Four young Native American men face combined sentences of more than 120 years for alleged child abuse. These convictions rest upon testimony of young children, the alleged victims, with some support from findings on medical examinations of the children.

These defendants are Jesse Rouse, Desmond Rouse, Garfield Feather and Russell Hubbeling. They and a fifth defendant, Duane Rouse, who was acquitted by the jury, faced twenty-three counts of aggravated sexual abuse of children under the age of twelve years in violation of 18 U.S.C. Section(s) 2241(c). The events allegedly occurred at family residences on a South Dakota Indian Reservation.

The jury found Jesse Rouse guilty of two counts of sexual abuse; Desmond Rouse guilty of three counts; Garfield Feather guilty of four counts and Russell Hubbeling guilty of two counts. These counts related to alleged abuse of five young Native American children. The children are referred to by initials in the text of this opinion. The jury acquitted the defendants of the remaining charges.

The appellants raise twelve allegations of error in the trial of the case.1 We grant relief on two issues: (1) refusal to allow expert opinion testimony by a court appointed psychologist that the children's evidence and testimony became tainted by suggestive influences to which the children were subject in the investigation and trial, which influences included taking the children (the alleged victims and nine other children) from their families and from their residences and (2) denial by the trial court of the defendants' motion for independent psychological examination of the allegedly abused children -- in light of the circumstances of the case.

Accordingly, the appellants are entitled to a new trial on these grounds.

Sufficiency of the evidence is not an issue. However, an examination of the record establishes that the medical evidence was inconclusive as to abuse or abuse by the defendants and that the children's reports of abuse may have been tainted by the influence of social workers and law enforcement officials who investigated and prepared the government's case.

The crucial issue for determination by the jury was this -- did the young children testify from their own memory of events or was a false memory induced during investigation by the methods by which those children were interrogated? Some of the evidence presented in the case suggests that the children may have had induced memories that sexual abuse occurred. While that issue remained for the jury, the jury evaluated the children's evidence of abuse without the benefit of a qualified defense child abuse expert who would have assisted the jury by explaining that the children had been subjected by state investigators to "powerful and coercive influences."

We briefly address the merits of some of the other errors alleged by the defendants. This commentary may be helpful with respect to the retrial of this case. See United States v. Azure, 801 F.2d 336, 341 (8th Cir. 1986).

First, the trial court's discretionary ruling which denied defendants a further independent medical examination leaves inconclusive the abuse conclusions resting on medical findings. This uncertainty in the evidence is further accentuated because the physician did not document findings with photographs.

Second, in light of the background of this case, the district court's exclusion of testimony relating to inter-child sexual activity among the alleged victims and other children on the reservation deprived the defendants of important evidence indicating another possible source relating to any physical manifestations of abuse. This issue may arise in the new trial.

Third, because of the possible influences on the memories of the children by social services personnel and other investigators and the lack of access to the children, the district court prejudicially erred in refusing to authorize an independent psychological examination.

Fourth, although not reversible error, the allegations of one juror's racism and that juror's contention that racial jokes regarding Native Americans were told in the jury room are troubling.

With respect to conditions that can influence children's memories, we are mindful of a historical event of some three hundred years ago (the Salem, Massachusetts witch trials) where child witnesses ages five to sixteen (the "circle girls") claimed to see persons (the defendants) flying on broomsticks and other envisaged celestial apparitions. Based on this testimony, nineteen alleged witches were put to death and a dozen others avoided executions by testifying to witchery, that which was not.2

This case, of course, is not a Salem witch hunt, but that history must remind us that memory, particularly children's memory, may be falsely induced. Where that occurs, the testimony may be true in the child's mind, but false in fact.

Five-year-old R.R. lived with her grandmother Rosemary Rouse on the Yankton Sioux Reservation in Marty, South Dakota, during the summer and fall of 1994. R.R. had been taken from her mother and placed in her grandmother's custody. She had been unhappy about living with her grandmother and had wanted to stay with her mother. After R.R. told a teacher that her grandmother was mean to her and was not feeding her, the Yankton Sioux Tribe's Department of Social Services (Department) removed R.R. from Rosemary Rouse's home for possible neglect and malnutrition and placed her with foster mother Donna Jordan.3

After living with her foster mother for several months, R.R. told Jordan that she was having nightmares and that she had been sexually abused. Jordan scheduled an appointment with counselor Ellen Kelson. After interviewing R.R. -- an interview which she did not audio or video tape -- Kelson immediately contacted the Department and reported that a number of children at the Rouse residence had been sexually abused.

The next day, on January 11, 1994, apparently without any additional evidence or investigation, the Department removed approximately thirteen children from the Rouse home and a nearby home.4 According to the evidence, squad cars pulled up and the children were physically removed while they cried and clung to their uncles' and other adults' legs. Jean Brock, a social worker for the Department, transported the children to Jordan's foster home, and told the children it was their uncles' fault that they were being taken away because the uncles were doing "bad things" to them. On her initial intake sheet, Brock noted that "the children love the adults," but that the home was messy.

Brock and Jordan told the children they could not go home until they told the "truth" about their uncles. Both alleged victims and non-victim children testified that Brock and Jordan repeatedly told them that they were taken away from their homes because their uncles did bad things to them and that they could safely go home only after they told the "truth" about their uncles and all these bad things got fixed. In fact, T.R. remembered later telling Kelson "that to say the truth is to say that your uncles did things to you," and that "[i]f you tell the truth you get to go home." When the district court asked J.R. what it means to tell the truth at her in camera competency examination, J.R. responded, "[y]ou mean you can go home."

Despite this encouragement to accuse their uncles, many of the children repeatedly denied being abused, and approximately nine children who consistently and adamantly denied being abused were allowed to go home to their parents.5 Those children who claimed that abuse occurred were not allowed to see their parents until approximately six months later in July 1994 (just before trial), despite repeated requests by the children and parents and a tribal court order. T.R. testified that she did see her mother once during this time period after repeated requests to do so. Nevertheless, both Brock and Kelson sat in on the parent-child meeting and took notes.

Beginning in January 11, 1994, the children all lived in Jordan's home. Jordan told them this situation was not their fault; she told them their uncles were at fault and she got very specific about the "bad" things their uncles were doing to them.

Dr. Richard Kaplan, a pediatrician, initially examined the children in Jordan's presence on January 15, 1994. At that time, he could not diagnose any of the children as having been abused, and he arranged for a subsequent examination by Dr. Robert Ferrell, a woman's obstetrician and gynecologist, which examination took place a month later. That doctor reported findings consistent with sexual abuse.

After the January 15 examination and after the group of children spent over a week in Jordan's home, FBI agent William Van Roe and BIA Criminal Investigator Dan Hudspeth interviewed the children on two separate days. The agents immediately identified themselves as police officers, and Jordan and Brock sat in on the interviews. At the initial interview, R.R. handed investigator Hudspeth a group of papers which contained statements written by Jordan which R.R. had allegedly made to Jordan about the abuse.

J.R. testified at trial that investigator Hudspeth "helped" her remember some things during the interview.6 Officers showed the children an anatomical drawing of a penis. Although agent Van Roe testified he did not ask the children leading questions, he later acknowledged that investigator Hudspeth had asked most of the questions. Investigator Hudspeth did not testify. The officers did not videotape any of these interviews.

Kelson counseled the children extensively from January through July 1994. She did not videotape a single one of these sessions. She forwarded her notes directly to the United States Attorneys Office. During her sessions with the children, Kelson used play therapy, art media and apparently dream journals. She met with the children first as a group ("talk circle") and then individually. She felt very strongly that these children decided before coming to see her what they would share with her because they all came in, directed the topic, and repeated the same or similar theme.

T.R., however, denied that the children ever discussed the sexual abuse outside the counseling sessions. Kelson reported that T.R., the eldest alleged victim, served as the "boss" or "leader"; T.R. was "very manipulative" and told the other children what to do. For example, R.R., the original complainant, identified Jesse Rouse as one of her abusers only after T.R. did so during a group session. Tabatha Smith, another child who lived in the Rouse home and denied any abuse, testified that before she left Jordan's foster care, T.R. told her she was "making stuff up" about Jesse Rouse and the uncles because she was mad at Jesse.7

The children's accusations expanded over time in a somewhat bizarre fashion. They accused various family members (additional to defendants), including their grandmother, of sexually abusing them. Several alleged victims claimed their uncles tied them up with ropes during the abuse. One child claimed the uncles locked him in the closet while abusing his sister. T.R. testified that the uncles locked all five other children in the pantry while they tied her up and abused her.

L.R. and J.R. also testified at trial that they had told Kelson that the uncles had tied up their aunts and grandmother and sexually abused them as well. Evidence at trial showed the closet had no locks, the pantry contained so much food that not even one child would fit inside, and no ropes were found in the house. In addition, Melanie Rouse testified that she had previously told T.R. about an unrelated, documented sexual abuse case in which an individual in Kentucky had tied up Melanie Rouse and sexually abused her after he locked her brother in a closet.

In March 1994, the children were again interviewed by law enforcement agents, and their charges had apparently expanded fantastically. Although the defense characterized these March interviews as "rife" with inconsistent statements by the child victims and essential to the defense, the district court characterized them as "unreliable" and would not allow either party to refer to the interviews at trial.

Most of the family involved in this case and also many members of the community who testified did not believe the children were abused and had never seen evidence of abuse, injury or fear. The children's mothers testified they had not observed any acts of sexual abuse or mistreatment; their children had never reported to them that they were being hurt or otherwise abused; for the most part, the uncles did not baby-sit the children or spend significant amounts of time with them; the children did not reveal to their mothers any fear of their uncles; only twice had the children ever complained of pain or discomfort in the vaginal or anal areas (L.R. had a yeast infection and J.R. had constipation -- which can cause symptoms similar to sexual abuse); and the doctor who had examined the children on those occasions revealed no suspicions of abuse.

Before trial, the defendants moved for independent psychological and medical examinations and for access to the children. The district court denied these motions. The court excluded the defense expert's testimony that there existed in the governmental custody and treatment of the children (through the social services agency) a "practice of suggestibility." The district court noted that access to the children would have to be obtained from the proper authorities. The defendants claimed that the United States Attorneys' Office maintained it was within the Department's discretion and the Department maintained it was within the United States Attorneys' Office discretion. As a result, access to the children was effectively denied.8

The district court also excluded testimony regarding inter-child sexual activity by and between the complaining witnesses and other children on the reservation in Marty, South Dakota, the place of residence of the family here involved. Particularly, Moses Rouse, an eleven-year-old boy who lived in the home, reported to investigators that he and T.R. had had sex over a long period of time, and other child witnesses testified that Moses and T.R. had engaged in sexual relations.9

Although defendants intimated they would be presenting evidence generally of inter-child sexual activity and accusations of sexual abuse of non-defendant family members in several motions and hearings weeks before trial, they never filed a formal motion as required by Federal Rule of Evidence 412. The day before trial, the government filed a motion in limine to prevent defendants from presenting such evidence at trial because they had not filed a Rule 412 motion. The district court granted the government's motion and disallowed such evidence.

The trial took place in August 1994. The children were promised picnics, vacations, and even a chance to return home as rewards for their "truthful," successful testimony at trial. In fact, they were told they could not go home until their uncles had been successfully removed.

At trial, the children were asked almost exclusively leading questions over closed circuit television. Rather than asking the children if the abuse occurred, the government asked them whether they had told various third persons that abuse had occurred. On redirect examination, J.R. basically denied that any abuse occurred.

The jury convicted Jesse Rouse of engaging in sexual acts with T.R. and J.R. The jury convicted Desmond Rouse of engaging in sexual acts with T.R., L.R., and R.R. The jury convicted Garfield Feather of engaging in sexual acts with T.R., L.R., and J.R. The jury convicted Russell Hubbeling of engaging in sexual acts with T.R. and F.R. The jury found the defendants not guilty of the various remaining charges and acquitted the remaining defendant Duane Rouse. The district court sentenced Jessie Rouse to 33 years imprisonment, Desmond Rouse to 32 years imprisonment, Garfield Feather to 30 years imprisonment, and Russell Hubbeling to 30 years imprisonment.

After the trial, a community member called the clerk's office and reported that she worked with one of the jurors who had often expressed a serious racial prejudice against Native Americans and had revealed that jurors told jokes regarding Native Americans in the jury room. The district court held a series of evidentiary hearings on this issue, but denied defendants' motion for a new trial on this issue. The court sentenced each defendant to approximately thirty years imprisonment.10

A. EXPERT TESTIMONY ON PRACTICES OF SUGGESTIBILITY IN THE INVESTIGATION

At trial, the defense offered the testimony of Dr. Ralph Charles Underwager. Dr. Underwager is a clinical psychologist and has been practicing his profession or teaching psychology for approximately twenty years. He has conducted extensive research and writing in the area of child sex abuse and is familiar with extensive psychological research into this subject during the past ten years. His expertise has not been challenged by the prosecutor, only the substance of his testimony.

The crucial question and answer (made by offer of proof) follows:

Q And based on your review of [the trial testimony] and your review of the records, all the files in this matter, is it your belief that there's been a practice of suggestibility employed in these techniques?

A Yes, sir.

(Trial Tr. Vol. IX at 1768.) The state objected to the offer as an area "within the province of the jury and not within something that an expert should testify on." (Trial Tr. Vol. IX at 1771.)

The court rejected the offer as essentially not the subject of expert testimony and not reliable or relevant under Federal Rule of Evidence 104(a) and confusing and misleading to the jury under Federal Rule of Evidence 403. The court rejected any proposed testimony directly relating to the credibility of the alleged abused victims as witnesses, but more than that barred the expert witness from testifying on whether or not the investigative practices constituted "a practice of suggestibility."

The court erred in its analysis. The jury needed and was entitled to have this evidence in evaluating whether the sexual abuse testified to by the children actually occurred. The testimony was relevant, proper, in keeping with our case law and crucial to the defense under the circumstances of this case. The denial of that testimony constituted prejudicial error.

In Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579 (1993), the Supreme Court addressed the standards of admissibility for scientific evidence under Federal Rule of Evidence 702.11 The Court rejected the general acceptance test for novel scientific testimony from Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), and asserted flexible guidelines for admissibility of scientific evidence under Rule 702.

Under Daubert, the trial judge plays a "gatekeeping" role, ensuring that all scientific testimony or evidence admitted is both reliable and relevant. Daubert, 509 U.S. at 589 n.7, 597.

The Daubert opinion emphasized first that the expert must testify to scientific knowledge. "[T]he requirement that an expert's testimony pertain to `scientific knowledge' establishes a standard of evidentiary reliability." Id. at 590 & n.9. Knowledge "applies to any body of known facts or to any body of ideas inferred from such facts or accepted as truths on good grounds." Id. at 590 (quoting Webster's Third New International Dictionary 1252 (1986)).

The Court explained scientific knowledge in terms of a theory or technique that (1) can be and has been tested, (2) has been subjected to peer review, (3) has a known or potential rate of error (when technique is scientific), and (4) has been generally accepted by the scientific community. Id. at 593-94.

The touchstone under Rule 702 is reliability. As the opinion states, "under the Rules the trial judge must ensure that any and all scientific testimony or evidence admitted is not only relevant, but reliable." Id. at 589.

Furthermore, the knowledge must "assist" the trier of fact. That is a relevance issue. Id. at 591. The key question for the trial judge in determining relevance under Federal Rule of Evidence 104(a)12 is whether the expert proposes 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.

The Daubert opinion, while dealing with scientific evidence, specifically noted that the discussion was limited to a scientific context that was the nature of the expertise offered in that case. The discussion in the case does not apply to "technical, or other specialized knowledge[,]" but only to "scientific knowledge." Id. at 590 n.8.

Here, we deal with a social science in which the research, theories and opinions cannot have the exactness of hard science methodologies such as blood tests, DNA, spectrographic evidence or chemical exposures with which Daubert dealt. As observed in a recent article, Daubert principles may not fully apply to certain social science evidence.

Application of Daubert criteria to behavioral and social science evidence, particularly psychological syndromes, is problematic for two reasons: (1) judges' level of understanding of scientific principles and methodology may ill prepare them to evaluate science, including social science, as now required by Daubert and (2) the nature of certain social and behavioral science theories may be inherently inconsistent with Daubert criteria such as "falsifiability" and "error rates."

James T. Richardson, et al., The Problems of Applying Daubert to Psychological Syndrome Evidence, 79 Judicature 10, 10-11 (July-Aug. 1995); see also Berry v. City of Detroit, 25 F.3d 1342, 1349 (6th Cir. 1994), cert. denied, 115 S. Ct. 902 (1995). But see Rincon v. United States, 114 S. Ct. 41 (1993) (summarily remanding case for reconsideration in light of Daubert where expert testimony about reliability of eye witness testimony at issue).

The standard of review for admission of expert testimony is abuse of discretion. See Cook v. American S.S. Co., 53 F.3d 733, 738 (6th Cir. 1995). The Supreme Court recently revisited the "abuse of discretion" standard in Koon v. United States, ___ U.S. ___, 64 U.S.L.W. 4512, 4517 (Jun. 13, 1996):

Little turns, however, on whether we label review of this particular question abuse of discretion or de novo, for an abuse of discretion standard does not mean a mistake of law is beyond appellate correction. Cooter & Gell, [496 U.S. 384] 402 (1990). A district court by definition abuses its discretion when it makes an error of law. 496 U.S., at 405. That a departure decision, in an occasional case, may call for a legal determination does not mean, as a consequence, that parts of the review must be labeled de novo while other parts are labeled an abuse of discretion. See id., at 403 (court of appeals should "apply a unitary abuse-of-discretion standard"). The abuse of discretion standard includes review to determine that the discretion was not guided by erroneous legal conclusions.

With this background, we examine Dr. Underwager's foundation and compare that foundation and his commentary on suggestibility with the status as of the time of trial of psychological research and writings concerning child witnesses and their susceptibility to faulty memory. As noted above, in the defense's offer of proof, Dr. Underwager testified outside the presence of the jury that from his review of the files, records and testimony in this matter, there had been "a practice of suggestibility employed in these techniques." (Tr. Vol. IX at 1768.)

He further testified outside the presence of the jury that Kelson's notes revealed she had exerted a massive influence over the children; she had a powerful prior assumption or conclusion that the children had been abused; and she engaged in highly suggestive and contaminating practices, such as the groups and questioning. Dr. Underwager testified the prosecutor asked the children only if they remembered reporting an incident to a particular individual (FBI agent, social worker, etc.), rather than whether they remembered the incident itself; the prosecutor used exclusively leading questions in the courtroom and the children's comfort level showed they were used to this type of questioning. He testified that studies show that adults almost always rely on leading questions given the task of finding something out from a child.

Dr. Underwager found the FBI's use of sexually explicit diagrams very suggestive and leading, and asserted the evidence does not show such diagrams accomplish anything other than to suggest to the child that the interviewer is interested in sexual behavior.

He testified that a large body of research shows that the presence at an interview of several adults -- people of relatively high status -- increases the conformity and compliance with what those adults expect from a child.

Dr. Underwager testified that the documents from the case files and courtroom testimony suggested to him that powerful and potentially coercive influences had been brought to bear on the small fourand five-year-old children who were taken without notice from their mothers, families and homes, without being told the reasons and kept incommunicado in a strange place where all the people around them urged them to talk about sex abuse. (Tr. Vol. IX at pp. 1768-74.)

The district court concluded that this expert testimony was not the sort even contemplated by Daubert, did not pass the initial Rule 104(a) threshold inquiry with regard to either reliability or relevancy, and could well mislead the jury.

Here, the court misinterpreted our precedent and applied Daubert incorrectly to bar this evidence. The defense fulfilled the requirements of Daubert. The witness did not purport to testify that witnesses had in fact succumbed to any suggestive aspects of the investigation; only that the investigative means in this case were consistent with the psychological studies that similar techniques operated suggestively on young children. In addition, every condition which Dr. Underwager attempted to testify to as creating a practice of suggestibility has been amply demonstrated in the psychological literature as producing undue suggestibility in children's testimony. The importance and relevance is apparent.

The remaining question is whether the answers assist the jury. By excluding the expert testimony, the district court assumed the jury could do without the informed opinion of the expert -- that from the files and records and testimony a practice of suggestibility has been employed in the investigative techniques used on young children. That assumption minimizes almost 100 years of extensive research in this area of psychology -- information which is beyond the knowledge or experience of the average individual.

We have examined both the evidence and the literature presented to the district court and conclude that both support the defendants' offer of proof. In particular, the district court made reference to a recent article by Stephen J. Ceci and Maggie Bruck, Suggestibility of Child Witnesses: A Historical Review and Synthesis, 113 Psychological Bulletin 403-439 (1993), which reviews the research and writing on the subject and supports the view that the very matters observed and testified to by Dr. Underwager can produce biased, untrue or false memories in children, and more particularly young children. Almost all the other literature presented to the court is consistent with the Ceci-Bruck article.

The Ceci-Bruck article does not state that young children should not testify but observes that many common interviewing practices can produce an altered memory. Among other things, the article documents adequate research indicating the following:

1. A subject's, particularly a child's, original verbal answers are better remembered than the actual events themselves, yes-no questioning leads to more error, and young children are particularly vulnerable to coaching and leading questions. Id. at 406-09.

A review of the record here reveals the children were asked entirely leading questions in court. Even though the children testified by television outside the presence of defendants, the prosecutor asked suggestive questions. Not only did the questions call only for yes or no answers, the children were asked only if they remembered reporting abuse to law enforcement officers, doctors, and their therapist, rather than whether they remembered the alleged abuse itself.

The questioning at trial represents a highly questionable aspect of testifying about an event. This is exactly what Dr. Underwager described in his offer of proof.

2. Children desire to comply or cooperate with the respected authority figure interviewer and will attempt to make answers consistent with what they see as the intent of the questioner rather than consistent with their knowledge of the event even if the question is bizarre. Id. at 418-19. Interviewer bias can skew results as a child will often attempt to reflect the interviewer's interpretation of events, particularly when more than one interviewer shares the same presuppositions. Id. at 422. If the interviewer's original perception is incorrect, this can lead to high levels of inaccurate recall.

Here, these children were taken from their homes on the basis of a five-year-old's statements, and were placed under the sole supervision and influences of Donna Jordan, Jean Brock, and Ellen Kelson -- interviewers who had decided at the outset that all the children had been sexually abused.

The FBI agents were also strong authority figures -- the kind of high status interviewers described by Dr. Underwager -- with preconceived notions about the facts of this case, and they did not interview the children until after the children had been with Jordan for over a week. Agent Van Roe testified that he had explained his status as an FBI agent at the initial interview and told the children that an FBI agent was like a policeman on the reservation. Van Roe testified that Jean Brock and foster mother Donna Jordan remained in the room while FBI agents conducted the initial interviews of the children on January 19 and 21, 1994 -- over a week after the children were taken from their parents' homes, told by Jordan and Brock that this was because their uncles had done bad things to them, and put into the care of Jordan.

At this initial interview, R.R. handed investigator Hudspeth a group of papers which reflected things she had previously told foster mother Donna Jordan which Jordan had written down for her. Thus, agents received a frame of reference which could produce bias, even before the start of the interviews.

3. Repeated questions can produce a change of answers as the child may interpret the question as "I must not have given the correct response the first time," and the child's answers may well become less accurate over time. Id. at 419-20. Repeated questioning of victims often results over time (or even within a single interview) in an inaccurate report.

A three-month hiatus existed from the time R.R. was taken from her home to the time of her complaints of sex abuse. These children were repeatedly questioned by Brock, Jordan, Kelson, doctors and law enforcement agents. By March 1994, the children's accounts of the familial sexual abuse were so skewed that the district court refused to admit these interviews into evidence.

4. Younger children are more susceptible to suggestibility than older children, especially in the context of stereotyping. Id. at 407, 417. Stereotypes organize memory, sometimes distorting what is perceived by adding thematically congruent information that was not perceived, and stereotype formation interacts with suggestive questioning to a greater extent for younger rather than older children. Id. at 416-17. Studies have shown children are particularly susceptible to an interviewer's "bad man" stereotype, and when repeatedly told the actor is a bad man, they may construct a false account of an event often embellished with perceptual details in keeping with the stereotype. Id.

Here, various persons told the children from the beginning that the defendants were "bad" and that it would not be "safe" to go home until the defendants were gone. The children remained isolated from their families and community.13 The "bad man-uncle" theme was replayed again and again, including at trial.14 In addition, the children testified via closed circuit television based on their "fear" of defendants. While closed circuit television, other security procedures at the courthouse, and disallowing the children to see any family members before the trial did not amount to trial error, those procedures served to reinforce the children's "bad men" stereotype of their uncles, the defendants.

5. The use of anatomical dolls or sexually explicit materials will not necessarily provide reliable evidence as children may be encouraged to engage in sexual play with dolls, etc., even if the child has not been sexually abused, and further no normative data exists on non-abused children's use of dolls. See id. at 423-25.

The second law enforcement (January 21) interview took place at the United States Attorney's Office with the Assistant United States Attorney present. The children saw an anatomical drawing of a penis. Later, Kelson utilized play therapy and art media, and apparently dream journals. Dr. Underwager testified that exposing children to these materials suggests to them that the authority figure wants information about sex.

6. "[A] major conclusion is that contrary to the claims of some, children sometimes lie when the motivational structure is tilted toward lying." Id. at 433. Patterns of bribes for disclosures, implied threats in nondisclosures, or insinuations that peers have already told investigators of suspects' abusive behavior are highly suggestive. Id. at 423. Children will lie for personal gain, and material and psychological rewards need not be of a large magnitude to be effective. Id.

Here, the children were promised picnics, vacations and even a chance to return home as a reward for their "truthful," successful testimony at trial. They were told they could not go home until their uncles had been successfully removed. Experts are critical of this kind of reward as "bribing" children to "admit" abuse or give abuse-consistent answers, such as promising to end the interview, or giving them other tangible rewards. Such techniques affect the accuracy of children's reports.

7. Dr. Underwager testified regarding the concept of "cross-germination" among the children. Children in studies and in actual cases have shown that peer pressure or interaction with other children has effects on the accuracy of their reporting: they will provide an inaccurate response when other children have "already told" in order to go along with a peer group or be part of the crowd. See id. at 423; see also Stephen J. Ceci, Jeopardy in the Courtroom: A Scientific Analysis of Children's Testimony 146-50 (American Psych. Assoc. 1st ed. 1995). In several cases where convictions have been overturned, children were shown to have talked with one another about the abuse, sometimes even siblings questioned siblings to get them to "open up" or provide incriminating evidence. Id. at 150-51.

As mentioned above, Kelson reported that she talked to the group in "talk circle"; that the group seemed to have discussed an agenda among themselves each week and that T.R. was the ringleader. Testimony at trial reflects that Jordan, Kelson, and FBI agents spoke to and questioned the children in groups about the abuse.

The Ceci-Bruck article's summary relating to interviewing of children stated:

The studies on interviewing provide evidence that suggestibility effects are influenced by the dynamics of the interview itself, the knowledge or beliefs possessed by the interviewer (especially one who is unfamiliar with the child), the emotional tone of the questioning, and the props used. Children attempt to be good conversational partners by complying with what they perceive to be the belief of their questioner. Their perceptions, and thus their suggestibility, may be influenced by subtle aspects of the interview such as the repetition of yes-no questions, but their compliance is evidenced most fully in naturalistic interview situations in which the interviewer is allowed to question the child freely; this gives the child the evidence to make the necessary attributions about the purposes of the interview and about the intents and beliefs of the interviewer.

Observations of interactions in the legal arena highlight the fact that children who testify in court are not interviewed in sterile conditions such as those found in many of the experiments we have reviewed. They are usually questioned repeatedly within and across sessions, sometimes about an ambiguous event by a variety of interviewers, each with their own agenda and beliefs. Children are sometimes interviewed formally and informally for many months preceding an official law-enforcement interview with anatomical dolls, providing an opportunity for the child to acquire scripted and stereotypical knowledge about what might have occurred.

Id. at 425. The authors conclude with these comments:

Our review of the literature indicates that children can indeed be led to make false or inaccurate reports about very crucial, personally experienced, central events.

. . . .

Therefore, it is of the utmost importance to examine the conditions prevalent at the time of a child's original report about a criminal event in order to judge the suitability of using that child as a witness in the court. It seems particularly important to know the circumstances under which the initial report of concern was made, how many times the child was questioned, the hypotheses of the interviewers who questioned the child, the kinds of questions the child was asked, and the consistency of the child's report over a period of time. If the child's disclosure was made in a nonthreatening, nonsuggestible atmosphere, if the disclosure was not made after repeated interviews, if the adults who had access to the child prior to his or her testimony are not motivated to distort the child's recollections through relentless and potent suggestions and outright coaching, and if the child's original report remains highly consistent over a period of time, then the young child would be judged to be capable of providing much that is forensically relevant. The absence of any of these conditions would not in and of itself invalidate a child's testimony, but it ought to raise cautions in the mind of the court.

Id. at 432-33.

Other psychological research and writing supports the Ceci-Bruck article and Dr. Underwager's offer of proof. See, e.g., Maryland v. Craig, 497 U.S. 836, 868-69 (1990) (Scalia, J., dissenting) (detailing injustice caused by erroneous testimony of children who were separated from their parents for months and repeatedly interrogated and noting "[s]ome studies show that children are substantially more vulnerable to suggestion than adults, and often unable to separate recollected fantasy (or suggestion) from reality"); Lindsay & Johnson, Reality Monitoring and Suggestibility: Children's Ability to Discriminate Among Memories From Different Sources, in Children's Eyewitness Memory 92 (S. Ceci, M. Toglia, & D. Ross eds. 1987); Christiansen, The Testimony of Child Witnesses: Fact, Fantasy, and the Influence of Pretrial Interviews, 62 Wash. L. Rev. 705, 708-711 (1987); Debbie Nathan, Justice in Wenatchee, N.Y. Times, Dec. 19, 1995, at A19 (testimony of children increasingly being discredited in sex-abuse cases; children who have not been abused sometimes re-enact purported sexual trauma with anatomically detailed dolls or adopt fantasies complete with visceral details when prompted; videotaped pretrial interviews in some cases have helped prompt jurors to acquit defendants); Daniel Goleman, Studies Reveal Suggestibility of Very Young as Witnesses, N. Y. Times, June 11, 1993, at A1.

Indeed, the prosecutor's child abuse expert, Tascha Boychuk of the Child's Advocacy Center, Phoenix, Arizona, who testified at a pretrial hearing stated, "[i]f the question is can a child's memory be falsified, certainly the probability and the likelihood is yes. We see situations of that. Yes."

The reality of children's susceptibility to suggestive interview practices is well-established in the literature and the necessary analysis is beyond the ken of a non-professional.15 The expert's foundation related the coercive factors that can influence testimony. The defense provided the court with an abundance of literature supporting the expert's explanation relating to the existence of coercive factors in this case. Yet the court declined to allow the testimony.

4. Eighth Circuit Caselaw Regarding Similar Testimony

Although the district court correctly precluded Dr. Underwager from testifying about the ultimate issue of the children's credibility, he should have been allowed to testify regarding the suggestibility of the techniques employed in this case and whether they could have affected these children's memories.

We see no essential difference in this testimony, and in a qualified expert opining that an abuse victim's symptoms are consistent with sexual abuse syndrome, battered woman syndrome, battered child syndrome and other recognized syndromes. See Estelle v. McGuire, 502 U.S. 62, 70 (1991) (evidence of battered child syndrome related to intent and its admission did not violate due process); United States v. Norquay, 987 F.2d 475, 479 (8th Cir. 1993) (affirming admittance of expert rape trauma syndrome evidence over defendant's objections that this amounted to admission of others' opinions of victim's credibility because witnesses were not allowed to state whether they believed the victim had indeed been raped), abrogated on other grounds, United States v. Thomas, 20 F.3d 817 (8th Cir. 1994) (en banc); United States v. Simpson, 979 F.2d 1282, 1287-88 (8th Cir. 1992) (recognizing battered woman syndrome), cert. denied, 507 U.S. 943 (1993); United States v. Whitetail, 956 F.2d 857, 859 (8th Cir. 1992) (same); United States v. St. Pierre, 812 F.2d 417, 419-20 (8th Cir. 1987) (expert can inform jury of characteristics found in sexually abused children and describe characteristics alleged victim exhibits).

In United States v. Johns, 15 F.3d 740, 743 (8th Cir. 1994), we rejected the defendant's argument that an expert impermissibly vouched for a sexual abuse victim's credibility because implicit in the expert's testimony was the opinion that the victim was telling the truth. We concluded that an expert may inform the jury of the characteristics of sexually abused children generally and may describe characteristics exhibited by the alleged victim, but may not state an opinion that abuse has in fact occurred. Id. Likewise, in United States v. Whitted, 11 F.3d 782, 785 (8th Cir. 1993), we determined that an expert may inform the jury of characteristics found in sexually abused children and describe characteristics the alleged victim exhibits. We stated that expert opinions are not inadmissible merely because they embrace the ultimate issue to be decided by the trier of fact, but they cannot be phrased in terms of inadequately explored legal criteria or merely tell the jury what result to reach. Id.

"Finally, in our landmark case of [United States v. Azure, 801 F.2d 336, 340 (8th Cir. 1986)], we stated that general testimony about a [child] victim's ability to separate truth from fantasy, the expression of an opinion on the similarities between a victim's claim and the evidence, and the comparison of behavioral and testimonial patterns of a particular victim with the behavioral patterns observed in victims in general, were all admissible in certain circumstances." Bachman v. Leapley, 953 F.2d 440, 442 (8th Cir. 1992); see also United States v. Plenty Arrows, 946 F.2d 62 (8th Cir. 1989) (no abuse of discretion where district court allowed health therapist to testify that victim's behavior consistent with that of other sexually abused children); Arcoren v. United States, 929 F.2d 1235, 1239-41 (8th Cir. 1991) (expert testimony readily admissible where psychologist testifies to mental aberrations in human behavior, when such knowledge will help jury to understand relevant issues in case, including helping jury to evaluate which of victim's conflicting statements were more credible, and expert does not express her opinion as to which statements were more credible or whether victim suffered from battered woman syndrome), cert. denied, 502 U.S. 913 (1991).

Here, Dr. Underwager was not testifying as to whether the children were credible, but rather to whether they were subjected to suggestive practices. The district court erred in excluding that important testimony.

In assessing the prejudice from the exclusion of this evidence, we do so against the backdrop of other alleged errors depriving these defendants of a fair trial.

Although the medical evidence was inconclusive and the examiners took no pictures, the district court denied defendants' requests for further medical examinations. This discretionary ruling is not error, but highlights the importance of the children's testimony and the prejudice to defendants caused by the court's refusal to admit Dr. Underwager's testimony.

Soon after these children were removed from their homes, the Department arranged for some of them to undergo a medical examination by Dr. Richard Kaplan, a pediatrician at the Yankton Medical Clinic who examines 500 to 600 children per month. Dr. Kaplan testified that the vaginal redness and possible trauma he observed could be consistent with abuse or any number of possible non-abuse causes; the conditions he observed were basically nonspecific as to cause; he could not conduct a thorough examination while the children were awake; and based on his limited examination, he could not positively diagnose any of the children as having been abused.

Thereafter, on February 11, 1994, his co-worker, Dr. Robert Ferrell, had the children placed under anesthesia and examined them with a colposcope. Although he had received some training in the sex abuse area seven years earlier while he was a resident, Dr. Ferrell had no special experience in sex abuse investigations. Dr. Ferrell had never testified in a criminal case. He did not take photographs of his colposcopic examinations, although the process would have been easy and helpful in this case.

Dr. Ferrell admitted that neovascularization (redness in the vaginal areas), decreased anal tone, and hymenal tags can be common place occurrences resulting from many different "everyday occurring" causes. He did make a post-operative diagnosis that F.R. indicated evidence of tearing and scarring of the anal mucosa but otherwise a normal anus and vagina; that R.R. had apparent damage to the hymenal ring consistent with vulvovaginal trauma, and possible anal trauma; that L.R. revealed a fusion and evidence of anal trauma; that J.R. had neovascularization, clue cells and a tag or scar on the hymen. On T.R., the anterior portion of the hymenal ring was essentially gone; he diagnosed vaginal and vulva trauma.

The defendants' pediatric expert, Dr. Robert Fay of Albany, New York, testified he had prior training and experience with Native American patients, sex abuse diagnosis and treatment, and that he had been previously retained by both defense and prosecution in other cases. In essence, Dr. Fay testified that the reported hymenal fusions in L.R., R.R. and J.R. are suspicious for sexually acquired trauma; that labial injury would be a significant finding in diagnosing sexual abuse, but that most of the conditions observed by the doctors offered by the prosecution -- such as redness, erythema, neovascularization, vaginal furrows and ridges, a "gaping hymen," a "hymenal notch," "clue cells," "relaxed anal tone," and "anal folds," were of no significance in evaluating whether sex abuse had occurred, and are found in a high percentage of non-abused children.

Dr. Fay testified he felt Dr. Ferrell was not qualified based on training or experience to perform a colposcopic examination of a child; his training was outdated; photographic evidence in such cases is frequently dispositive, very helpful, and perhaps essential; and a further physical examination of the children would be very helpful.

The literature in this area, see Jan Bays & David Chadwick, Medical Diagnosis of the Sexually Abused Child, 17 Child Abuse & Neglect 91, 92, 95, 103 (1993), indicates that frequently findings on examination of children allegedly sexually abused are no different than similar findings on children who most likely have not been subject to sexual abuse. That work indicates that a number of factors or conditions may mimic findings caused by sexual abuse or wrongly produce a history suggestive of child sexual abuse, including adults misinterpreting normal masturbation or sexual play between children and a variety of other dermatologic, congenital, traumatic and infectious conditions.16

Even in our present state of knowledge, it is becoming increasingly evident that, as a consequence of naturally occurring physical changes, there will always be an overlap in findings between nonabused children and the victims of sexual misuse. The appreciation of this reality should serve as a constant reminder that the determination of sexual abuse can rarely rely on a physical examination alone and that consideration of all the components of the investigation -- especially the information obtained from the child -- is essential.

John McCann, M.D., et al., Genital Findings in Prepubertal Girls Selected for Nonabuse: A Descriptive Study, 86 Pediatrics 428, 438 (Sept. 1990).

We agree that, as a matter of discretion, the district court need not have required more invasive procedures on these small children. But we must observe that the medical testimony, while consistent with possible sexual abuse, is inconclusive in light of other matters discussed herein. In addition, some of that alleged trauma may have occurred from sexual interplay and activity between and among the victims and other young children.

2. Rejection by the Trial Court of Testimony regarding inter-child sexual activity

The defendants sought to introduce testimony regarding substantial inter-child sexual activity by and between the children in question and other children in Marty, South Dakota on the reservation. Particularly, an eleven-year-old boy who lived in the Rouse home told investigators that he and T.R. had sex for a long time, and other children corroborated this testimony. Also, during the trial, government evidence surfaced showing that T.R. also had engaged in sexual relations with another boy. This evidence was not placed in front of the jury however.

The children themselves provided a source of this information because after weeks of interrogation and "counselling," the sex abuse accusations expanded to include all sorts of family members including the grandmother. This well may have been fantasy and bears on the reliability of the government's case against these defendants.

Although defendants intimated that they would be presenting evidence generally of inter-child sexual activity and accusations of sexual abuse by non-defendant family members in several motions, the defendants did not file a formal motion as required by Federal Rule of Evidence 412.17

The day before the trial, the government filed a motion in limine to prevent defendants from presenting such evidence because of the absence of a formal Rule 412 motion. The defendants responded by then filing the 412 motion, claiming they had not received some of the relevant evidence, including L.R.'s representations that everybody was having sex with everybody else in the house -- particularly that she was having sexual relations with her grandmother -- until the government filed its motion. The district court denied the motion for absence of timeliness.

While that ruling can be approved based on the record in this case, the issue is troubling because that sort of evidence would have cast additional light on whether the alleged medical evidence of sexual abuse could be attributed to sources other than the charged defendants. This sort of evidence may be constitutionally required. United States v. Bear Stops, 997 F.2d 451 (8th Cir. 1993) (reversing conviction where district court limited admission of evidence relating to previous sexual assault on victim to establish alternative explanation for why victim exhibited behavior of sexually abused child).

Turning to Federal Rule of Evidence 412, that rule permits the filing of a 412 motion during trial for good cause. See Rule 412(c)(1)(A). Where the government only gave the defendants the FBI report regarding the second boy during the trial, the district court's refusal to allow the defendants to present testimonial evidence kept important information, helpful to defendants, from the jury.

The trial court might have granted the belated Rule 412 motion. Nevertheless, we cannot say that the trial judge abused his discretion in strictly following the language of the statute in requiring fifteen days advance notice to present the evidence. Inasmuch as we grant a new trial on other grounds, the untimeliness issue should not arise on the new trial.

3. Denial of Independent Psychological Examination

The defense moved for a psychological examination of the children stating that this examination would be crucial to the defense in preparation of the case and observing:

The children have been subjected to countless, unrecorded interviews by social workers, FBI and tribal officers, the U.S. Attorneys Office and others. There is simply no way of telling what occurred at these interviews, the nature and form of the questions, or whether or not some children who are now complaining witnesses, at first denied the abuse occurred, or whether there have been retractions or recantations.

In response to the government's objections, the movants also asserted "the stated claim in preventing contact with the child by the accused or the appointed psychologist for the accused is to protect the children from any further abuse or possible intimidation or harassment by the accused or the appointed psychologist."

The movants further argued that social workers interrogated the children several times; that the police and the FBI interrogated the children several times; that the children were told to keep dream journals by the foster care provider; that the children had been the recipients of constant interrogation by the foster care provider; that the counselor had interrogated the children on numerous occasions; that the prosecuting attorney had interviewed the children several times; and that there had been dozens of interrogations by various agencies involved. The movants noted that at no time was the cry of trauma, embarrassment, invasion of privacy and possible harassment by these interrogations raised, and that perhaps this was because the various agencies were building their cases.

This motion was well taken.

Although it is addressed to the discretion of the district court and the district court denied the motion on the grounds that such examination was unnecessary and intrusive, the record in this case amply indicates that the defendants suffered substantial prejudice by the nature of the case against them without the opportunity to indeed show that possible falsification in testimony had occurred because of the nature of the government's investigations. Given their lack of access to the children and the amount of suggestive interviewing done to support the prosecution, we believe the defendants were entitled to an independent psychological examination.

This belief is strengthened by the failure to videotape or audiotape any of the investigatory or counseling interviews. Electronic recording of child witness interviews (particularly, the preliminary interviews) is crucial to rule out the potential influences of coaching and interrogative suggestion. Written summaries by the adult interviewers (be it Kelson, Jordan or law enforcement agents) are no substitute for electronic recordings of these interviews, particularly in legal proceedings:

Although one would excuse such missing data when the allegation was first made to parents, one would hope that it would be normal procedure for the police, social workers, and therapists to have recorded all interviews with the children, if the purpose of the interview could -- even remotely -- be considered "forensic."

Ceci, supra, at p. 242. No tape or audio recordings were taken of any of the multitudinous interviews which took place in this case. Many of the discrepancies in testimony in this case might have been resolved by a taped record of these interviews.

In addition, some of the children's testimony reflects an element of fantasy -- possibly the tying up of practically every member of the household and locking up multiple children in closets in light of the previous abuse experience Melanie Rouse had shared with the other children.18 Studies show that children will fantasize --telling elaborate stories about an event that never happened or fabricating an entire episode or sequence of events within a larger episode, particularly over time on the basis of acquired interviewer stereotypes, or they may produce convincing false narratives to explain fictitious events suggested to them. See Ceci, supra, at 133-34, 218-222, 227; see also, Ceci & Bruck, supra, at 407, 417 (boundaries of children's fantasy-reality distinctions can be fragile; children's disclosures may become increasingly bizarre and incredible, sometimes caused by interviewers not drawing children back to reality when they made fantastic claims; and children may have trouble distinguishing what they experienced through perception and what they only imagined they experienced). Defendants had the right to have some of these stories explored by an independent psychological examiner.

In light of the manner in which the prosecution, state agencies and others had proceeded in the investigation, the district court abused its discretion in denying the defense a fair opportunity to determine whether the children had, in fact, been influenced by the manner in which the investigation had taken place.19

After the verdict in this trial, an allegation of juror misconduct was brought to the court's attention. Verna Severson contacted the clerk's office and said that she was surprised her co-worker (at a preschool) Pat Pickard was allowed to serve on the jury of a case involving Native Americans because she believed Pickard was prejudiced against them. The district court held a number of hearings and heard testimony from a number of witnesses.

Everyone who testified (mostly co-workers from the preschool) except Severson, unequivocally stated that Pickard is not racist and has not demonstrated a bias against Native Americans, although the school is now instituting workshops on racial sensitivity. Severson had a long-standing animosity toward Pickard and some of her testimony was contradicted by other witnesses.

On the other hand, Severson testified that she had listened to Pickard's racist statements for many years (including one to the effect that adult males often have sex with young girls as part of the Native American culture), and that she had engaged in arguments with Pickard over the years on the subject of Pickard's racism and bias against Native Americans. Severson testified Pickard told her that Pickard and two other jurors in this case enjoyed making racial jokes about Indians. The district court precluded Severson from telling what Pickard said went on while the jury was in the jury room pursuant to Federal Rule of Evidence 606(b).

The defendants assert Pickard's own testimony and efforts to avoid answering questions -- "Not in my opinion," "I don't know," "I don't remember," "I don't think so," and "I may have" -- were most revealing of her bias and her equivocal answers to most of the questions impeached her specific denial of bias. Pickard specifically testified she heard and laughed at a comment (rather than joke) in the jury room about an Indian (but she could not reveal more under Federal Rule of Evidence 606(b)), and she told a fellow juror after the trial that "Well, you know what to say [regarding racial prejudice] if you want to be on or off the jury."

Pickard also acknowledged that her sister-in-law, a social worker who worked with abused children, told her that it is terrible to be born an Indian baby girl; that she had repeated this remark to Severson and other individuals; and that her intention was to repeat a statement of an experienced social worker.

Defendants argue the court's order denying the motions for a new trial incorrectly narrowed the focus of the juror misconduct inquiry to voir dire only and not into comments which took place in the jury room. United States v. Heller, 785 F.2d 1524 (11th Cir. 1986) (overturning conviction where jurors made anti-semitic jokes at trial and others reacted to them with gales of laughter).

We do not quarrel with the credibility determination of the district court rejecting serious charges against Pickard, but the evidence relating to this issue erodes confidence in the result. We cannot ignore the existence of racial prejudices in our society and as against Native Americans in areas near reservations. Pickard's statements relating to Native American racial jokes or comments raises a matter of grave concern. Racial prejudice in the jury room cannot and will not be tolerated or condoned. Here four Native Americans placed their liberties in the hands of all whites: prosecutors, defense counsel, judge and jury. The law requires that they receive a fair trial without the impact of racial bias.

Abuse of young children is a serious crime. Here, only five of thirteen children brought forth tales of abuse and told their versions in an atmosphere that could be coercive. These circumstances raise a serious and a close question to the validity of the verdict. The trial court barred crucial defense evidence relating to the practice of powerful and coercive suggestibility relating to child witnesses. Such evidence wrongfully excluded could have made a difference. We reverse and remand for a new trial.20

LOKEN, Circuit Judge, dissenting.

I respectfully dissent. In my view, the majority's opinion misrepresents the factual record, misstates the district court's procedural and evidentiary rulings, and ignores the context of those rulings. The majority accepts defendants' view of numerous issues of disputed fact, violating our duty to respect the jury's verdict. I leave to the district court the task of separating appellate fact from fiction when the case is retried. But I will explain the reasons why I believe this decision is very wrong.

I. Convincing Evidence of Guilt.

The majority takes many liberties with the trial record in attempting to cast doubt on the jury verdict that defendants committed some, but not all, of the alleged acts of criminal sexual abuse of five young girls. The government based its case on the testimony of two physicians, the four oldest victims, another child who witnessed sexual abuse, and FBI Agent Van Roe. A brief review of that evidence is needed to set the record straight.

A. The Basic Chronology. The South Dakota Department of Social Services ("DSS") placed R.R. in Donna Jordan's foster home on November 9, 1993, because of neglect and malnutrition. In early January, Jordan reported to DSS (as she was required to do) that R.R. said she had been sexually abused. On January 10, DSS told Jordan to take R.R. to therapist Ellen Kelson. After an initial interview, Kelson reported to DSS (as she was required to do) that R.R. had reported acts of sexual abuse against herself and other children in the Rouse home. On January 11, DSS removed children living in the Rouse home and placed them in Jordan's foster home. On January 15, Dr. Richard Kaplan examined the children. On January 19, FBI Agent Van Roe and BIA Agent Hudspeth interviewed the children. That evening, they were seen by a psychiatrist, who referred them to Kelson for therapy. Kelson first saw the children in a group on January 22. The majority frequently misstates or obfuscates these undisputed chronological facts.

B. The Medical Evidence. After performing the initial medical examinations on January 15, Dr. Kaplan reported to DSS his medical findings and what the children had told him about sexual abuse. J.R. told Dr. Kaplan, "Uncle Jess hurt me," pointing to her left labia; Dr. Kaplan found a recent bruise or contusion consistent with that kind of abuse. L.R. had "a fairly acute injury" on the right side of her labia majora which "really hurt her." R.R. told Dr. Kaplan, "I have a bruise where my uncle put his private spot," and Dr. Kaplan found a sagging vagina and a scar on her anus. Dr. Kaplan found that T.R. had "obvious trauma and contusion . . . and very, very much tenderness" on her labia majora; T.R. told him, "Uncle Jess hurt me there."

On February 11, Dr. Robert Ferrell conducted a colposcopic examination of the five victims. Dr. Ferrell found "very significant" damage to R.R.'s hymenal ring and tearing in her anal area consistent with anal intercourse. He noted a "whole constellation of findings" indicating L.R. had been abused -- damage to her hymenal area, furrowing on either side of her vagina, chronic irritation or trauma, and "clue cells" that are "known to be sexually transmitted." To Dr. Ferrell, a scar on J.R.'s hymen where a tear had healed was an "important finding," while T.R.'s "hymenal ring was essentially gone," the entire area was irritated, and she had furrows in her vagina. Infant F.R. had "tearing and scarring of the anal mucosa."

Defendants' medical expert, Dr. Fay, admitted that the reported hymenal scarring on L.R., R.R., and J.R. "certainly . . . leads you to think about sexual abuse," and that "a labial injury . . . is a very significant finding" of abuse. In its rebuttal, the government called Dr. Randall Alexander, a member of the Board of Governors of the National Committee to Prevent Child Abuse. Dr. Alexander testified that it takes considerable force to inflict labial injuries like those exhibited by three of the victims. "It's rare to see one [in young girls] and to see three of them show up is just . . . rareness to the third power."

On this record, the majority dissembles when it repeatedly opines that the government presented "inconclusive" medical evidence of sexual abuse.

C. The Victims' Testimony. At trial, four victims testified that defendants sexually abused them (the fifth victim, infant F.R., was too young to testify). Their nine-year-old cousin testified that three defendants shut him in the attic when he saw them abusing T.R. A few observations about this testimony.

First, the victims' trial testimony was consistent with their "free recall" -- the reports of abuse R.R. volunteered to Donna Jordan in early January, and the four oldest victims made to Dr. Kaplan during his January 15 medical examinations. These unprogrammed reports preceded the FBI interviews and Ellen Kelson's therapy. Is that significant? I refer to the majority's non-testifying experts, Ceci and Bruck, in the "Conclusions" portion of their Suggestibility article, 113 Psych. Bulletin at 433:

[I]t is of the utmost importance to examine the conditions prevalent at the time of a child's original report about a criminal event . . . . If the child's disclosure was made in a nonthreatening, nonsuggestible atmosphere, if the disclosure was not made after repeated interviews, if the adults who had access to the child prior to his or her testimony are not motivated to distort the child's recollections through relentless and potent suggestions and outright coaching, and if the child's original report remains highly consistent over a period of time, then the young child would be judged to be capable of providing much that is forensically relevant.

Likewise, Dr. Underwager testified for the defense:

Q What has the research told us about the types of questions that should be asked?

A Basically, the most reliable information is obtained from free recall.

Second, it is certainly true that the prosecutor asked the children leading questions at trial. When the first child witness (the nine-year-old male cousin) froze on the stand in open court, the district court, consistent with numerous Eighth Circuit cases, ruled that leading questions could be asked of reticent child witnesses. Defendants did not object to this ruling nor raise the issue on appeal.

Third, defendants did object to permiting three of the child witnesses to testify by closed circuit television. The district court questioned each child in chambers, in the presence of defense counsel, one prosecutor, the child's guardian ad litem, and a court reporter. See 18 U.S.C. Section(s) 3509(b)(1)(C). Five-year-old J.R. was unable to speak when called to testify and stated in chambers that she was afraid to speak in front of her uncles. Six-year-old R.R. was found sobbing outside the courtroom and affirmed in chambers that she was crying out of fear of her uncles. Nine-year-old T.R. became so fearful before testifying that "the guardian ad litem would have had to physically pull her into the courtroom." The court found that defendants' presence in the courtroom would prevent these three children from testifying and permitted them to testify in chambers by closed circuit television.21 Though I share the concerns expressed by Justice Scalia in his dissenting opinion in Maryland v. Craig, 497 U.S. 836, 867-69 (1990), I conclude the court properly resolved this issue under Maryland v. Craig, Hoversten v. Iowa, 998 F.2d 614 (8th Cir. 1993), and 18 U.S.C. Section(s) 3509(b). The majority apparently agrees.

D. The FBI Interviews. After a hearing outside the jury's presence, the district court permitted FBI Agent Van Roe to testify to what the three oldest victims said during his January 19, 1994, interviews because those hearsay statements were spontaneous and trustworthy. Though defendants challenge this ruling on appeal, "a formidable line of Circuit precedent . . . sanctions the use of hearsay testimony in child sexual abuse cases." United States v. St. John, 851 F.2d 1096, 1098 (8th Cir. 1988).

Accepting this ruling, the majority instead chides the district court for excluding evidence of March 1994 interviews that were "essential to the defense."22 But the district court did not exclude that evidence. When Desmond Rouse's attorney asked if he could cross-examine Agent Van Roe regarding the March interviews, the court responded that such questions were beyond the scope of Van Roe's direct testimony and raised distinct reliability questions. However, said the court, "If you want to do something later, that is up to you. I'm not trying to tell you how to try your case." Though Van Roe was recalled as a defense witness, he was not questioned about the March interviews.

Finally, it is worth noting in surveying the evidence that the government's rebuttal included testimony by another FBI agent that Jessie Rouse and Desmond Rouse made damaging admissions when interviewed on January 25, 1994.

E. The Majority's Riposte - (1) Donna Jordan as the Grand Inquisitor. Jordan was the victims' foster mother, a function she has performed for over ninety children for seventeen years. She was not involved in law enforcement. She never "interviewed" the children. She was not asked as a witness to relate what the victims told her. Yet the majority asserts that Jordan "told them their uncles were at fault and she got very specific about the 'bad' things their uncles were doing to them." Supra, at p. 8. Let us see how Jordan described these supposedly sinister conversations during her cross examination at trial:

Q You have gone to each one of the girls one by one and in groups, and you have told them that their uncles have been raping them?

A No. Never. No.

Q What specific bad things have you told the girls that the uncles were doing?

A When they were first talking about their sexual abuse -- Q I'm talking about what you have said to the kids.

A I just say it's wrong when they bring up the subject, and I say it's wrong, it's wrong, it's bad. But there is nothing, you know -- that's it.

Q Have you talked to these kids both alone and in groups?

A No.

Q Well -- A Unless they come -- I mean things can happen at home. Someone can come up and say something while you are fixing supper or whatever you are doing, and start talking about something. But, no, not deliberately talking about this. No. I haven't.

The majority's attempt to label Jordan as a forensic interviewer who bombarded the victims with "a practice of suggestibility" in effect accuses Jordan of lying under oath and improperly performing her duties. It also twists the record to fit the majority's unsupportable view of the case.

(2) Ellen Kelson as the Grand Inquisitor. After initial medical examinations and FBI interviews, the children were referred by a psychiatrist to Ms. Kelson for therapy. Because the children were Medicaid patients, DSS was Kelson's client, and she was required to submit her session notes. When Jean Brock suddenly left DSS in March 1994,23 Kelson began sending these confidential notes to an Assistant U.S. Attorney to ensure that only authorized persons saw them, a fact the majority uses to place Kelson in the prosecutor's camp. But she conducted no forensic interviews of the children, and the government did not call her as a witness.

Defendants did call Kelson at trial and questioned her for some 150 transcript pages in an unsuccessful effort to convince the jury that Kelson had "implanted" the victims' memories, thereby contaminating their trial testimony. The majority accepts this theory, calling Kelson primarily responsible for the "practice of suggestibility." The majority encourages defense counsel to call a victim's pretrial therapist, question about the nature of the therapy, and then call a psychologist who will opine that the therapist's counseling destroyed the credibility of the child's trial testimony. This tactic if widely used would force the government to choose between timely therapy for the child victim, and effective prosecution of the child abuser. This is terrible public policy, contrary to congressional mandates. District courts should block such injustice in future cases by exercising their discretion to preclude defendants from introducing evidence of pretrial therapy or counseling by a professional who has not testified in the government's case-in-chief.

(3) Excluded Evidence of Other Sexual Activity. Defendants filed Rule 412 motions to introduce evidence that one victim had been sexually active. The district court denied the motions because the defense learned about this evidence from an interview with a young boy almost three months before trial. I conclude, and the majority apparently agrees, that the district court did not abuse its discretion by excluding this evidence as untimely. See Rule 412(c)(1)(A); United States v. Provost, 875 F.2d 172, 177-78 (8th Cir.), cert. denied, 493 U.S. 859 (1989).

Defendants' Rule 412 evidence consisted primarily of a claim by one defendant's eleven-year-old son that he had "sex" with the victim, which she denied. After reviewing interview reports, the district court expressed concern that this evidence "could wind up creating a mini trial as to whether . . . this experimentation took place." But the court refrained from ruling on whether this evidence of sexual activity would be admissible if the government presented medical evidence of genital injuries. That is the relevant inquiry. See Rule 412(b)(1); United States v. Bear Stops, 997 F.2d 451, 454-56 (8th Cir. 1993); United States v. Eagle Thunder, 893 F.2d 950, 954 (8th Cir. 1990). Though the government offered such injury evidence, when defendants called the boy as a defense witness, they did not question him regarding this alleged sexual activity. Thus, the district court dealt with the merits of this issue correctly, and the majority's suggestions to the contrary reflect a distressing ignorance of the trial record.

II. Dr. Underwager's Testimony.

Dr. Underwager's testimony was the culmination of the defense strategy to destroy the credibility of the victims' trial testimony by proving that this testimony was the product of "implanted" memories. The exclusion of some of Dr. Underwager's proferred testimony provokes a dissertation by the majority on Daubert and "soft science."24 Proper review of this issue instead requires careful attention to its procedural context.

Before Dr. Underwager testified, the district court held a hearing to determine whether this expert would provide sufficiently reliable scientific evidence that would assist the jury to understand or determine a fact in issue. See Fed. R. Evid. 702; Daubert, 113 S. Ct. at 2796. When the court asked what opinions Dr. Underwager proposed to express at trial, defense counsel submitted a letter from Dr. Underwager stating:

Based upon my review of the documents you have supplied, including but not limited to, therapist notes, FBI reports, and other documents, it is my opinion that the children in this case have been subjected to massive and coercive social influence by adults. . . . The level of adult influence is such as to make it highly likely any statements are so contaminated by adult behaviors as to be unreliable . . . . If asked, my opinion is that there is such a low probability of any sexual abuse by the defendants, that a reasonable person must conclude it did not take place.

The court also heard Dr. Underwager describe his theories of "learned" or "implanted" memory, and it reviewed some of the literature to which he referred (including the Ceci and Bruck article extensively cited by the majority). The court ruled:

I'm not going to allow Dr. Underwager to testify as to whether or not the [child] witness's testimony is believable or not, or telling the truth or not. . . . [T]here may be other areas that Dr. Underwager may be proffered to testify on, and those will have to be, or may be anyway, the subject of an offer of proof when we get to that point.

The majority concedes, as it must, that this preliminary ruling was correct. Assessing the reliability or credibility of a victim's accusations is the exclusive function of the jury; it is not a proper subject of expert testimony. See Westcott v. Crinklaw, 68 F.3d 1073, 1076 (8th Cir. 1995); United States v. Witted, 11 F.3d 782, 785-86 (8th Cir. 1993); United States v. Azure, 801 F.2d 336, 339-40 (8th Cir. 1986). Dr. Underwager's attempts to express such opinions in other child abuse cases have been consistently rejected. See State v. Swan, 790 P.2d 610, 632 (Wash. 1990) (en banc), cert. denied, 498 U.S. 1046 (1991); State v. Erickson, 454 N.W.2d 624, 627-29 (Minn. App. 1990).

The district court also made a second preliminary ruling. Because this is an area of valid scientific inquiry, the court ruled that Dr. Underwager could express his own expert opinions and explain his own prior research. But regarding the theories and writings of other psychologists, the court concluded:

there is not anywhere near yet the agreement in the [scientific] community as to methods, techniques, testings or reliability that would warrant the admissibility before a jury of these matters . . . . It would result in a confusion of the issues, a possible misleading of the jury . . . . So, for these reasons, under Daubert, I'm not going to allow evidence with regard to the different . . . psychological methods of evaluating the reliability of witnesses.

The majority does not question this exercise of the court's discretion, perhaps because the Ceci and Bruck article is itself a compendium of conflicting theories and opinions on this subject dating back to 1900. As an aside, many of the majority's citations are to Ceci and Bruck's summaries of other experts' work, some of it more than ninety years old.

With that procedural background, I will summarize Dr. Underwager's trial testimony for the defense, organizing that testimony into the categories of "suggestibility" evidence that the majority claims were totally excluded (supra, at pp. 20-27):

1. Asking children leading questions at trial. After opining that open questions produce more reliable information, Dr. Underwager asserted, "an adult who has a bias, a preconceived assumption, tends very quickly to go to leading questions, go to coercive questioning in order to get what they think that they need or want." The district court then sustained an objection to a later question, "What would be your comment in regard to the form of the questions [asked of the children at trial]?"

2. A child will give authoritative interviewers what the child perceives is desired. "[I]t's very clear that when the person doing the questioning starts with the assumption that there's been abuse, that's what you get, that's what you produce from the child."

3. Repetitive questioning can change a child's answers. "[T]he repeating of questions is one of the most powerful ways that adults influence children to produce the answers the adult wants. Parents know this."

4. Younger children are more susceptible to suggestibility than older children. "The younger the child, the greater the suggestibility, the more vulnerable they are to the influences."

5. Anatomical dolls and sex play will not necessarily produce reliable evidence of sex abuse. "Play therapy has no therapeutic value whatsoever. In fact, the research evidence suggests that it's harmful . . . . [Play therapy is] all Freudian stuff, and there's no scientific support whatsoever for those concepts." Dr. Underwager was then asked about research "into the effects of play therapy where the allegations of sexual abuse may not be true," an objection was sustained, and defense counsel dropped the issue.

6. Children lie when motivated to lie. "[V]ery frequently the adult will give some kind of promised reward . . . to shape the behaviors of children." Adults also use "what psychologists know as negative reinforcement; that is, the removal of an adverse stimulus . . . . [These techniques are] very powerful and very often used." 7. "Cross germination" amoung a group of children will produce inaccurate memories. "When children talk to each other, they have an effect on each other, and they can communicate . . . stories that are picked up on." "[Y]ou can produce changes, and accounts shift and move and all kinds of things happen."

8. On the general subject of implanted memory. "[M]emories are now shown to be implanted[.There] can be a complete nonevent, but a memory can be created . . . by questioning someone."

In my view, this summary conclusively demonstrates that every category of "suggestibility" evidence identified by the majority was the subject of an opinion expressed by Dr. Underwager to the jury. That should be the end of the matter on appeal. The jury learned how the child victims were medically examined, interviewed, and counseled before trial. It saw and heard the children testify. And it heard Dr. Underwager's opinions as to factors that might influence the reliability of that testimony. The district court correctly concluded that this was an adequate evidentiary basis for the jury to make its ultimate credibility findings.25

Rather than examine what the jury in fact heard, the majority places the cart before the horse by starting with the offer of proof made by defense counsel at the conclusion of the above-summarized testimony. That offer was made because the district court precluded Dr. Underwager from commenting on the form of the questions asked the child witnesses at trial.26

The majority states that the crucial question and answer in the offer of proof was, "is it your belief that there's been a practice of suggestibility employed," and Dr. Underwager's cryptic answer, "Yes, sir." Supra, at p. 13. I disagree. What was "crucial" to the district court was the very next question, "could you explain to the Court how you observed that," and Dr. Underwager's three-page narrative answer. In that answer, Dr. Underwager opined (i) that therapist Kelson had exerted "massive social influence" on the victims; (ii) that Kelson engaged in "highly suggestive and highly contaminating" practices; (iii) that the prosecutor used leading questions at trial and the children "were comfortable doing the yes/no bit," showing "they'd learned" to answer yes; (iv) that Van Roe's use of diagrams was "very suggestive and very leading"; (v) that the children "were kidnapped . . . taken from their families, taken to this strange place where all of the people are concerned that they talk about sex abuse"; and (vi) that the "total environment [was] one of the most powerful and coercive influences upon children that I've seen."

That was totally improper "expert" testimony. If Dr. Underwager may not opine that the children's trial testimony was not credible, as we have often ruled, then the district court properly precluded him from indirectly stating that same opinion. For example, if Dr. Underwager has expressed the opinion that leading questions produce testimony that is not credible, then his opinion that the prosecution asked leading questions at trial is, quite obviously, an opinion that the children's testimony was not credible. Defendants' offer of proof reflected Dr. Underwager's passionate attempt to find a court -- any court -- that will allow him to opine that particular child witnesses have not told the truth. Those opinions have been rejected by every court which has considered them, they were rejected by the district court, and they are rejected in theory by the majority. Yet these convictions are reversed because defendants' offer of proof was denied! The district court properly "circumscribed [Dr. Underwager's testimony] so as to educate rather than to usurp the role of the jury." United States v. Johns, 15 F.3d 740, 743 (8th Cir. 1994). I have no doubt that if defendants had asked Dr. Underwager to describe a "pattern of suggestibility," without opining whether one occurred in this case, the district court would have allowed that addition to the opinions he did express to the jury. In other words, the majority's stated reason for reversal is a contrivance, apparently born of a desire to publish a lecture on Daubert and a distaste for either this type of prosecution or the long prison sentences it has produced.

III. Denial of Defense Psychological Interviews.

Prior to trial, the victims were in DSS's custody. The majority criticizes DSS's efforts to isolate the young victims27 and holds that the district court abused its discretion in denying defendants' motions to subject the victims to adversarial psychological examinations by Dr. Underwager. I disagree.

Defendants and Dr. Underwager had available to them the reports of the victims' medical examinations, Agent Van Roe's interview reports, and therapist Kelson's extensive notes of her sessions with the children. Dr. Underwager stated at the motion hearing that he had sufficient information to assess whether the children had been sexually abused. He observed the trial testimony of the victims and therapist Kelson, assisted defendants at trial, testified regarding the effects of child interview techniques, and was prepared to express opinions on the suggestibility of the investigative and therapeutic practices employed. In these circumstances, defendants' motions for further psychological examinations were properly denied. See United States v. Spotted War Bonnet, 882 F.2d 1360, 1362 (8th Cir. 1989)(subsequent history omitted) (interview properly denied because defense expert reviewed other interview records and was present when victim testified).

In denying these motions, the district court properly considered the victims' interests, requiring defendants to show good cause for this "additional intrusion into the alleged victims already troubled lives." An adult witness may simply refuse to undergo adversarial medical or psychological examinations. See United States v. Bittner, 728 F.2d 1038, 1041 (8th Cir. 1984). With children in protective custody, the issue is more complex because they are not able to make such decisions for themselves. The trial court must protect a criminal defendant's right to a fair trial, but it must also protect the State's paramount interest in the welfare of the child. At a minimum, the court should heed a custodial agency's opinion that an investigative or adversarial examination is unnecessary or unwise.28

Here, the children's guardian opposed further psychological examinations, particularly by adversarial experts, and defendants did not show need for the requested examinations. On this record, the majority's decision that the district court nonetheless abused its discretion in declining to order the examinations raises a barrier to the prosecution of these kinds of crime by maximizing the trauma that victims must routinely endure. Congress has repeatedly legislated the opposite public policy, for example, in enacting Rules 412, 413, and 414 of the Federal Rules of Evidence. This aspect of the majority's decision is, in a word, lawless.

IV. Conclusion.

I have carefully reviewed defendants' other contentions on appeal and conclude that each is without merit. One -- the allegedly biased juror -- is mentioned by majority in yet another gratuitous slap at the way the district court conducted this trial. In fact, after a thorough evidentiary hearing on this issue, the court found that juror Pickard was the target of a spiteful co-worker whose testimony was not credible. It further found that juror Pickard had not concealed "any racially prejudiced attitudes, beliefs, or opinions" and that "no improper outside influence affected the jury." These findings established that defendants were not entitled to a new trial on this ground. See United States v. Tanner, 483 U.S. 107, 120-27 (1987); McDonough Power Equip., Inc. v. Greenwood, 464 U.S. 548, 556 (1984); United States v. Whiting, 538 F.2d 220, 222-23 (8th Cir. 1976); Fed. R. Evid. 606(b). On this record, for the majority to suggest that defendants did not "receive a fair trial without the impact of racial bias" is outrageous.

This was a difficult case to try. The record reflects that the district court dealt carefully, fairly, and impartially with the many issues that arose before, during, and after the trial, and that the jury deliberated carefully in convicting defendants on some counts and acquitting them on many others. The majority now likens this to the Salem Witch Trials! That is an indictment of all the government officials involved in these proceedings -- and most particularly, of an exceedingly competent and fair United States District Judge -- that I cannot abide. I would affirm the judgments of the district court.