" went too far. Id. at 786. Since his diagnosis rested solely on his acceptance of the victim's account, the testimony amounted to nothing more than an invitation to the jury to believe his assessment of the victim's truthfulness. Conclusory statements without any explanation why the expert can contribute to the jury's understanding of the subject are also subject to exclusion. See Miller v. Pate, 300 F.2d 414, 420 (7th Cir.1962). Even though experts are entitled to give their opinion on an ultimate issue in the case, see Rule 704(a), this does not mean that the opinion may be given divorced from the scientific, medical, or other expert basis that qualified the witness in the first place. See also United States v. Reno, 992 F.2d 739, 743 (7th Cir.), cert. denied, 510 U.S. 971, 114 S. Ct. 458, 126 L. Ed. 2d 390 (1993) (testimony regarding whether particular characteristics of mental disease render defendant able to appreciate the wrongfulness of his actions is admissible); United States v. Roberts, 887 F.2d 534, 536 (5th Cir.1989) (expert psychiatric testimony about defendant's naive and autocratic personality traits is admissible to demonstrate defendant claiming reverse-sting defense did not intend to violate the law); United States v. Newman, 849 F.2d 156, 165 (5th Cir.1988) (expert psychiatric testimony is admissible to demonstrate that a mental disease, defect, or subnormal intelligence makes a defendant peculiarly susceptible to inducement when an entrapment defense is raised).

Within this general framework, each case necessarily turns on its own facts. It is important in this connection to note that once the judge decided that Hall's confession was voluntary, the jury was entitled "to hear relevant evidence on the issue of voluntariness and [the trial judge was to] instruct the jury to give such weight to the confession as the jury feels it deserves under all the circumstances." 18 U.S.C. § 3501. By analogy, it was certainly within the jury's province to assess the truthfulness and accuracy of the confession. In Hall's case, because the prosecutor did not challenge the scientific basis of the proffered testimony, we assume for the sake of argument that it was valid, without, of course, prejudging the outcome of that inquiry in any further proceedings. The court indicated that it saw no potential usefulness in the evidence, because it was within the jury's knowledge. This ruling overlooked the utility of valid social science. Even though the jury may have had beliefs about the subject, the question is whether those beliefs were correct. Properly conducted social science research often shows that commonly held beliefs are in error. Dr. Ofshe's testimony, assuming its scientific validity, would have let the jury know that a phenomenon known as false confessions exists, how to recognize it, and how to decide whether it fit the facts of the case being tried.

The district court's conclusion therefore missed the point of the proffer. It was precisely because juries are unlikely to know that social scientists and psychologists have identified a personality disorder that will cause individuals to make false confessions that the testimony would have assisted the jury in making its decision. It would have been up to the jury, of course, to decide how much weight to attach to Dr. Ofshe's theory, and to decide whether they believed his explanation of Hall's behavior or the more commonplace explanation that the confession was true. See 18 U.S.C. § 3501. But the jury here may have been deprived of critical information it should have had in evaluating Hall's case.

In these circumstances, we cannot conclude that the court's failure to conduct a full Daubert inquiry, applying the correct legal standards under Rule 702, was harmless error. Dr. Ofshe's testimony went to the heart of Hall's defense. The confession itself, it is important to recall, was written down by Agent Randolph and simply given to Hall to sign. When we asked the government at oral argument what new information about the Roach murder the police learned during the interrogation, counsel replied that Hall knew how, where, and when Jessica Roach was murdered. On rebuttal, Hall's counsel explained that Hall "knew" Roach was strangled. In fact, the forensic experts were unable to confirm strangulation or any other particular cause of death, because her body was found after it had been badly decomposed and decapitated by a farm harvester. Without corroborating evidence, it is possible that Hall may have just picked a common method of killing from the stories with which he was obsessed. Hall placed a pin on a state-wide map identifying the area where her body was found. He knew, however, that this was the area where Miller came from (with whom he had spoken two weeks before the confession, and who he knew was interested in the Roach case). Finally, he named a day during the four-day period when the state believes the murder was committed. The government is entitled to argue its version of these facts to a jury, in support of its theory about the validity of the confession, but Hall was entitled to present his own theory to them as well, including the likelihood that the "confession" added nothing to what the government already knew.

We have the same concerns about the court's decision with respect to Dr. Traugott, although this is a closer case, because the judge permitted him to testify about Hall's mental condition. He was also allowed to testify that someone interrogating Hall would experience difficulty obtaining reliable answers, because Hall was easily led. On remand, it is clear that at least this much would be admissible again from Dr. Traugott. But more was needed to link the condition Dr. Ofshe identified to Hall himself, and Dr. Traugott was prepared to do this. Based on his personal examination of Hall and Hall's medical records, he would have testified about Hall's susceptibility to various interrogation techniques and his propensity to give a false confession. Even if the testimony about propensity touched on a key issue in the case, experts are entitled under Rule 704(a) to offer an opinion on the ultimate issue in a case. The fact that there was a dispute between Hall and the interrogating officers about the nature of the questioning itself provides no reason to exclude the expert testimony; it is a rare case where everything is agreed except the subject matter for which the expert is presented. It is enough if the expert makes clear what his opinion is, based on the different possible factual scenarios that might have taken place.

Even though the exclusion of the expert testimony is enough to require a remand, we think it best to address Hall's other two arguments now. He claims that his incriminating statements were the product of police coercion and were therefore involuntary for that reason. We apply a totality of the circumstances standard to these claims. See United States v. Montgomery, 14 F.3d 1189, 1194 (7th Cir.1994). If a defendant has a mental impairment that was known or reasonably apparent to the interrogators, they must exercise special care. United States v. Sablotny, 21 F.3d 747, 752 (7th Cir.1994). Here, the district court credited the testimony of Miller and Randolph to the effect that they did not have special knowledge of Hall's condition and that in any event they did not threaten him, make misrepresentations or attempt to deceive him, or structure their interrogation to take advantage of him. Giving due deference to the district court's ability to observe the witnesses, we see nothing in the record to cast doubt on this conclusion.

Second, Hall complains that the district court erred in allowing the government to introduce "other crimes" evidence under Rule 404(b). In his statement, he admitted to abducting and murdering Tricia Reitler and two other unnamed girls. This evidence came in, along with evidence of several of his stalking incidents. In reviewing the admission of evidence under Rule 404(b), this court considers whether "(1) the evidence is directed toward establishing a matter in issue other than the defendant's propensity to commit the crime charged; (2) the evidence shows that the other act is similar enough and close in time to be relevant to the matter in issue; (3) there is sufficient evidence to support a finding by the jury that the [defendant] committed the similar act; and (4) the probative value of the evidence is not outweighed by the danger of unfair prejudice." United States v. Lampkins, 47 F.3d 175, 179 (7th Cir.), cert. denied, --- U.S. ----, 115 S. Ct. 1440, 131 L. Ed. 2d 319 (1995).

The district court reasonably could have concluded that elements (1) and (2) of this test were satisfied. We are troubled, however, by element (3) for many of the same reasons that we are concerned with the exclusion of the expert testimony for the crime charged. The Reitler case had been discussed extensively in the press, and Hall's statement contained facts that could have been gleaned entirely from news reports. It offered nothing that could have been known only to the guilty party. The government presented no other evidence linking Hall to the Reitler murder or the other two murders, either at the district court or before this court. In another rejected proffer of testimony, Hall presented statements from the Chief of Police of Marion and a representative of the Grant County Sheriff's Department, in which they both said that Hall was not a suspect in the Reitler case, and in which they both specifically named a different suspect. Considering the potential unreliability of the confession, and the massive prejudice that inevitably attends evidence suggesting Hall had committed three similar murders, we conclude that the district court should not have admitted that portion of Hall's statement. (The evidence of stalking, of course, was well documented and clearly admissible to show intent, preparation, or plan under Rule 404(b).)

We regret the necessity of putting all involved in this case through another trial. Nevertheless, Hall was entitled to present his defense to the jury with the use of expert testimony meeting the standards of Rule 702 and Daubert. The district court's failure to test the Ofshe and Traugott proffers under that framework may have led to the exclusion of critical testimony for Hall. A jury with all the evidence before it might have convicted, but it might have concluded that Hall was a "wannabe," and that the true kidnapper/murderer is still at large. We accordingly VACATE the conviction and REMAND for further proceedings consistent with this opinion.