Jurisdiction for this appeal is under 18 U.S.C. § 3595 as to the capital counts; jurisdiction for the non-capital counts is based upon 28 U.S.C. § 1291 and 18 U.S.C. § 3742(a). Jurisdiction in the District Court was based on 18 U.S.C. § 3231. Notices of Appeal were filed on February 20 and April 13, 1998, following denials of Motions for a new guilt phase hearing and a new sentencing phase hearing.

I.  DID THE DISTRICT COURT COMMIT PREJUDICIAL ERROR IN THE SENTENCING PHASE?

A. Did the District Court Err in Refusing to Permit the Defendant to Rebut a New Diagnosis of Psychopathy First Offered by the Government in Rebuttal?

B. Did the District Court Err in Refusing to Make a Determination that the Bases for the Government’s Expert’s Opinion -- the Use of a Questionable Screening Test and the Defendant’s Race, Sex, Age and Poverty -- Were Probative and Reliable?

C. Did the District Court Err in Incorrectly Instructing the Jury that if it Could not Agree Between Sentences of Death and Life Without Release that the Defendant Could be Given a Lesser Sentence?

D. Did the District Court Err in Submitting an Unconstitutionally Vague Non-Statutory Aggravating Factor, "Victim Impact," and in Permitting Evidence on this Factor that Violated Due Process?

E. Did the District Court Err in Submitting as a Statutory Aggravating Factor in Robin Williams’ Death that her Murder Created a "Grave Risk of Death" to Others that was not Supported by the Evidence?

F. Did the District Court Err in Denying the Defendant’s Right of Allocution Before His Sentencing Jury?

II. DID THE DISTRICT COURT COMMIT PREJUDICIAL ERROR AT THE GUILT PHASE?

A. Did the District Court Err in Denying the Defendant’s Motion to Dismiss Counts 1, 2, 3, 10 and 11 -- All Involving Crimes That Occurred in Roanoke, Virginia -- for Improper Venue?

B. Did the District Court Err in Submitting Counts 1, 2, 3, 10 and 11 Based Upon Violation of the Violence Against Women Act When Robin Williams and the Defendant Were Not "Intimate Partners" as Defined by the Act?

III. DID THE DISTRICT COURT COMMIT PREJUDICIAL ERROR DURING THE SELECTION OF THE JURY?

A. Did the District Court Err in Excluding for Cause Under Witherspoon- Witt A Juror Whose Personal Beliefs Did Not Prevent or Substantially Impair His Ability to Consider the Death Penalty as a Sentence Option?

B. Did the District Court Err in Permitting the Government to Peremptorily Exclude an African-American Juror Based Upon Her Race?

IV. WERE THE SPECIAL VERDICTS IMPOSED UNDER THE INFLUENCE OF PASSION, PREJUDICE OR ARBITRARINESS? 

III.  STATEMENT OF THE CASE

This is an appeal from three sentences of death and a sentence of life plus 540 months rendered on 11 convictions contained in a single Bill of Indictment against the defendant, Aquilla Marcivicci Barnette ("Mark Barnette"). On February 4, 1997, the defendant was indicted on 11 counts relating to the deaths of Robin Williams and Donald Allen including two counts of Interstate Domestic Violence, 18 U.S.C. § 2261, one count of possessing a firearm during a crime of Interstate Domestic Violence, 18 U.S.C. § 924(c)(1), one count of arson in the commission of a felony, 18 U.S.C. § 844(d), one count of providing false information to acquire a firearm, 18 U.S.C. § 922(a)(6), one count of making a firearm, 26 U.S.C. §§ 5861(f) and 5871, one count of possession of a firearm by a felon, 18 U.S.C. § 922(g)(1) and 924(e), one count of carjacking resulting in death, 18 U.S.C. § 2119(3), one count of using a firearm in a carjacking resulting in death, 18 U.S.C. § 924(c) and (j), one count of interstate transportation of a stolen vehicle, 18 U.S.C. § 2312 and one count of using a firearm in an act of Interstate Domestic Violence resulting in death, 18 U.S.C. § 924(c) and (i). Counts Seven -- Carjacking resulting in death -- Eight -- Use of Firearm in Carjacking resulting in death -- and Eleven -- Use of firearm in commission of act of Interstate Domestic Violence resulting in death -- all carry the death penalty.

Trial was had before the Hon. Robert Potter with Jury in the Western District of North Carolina beginning on January 5, 1998. The defendant was convicted on all counts of the indictment on January 27, 1998. The defendant filed a Motion for a New Trial on January 28, which was denied by written Order on February 18, 1998.

Pursuant to 18 U.S.C. § 3591 et. seq., a sentencing proceeding was held beginning on January 29, 1998, before the same Jury. On February 10, 1998, the defendant was sentenced to death on Counts Seven, Eight and Eleven of the indictment by special verdict. The defendant requested an extension of time within which to file a Motion for a New Sentencing Hearing on February 12, which was granted by written Order on February 17. On February 20, Mark Barnette was sentenced to a term of life imprisonment plus 540 months on the non-capital counts and a Notice of Appeal was filed. On April 1, 1998, the District Court denied the defendant’s Motion for a New Sentencing Hearing by written order. A Notice of Appeal was filed on April 13, 1998.¹

On June 25, 1996, in his mother’s West Boulevard home and surrounded by his family, Mark Barnette voluntarily gave himself up to agents of the Federal Bureau of Investigation and investigators of the Charlotte-Mecklenburg Police Department ("CMPD"). GP Tr. at 410.² Wearing a tie, white shirt, dark pants and holding a Bible, Barnette said goodbye to his mother, younger brother, aunt and grandfather, and then voluntarily accompanied these agents and investigators outside to their cars where he was handcuffed and advised of his rights under Miranda.³ GP Tr. at 411. His surrender came three days after he was identified as the person who had shot and killed Robin Williams in Roanoke, Virginia. GP Tr. at 405-6. During these three days, Barnette had traveled from Roanoke to Knoxville, Tenn., where, after seeking counseling at a church, he twice attempted to take his own life. Gov’t Ex. 36 at 1082; Gov’t Ex. 51B at 1423-1434. Failing in these efforts, Barnette returned to his family in Charlotte to surrender to police.

Once at the F.B.I. offices, Barnette was again advised of his rights. GP Tr. at 412-13. He acknowledged Robin Williams’ murder in Roanoke and was asked about the location of the driver of a blue Honda automobile, the automobile that had been identified as the car used during the Roanoke murder:

A. (Special Agent Womble): I told Mr. Barnette that he didn’t have to answer the questions if he didn’t want to, but that we were concerned about the vehicle that had been recovered the previous day, the blue Honda, we were concerned about the whereabouts of the individual that the vehicle belonged to, we were concerned about his safety, and if he had any idea where the person may be, if he was in need of medical assistance or if he was bound and gagged someplace, please let us know what the story was.

Mr. Barnette asked for a piece of paper from the note pad that I was writing my notes on. I gave him a piece of paper. Investigator Mike Graham gave him a pen and he wrote, corner of Billy Graham Parkway and Morris Field, left side drainage ditch, on a piece of paper and slid it back on the table to us.

GP Tr. at 413-14. At the request of law enforcement, Barnette lead investigators to the body of Donald Allen, the owner of the blue Honda. Once there, "[t]he defendant started crying and pointing across Morris Field" to the location of the body. Gov’t Ex. 36 at 1079.

As Barnette was being taken from the scene, he was asked by investigators to identify the car that he used to commit Robin Williams’ murder, the car that belonged to Donald Allen. Barnette agreed and was taken to the CMPD Law Enforcement Center where he identified Donald Allen’s blue Honda. Gov’t Ex. 36 at 1079. Among the items recovered from the car were a church bulletin from June 23, duct tape taken from the tailpipe, a garden hose consistent with a failed attempt at carbon monoxide poisoning, and love letters between Barnette and Williams. Def. Ex. 13, 14, and 16.

Shortly thereafter, while still at the Law Enforcement Center, Barnette was interviewed by investigators for the CMPD, first on videotape and then by audiotape. Three days later, Barnette again was interviewed on videotape by a CMPD investigator. These tapes, and the information supplied by Barnette after surrendering to law enforcement, provide the bulk of the facts surrounding these crimes. They are, in many instances, disjointed because Barnette’s interrogators were forced to stop the tapes on numerous occasions to allow Barnette to gain control of his emotions and grief. GP Tr. at 472, 477; Gov’t Ex. 37B at 1603; Gov’t Ex. 51B at 1418, 1419, 1421.

Mark Barnette and Robin Williams met in the Spring of 1994. GP Tr. at 311. They were both almost 21. By March 1995, Barnette had moved from Charlotte to Roanoke -- Williams’ hometown -- to live in Robin Williams’ apartment. Williams’ family testified at trial that she was in love with Barnette and thought "the world" of him. GP Tr. at 312. Barnette found work in Roanoke, as did Williams. However, within a year, largely as a consequence of Barnette’s irrational and obsessive jealousy about Williams, their relationship ended and Barnette moved back to his mother’s house on West Boulevard in Charlotte on April 10, 1996. GP Tr. at 319; Gov’t Ex. 37B at 1608-9.

Barnette’s irrational obsession with Williams did not stop after he moved from Roanoke. He continued to talk about her and, according to testimony presented by the Government, she was "always" on his mind. GP Tr. at 157. Barnette called Williams and pleaded with her to reestablish their relationship; he also threatened her and accused her of infidelity. In response, Williams changed the locks on her apartment and, for a time, lived with her mother.

By April 30, 1996, Williams had moved back into her apartment, accompanied by a male friend, Benjamin Greene. In the early morning hours of April 30, Greene was awakened by Williams, who told him that "he was here." GP Tr. at 85. When Greene went into the living room of the apartment, he saw Barnette striking at the windows with a baseball bat, screaming obscenities and threatening to kill Williams. GP Tr. at 87. The telephone lines to the house had been cut. A container with gasoline had been thrown into the apartment and the front part of the living room was in flames. Greene retrieved a 9mm Ruger and shot at Barnette, who fled. GP Tr. at 86-9. Greene and Williams escaped the burning apartment by jumping out of a rear window. GP Tr. at 91. Greene was uninjured, but Williams, in attempting to put out the fire, suffered second and third degree burns on her hands and arms. GP Tr. 89.

Barnette was identified by Williams to the Roanoke police, who then issued warrants for his arrest. By Telex, the CMPD was notified of Barnette’s crimes, the outstanding arrest warrants, and his home address on West Boulevard.⁴ Testimony presented by the Government revealed that Barnette, after confessing this incident to his mother and best friend, waited at his mother’s home on West Boulevard for the police to arrest him. GP Tr. at 161-62. However, the CMPD reported to the Roanoke police that this address -- an address that had existed for more than 30 years, one at which CMPD had answered a number of calls, and the address at which Barnette would surrender on June 25 -- did not exist. Def. Ex. 1, 3, 4, 5, 7, 8, 10, and 18. Barnette continued to call Williams and her family and, on at least one occasion, left cards that she had sent to him on the windshield of her car. Then, in late June, Barnette purchased a semi-automatic shotgun, using his brother’s identification and name. Gov’t Ex. 37B at 1613; Gov’t Ex. 51B at 1412. Barnette told police that he sawed off the barrel of the gun, taped a flashlight to it, and packed a bag with bolt cutters, a crow bar and screwdrivers. Gov’t Ex. 51B at 1412. After consuming at least six beers, Barnette walked to a dark intersection on Billy Graham Parkway and waited for a car to stop at the light. Donald Allen stopped at the intersection after midnight on June 22.

Mark Barnette approached Allen’s car with the shotgun and told him to get out. He told Allen to throw down his wallet at the side of the road, and then he and Allen walked to a drainage ditch beyond sight of the road. Barnette told police that he told Allen to turn around and then shot him in the back, firing three times.⁵ Gov’t Ex. 51B at 1415-16. The CMPD Investigator who interviewed Barnette asked: "Mark, Why did you shoot him?" Barnette’s reply was, "AI don’t know. I still don’t know (CRYING)." Gov’t Ex. 51B at 1419. Barnette took Allen’s wallet, retrieved his bag, and drove away in Allen’s car. He then drove to Roanoke to Robin Williams’ mother’s house.

Barnette told police that he arrived in Roanoke at about 5:00 am, and sat in the car until 7:00 am. He then drove the car to the rear of the house, entered the property by the rear gate and cut the telephone lines to the house. Gov’t Ex. 37B at 1603-4. Barnette went to the kitchen door and, finding it locked, shot and kicked the door in. Gov’t Ex. 37B at 1605. As Barnette entered the house, he saw Williams leaving through the front door and encountered her mother, who was holding a baby. Barnette ran through the house, out the door, and chased Williams across the street, catching her after she had fallen down. Gov’t Ex. 1606-7; Gov’t Ex. 51B at 1422. He then grabbed her by the arm and dragged her back to the house telling her to "come with me and ask her why did she do this, why did she take everything away from me. . . . asking her why Benny [Benjamin Greene]." Gov’t Ex. 37B 1607. Barnette told Williams that he had one shell for her and another for himself. Gov’t Ex. 37B at 1609. When Williams broke free, and as she was running towards her mother, Gov’t Ex. 37B at 1609, Barnette shot her twice:

H(Inv. Holl): Okay, how many times did you shoot at Robin?

B(Barnette): (CRYING) . . . . Two . . . . (CRYING)

H: Twice? Yes or no.

B: Yes.

H: What did you do after the shooting occurred in Roanoke?

B: I . . . (CRYING).

Gov’t Ex. 51B at 1421.⁶ On the day of his arrest, Barnette was asked, "if you know, what do you feel your main reason was for shooting her?":

I’m still trying to find that out. I wanted to die. I wanted to kill myself and I wanted her with me, just one part of me, another, is that I was mad that she had threw our relationship out the window for somebody who I didn’t know existed and another part was that I was just so upset that I was depressed I didn’t know what I was going to do about the relationship we had because it was over in live [sic] one night and we had been together for two years . . . I don’t know why

Gov’t Ex. 37B at 1614.

Barnette left the scene in the stolen car and drove to Knoxville, Tennessee. He went to stores in the Knoxville area to purchase a garden hose and attempted to kill himself with carbon exhaust poisoning. Gov’t Ex. 51B at 1423. After this attempt failed, Barnette stole a license plate so that he would not be stopped before he tried to kill himself again. Gov’t Ex. 51B at 1426-27. He then went to church:

H: Why did you go to church?

B. Because I knew I was wrong I shouldn’t have done that stuff and I, I felt like that I needed to ask the Lord to forgive me for what I had did . . . I don’t know what had happened to me, why I did that, to where I let my anger and hurt feelings and all that stuff build up inside me to where I would hurt anybody like that and I wanted Jesus to forgive me and to take care of them because I was wrong and I know better. (CRYING).

Gov’t Ex. 37B at 1614.

Barnette called his mother in Charlotte to tell her that he was sorry, Gov’t Ex. 51B at 1428, and again attempted suicide using a combination of sleeping pills and exhaust poisoning. Gov’t Ex. 51B at 1429-1434. When this attempt was interrupted, he returned to Charlotte and surrendered.

The Prosecution of Barnette

Barnette was arrested by federal agents acting on UFAP warrants from Virginia for the murder of Robin Williams. However, within three days of his arrest, the federal charges against Barnette were dismissed; he was subsequently served with murder warrants from North Carolina for the death of Donald Allen. For the next eight months, proceedings in both Virginia and North Carolina went forward on the capital murder charges.

On February 4, 1997, the federal government asserted its jurisdiction in this case, indicting Barnette in the Western District of North Carolina on eleven counts involving the murders of Donald Allen and Robin Williams, the arson attack on Robin Williams’ apartment in April, Barnette’s purchase of the shotgun using false identity, his possession of the shotgun as a convicted felon, and his alteration and use of the shotgun in committing these killings. Ten days later, the Government served its Notice of Consideration for Death Penalty. On February 28, defense counsel was appointed for Barnette.

On March 26, 1997, Barnette was arraigned on these counts. Counsel filed a number of motions including motions raising mental illness as a potential issue, contesting the jurisdiction of the federal government to try Barnette for violations of the Interstate Domestic Violence Act, the carjacking statute, and challenging the venue of the Western District of North Carolina to try the offenses that took place in Roanoke, Virginia.

On August 7, 1997, the Government filed its Notice of Intent to Seek Death Penalty. The Government’s Notice was based upon Counts 7, 8 and 11 of the Indictment -- carjacking resulting in the death of Donald Allen, using and carrying a firearm in the carjacking of Donald Allen, and using and carrying a firearm during a crime of violence -- interstate domestic violence -- in the death of Robin Williams. Notice of Intent to Seek the Death Penalty 8/7/97 at 1; 18 U.S.C. §§ 2119(3), 924(c) and (j), 2261(a)(1) and (b). The Notice specified three statutory aggravating factors in Allen’s death -- pecuniary gain, 18 U.S.C. § 3592(c)(8), substantial planning and premeditation, 18 U.S.C. § 3592(c)(9), and multiple killings, 18 U.S.C. § 3592(c)(16) -- and three non-statutory aggravating factors -- the impact of Allen’s death on his family ("victim impact"), the future dangerousness of Barnette premised upon specified past acts of violence, and Barnette’s killing of two people. Notice at 2-7. It also specified three statutory aggravating factors in the death of Robin Williams -- grave risk of death to more than one person, 18 U.S.C. § 3592(c)(5), substantial planning and premeditation, and multiple killings -- and three non-statutory aggravating factors -- victim impact, Barnette’s future dangerousness and his killing of two persons. Notice at 3-7. Following the filing of this Notice, the District Court denied Barnette’s jurisdictional, venue, and dismissal motions and scheduled his trial to begin on January 5, 1998.

The Guilt Phase

Jury selection proceeded in three parts: first, prospective jurors were brought to the court to fill out a Juror Questionnaire created by the District Court and counsel; second, the prospective jurors were examined for the purpose of exercising "for cause" challenges; and, third, the parties exercised peremptory challenges until satisfied with the jury. Three potential jurors were African-American; two were struck by the Government and the jury that was eventually empanelled to try Barnette was all white.⁷

The Guilt Phase began on January 21, 1998. Over the next three days the Government presented the testimony of 44 witnesses and introduced 212 exhibits. In light of Barnette’s statements to police and the testimony of eyewitnesses to the Roanoke killing, much of the Government’s case was not seriously contested. The defense presented no witnesses and introduced 14 exhibits, most relating to the failure of the CMPD to locate Barnette and evidence of his suicide attempts.

The Jury deliberated two hours and five minutes before finding Barnette guilty on all 11 counts of the indictment.

The Sentencing Phase

1. The Government’s Case in Aggravation.

The Government began the Sentencing Phase with the presentation of Barnette’s videotaped confession, a videotaped confession that it had objected to at the Guilt Phase because it demonstrated excessive remorse. GP Tr. at 544-48. The next 11 witnesses called by the Government testified solely on the non-statutory aggravating factor of "future dangerousness"; all were called to testify concerning prior acts of violence by Barnette. These acts dealt with both adjudicated felonies -- for cruelty to children involving Barnette’s use of a coat hanger to spank children when he was 19 and for felonious restraint involving an assault on a girlfriend -- and unadjudicated acts -- an assault on his girlfriend in Georgia over his discipline of her children, and charges of rape and attempted kidnapping relating to the felonious restraint conviction. Significantly, none of these acts were prosecuted by the local authorities and, indeed, the claims for rape and attempted kidnapping were rejected for prosecution. The most violent of these acts involved an individual named Alesha Chambers who accused Barnette of rape and attempted kidnapping; she admitted on cross-examination that shortly after those events she called Barnette and spent the day in a hotel room with him having consensual sex. SP Tr. at Tr. 181-83. Significantly, at closing, the Government had conceded that Chambers was not credible. SP Tr. at 1127. Crystal Dennis, who was called to testify concerning Barnette’s unprosecuted assault against her and his guilty plea for cruelty to her children, twice told the jury that Barnette was "cold-hearted" and called him a "cold-hearted motherfucker."⁸ Tr. 112-23. No determination was made by the Court as to the trustworthiness of the unadjudicated acts and defense counsel’s objections to this testimony were overruled. SP Tr. at 143, 145, 218.⁹

The Government’s final seven witnesses were members of the families of Donald Allen and Robin Williams; all were called in support of the non-statutory aggravating factor of "victim impact." Over objection, SP Tr. at 303, 306-7, members of Donald Allen’s family were permitted to show home videos and family pictures to the Jury, Gov’t Ex. 62, SP Tr. at 261-62, and were allowed to read poetry to the Jury concerning Donald Allen’s death. SP Tr. at 263. Members of Robin Williams’ family were permitted to display her hospital gown to the Jury, Gov’t Ex. 63C, and her mother was allowed to testify as to her pride in her accomplishments, to read poetry to the Jury and to introduce a "Memorial" advertisement placed in the newspaper on the anniversary of her death. SP Tr. at 292-93, 295-96, 298-300; Gov’t Ex. 63B, 63E and 63F.

Thus, of the 19 witnesses called in the Government’s case-in-chief in aggravation, none testified concerning a statutory aggravating factor; 18 of the 19 testified concerning two non-statutory aggravating factors -- future dangerousness (based on past acts) and victim impact.

2. The Defendant’s Case in Mitigation.

Barnette presented the testimony of 22 witnesses including family members, jailers, psychiatrists and psychologists. This testimony revealed that Mark Barnette was born to a 14 year old and into a family marked by alcohol abuse, drug abuse and violence. SP Tr. at 337, 636-38. Barnette’s maternal grandmother -- the only stable provider in the family -- was murdered by her second husband when Barnette was 10 months old. SP Tr. at 329 Thereafter, Barnette and his then 15 year old mother were shuttled among a series of relatives including his disabled and alcoholic maternal grandfather and his aunt. When Barnette was approximately three years old, his mother married his putative father who had just turned 19. The marriage lasted approximately 9 years during which Barnette’s younger brother, Mario, was born. Testimony at trial revealed that Barnette and his mother were beaten by his father, SP Tr. at 347-48, both with his hands and with objects, typically a belt. During Mark’s beatings, his mother would often lock herself in the bathroom so that she could not hear his screams. SP Tr. at 351. Mark observed his mother being beaten and struck by his father, violence that escalated as their marriage dissolved. SP Tr. at 348-51. Much of the violence against Barnette’s mother followed his father’s repeated suspicions and accusations of infidelity.

In approximately 1984, when Barnette was 11 or 12 years old, his parents separated. Shortly thereafter, as part of the divorce proceedings, Barnette’s father insisted on paternity testing; the results revealed that neither Barnette nor his younger brother were the biological children of his father. SP Tr. at 357-58. Upon learning these results, Barnette’s father informed Mark and his brother that they were not his children and that their mother had been "running around having other men’s babies"; he then terminated all child support, moved to Philadelphia and lost contact with Mark and Mario. SP Tr. at 700. Following this revelation, Barnette’s mother began to use drugs and alcohol and embarked on a series of moves that injected even more chaos into Barnette’s family. SP Tr. at 354-56. Over the course of his childhood, Barnette was moved some 22 times and enrolled in nearly a dozen different schools. During this time Barnette witnessed a number of violent acts against his family and in the areas in which he lived; he attempted suicide and made suicidal threats on a number of different occasions, starting at the age of 8. SP Tr. at 700-03 Barnette finally dropped out of the 11^th grade at 17; by this time he had fathered a child and moved in with the mother and away from his family. Each relationship that Barnette established from that time until the time of the crimes ended in violence or accusations of violence by Barnette based upon his suspicions of infidelity.

Psychiatric and psychological examinations and tests presented by the defense revealed that Barnette was suffering from a Personality Disorder, Borderline Personality Disorder, a Mood Disorder characterized by depression and suicidality over time, an intermittent explosive disorder, and aspects of manic-depression characterized as approaching a Bipolar disorder with narcissistic and antisocial elements. SP Tr. at 644-47, 706-14. As described by these witnesses, Barnette’s upbringing -- characterized as it was by violence against women, infidelity, abandonment, instability and chaotic moves -- led to the development of an inadequate personality structure in which he perceived that violence was an acceptable way of dealing with women and an acceptable reaction to rejection, in which women were always perceived as unfaithful, and in which paranoid fears of jealousy, rage and infidelity dominated his perception of the world. His multitude of psychiatric disorders meant that he was impulsive, emotionally unstable, experienced transient paranoia and would engage in frantic, almost manic activity to maintain female relationships. SP Tr. at 702-04, 645-47.

Defense counsel also presented the testimony of a number of jailers and other correction officials concerning Barnette’s ability to adapt to prison; all testified that in the nearly two years Barnette had been imprisoned awaiting trial he had never received an infraction write-up and was not considered a threat to himself, to others in the inmate population, or to corrections officials. Barnette’s final witness in mitigation, Dr. Mark Cunningham, utilized a statistical analysis of prison violence, violence by prisoners convicted of murder, and Barnette’s psychological profile, diagnoses and history, in an effort to predict his "future dangerousness." While acknowledging the impossibility of definitively predicting whether any individual will or will not be violent in the future, SP Tr. at 797-98, Dr. Cunningham presented statistics generated from, among other sources, the Bureau of Prisons, that demonstrated that persons convicted of homicide had .9 infractions per inmate per year, that the serious assault rate on prison staff was approximately .019% per year and that the serious assault rate on other inmates as .22% per year. SP Tr. at 814-27. Utilizing these sources, Dr. Cunningham testified that the inmate homicide rate in federal prison is approximately 7 per 100,000 per year and the staff homicide rate is approximately 1 per 2 to 3 million inmates per year,¹⁰ that violence and assault rates are lower for long-term prisoners than short-term prisoners, that these rates are lower for those convicted of homicide than for other crimes, and that these rates lower with age. SP Tr. at 823-29. In light of Barnette’s psychological disturbances and his prior history of violence -- all of which centered on his relationships with women -- Dr. Cunningham testified that, in his opinion, the likelihood that Barnette would engage in serious violence if imprisoned for life was low. SP Tr. at 838.

3. The Government’s Case in Rebuttal.

The Government sought to meet this evidence in rebuttal through the testimony of 5 witnesses: 3 former co-workers of Barnette’s, his probation officer and Dr. Scott Duncan, a Bureau of Prisons psychologist. Barnette’s co-workers testified about incidents in which Barnette was accused of sexual harassment, SP Tr. at 883-94, and had made lewd sexual suggestions, SP Tr. at 899-909; one testified that she engaged in a consensual sexual relationship with Barnette during the time he lived with Robin Williams, SP Tr. at 894-99. Barnette’s probation officer testified about his unsuccessful attempts to monitor Barnette during the probationary term he received in North Carolina for the conviction for felonious restraint against Alesha Chambers, and Barnette’s deceptions concerning his employment and place of residence.¹¹ SP Tr. at 909-50.

As its last witness, the Government presented Dr. Scott Duncan, a Forensic Psychological Coordinator with the Bureau of Prisons in Atlanta, Georgia. Dr. Duncan employed as the basis for his opinion an "instrument" known as the PCL-R checklist -- Psychopathy Checklist Revised -- which he scored based upon two interviews with Barnette performed in the two weeks before trial. Defense counsel had moved prior to this testimony to bar the use of this instrument based upon its lack of scientific reliability. Motion in Limine, 2/5/98. The PCL-R was developed based upon studies of white males in the Canadian criminal justice system and has never been standardized or found scientifically applicable to African-American males. SP Tr. at 1019-22. Its purpose is to diagnose whether a person is a "psychopath," a mental condition that Dr. Duncan admitted was a "theoretical construct" which has never been established as a diagnostic condition. SP Tr. at 1012, 973. The PCL-R consists of 20 "categories" including "glibness and superficial charm," "lack of remorse," "promiscuous sexual behavior," "irresponsibility," and whether a person has violated probation. SP Tr. at 976-77. The "examiner" scores the subject from 0 to 2 in each category, with 2 being the highest. A total score of 30 -- with an upwards adjustment of 2 points for African-Americans -- indicates psychopathy. Dr. Duncan scored Barnette initially at 32 and then, prior to his testimony, re-scored Barnette based upon a review of his notes to 34. Despite the fact that Dr. Duncan had never used this "instrument" in a death penalty case and that it has never been standardized for such use, the District Court denied defense counsel’s motion in limine without a hearing and permitted Dr. Duncan, over objection, to testify about the PCL-R as a scientifically valid and acceptable test.¹²

Dr. Duncan testified that Barnette fit the "theoretical construct" of a psychopath; this meant that Barnette was callous and cunning and able to "look normal," like "fake fruit in a bowl." SP Tr. at 977. According to Dr. Duncan, this diagnosis, when combined with Barnette’s criminal behavior, meant that Barnette was twice as likely to engage in future criminal behavior and three times as likely to engage in violence. SP Tr. 974. Dr. Duncan’s prediction of future violence by Barnette was also based upon studies done by John Monahan which list 11 risk factors for future violence; four of these factors are race, sex, age and poverty, with the conclusion that poor, young, black males are at the highest risk of future violence. SP Tr. at 984. Dr. Duncan’s ultimate conclusion was that "I think I could say with a reasonable degree of medical or psychological certainty that I feel that the defendant is at a much greater risk of being violent in the future." SP Tr. at 988.

To rebut this newly-presented diagnosis of "psychopathy" and the use of the PCL-R as a valid scientific instrument, defense counsel sought to recall Dr. Cunningham.¹³ The Government objected, arguing that since Dr. Duncan’s opinion had been made known to defense counsel in his report, the defendant should have anticipatorily rebutted this diagnosis and this instrument during the course of his evidence. SP Tr. at 1027. The District Court sustained the Government’s objection and barred Dr. Cunningham from testifying on the ground that, since defense counsel had cross-examined Dr. Duncan on the validity of the PCL-R using a published article, then Dr. Cunningham’s testimony would be cumulative. SP Tr. at 1029-30. The District Court permitted defense counsel to introduce the article into evidence, but indicated that it would not let the jury read the article; the Court also refused to permit defense counsel to call Dr. Cunningham for the purposes of making a record of his testimony, SP Tr. at 1041-42, instead instructing counsel to summarize what that testimony would have been. SP Tr. at 1042-43.

4. The Charge Conference and Closing Argument.

Following the close of the evidence, the District Court held a charge conference on the instructions to be given to the Jury. Both the defendant and the Government tendered instructions to the Court. Over the defendant’s objection, the District Court submitted as a non-statutory aggravating circumstance "victim impact," the statutory aggravating circumstance in Robin Williams’ murder that a "grave risk" of death was posed to others, and instructed the Jury that, should it be unable to determine whether death or imprisonment for life without the possibility of parole was the appropriate sentence, then the Court would sentence the defendant to a "lesser sentence." Barnette moved for, and the District Court denied him, an opportunity to allocute before the Jury prior to its imposition of sentence. SP Charge Tr. at 6-11.

Both counsel for the Government and for the defendant argued in closing; the Government took the first and last argument. During the course of the Government’s last argument, the Assistant United States Attorney labeled Dr. Cunningham as a "liar," who would testify to anything for a fee. SP Tr. at 1108, 1109. He personally attacked defense counsel, asking the Jury to "consider whether the lawyers who present that kind of evidence to you really want you to base your decision on the evidence in this case, on the instructions of the Court, on your oath," SP Tr. at 1109-10, and urged, "[d]on’t let these defendants -- these defense attorneys make you responsible for the defendant’s death." SP Tr. at 1112. Over objection, the "US" told the jurors that, if they believed that the aggravating circumstances had been proven and outweighed the mitigating circumstance and they did not vote for death, "you have not been true to your oath, and I don’t know how you are going to be able to live life with not being true." SP Tr. at 1112. Telling the Jury that they were a "voice for the voiceless," counsel argued that when Robin Williams’ mother screamed "Jesus" as her daughter was shot, "[t]hat was a prayer for her daughter’s life. I contend to you that was a prayer for justice today"; he concluded with an appeal that God required the ultimate punishment: "For God’s sake, judge this man as a fully human being and impose on him the ultimate punishment for the ultimate crime of premeditated murder of Donnie and Robin." SP Tr. at 1121. Defense counsel’s Motion for a Mistrial at the conclusion of this argument was denied. SP Tr. at 1138-39.

5. The Special Verdicts.

The Jury deliberated for more than 7 hours over the course of two days before returning its special verdicts. The Special Verdicts reveal that the Jury unanimously found each of the two statutory and three non-statutory aggravating factors submitted to it in each murder: pecuniary gain, substantial planning and premeditation, victim impact future dangerousness and multiple killings¹⁴ in the death of Donald Allen, and grave risk of death, substantial planning and premeditation, victim impact, future dangerousness and multiple killings in the death of Robin Williams.

Against these aggravating factors, the jury weighed 29 statutory and non-statutory mitigating factors found by at least one juror. These mitigating factors included the factors that: Barnette assisted police in locating Donald Allen’s body (12), he voluntarily turned himself in (9), he was physically and emotionally abused by his father (7), he was neglected by his mother when she was drunk and distraught over the break-up of her marriage (9), he grew up in a home where alcohol and drugs were used frequently (12), he grew up in a home where violence was prevalent (11), he confessed to the police (12) and he cooperated with police (11), he had been a model inmate since his arrest (12), he had exhibited psychological problems since he was a minor (6), and he would do well in the structured environment offered by prison (6). Despite the Jury’s unanimous finding that Barnette was "likely to commit criminal acts of violence in the future which would be a continuing and serious threat to society," not only did the Jury unanimously conclude that Barnette had been a model inmate -- with an additional finding by 6 jurors that he would do well in a prison’s structured environment -- but 3 jurors found that if "sentenced to life without the possibility of release he will not be a future danger."¹⁵

In accord with the Jury’s recommendations, the District Court sentenced Barnette to death on Counts 7, 8 and 11. At that time, Barnette addressed the Court:

Since this trial started, I haven’t had a chance to address the families, and I wanted to very badly. And I would just like to tell you that I’ve stared in your eyes and I’ve felt your pain. No matter what these guys say, I’m hurting. This shouldn’t have happened, none of this should have happened. I’m ashamed of myself. I used to be able to look people in the eye, but I can’t do it anymore.

. . . . you’ve heard a lot of bad things about me, but you haven’t heard any good, and that’s fine, because nobody deserves to hear any good about what I did to you. I took something from you that you very much deserved to keep, and I hurt because I know you hurt, and I know what you lost.

They’ve sentenced me to death, that’s fine. The worst thing they could have done is sentence me to life, because I live every day in a little block and think about what I did, and that is painful. That hurts worse than anything I’ve ever felt, and I feel it every day. And I can only imagine what you are going through.

I may have done a lot of bad things, but I’ve never done anything like that. And to think about what I’ve done to you, both families, and to my family, it hurts, and I’m very sorry.

I can’t and I will never ask you to forgive me, because I don’t deserve to be forgiven. I would never ask you for mercy because I don’t deserve it. You don’t deserve a second chance.

They have said a lot of things that I agree with. Regardless of what my attorneys have tried to do, I’ve agreed with the prosecution on a lot of stuff, because I was wrong.

And I want you to know straight from me that I do feel for you, and I do feel for the pain, and I still pray every day, and I will never stop because I know you won’t.

SP Tr. At 1267-69.

Subsequently, on February 20, the District Court sentenced Barnette to a term of imprisonment for life plus 540 months on the remaining counts of the indictment.

V.  SUMMARY OF ARGUMENT

The Sentencing Phase

This case was the first tried in this Circuit under the provisions of the Federal Death Penalty Act, and one of the first in the country. While both the District Court and counsel necessarily operated in a vacuum of federal case law on the specific provisions of the FDPA, the guiding principles of Eighth Amendment jurisprudence are well-settled. In their essence, they demand a 'heightened reliability" when death is imposed. That reliability did not occur in the sentencing phase of this case.

1. The Refusal to Permit Surrebuttal Evidence.

The FDPA specifically grants both the Government and the defendant the right to "rebut any information received" in the penalty phase; the Eighth Amendment requires the grant of this right. See Buchanan v. Angelone, 522 U.S. 269, 118 S. Ct. 757 (1998). The District Court violated these rights, in the process abusing its discretion, when it barred the defendant from presenting surrebuttal testimony to meet a diagnosis of "psychopathy" first offered by the Government in its rebuttal evidence. This Court has, in the context of pleas of insanity, specified the order of proof where a defendant raises his mental condition in his defense. See United States v. Burgess, 691 F.2d 1146 (4^th Cir. 1982). That order specifically contemplates that, when a defendant raises his mental condition, the Government is entitled to meet this proof in rebuttal, and the defendant is entitled to rebut new matters raised by the Government.

The District Court ignored this procedure. At the time it did so, it ruled that the defendant’s proffered testimony was "cumulative" of the cross-examination of the Government’s expert. Later, it ruled that the defendant was required to anticipatorily rebut in his case-in-chief evidence that the Government might later present. Neither ruling was correct nor accurately states the law.

2. The Refusal to Determine the Reliability of the Government’s Expert’s Opinions.

The District Court’s error in denying the defendant an opportunity to present evidence to rebut the new diagnosis offered by the Government was compounded by it refusal to conduct any inquiry into the bases underlying this diagnosis. The District Court’s refusal to inquire into the reliability of this diagnosis as required by Daubert v. Merrill-Dow Pharmaceuticals, 509 U.S. 579, 113 S. Ct. 2789 (1993). The evidence reveals that the Government’s only expert relied upon an interview screening "instrument" that is not generally-accepted, that seeks to test for a mental condition which is merely "theoretical," that has a false positive rate potentially as high as 40%, and that fails to distinguish among nearly 15,000 combinations of factors. More fundamentally, the Government’s expert incorrectly scored the defendant -- a fact that the Jury was not permitted to hear because the District Court barred the defendant’s surrebuttal testimony. In addition, the Government’s expert explicitly relied in part upon the fact that the defendant was a poor, young, black, male in arriving at his opinion that the defendant would be dangerous in prison in the future. These bases are constitutionally impermissible.

3. Misinforming the Jury on the Consequences of Non-Unanimity.

Under the FDPA, a jury may choose between three sentencing options: death, life without release, and a "lesser" sentence. A sentence may be chosen only by a unanimous jury. No reasonable interpretation of the FDPA would conclude that, if a jury could not unanimously decided between death and life without release, that the defendant can receive a "lesser" sentence. Yet, the District Court, over objection and at the urging of the Government, instructed the Jury that if it could not "unanimously agree on any recommendation" then "the Court will sentence the defendant to a sentence other than death . . . which for that count may be imprisonment for up to life without the possibility of release." This instruction misinformed the Jury about the consequences of nonunanimity and necessarily resulted in an unconstitutionally coerced unanimity: pro-death jurors could argue to pro-life holdouts who did not want the defendant to be released that, by continuing their holdout, the defendant could receive a sentence of less than life without release.

4. The Non-Statutory Factor of "Victim Impact" Was Unconstitutionally Vague and Was Supported by Impermissible Evidence.

The FDPA specifically provides that a jury may consider "the effect of the offense on the victim and the victim’s family," though it does not list "victim impact" as a statutory aggravating factor. The Government requested and the District Court submitted the issue of whether these murders "caused harm" to the families of the victims as a non-statutory aggravating factor. This factor as submitted was unconstitutionally vague as it did nothing to "narrow" or "channel" the jury’s discretion in determining whether Barnette should die. See Godfrey v. Georgia, 446 U.S. 420, 100 S. Ct. 1759 (1980). Similar submissions under the FDPA have been found unconstitutionally vague by other courts. United States v. Jones, 132 F.3d 232 (5^th Cir. 1998). Compounding this error, the District Court permitted evidence of "victim impact" which went well beyond the constitutionally permissible "quick glimpse" into the lives of the victims. Rather, the District Court permitted extended and prejudicial testimony, including the showing of home movies, photographs, and the reading of poetry and memorials to the Jury.

5. The Evidence Was Insufficient to Establish That a "Grave Risk of Death" to Others Was Created During Robin Williams’ Murder.

The Government requested the submission of a statutory aggravating factor that the defendant created a "grave risk" of death to others when he murdered Robin Williams. The evidence, however, failed to establish any objective "grave risk" of death to others and, in fact, no others were injured during Williams’ murder. In order for this aggravating factor to act as a sufficient narrowing device -- in order for it not to be vague -- its application must be strictly circumscribed. Under these facts, this factor was not raised; its submission was both error and renders this aggravating factor unconstitutionally vague as applied to these facts.

6. The Denial of the Defendant’s Request to Allocute.

Finally, the District Court erred in denying the defendant an opportunity to allocute to the Jury before it determined his sentence. His request to do so was specific and unambiguous. By denying that request, the District Court failed to follow this Court’s holding in Ashe v. North Carolina, 586 F.2d 334 (4^th Cir. 1978).

The Guilt Phase

While much of the Government’s evidence against the defendant was uncontested and, indeed, was built upon his own words in confession, two serious errors relating to the Roanoke crimes were committed, mandating reversal of Counts 1, 2, 3, 10 and 11.

1. The Impropriety of Venue.

The venue for the crimes committed against Robin Williams was not proper in North Carolina. Rather, as a matter of statutory interpretation and constitutional law, venue was proper only in Roanoke, Virginia. For the capital count, Count 11, Congress has prescribed the place of trial as the district in which the offense occurred, 18 U.S.C. § 3235, and has defined the offense of murder as occurring where the injury was inflicted, 18 U.S.C. § 3236. This place is unquestionably Roanoke, Virginia. As to the non-capital counts, for which no venue is provided, analysis of the statute at issue -- the Violence Against Women Act -- reveals that the offense consists of committing a crime of violence against an intimate partner inflicting bodily injury. All such acts, again, occurred in Roanoke. The District Court thus erred in refusing to dismiss these counts for improper venue.

2. Whether Barnette and Williams Were "Intimate Partners" Under the Violence Against Women Act.

The counts predicated upon the Violence Against Women Act, 18 U.S.C. § 2261 et seq., were further flawed because Barnette was not an "intimate partner" at the time of these crimes as required by this Act. Relying exclusively on the definition of "intimate partner" contained in 18 U.S.C. § 2266 (A), the Government urged that, because Barnette and Williams had cohabited in the past, this fact alone made them intimate partners always for purposes of the Act. This interpretation is wrong. The Violence Against Women Act applies only to those who are or were married, share a child, are presently cohabiting, or cohabited in the past "as a spouse." The Government introduced no evidence that Barnette and Williams lived together as the functional equivalent of husband and wife -- there was no evidence that they shared assets, that they shared obligations, that they held themselves out to the community as the equivalent of husband and wife, that their families considered them the equivalent of husband and wife, or that they even had a sexual relationship. Mere prior cohabitation, devoid of proof that the cohabitation was "as a spouse," fails to meet the requirements for this necessary element of the crimes under the Violence Against Women Act.

Jury Selection

1. The Trial Court Improperly Excluded a Potential Juror For Cause Under Wainwright-Witt.

Potential juror Rodney Bell was questioned extensively by the Government and defense counsel concerning his beliefs about the death penalty. He consistently indicated that, while he did not favor the death penalty, he could nonetheless consider it. The District Court nonetheless struck him for cause based upon an misimpression of what he said and a misunderstanding of the standards enunciated by Wainwright-Witt. The District Court excluded Bell not because his beliefs would prevent or substantially impair his ability to consider death as an appropriate punishment in the sentencing phase, but because he did not favor death when given the option of imprisonment for life without release. In so doing, it ignored then-Associate Justice Rehnquist’s admonition to the contrary in Lockhart v. McCree, 476 U.S. 162, 106 S. Ct. 1758 (1986): "It is important to remember that not all who oppose the death penalty are subject to removal for cause; those who firmly believe that the death penalty is unjust may nevertheless serve as jurors . . . so long as they state clearly that they are willing to temporarily set aside their own beliefs in deference to the rule of law."

2. The Government Improperly Exercised a Peremptory Challenge Based Upon the Race of a Potential Juror.

Barnette, who is African-American, was ultimately tried for his life by an all-white jury. Of the three African-American jurors who survived challenges for cause, the Government struck two peremptorily. Its reason for doing, at least as to potential juror Stephany Jones, was plainly based on race, the Government’s proffered explanation notwithstanding. While the Government justified its strike based upon Jones’ alleged opposition to the death penalty, the record reveals clearly that she expressed no such opposition, unequivocally answered that she could impose the death penalty, and stated specifically that she would not even favor its abolition. In light of the pretext asserted by the Government, its strike of this potential juror violated the precepts of Batson v. Kentucky, 476 U.S. 79, 106 S. Ct. 1712 (1986).

The Sentences of Death Were Imposed Under the Influence of Passion, Prejudice and Arbitrariness

Under the FDPA, this Court is required to conduct an independent review of the evidence in this case to determine whether the sentences of death were the product of passion, prejudice and arbitrary choice. The errors brought forward in this case demonstrate that the sentencing phase of this proceeding was fundamentally flawed; these flaws necessarily undermine any confidence that the heightened reliability mandated by the Eighth Amendment existed in this proceeding. In addition, numerous other errors in this trial contributed to the passion, prejudice and arbitrary choice that led to these sentences. Under the standards enunciated by the FDPA, it is not necessary that these individual errors be sufficient, standing alone, to justify reversal; it is only sufficient that they undermine the heightened reliability and fairness of the proceeding itself. These errors, under these circumstances, do precisely that.

VI.  ARGUMENT

I. THE DISTRICT COURT COMMITTED PREJUDICIAL ERROR AT THE SENTENCING PHASE

The statutory scheme under which Barnette has been sentenced to die, the Federal Death Penalty Act ("FDPA"), was the first general death penalty statute enacted by Congress since the decision in Furman v. Georgia, 408 U.S. 238 (1972). It attempts to create a system of guided discretion through the use of a "weighing" process, one at which the normal rules of evidence do not apply and one which requires consideration by the Jury of "any information relevant" to any aggravating or mitigating factor. 18 U.S.C. § 3593(c). Under this structure, the Jury is to first consider whether the Government has established the existence of any aggravating factors beyond a reasonable doubt; second, it determines whether any one or more jurors believe that the defendant has proven the existence of mitigating factors by a preponderance of the evidence. Then the Jury engages in a "weighing" process in which it is to determine whether the aggravating factors proven "sufficiently outweigh" the mitigating factors proven "to justify a sentence of death." 18 U.S.C. § 3593(e).¹⁶

For purposes of appeal, the FDPA specifies three grounds upon which a sentence of death must be reversed:

(2) Whenever the court of appeals finds that --

(A) the sentence of death was imposed under the influence of passion, prejudice, or any other arbitrary factor;

(B) the admissible evidence and information adduced does not support the special finding of the existence of the required aggravating factor; or

(C) the proceedings involved any other legal error requiring reversal of the sentence that was properly preserved for appeal under the rules of criminal procedure,

the court shall remand the case for reconsideration...or imposition of a sentence other than death.

18 U.S.C. § 3595(c)(2)(A)-(C). The occurrence of error at the sentencing phase of a capital proceeding under the FDPA mandates remand unless "the Government establishes beyond a reasonable doubt that the error was harmless." Id. This standard is, of course, the standard established for constitutional error under Chapman v. California, 386 U.S. 18, 87 S. Ct. 824 (1967). Thus, for any error under the FDPA, the Government’s proof must leave no reasonable doubt that the error "did not contribute to the verdict." 87 S.Ct. at 828.

A. The District Court Committed Prejudicial Error in Refusing to Permit the Defendant to Present Evidence to Rebut the Diagnosis of Psychopathy Raised by the Government in Rebuttal.

The deliberative weighing scheme established by the FDPA provides not only that the defendant and the Government "may present any information relevant" to a mitigating or aggravating factor, but also that the "government and the defendant shall be permitted to rebut any information received at the hearing." 18 U.S.C. § 3593(c). Only information whose "probative value is outweighed by the danger of creating unfair prejudice, confusing the issues, or misleading the jury" may be excluded in the sentencing phase. 18 U.S.C. § 3593(c). In large measure, this statutory right of a defendant to present "any information relevant to a mitigating factor" and to rebut "any information received at the hearing" established by the FDPA is compelled by constitutional considerations. As noted by the Supreme Court last term in Buchanan v. Angelone, "[i]n the selection phase [of a capital sentencing hearing], our cases have established that the sentencer may not be precluded from considering, and may not refuse to consider, any constitutionally relevant mitigating evidence. Penry v. Lynaugh [ ], Eddings v. Oklahoma [ ], Lockett v. Ohio [ ], . . . . Our consistent concern has been that restrictions on the jury’s sentencing determination not preclude the jury from being able to give effect to mitigating evidence." 522 U.S. 269, 118 S.Ct. 757, 761 (1998) (citations omitted).

1. The Sequence of Evidence.

In this case, the Government -- as the party with both the initial burden of going forward and the ultimate burden of proof ¹⁷-- presented its evidence in aggravation first. It did so by relying on the evidence introduced at the Guilt Phase to the statutory aggravating factors and presenting only evidence relating to two non-statutory aggravating factors -- "victim impact" and Barnette’s "future dangerousness." Neither obviously placed Barnette’s mental condition into issue. Rather, Barnette’s mental condition -- including both the nature, the number, and the extent of those conditions -- was raised first in the defendant’s case-in-chief in mitigation.

As in a plea of insanity, notice of Barnette’s mental condition as a mitigating circumstance had been made well in advance of the sentencing hearing. One of the first pleadings filed by defense counsel was a Notice of a possible mental illness defense. Notice of Insanity, 3/26/97. Barnette, on motion of the Government, had been examined for both competency and insanity, and the District Court had permitted an additional psychological examination of Barnette by two Government experts, Drs. Scott Duncan and William Grant. The report of Barnette’s competency exam had been filed with the District Court and sealed until defense counsel confirmed their intention to use mental illness as a mitigating factor at sentencing; expert reports and evaluations were similarly shared at that time between the Government and defense counsel. In structure then, the presentation of Barnette’s mental condition in the sentencing phase closely resembled the structure of an insanity defense at a guilt proceeding:

The normal approach in a case where a defense of insanity is raised contemplates:

1) The Government’s establishment of the performance by the defendant of the criminal acts.

2) The defense-in-chief, in which evidence of insanity is introduced.

3) Rebuttal by the Government, in which an attempt to prove sanity beyond a reasonable doubt is made.

4) Surrebuttal, in which the defendant seeks to produce testimony to contradict the Government’s evidence introduced during rebuttal, to the extent that such testimony should not already have been introduced during the defense-in-chief.

United States v. Burgess, 691 F.2d 1146, 1149 n. 5 (4^th Cir. 1982) citing United States v. Green, 497 F.2d 1068, 1083 (7^th Cir. 1974).

Barnette’s psychiatric and psychological expert witnesses -- Dr. Faye Sultan, Dr. Seymour Halleck, and Dr. Mark Cunningham -- testified that he suffered from a wide-range of mental conditions set forth in the Diagnostic and Statistical Manual of Mental Disorders ("DSM 4"). The DSM 4 was described at the sentencing hearing as "essentially the Bible, the diagnostic manual that psychologists and psychiatrists and other mental health professionals are told to use to choose their diagnoses. . . . it is pretty much the standard text in mental health for how we make differential diagnoses." SP Tr. at 643. Dr. Seymour Halleck, a medical doctor, Professor of Psychiatry at the University of North Carolina at Chapel Hill, and a former Chief of Corrections for the State of Wisconsin, testified that the DSM 4 "was created to give us what we call a reliability, so that if two people looked at the same things they would come up with hopefully the same answer." SP Tr. at 704-5.

Based upon the widely and generally-accepted diagnostic criteria established and recognized in the fields of psychiatry and psychology as set forth in the DSM 4, Barnette’s experts diagnosed him as suffering from a Personality Disorder, from a Borderline Personality Disorder, and from elements of manic-depression approaching a BiPolar Disorder. As described by Dr. Halleck:

People with borderline personality disorder are often suicidal, their moods change constantly and mostly their moods go downward. They have great difficulty in getting along with people, and their problems lead to them getting into states of high, very high anxiety and depression. And in my work, I find that these people probably suffer as much as any group of patients we work with in psychiatry. There are certainly times when depressed patients suffer more. People with borderline disorder can get very, very depressed. People with borderline disorder can also become transiently psychotic so that they can distort reality and at those points they are suffering as much as the schizophrenic patient. So even if it’s a personality disorder diagnosis, I consider it a major mental illness.

SP Tr. at 712. Dr. Mark Cunningham, a Board-certified Forensic Psychologist with a Ph.D. in psychology who was accepted as an expert in violence risk assessment without objection by the Government, described this Disorder and its implications for Barnette’s future in prison:

The borderline personality disorder that Dr. Halleck diagnosed is a disorder that is characterized by profound anxiety and distress at the prospect of being abandoned. And he [Barnette] becomes excessively preoccupied about this abandonment and begins to experience a mix[ture] of depression, rage, desperate futility that his life is over and that nothing has meaning, begins to have suicidal impulses and homicidal ideas towards his female partner.

Now, other part's of this are feeding it as well. The hurt and anger that he experienced is part of what feeds this obsessive possessiveness and controlling behavior, it feeds this excessive preoccupation . . . .

SP Tr. at 833-34. For Barnette, however, the ultimate factor in "his history of violence, it virtually all seems to surround a domestic situation with a woman." SP Tr. at 834. Dr. Cunningham’s ultimate conclusion was that Barnette’s mental illness, combined with other aspects of his personality made the risk that he would commit serious violence in prison Avery low." SP Tr. at 838.

To rebut this testimony from Drs. Sultan, Halleck and Cunningham, the Government offered only a single witness, Dr. Scott Duncan. Dr. Duncan is the Forensic Coordinator for the Bureau of Prisons in Atlanta, and holds a Ph.D. in psychology from Florida International University. Dr. Duncan also utilized the DSM 4 for diagnostic purposes and agreed with Dr. Halleck’s diagnosis of a Borderline Personality Disorder. However, Dr. Duncan introduced a new diagnosis into evidence, one based upon the PCL-R and "from observed behavior": Psychopath. SP Tr. at 967-68. The diagnosis of "psychopath" is not contained in the DSM 4, is not a formal diagnosis, and is a "theoretical construct." SP Tr. at 973, 1012. The "instrument" used by Dr. Duncan, the PCL-R, to arrive at this diagnosis is a controversial and subjectively based scoring assessment of 20 different diagnostic categories and has never been standardized for this situation. A score of 30 allows the evaluator to classify the subject as a "psychopath"; Dr. Duncan testified that Barnette scored 34 after he added two additional points based on his second interview. SP Tr. 976-77. As explained to the Jury by Dr. Duncan, a diagnosis of "psychopathy" was devastating for Barnette:

In general, a psychopath is an individual who lacks the ability to feel at the same level and have the intensity of what feelings are as compared to non-psychopathic individuals. Typically they are very callous, manipulative, calculating, individuals that will often exploit other people. There is research to suggest that biologically, they do not respond to what non-psychopaths view as fear and anxiety which are two emotions that make up what we refer to as remorse or guilt. The psychopath is an individual that has little if any ability to feel remorse or guilt for the behavior they engage in.

Well . . . criminal psychopaths are twice as likely to engage in future criminal behavior when compared to non-criminal psychopaths. Criminal psychopaths are three times as likely to engage in violent future criminal behavior when compared to non-psychopathic criminals. Although in any prison population only about 20 percent of that population typically are psychopaths, they are responsible for over 50 percent of violent crimes that are committed. . . .

SP Tr. at 973-74. Dr. Duncan continued, noting that a "psychopath, probably one of their best assets and probably one of the greatest fears of non-psychopaths is their ability to look normal. We would all like to be able to think that we can pick out the psychopaths . . . but none of us are immune to that. . . . it is kind of like seeing a bow[ ]l of fruit, and you say to yourself, gosh that bowl of fruit looks wonderful . . . . But when you get close to the bow[ ]l of fruit and pick it up you realize it’s fake fruit." SP Tr. at 977.

In order to meet this new diagnosis of "psychopath," defense counsel sought to recall Dr. Mark Cunningham in "surrebuttal."¹⁸ Their efforts to do so came as no surprise to the Government or District Court as, before Dr. Duncan testified, the District Court ruled that Dr. Cunningham had to remain sequestered since he was to be called in surrebuttal.¹⁹ SP Tr. at 882 ("We want to make sure it’s okay he stays in the courtroom in case we want to offer surrebuttal to the government’s rebuttal. We have a right to offer surrebuttal."). When offered, however, the Government objected, arguing that since the defendant had received Dr. Duncan and Grant’s report 5 days earlier and therefore knew that he had arrived at a diagnosis of "psychopath," he was required to have "rebutted" this diagnosis in his defense-in-chief -- before the Government first raised it through Dr. Duncan’s testimony. SP Tr. at 1027 The District Court sustained the Government’s objection; it did so not based on the Government’s argument that Dr. Cunningham’s testimony should have been presented as anticipatory rebuttal, but on the grounds that "Mr. Williams [defense counsel] thoroughly cross-examined him, this witness . . . . All of that was put into evidence before the jury by Mr. Williams, and I just don’t see that anything additional needs to be said about it at this time." SP Tr. at 1028-30. The basis for the District Court’s ruling that the evidence was cumulative rested solely on defense counsel’s use of an article which discussed whether it was appropriate to use the PCL-R in African-American males and came before any proffer of the testimony was made. SP Tr. at 1029. The District Court, despite its ruling that the information had already been put before the Jury, did allow the article used on cross-examination to be put into evidence. SP Tr. at 1031; Def. Ex. 64.²⁰

Defense counsel then sought to call Dr. Cunningham outside the Jury’s presence for the purpose of preserving his testimony; the District Court refused, stating that "I’m not going to have going through the whole testimony on this witness because it’s not going to add anything to this at all" and directing defense counsel to "proffer the evidence, the testimony that he is going to give." SP Tr. at 1041, 1042. That proffer indicated that Dr. Cunningham would have testified as to the scientific short-comings of the PCL-R and its lack of standardization for African-American males in general and for Barnette in particular, that Barnette "was not a psychopath and did not meet the standards of the underlying opinion of psychopath," that the additional research on this issue cited by Dr. Duncan did not exist, and that Dr. Cunningham had reviewed Dr. Duncan’s scoring on the PCL-R, found it incorrect, and would testify that Barnette’s true score was 20 -- well below the cut-off for psychopathy -- and not 34 as testified to by Dr. Duncan. SP Tr. at 1042-43.

Having been denied the opportunity to actually proffer Dr. Cunningham’s testimony at the time of his exclusion, counsel later submitted an affidavit from Dr. Cunningham in connection with the defendant’s motion for a new trial, one which set forth his testimony in more detail than the "three minutes" provided by the District Court to defense counsel. That affidavit reveals that Dr. Cunningham, if permitted by the court, would have testified that "Barnette is not a psychopath," that Dr. Duncan’s scoring on the PCL-R was both deeply flawed and wrong and that properly scored, Barnette had a score of 20, rather than the revised score of 34 testified to by Dr. Duncan. Motion for New Sentencing Hearing, Cunningham Aff. at & 6. This revised score placed Barnette in the lower two-thirds of male prison inmates. Cunningham Aff. at & 6. Dr. Cunningham’s affidavit also reveals that there is no valid and peer-reviewed substantiation of the use of the PCL-R for African-Americans, Cunningham Aff. at & 7, and that there is no research linking prison misconduct and psychopathy. Far from being "heads above" other techniques used to predict future violence, Dr. Cunningham would have testified that a diagnosis based on the PCL-R is of unproven value. In sharp contrast to Dr. Duncan’s portrayal of a callous and manipulative individual without remorse, Dr. Cunningham’s testimony would have revealed that Barnette has never denied his crimes and "did not make excuses for the murders or claim amnesia, black-outs, multiple personality, or temporary insanity," and never attempted to manipulate or "con" either Dr. Cunningham or any prison personnel. Cunningham Aff. at & 6. Dr. Cunningham’s affidavit also reveals the deep flaws in Dr. Duncan’s utilization of the PCL-R and that there is no validated peer-reviewed scientific substantiation of the use of this "instrument" for African-Americans. Cunningham Aff. at & 7.

2. Dr. Cunningham’s Testimony Was Proper Surrebuttal

In United States v. Burgess, which involved an insanity defense, this Court defined surrebuttal evidence: "surrebuttal . . . is limited to refutation of the testimony of an opposing witness . . . . [and] concerns refutation of the specific evidence of a witness. It is not a basis for a general attack on the overall conclusory inference to be drawn from his evidence." 691 F.2d at 1149 n. 7. In this Circuit, "[s]urrebuttal evidence, though rarely offered, is admissible to respond to any new matter brought up on rebuttal." United States v. King, 879 F.2d at 137, 138 (4^th Cir. 1989) citing United States v. Burgess. While, in non-capital contexts, the decision to limit surrebuttal evidence is within the discretion of the court, it is nonetheless an abuse of that discretion to refuse to permit surrebuttal evidence "where (1) the government’s rebuttal testimony raises a new issue, which broadens the scope of the government’s case, and (2) the defense’s proffered surrebuttal testimony is not tangential, but capable of discrediting the essence of the government’s rebuttal testimony." United States v. Moody, 903 F.2d 321, 331 (5^th Cir. 1990). See United States v. Wilson, 134 F.3d 855, 866-67 (7^th Cir. 1998)(abuse of discretion not to permit surrebuttal testimony to rebut identification made by Government in rebuttal).

It is plain that Dr. Duncan’s testimony in rebuttal raised a new matter by introducing both a new testing instrument and a new diagnosis not previously placed before the Jury. The import of these items was not only to rebut the mitigating weight to be given to the defendant’s mental condition, his ability to adapt to a prison environment, and his remorse, but also to provide new evidence in support of the Government’s nonstatutory aggravating factor of "future dangerousness." It is also plain that Dr. Cunningham’s testimony was proper surrebuttal for it was directed to rebutting the essence of the Government’s new diagnosis and the testing instrument and methodology used to arrive at that diagnosis. Dr. Cunningham’s profferred testimony was a focused refutation of this specific new evidence, rather than a general attack on his conclusory inferences.

Moreover, it is clear that Dr. Cunningham’s proffered testimony could not have been part of the defense case-in-chief. Requiring such anticipatory rebuttal would be to redefine the defense-in-chief in mitigation as obliging the defense experts to refute diagnoses that they do not believe exist and based on instruments that these experts do not believe to be reliable, solely because the Government might later introduce evidence on this point. It would in essence require the defendant to prove the negative of what mental disorders or "theoretical constructs" do not apply to him in addition to proving those that do.

It must be recognized that, at the time of the defense case-in-chief, the defense only had the report of Dr. Duncan and Dr. Grant, a report that counsel received during the Government’s case in aggravation; the defense had no way of knowing with any certainty that Dr. Duncan or Dr. Grant would actually testify. Counsel for the Government was undoubtedly evaluating whether to call Dr. Duncan and Dr. Grant based upon its own cross-examination of the defense experts; they could certainly have concluded there was no need to call either expert if they believed their cross-examination was sufficient. In fact, of the two experts employed by the Government to conduct Barnette’s evaluation, only Dr. Duncan was called to testify. As noted by the Seventh Circuit in a similar context:

We agree that anticipatory impeachment and contradiction would here have been unwise. Among other aspects of the matter, anticipatory contradiction testimony would of necessity have had to be directed to examination reports. The testimony of the preparer of a report, even though based upon the report, might well have been broader or narrower and almost certainly would have had explicatory aspects. If the jury is to hear refuting testimony, it ordinarily appears to be preferable that such testimony be directed toward the actual testimony heard by the jury. That which will be heard must necessarily remain somewhat in the speculative area until it is articulated from the witness stand.

United States v. Portis, 542 F.2d 414, 417 (7^th Cir. 1976)(finding reversible error in the District Court’s failure to permit expert surrebuttal testimony in insanity defense to meet opinions of Government’s expert).

The Government’s position that rebuttal of its experts must occur before they testify is belied by the fact that, when defense counsel attempted to attack Dr. Duncan’s report in its case-in-chief, the Government objected to this evidence. SP Tr. at 714-15. The District Court sustained these objections to questions concerning Dr. Halleck’s opinions about Dr. Duncan’s inherent bias as a federal prison employee in performing this evaluation and his conflicts of interest.

The District Court’s stated basis for excluding Dr. Cunningham’s testimony at the time it was offered -- that it would be merely cumulative of the cross-examination -- is necessarily wrong. While impeachment may occur in cross-examination, the questioning of an expert witness by an attorney is a wholly inadequate substitute for the testimony of an expert witness to actually verify what the attorney implies:

Portis complains that . . . [t]he psychiatrist as a medical doctor, could have handled in a better fashion the testimony of the Government’s psychologist, and the psychiatrist was the only witness for the defense who could have dealt with the testimony of the Government’s psychiatrist.

The trial court, however, should have given Portis more leeway and permitted him to call Dr. Finkel in surrebuttal. . . [T]he jury’s comprehension of the validity of the claim of insanity would be much more likely to be aided by knowing the why’s and wherefore’s of the disagreement rather than by merely being aware that there was a bare bones difference of opinion.

Portis, 542 F.2d at 417-18. Indeed, the District Court instructed the Jury as much in its Charge at the Guilt Phase: "Let me emphasize that a lawyer’s question is not evidence. At times, a lawyer may have incorporated into a question a statement that assumes certain facts to be true and asked the witness if the statement was true. If the witness does not answer or denies the truth of the statement, and if there is not evidence in the record proving the assumed fact is true, then you may not consider the fact simply because it is contained in the lawyer’s question." GP Tr. at 658.

Moreover, it is plain from both counsel’s proffer and Dr. Cunningham’s affidavit that the evidence sought to be introduced went beyond the statements in the article about which Dr. Duncan was cross-examined. No portion of the article cited by the District Court concluded that Barnette was not a psychopath; no portion of the article challenged the validity of the scores assessed by Dr. Duncan in this case. The rebuttal of Dr. Duncan’s opinion and scoring through an opposite conclusion about Barnette and his "score" could not be accomplished through an article, but could occur only through a live expert witness in surrebuttal to these facts. It must be remembered that of the more than a dozen witnesses called by the Government to establish Barnette’s "future dangerousness," only one -- Dr. Duncan -- actually testified that Barnette would be dangerous in a prison setting in the future; every other witness testified solely about acts done by Barnette in the past when he was not incarcerated. Thus, it was only Dr. Duncan’s testimony that directly supported the Government’s contention that Barnette would be dangerous in prison in the future; his importance to the Government’s case on this issue cannot be underestimated. Similarly, by excluding Dr. Cunningham’s testimony, the District Court deprived Barnette of sharply-focused, unique and non-cumulative evidence refuting the Government’s contention.

The mandates of the FDPA, the Eighth Amendment, and ultimately of due process, are central to any analysis of this issue in this capital case. The language of the FDPA could not be clearer: "The government and the defendant shall be permitted to rebut any information received at the hearing. . . ." 18 U.S.C. § 3593(c) (emphasis supplied). Moreover, under these circumstances rebuttal of the Government’s evidence of psychopathy supported several of the mitigating factors offered on Barnette’s behalf, including his ability to adapt to prison, his remorse, his ability to cope in the structure, his capacity to appreciate the wrongfulness of his conduct at the time of the crimes, and the fact that he would not be a future danger if sentenced to life without the possibility of parole. The District Court’s refusal to permit this evidence effectively foreclosed the Jury from considering this evidence on these mitigating factors. This violated the Eighth Amendment’s requirement that "the sentencer may not be precluded from considering, and may not refuse to consider, any constitutionally relevant mitigating evidence. . . Our consistent concern has been that restrictions on the jury’s sentencing determination not preclude the jury from being able to give effect to mitigating evidence." Buchanan v. Angelone, 118 S. Ct. at 761. By preventing the Jury from considering the evidence that Barnette was not a psychopath, and the reasons underlying that conclusion, the District Court necessarily both barred mitigating evidence from the Jury’s consideration and prevented the Jury from giving full effect to the mitigating evidence that it was permitted to hear. Finally, the complete denial of an opportunity to offer rebuttal evidence to a criminal defendant necessarily violates due process, for "[d]ue process requires a fair opportunity to defend against the charges, including calling and cross-examining witnesses; keeping information, including witnesses, from the jury may violate due process." United States v. Webster, 162 F.3d 308, 350 (5^th Cir. 1998)(finding that restricting the defendant to one expert in surrebuttal was not due process violation). See United States v. One Single Family Residence, 778 F. Supp. 1215, 1219 (S.D.Fla. 1991)(due process violated with complete denial of surrebuttal).

It is clear then, as a matter of decisional law, as a matter of statutory law, and as a matter of constitutional law, that the District Court abused its discretion in barring the defendant’s surrebuttal. In so doing, the District Court violated not only the express mandate of the FDPA, but also the defendant’s right to have his mitigating evidence considered by the Jury under the Fifth and Eighth Amendments. Its error in so doing left devastating evidence tendered by the Government unchallenged by any expert and necessarily had a profound impact on the Jury’s verdict.

Counsel for the defendant recognizes that this was a lengthy trial involving complex issues. However, it is precisely because the issues were complex, and because the ultimate penalty was at stake, that admission of this evidence was required. As noted by this Court in an analogous context:

Undoubtedly, the trial court in its rulings was seeking to expedite the trial. This procedure is normally to be commended. But where the plea is insanity, the goal of expediting the trial must not be allowed to interfere with the defendant’s right to develop fully and completely the many complex and often tenuous circumstances that may shed light on his plea.

United States v. Smith, 507 F.2d 710, 712 (4^th Cir. 1974).

B.  The District Court Committed Prejudicial Error in Permitting Dr. Duncan to Offer His Opinion That Barnette Was a Psychopath and Would be Dangerous in the Future Based Upon the PCL-R and Barnette’s Race, Wealth, Age and Sex.

The District Court Committed Prejudicial Error in Permitting Dr. Duncan to Offer His Opinion That Barnette Was a Psychopath and Would be Dangerous in the Future Based Upon the PCL-R and Barnette’s Race, Wealth, Age and Sex.

Prior to Dr. Duncan’s testimony, defense counsel moved in limine to "restrict the testimony of Drs. Duncan and Grant and to disallow any testimony by them with regard to any testing utilizing the PCL-R . . . or any references or opinions to a diagnosis of psychopathy." Motion in Limine, 2/5/98. The basis for the Motion rested upon the facts that the PCL-R had "not been sufficiently standardized to the black population [or] . . . an institutionalized population," and that permitting testimony that Barnette was a psychopath without an adequate scientific basis was misleading and highly prejudicial. Defense counsel sought to present testimony on the PCL-R’s lack of scientific validity in support of the Motion; the request was denied by the District Court. As counsel sought to present the bases in the literature for the Motion, the District Court cut him short, admonishing counsel that "We have got a short time, Mr. Williams, we have a short time. I don’t have juries come up here 50 or 60 miles and sit here and wait while we argue." SP Tr. at 881.²¹ The Government’s response to the Motion was "that’s all proper stuff for cross-examination as this is testing that’s reasonably relied upon in the field. Dr. Duncan relied upon it. He was able to testify to it. It’s relevant . . . " SP Tr. at 881. The District Court denied the Motion based upon this response, SP Tr. at 881; it made no findings as to the scientific basis or reliability of either this test, Dr. Duncan’s diagnosis, or the basis for his opinion that Barnette was at a higher risk for future violence.

When Dr. Duncan testified, he informed the Jury that his opinion was based on "three different ways to establish . . . whether Mr. Barnette might be at increased risk for being violent in the future. One was the use of the Psychopathic Checklist Revised." Sp. Tr. at 983. Dr. Duncan testified extensively using the PCL-R. SP Tr. at 965-67, 972-78. He further testified that his opinion that Barnette was a "psychopath" based upon his score on the PCL-R, his "observed behavior," and his review of the evidence in this case. SP Tr. at 967-68, 972, 982, 988-90, 1022-24. In the course of this testimony, Dr. Duncan testified that the PCL-R was developed by Dr. Robert Hare in Canada, that Dr. Hare "has spent over 25 years researching psychopaths," that he had spoken with Dr. Hare who "told me of a great deal of research that he has done in the last two years that suggests that although African American individuals on average will score two points higher . . . when you take a look at the PCL-R’s ability to predict whether they will be violent or not and use a cut off score of 30, it predicts just as well," and that Dr. Henry Richards of the Patuxent Institute had performed additional research that would be presented on "NBC’s Dateline" indicating "that although you do see a two point increase with African American males you can still use a cut off score of 30 to be able to predict future dangerousness." SP Tr. at 989-90. Dr. Duncan further testified that the PCL-R was "heads above any other instrument that we currently have in predicting future dangerousness." SP Tr. at 990. Closer analysis of the PCL-R in light of the actual science underlying it reveals that these claims are not true.

An additional basis for Dr. Duncan’s opinion was "look[ing] at the particular items that had been borne out in the research." These items were found in research conducted by Dr. John Monahan who, according to Dr. Duncan, had identified A11 items . . . as being items to look at in predicting whether someone will be a future danger." Four of those eleven items were Barnette’s poverty, young age, the fact that he was male, and the fact that he was an African-American. SP Tr. at 984. Thus, part of the basis for Dr. Duncan’s opinion that Barnette would be dangerous in the future -- and thus deserves execution -- was the fact that Barnette was a poor young black male: "The research suggests that males are much more likely to be violent than females. The next was race. The research suggests that African Americans are at a higher risk than other races to engage in violent behavior . . . an individual that comes from higher -- from a higher SES or social economic status and has more stable employment history is less likely to engage in violent behavior in the future." SP Tr. at 984-85.

1. Daubert Analysis

When presented "with a proffer of expert scientific testimony," wrote the Supreme Court in Daubert v. Merrill Dow Pharmaceuticals, "the trial judge must determine at the outset, pursuant to [Fed. R. Evid.]Rule 104(a) 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." 509 U.S. 579, 592, 113 S. Ct. 2789, 2796 (1993). This determination "entails a preliminary assessment of whether the reasoning or methodology underlying the testimony is scientifically valid and whether that reasoning or methodology properly can be applied to the facts in issue." Contrary to the Government’s assertion -- accepted by the District Court -- that these matters are for cross-examination, it has been plain since Daubert that it is the trial court in the first instance that must act as the "gatekeeper" in determining whether the "science" sought to be offered is reliable and thus probative of the fact in issue. In performing this gatekeeping function, the District Court must determine whether the theory has, or can be, tested, whether it has been subject to peer review and publication, the known or potential rate of error, and whether the theory has been met with widespread acceptance. As this Court has interpreted Daubert, "a trial judge must ensure that all scientific testimony or evidence admitted is both relevant and reliable, and that its evidentiary reliability is based upon scientific validity." United States v. Powers, 59 F.3d 1460, 1470 (4^th Cir. 1995). See United States v. Dorsey, 45 F.3d 809, 814-15 (4^th Cir. 1995)(Trial court must analyze proposed expert testimony and testing by four criteria); United States v. Bynum, 3 F.3d 769, 773 (4^th Cir. 1993)("The Daubert Court did make clear, though, that district courts need not admit all >scientific’ evidence without any regard for its reliability. Trial judges may consider whether the particular opinion or technique has been subjected to peer review, what the known rate of error of the technique is, and recalling Frye, whether it enjoys >widespread acceptance’ in the community.").

The evidentiary issue of Daubert’s applicability to scientific or expert testimony at the sentencing phase of a capital trial under the FDPA has never been addressed. Certainly, the FDPA contemplates that the Rules of Evidence generally do not apply to such a proceeding. However, the FDPA does specifically provide that information may be excluded from the Jury’s consideration "if its probative value is outweighed by the danger of creating unfair prejudice, confusing the issues, or misleading the jury." 18 U.S.C. § 3593(c). This language finds its source in the standard of relevancy under Fed. R. Evid. Rule 403. However, unlike Rule 403, which requires the exclusion of relevant evidence only when it is "substantially outweighed" by prejudice, '3593(c) omits any requirement that the evidence’s prejudice "substantially" outweigh its probative value. See United States v. Frank, 8 F. Supp.2d 253, 268 (S.D.N.Y. 1998). This balancing is also the standard underlying Daubert. That is, while Daubert involved interpretations of the trial court’s role as "gatekeeper" under Rule 104(a) and the standards for expert testimony under Rule 702, it was ultimately relevance under Rule 403 which drove its conclusion:

Finally, Rule 403 permits the exclusion of relevant evidence 'if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury . . .’ Judge Weinstein has explained: 'Expert evidence can be both powerful and quite misleading because of the difficulty in evaluating it. Because of this risk, the judge in weighing possible prejudice against probative force under Rule 403 of the present rules exercises more control over experts than over lay witnesses.’

Daubert, 509 U.S. at 595, 113 S. Ct. at 2798. This Court echoed Daubert’s conclusion in Dorsey: "the Supreme Court warned that throughout an admissibility determination a judge must be mindful of other evidentiary rules, such as FRE 403, which permits the exclusion of relevant evidence. . . ." 59 F.3d at 1471. As such, Daubert is ultimately nothing more than an inquiry into the probative value of testimony and evidence offered as "science." Indeed, at least one court has adopted the concept of Daubert’s "gatekeeper" as the standard by which evidence in the penalty phase must be judged:

Although Section 3593 expressly suspends the formal rules of evidence at the sentencing hearing, it does not suspend all sense of order. Explicit in the Section’s recitation of what may be admitted -- as is the case throughout the FDPA -- is a role for the district court as a gate-keeper, as well as the requirement that all information presented be relevant.

Frank, 8 F. Supp.2d at 268.

Daubert’s rules must necessarily be part of the inquiry in a sentencing phase in measuring the "probative value" of evidence. Indeed, Daubert is nothing more than a declaration that if scientific evidence and testimony based on such evidence cannot be tested, is not subject to peer review, contains a high rate of error, and does not enjoy widespread acceptance, then it is not probative and the danger of its admission as "expert opinion" mandates its exclusion. Moreover, if the underpinnings of the opinion are based upon factors that are not constitutionally or legally relevant -- if it cannot assist the Jury -- then it is plainly not "probative" of anything. This application of the principles of Daubert is particularly appropriate in the context of a capital sentencing proceeding where the Eighth Amendment and due process have always demanded a "heightened reliability." Woodson v. North Carolina, 428 U.S. 280, 305, 96 S. Ct. 2978, 2991 (1976)(plurality opinion); Caldwell v. Mississippi, 472 U.S. 320, 323, 105 S. Ct. 2633, 2636 (1985) (referring to "the Eighth Amendment’s heightened 'need for reliability in the determination that death is the appropriate punishment in a specific case.’")(quoting Woodson).

2. The Scientific Reliability of the PCL-R and the Diagnosis of "Psychopath."

As conceded by Dr. Duncan, the diagnosis of "psychopathy" is only theoretical. Put differently, there is no generally or widely-accepted diagnostic category of psychopathy in the DSM 4, the repository for generally recognized mental disorders. Consequently, the ultimate diagnosis of "psychopath" is open to scientific dispute as it is not yet statistically demonstrable in a reliable fashion. The instrument used by Dr. Duncan to arrive at this "diagnosis" is even more questionable. As set forth in Defendant’s Exhibit 64, AA Review and Meta-Analysis of the Psychopathy Checklist and Psychopathy Checklist-Revised: Predictive Validity of Dangerousness," the PCL-R has at least four major problems as a Avalid and reliable clinical method for determinations of dangerousness": "(a) the PCL/PCL-R is an atheoretical measure of psychopathy, (b) difficulties may arise when clinicians attempt to identify an appropriate cutoff score to separate psychopaths from non-psychopaths, (c) factor scores may predict different levels of dangerousness, and (d) the generalizability of the PCL has not been extensively studied." Def. Ex. 64 at 205. As explained by Def. Ex. 64, the PCL-R has no generalized and acceptable theory underlying it -- since psychopathy is a "theoretical construct," the PCL-R is testing, and employing different factors to test for, a mental disease entity which is unsupported by any widely accepted scientific data. Def. Ex. 64 at 205-06. As such, the factors used, and the significance of any single factor and its weight with respect to other factors, cannot be accurately judged. The score at which a "diagnosis" of psychopathy is made has been the subject of widespread debate. To achieve accuracy within two standard deviations, a cutoff of at least 37 should be used; the cutoff employed by Dr. Duncan -- 30 -- is accurate to only one standard deviation, resulting in a specificity level that has been criticized as generating too many "false positives." Def. Ex. 64 at 206-07. Indeed, the analysis undertaken in the article revealed that "41% of the nonviolent offenders were incorrectly classified as violent" when the literature and studies were reviewed. Def. Ex. 64 at 211.

Moreover, the way in which a particular "score" is achieved necessarily has an impact on the prediction and level of any future dangerousness. For example, there are more than 15,000 variations of psychopathy based upon the PCL-R; Dr. Duncan’s classification assumed that all are the same. However, as noted in Defendant’s Exhibit 64, "different combinations of characteristics [ ] do not necessarily represent the same risk. Psychologists need to be aware that F1 [items in the PCL-R] and F2 differ in their relationship to future criminal behavior." Def. Ex. 64 at 207. In fact, there is no "consistent description on how psychologists could make differential use of factor scores in their assessment of psychopathy." Finally, the "efforts to evaluate the reliability and validity of the PCL/PCL-R are almost exclusively limited to forensic populations within the Canadian criminal justice system . . . primarily composed of white males." Def. Ex. 64 at 207. The article concluded that Awe do not believe the existing validity data justify the use of the PCL-R with minority groups." Def. Ex. 64 at 208.

The District Court in this case failed to conduct any inquiry whatsoever into the scientific underpinnings of Dr. Duncan’s expert opinions, and specifically the PCL-R and its validity as a scientific testing instrument. Had it done so by permitting the hearing and receiving the testimony proffered by defense counsel and contained within Def. Ex. 64, it would have discovered that the PCL-R as a testing instrument fails all of Daubert’s criteria. As noted in Def. Ex. 64, the testing necessary to replicate the results of the PCL-R and its ability to predict future violence are hampered by the facts that: (1) it seeks to test for a mental condition which is merely theoretical; (2) it seeks to test for a mental condition which has no underlying theoretical basis; (3) it seeks to test for a mental condition using a screening interview that yields nearly 15,000 separate combinations of scores; and, that it seeks to test for a mental condition using a screening interview that incorrectly identifies nonviolent offenders as violent offenders potentially 40% of the time using Dr. Duncan’s cutoff score of 30. While the validity of the test appears to be greater in white males, the only peer-reviewed literature offered indicated that it was not to be applied in the evaluation of minority populations, as Dr. Duncan did in this case. While Dr. Duncan did testify that he "spoke" with others who were conducting research -- some of whom were going to appear on "Dateline" -- Dr. Duncan did not produce a single peer-reviewed article that supported the use of this instrument in Barnette’s case; indeed, there was no evidence presented that this instrument had ever been used to evaluate a person on trial for his life in conjunction with a penalty proceeding.

Moreover, as demonstrated by Def. Ex. 64, there is no widespread acceptance in the scientific community for the use of this instrument with minority populations. Indeed, the fact that, of the four experts who testified in this field, only Dr. Duncan employed the PCL-R to reach a diagnosis demonstrates that it is not widely accepted; certainly it is not accepted in the same way the DSM 4 used by every expert in this case is accepted as a diagnostic guide. Significantly, the only peer-reviewed literature introduced concerning the PCL-R indicated that its use was discouraged and suspect with minority groups -- the exact opposite of the "peer review" testing and "widespread acceptance" required under Daubert. In short, the use of the PCL-R is a subject of controversy in the scientific community and no evidence in the form of valid scientific and peer-reviewed opinion that it could be reasonably applied to this defendant in this case was ever presented.

The Government, as the proponent of this evidence, was under an obligation to present information to the District Court from which it could find that Dr. Duncan’s testimony that Barnette was a psychopath and would be dangerous in the future was both scientifically reliable and relevant in assisting the trier of fact in concluding that Barnette should die. It presented nothing in response to the defendant’s motion other than to argue that the widespread criticism of the PCL-R as a valid instrument was a matter for Across-examination." While the District Court’s decision in admitting this evidence and Dr. Duncan’s opinion is, admittedly, discretionary, see Powers, 59 F.3d at 1470, this Court has not hesitated to exclude purportedly scientific evidence under similar circumstances. Thus, in Dorsey, this Court upheld the exclusion of testimony by forensic anthropologists concerning the defendant’s photographic identification because the method used by the anthropologists had never been tested, their conclusions and methods not subjected to "peer review and publication," the known rate of error "[was] potentially very high," and "there [was] no indication that the type of evidence offered here has been widely accepted in the relevant scientific community." 45 F.3d at 815. Similarly, in Powers, a case involving a claim of child sexual abuse and incest, this Court upheld the exclusion of the testimony by two experts, one based upon a penile plethysmograph and the other based upon a psychological profile of a fixated pedophile. The plethysmograph was excluded based upon literature showing that it was not regarded "as a valid diagnostic tool because, although useful for treatment of sex offenders, it has no accepted standards in the scientific community," and "false negatives render the test unreliable." 59 F.3d at 1471. The profile was excluded because no scientific link was presented between "a non-proclivity for pedophilia with a non-proclivity for incest abuse." 59 F.3d at 1460. Both cases demonstrate aptly the rigorous examination that the District Court is required to make of scientific evidence.

There can be little question that the error in permitting Dr. Duncan to introduce the results of his scoring of the PCL-R, and his diagnosis of psychopathy based upon the PCL-R, substantially influenced the Jury’s verdict in this case. The devastating impact of a diagnosis of "psychopath" is plain, as are the implications of that label for future dangerousness. It was the PCL-R that permitted Dr. Duncan to testify that Barnette, as a criminal psychopath, was twice as likely to commit crime and three times as likely to commit violence; it was the PCL-R that permitted Dr. Duncan to tell the Jury that while, in his estimation, psychopaths made up only 20% of the prison population, they were responsible for 50% of the violence. SP Tr. at 974. And it was the PCL-R that permitted Dr. Duncan to tell the Jury that, as a psychopath, Barnette:

lacks the ability to feel at the same level and have the same intensity of feelings . . . . they are very callous, manipulative, calculating individuals that [sic] will often exploit other people. . . . they do not respond to what non-psychopaths view as fear and anxiety which are two emotions that make up what we refer to as remorse or guilt. The psychopath is an individual that [sic] has little if any ability to feel remorse or guilt for behavior they engage in.

SP Tr. at 973.

3. The Use of Race, Sex, Age and Poverty to Support the Opinion of Dangerousness.

Had the District Court probed further into the bases for Dr. Duncan’s opinion, it would have learned that this opinion was supported in part based upon the fact that Barnette was poor, young, male and black. In short, part of Dr. Duncan’s opinion was not based on the individual character of this defendant -- an individuality that lies at the core of the Eighth Amendment -- but on immutable racial, sex and economic circumstances over which Barnette had little if any control. The Eighth Amendment plainly prohibits the use of these facts; indeed, "in the realm of capital sentencing . . . race is a 'totally irrelevant factor.’" United States v. Webster, 162 F.3d 308, 355 (5^th Cir. 1998) quoting Zant v. Stephens, 462 U.S. 875, 885, 103 S.Ct. 2733 (1983). Rather, as observed by the Fifth Circuit, "the use of race in sentencing determinations is particularly invidious . . . And, in capital sentencing, the use of race becomes more offensive still:

"Considering the race of a defendant or victim in deciding if the death penalty should be imposed is completely at odds with the concern that an individual be evaluated as a unique human being. Decisions influenced by race rest in part on a categorical assessment of the worth of human beings according to color, insensitive to whatever qualities the individual in question may possess."

United States v. Webster, 162 F.3d. at 356 (rejecting constitutional challenge to the FDPA because it precludes the consideration of race) quoting McCleskey v. Kemp, 481 U.S. at 336)(Brennan, J. dissenting). The court continued: "In sum, a long line of Supreme Court precedent admonishes that the guillotine must be as color blind as the Constitution. [ ] Today’s decision recognizes this precedent in interpreting the FDPA in the only way constitutionally permissible: as prohibiting the consideration of race in sentencing." Webster, 162 F.3d at 356.

Thus, the very factors - - poverty, race, and sex - - that Dr. Duncan relied upon in part for his opinion are not constitutionally relevant in the sentencing phase of a capital trial. Since these factors cannot be considered by the Jury under either the Eighth Amendment or due process, then they cannot possibly assist the Jury in making its determination that Barnette should or should not die. As such, they cannot be relevant under Daubert. See Powers, 59 F.3d at 1471-73 (excluding evidence of "profile" of pedophile on basis that it would not assist Jury in determining whether the defendant was more likely than not to have committed incest).

There can be little question that allowing Dr. Duncan to base his prediction that Barnette was at a higher risk for violence in the future because he was poor, young, male and black was prejudicial. In a case where one of the victims was white, it was nothing more than an invitation to the all-white jury, under the guise of an expert opinion on "future dangerousness" to indulge in the worst of stereotypical assumptions. Perhaps in recognition of this, the District Court, in denying the defendant’s Motion for a new sentencing hearing, sought to use the certificate of "no discrimination" mandated by 18 U.S.C. § 3593(f), as proof that these stereotypes were not indulged. The problem with this rationalization, however, is that Dr. Duncan’s "expert" opinion permitted the Jury to find that Barnette was dangerous because he was a young, black, poor male; this aggravating factor, not the racial and economic stereotypes on which it was based, weighed in favor of the death penalty. Thus, a juror could reasonably indulge in these stereotypes labeled as "future dangerousness" and truthfully answer the certificate. Of course, had Barnette been older, white, and rich, he would have been, according to Dr. Duncan, less dangerous, enabling jurors not to find this aggravating circumstance at all.

Dr. Duncan’s opinions, presented to this Ju