Edward Horsley v. State of Alabama

45 F.3d 1486 | 11th Cir. | 1995

Lead Opinion

EDMONDSON, Circuit Judge:

Edward Horsley, a prisoner of the state of Alabama, appeals the district court’s denial of his petition for a writ of habeas corpus pursuant to 28 U.S.C. § 2254. We affirm the judgment.

BACKGROUND

In 1977, Horsley was convicted and sentenced to death in Monroe County, Alabama for the capital offense of robbery in which the victim is intentionally killed, Ala.Code § 13 — 11—2(a)(2) (1975).

Horsley’s guilt is not disputed. Both Horsley and his co-defendant Brian Baldwin confessed. We briefly review the facts. On Saturday, March 12, 1977, Horsley (then nineteen years old) and his co-defendant (then eighteen years old) escaped from a North Carolina prison camp. Horsley had been convicted of four counts of robbery during the course of which a police officer was shot. That evening the murder victim, Naomi Rolon, sixteen years old, left her home in North Carolina to visit her father who was in the hospital. , Horsley and Baldwin forcibly seized Rolon and her car and drove to Charlotte, N.C., despite pleas and prayers from the victim. There, both men attempted to rape her and attempted to choke her to death. She was stripped, stabbed with a knife in different parts of her body, run over with the car at least once, and locked in the trunk while they drove to Alabama. On Monday afternoon, Horsley and Baldwin stole a pickup truck and drove both vehicles to a secluded wooded area. Baldwin took Naomi from the trunk and told Horsley to back over her with the car. Horsley tried twice, but the ear became stuck. Baldwin then cut Naomi’s throat with a hatchet. She died after this 40 horn’ ordeal.

Horsley was tried separately and was found guilty as charged by a jury which fixed his punishment at death by electrocution.¹ Following Horsley’s capital conviction, a sentencing hearing was conducted before the court. After the hearing, the trial court “having considered the evidence presented at the trial and at said sentence hearing,” entered a sentence order- finding the following aggravating circumstances: 1) the capital offense was committed by a person under sentence of imprisonment; 2) the defendant was earlier convicted of a felony involving robbery, in the course of which a police officer was shot; 3) the capital felony was committed while defendant was engaged in commission of or flight after committing a robbery; and 4) the capital felony was especially heinous, atrocious or cruel. The court found Horsley’s age to be a mitigating circumstance.² The trial court then found that the aggravating circumstances “far outweigh[ed]” the mitigating circumstances and sentenced Horsley to death.

In 1989, after challenging the conviction and sentence in state court,³ Horsley petitioned for writ of habeas corpus in the federal district court. The district court entered a memorandum order in December 1991, denying certain claims and granting an evidentia-ry hearing on four specified claims.⁴ After a four-day evidentiary hearing, the district court denied all claims.

On appeal, Horsley raises two claims that merit discussion: 1) the claim that his sentence violated the Eighth Amendment because the trial judge in this case expressly limited his consideration of mitigating circumstances to those enumerated in the Alabama death penalty statute; and 2) the claim that his counsel provided ineffective assistance at sentencing by presenting no expert testimony to demonstrate Horsley’s alleged vulnerability to domination by his co-defendant Baldwin.⁵

CONSIDERATION OF NONSTATUTORY MITIGATING EVIDENCE

Horsley argues that the state trial judge expressly limited his consideration of mitigating circumstances to those set out in the Alabama death penalty statute in force at the time of Horsley’s conviction. As a result, he contends the trial judge considered only Horsley’s youth in mitigation and excluded other mitigating circumstances that were before him. Thus, Horsley claims that he was sentenced to death in violation of Hitchcock v. Dugger, 481 U.S. 393, 107 S.Ct. 1821, 95 L.Ed.2d 347 (1987).⁶ Horsley bases this argument on the similarity between the trial court’s order in this case and the trial court’s order in Hitchcock.

The district court held no evidentiary hearing on this issue. Based upon the similarity between the sentencing order in this case and the order in Hitchcock, the district court said, in a preliminary comment, that the procedure “does not appear to satisfy Hitchcock.” The court, however, further concluded that “no evidentiary hearing is required on this issue as the merits can be considered without further evidence. Whether this claim is proeedurally barred is a matter that will be addressed in the Court’s final Order following the evidentiary hearing.” In its final order, the court held that, because Horsley failed to raise this claim on direct appeal or on coram nobis and because he had failed to show cause and prejudice, the claim was proeedurally barred.

On appeal, Horsley contends that the district court, in the initial order, made findings of fact and held that there was a Hitchcock error.⁷ We disagree. When read in the context of case law on procedural default, the district court’s comments seem to be preliminary and passing in nature. The Supreme Court has held that, “[ujnless a habeas petitioner shows cause and prejudice, a court may not reach the merits of ... proeedurally defaulted claims in which the petitioner failed to follow -applicable state procedural rules in raising the claim.” Sawyer v. Whitley, — U.S. -, -, 112 S.Ct. 2514, 2528, 120 L.Ed.2d 269 (1992). We also have held that a federal habeas court “will not ” consider an issue that is proeedurally barred unless the petitioner can show cause and prejudice. Amadeo v. Kemp, 816 F.2d 1502, 1505 (11th Cir.1987). Given this legal background, we conclude that the district court quite properly held in abeyance ruling on the merits of the claim until the procedural default issue was resolved.

First, we address the procedural bar. Horsley argues that the district court erred in finding the Hitchcock claim to be proee-durally barred. He does not dispute that he never directly presented the issue on appeal or in collateral proceedings; nor does he contend that his collateral attacks in the Alabama courts in any way raised this claim. Instead, he argues that this claim is not proeedurally defaulted because the Alabama Court of Criminal Appeals sua sponte raised and answered the question of whether the sentencing court’s consideration of mitigating evidence complied with requirements of Lockett. We agree.

When a state court decides a constitutional question, even though it does not have to, the considerations of comity and federalism which woidd ordinarily preclude federal review of procedurally defaulted issues no longer apply. Cooper v. Wainwright, 807 F.2d 881, 886 (11th Cir.1986). We have said that “a state court’s decision to raise and answer a constitutional question sua sponte will ... permit subsequent federal habeas review.” Id.

In Horsley’s direct appeal, the Alabama Court of Criminal Appeals stated:

“We have reviewed the aggravating and mitigating circumstances set out in the record and the trial court’s findings relative to those circumstances. The appellant was .given an opportunity to present any mitigating circumstances he desired. Alabama’s capital felony act ... fully comports with Lockett v. Ohio, 438 U.S. 586 [98 S.Ct. 2954, 57 L.Ed.2d 973] ... (1978), in that it does not preclude from consideration as a mitigating factor ‘any aspect of a defendant’s character and record and any of the circumstances of the offense that the defendant proffers as a basis for a sentence of less than death.’ ”

Horsley, 374 So.2d at 375. In this ease, we conclude that the Alabama courts, even though they did not have to, raised and answered the Lockett issue. As a result, Horsley’s claim is properly before us for a decision on the merits.

We review Lockett-Hitchcock claims by matching the records in the case under consideration with the Hitchcock record. Hargrave v. Dugger, 832 F.2d 1528, 1533 (11th Cir.1987) (in banc). Relevant factors may include: 1) statements made by the sentencing judge; 2) comments made by the prosecutor and defense counsel; 3) the mitigating evidence presented; and 4) the state law at the time of sentencing. E.g., Knight v. Dugger, 863 F.2d 705, 708-10 (11th Cir.1988).

The Supreme Court noted that the trial judge in Hitchcock expressly weighed, in imposing sentence, only those mitigating factors enumerated in the death penalty statute:

[T]he sentencing judge found that ‘there [were] insufficient mitigating circumstances as enumerated in Florida Statute ... to outweigh the aggravating circumstances.’ He described the process by which he reached his sentencing judgment as follows: ‘In determining whether the defendant should be sentenced to death or life imprisonment, this Court is mandated to apply the facts to certain enumerated “aggravating” and “mitigating” circumstances.’

Hitchcock, 481 U.S. at 398, 107 S.Ct. at 1824 (record citations omitted; emphasis in original). In Hitchcock, the Court also considered that the jury was instructed to consider only those mitigating circumstances that were listed in the statute.⁸ Based upon these facts, the Court held that the sentencing proceeding in Hitchcock had not complied with Lockett.

The Alabama death penalty statute in effect at Horsley’s sentencing permitted the defendant to present evidence on any matter that the court deemed relevant to the sentence. Ala.Code § 13-11-3. Horsley does not contend that the trial court denied him the opportunity to present mitigating evidence.

At the sentencing hearing, the court said to Horsley:

Edward, this is your day in court to tell me what you have in your own behalf in the way of mitigating the sentence that has been imposed on you and I want to give you an opportunity to tell me whatever you feel like might be helpful to you to get this sentence reduced from electrocution to life imprisonment. Do you want tell me anything else other than what you’ve been asked about?

I am giving you now the opportunity to tell me anything in your whole life which you feel like might be helpful to you to get this sentence reduced....

Thus, the sentencing court invited Horsley to present any mitigating evidence he desired. Implicit within this invitation is the notion that the court would consider the evidence offered. The record also contains this statement by the sentencing court:

The Court, having conducted a hearing pursuant to ... the 1940 Code of Alabama, as amended, to determine whether or not the Court would sentence Edward Horsley to death or life imprisonment without parole and the Court having considered the evidence presented at the trial and at said hearing, makes the following findings_

Horsley, 476 So.2d at 625 (emphasis added).

We also note that, during Horsley’s testimony at the sentencing hearing, defense counsel asked him on at least four occasions whether there was anything else Horsley wanted to tell the judge. He was urged, “[tjell the judge whatever you want to.” The statements of defense counsel, therefore, also indicate that they believed that the sentencing court would consider petitioner’s testimony.

Horsley’s best argument that the eourt did not consider the nonstatutory mitigating evidence is the similarity of one of the trial judge’s statements to one statement of the sentencing judge in Hitchcock. As noted above, the sentencing judge weighing the aggravating and mitigating circumstances in Hitchcock said that there were insufficient mitigating circumstances as enumerated in the Florida statute to outweigh the aggravating circumstances. In this ease, the court’s sentencing order said: “The Court now considers mitigating circumstances as described in Title 15, Section 342(9) of the 1940 Code of Alabama_” Id. (emphasis added). The sentencing court then found that only Hors-ley’s age was a mitigating factor under the statute. The court’s next statement, however, was: “having considered the aggravating circumstances and the mitigating circumstances [plural] and after weighing the aggravating and mitigating circumstances, it is the judgment of the Court that the aggravating circumstances far outweigh the mitigating circumstances.... ” Id.

We see substantive differences in the sentencing record in the two eases. In Hitchcock, the trial court said it weighed only the enumerated factors; and the Supreme Court found other plain statements in the record indicating that the trial court believed that it could consider no nonstatutory mitigating factors. Hitchcock, 481 U.S. at 397-99, 107 S.Ct. at 1824. In this case, statements exist by the judge and Horsley’s defense counsel indicating that the court would consider non-statutory mitigating evidence. In its balancing of the aggravating and mitigating evidence, the trial court here referred to mitigating circumstances (again plural), even though it had found only one statutory mitigating circumstance.

There is a reason, moreover, that the sentencing order only contains detailed findings on statutory mitigating circumstances but makes no explicit reference to nonstatutory mitigating circumstances. The pertinent state statute may require the sentencing judge to set forth explicitly his findings for only the statutory mitigating circumstance that he found insufficient to outweigh the aggravating circumstances. Ala.Code § 13-11-4. As the Supreme Court has observed, non-statutory evidence, “precisely because it does not fall into any predefined category is considerably more difficult to organize into a coherent explanation.” Parker v. Dugger, 498 U.S. 308, 318, 111 S.Ct. 731, 738, 112 L.Ed.2d 812 (1991). We believe that Hors-ley’s focus on one isolated statement of the sentencing judge places far too much stress on just one statement in a long discourse, a discourse which, taken as a whole, shows that the sentencing court probably considered nonstatutory mitigating evidence.⁹ But, for the purpose of deciding this appeal, we will assume that nonstatutory mitigating evidence was not considered by the sentencing judge.

Assuming a Lockett/Hitchcock error in this case, we nevertheless affirm the district court’s denial of relief under the doctrine of harmless error. Because Horsley, relying on Chapman v. California, 386 U.S. 18, 87 S.Ct. 824, 17 L.Ed.2d 705 (1967), and decisions of this court applying that standard, argued in his original briefs that any Hitchcock error was not harmless and because the Supreme Court has recently held that different harmless error standards are to be applied for direct and collateral review of state court convictions and sentences, we ordered supplemental briefing on this issue.¹⁰

The Supreme Court recently has held that a “less onerous harmless error” standard is appropriate on habeas review of constitutional error. Brecht v. Abrahamson, — U.S. -,-, 113 S.Ct. 1710, 1721-22, 123 L.Ed.2d 353 (1993). On collateral review federal courts shall apply the less onerous standard announced in Kotteakos v. United States, 328 U.S. 750, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946), which is, “whether the error ‘had substantial and injurious effect or influence in determining the jury’s [or the court’s] verdict.’ ”¹¹ Brecht, — U.S. at -, 113 S.Ct. at 1722 (quoting Kotteakos, 328 U.S. at 776, 66 S.Ct. at 1253). The Court expressly said that, to meet this test, a habeas petitioner alleging constitutional trial error is entitled to no habeas relief in federal court unless he can establish actual prejudice. Id.¹² We have held that Lockett and Hitchcock errors are “trial type” errors governed by Brecht. Bolender v. Singletary, 16 F.3d 1547 (11th Cir.1994).

The mitigating circumstances before the sentencing court which, according to Horsley, were not considered include: the absence of earlier criminal activity (other than the conviction for armed robbery in which a police officer was shot which comprised the second aggravating factor); a childhood head injury which left Horsley suffering from headaches and spasms and made him easier to dominate; and the fact that Horsley grew up without a father. But we are struck by the full force of the savage facts of this crime. We remember also the imbalance between aggravating and mitigating (including non-statutory factors) circumstances marked by Alabama’s appellate courts (which we see as persuasive, as a factual matter, in our weighing of injurious effect on the senteneer— although in no way controlling, a la Clemons ). And, after making our own examination of the record anew, we conclude independently that any error was harmless under the Brecht standard.

In the past, we have not flinched from concluding that Hitchcock errors were harmless even when applying the more strict Chapman standard for harmlessness, e.g., Demps v. Dugger, 874 F.2d 1385, 1890 (11th Cir.1989); and we think we could conclude so here. But, considering that the Brecht rule requires substantial and injurious effect, that is, more than a reasonable possibility that the error contributed to the sentence, we can and do say with greater reason that petitioner is due no relief. Accordingly, we affirm the district court’s denial of relief on the Hitchcock grounds.

INEFFECTIVE ASSISTANCE OF COUNSEL

To demonstrate ineffective assistance of counsel at sentencing, Horsley must establish both (1) that identified acts or omissions of counsel were deficient, and (2) that the deficient performance prejudiced the defense such that, without the errors, there is a reasonable probability that the balance of aggravating and mitigating circumstances would have been different. Strickland v. Washington, 466 U.S. 668, 686-87, 104 S.Ct. 2052, 2064, 80 L.Ed.2d 674 (1984). Courts need not address both of these components “if the defendant makes an insufficient showing on one.” Id. at 697, 104 S.Ct. at 2069. Because Horsley has made an insufficient showing to meet the prejudice element we do not address performance.

In this ease,- Horsley claims that his counsel provided ineffective assistance at sentencing by failing to present expert mental health testimony to demonstrate Horsley’s vulnerability to domination by his eodefendant Baldwin.¹³ Horsley argues that, before trial his counsel learned or should have learned that Horsley had suffered a head injury as a four-year old, but counsel failed to investigate this matter and sought no assistance from mental health experts. Horsley contends that had defense counsel pursued these matters and sought the help of a mental health expert, they would have found that “Horsley suffered neurological impairments from. the head injury that made him a slow learner and that made him vulnerable to the domination of [Baldwin].... ” Horsley claims that, if the sentencing judge had heard this evidence, Horsley would not have received the death penalty.

.After reviewing the record of the state trial, the sentencing hearing and the coram nobis proceedings, and the evidentiary materials on file, the district court concluded that the state record was insufficient to determine whether or not counsels’ decision to present no psychiatric evidence was strategic and concluded that further inquiry into this claim was warranted. At the sentencing hearing, Horsley, his mother, and his grandmother testified about his childhood head injury.¹⁴ The district court found that little of the lay testimony presented at the sentencing hearing was useful in determining the effect of the fall on Horsley’s state of mind or mental competency. The district court, therefore, granted an evidentiary hearing on Horsley’s claim of “ineffective assistance of counsel for failing to present psychiatric evidence at sentencing.” ¹⁵

At the four-day federal evidentiary hearing, the petitioner presented lay witnesses who gave testimony about Horsley’s impoverished background, possible fetal exposure to alcohol, poor performance at school, and exposure to violence. Two mental health experts, Dr. Phillips, a psychiatrist from Connecticut, and Dr. Lyman, a psychologist from Tuscaloosa, Alabama, testified on behalf of petitioner.¹⁶ The district court held that petitioner failed to satisfy the prejudice element of the Strickland test. The court found that Horsley had failed to meet the Strickland test because he “presented no evidence that it was reasonably probable that experts such as those who testified at the evidentiary hearing were available” at the time of the 1977 trial. The district court also found that there was no evidence that state funds were available for such experts in 1977 and that no evidence was presented about the practice in 1977 for presentation of mental state evidence when the defendant appeared to be lucid and attentive with normal intelligence. Even if Horsley had made such a showing, the district court held that counsel was not ineffective because Horsley failed to demonstrate a reasonable probability of a different sentence had the expert testimony been offered. As a result, Horsley was not prejudiced by the absence of expert mental health evidence.¹⁷

In Elledge v. Dugger, 823 F.2d 1439 (11th Cir.1987) (per curiam), we wrote:

“[s]imply put, the Strickland test requires a habeas petitioner ... to show: a) that it was professionally unreasonable for counsel not to investigate; b) what kind of, and how much, investigation an ordinary, reasonable lawyer would have undertaken; e) that it is reasonably probable that a reasonable investigation would have turned up an expert who would have presented testimony similar to that which was eventually adduced; and d) that it is reasonably probable that this testimony would have affected the sentence eventually imposed. Failure to meet any of these steps defeats the ineffectiveness claim.

Id. at 1447 n. 15 (emphasis added).

So, to prove prejudice by failure to investigate and failure to produce a certain kind of expert witness, a habeas petitioner must demonstrate a reasonable likelihood that an ordinarily competent attorney conducting a reasonable investigation would have found an expert similar to the one eventually produced. Id., at 1446. In the absence of such a probability, the petitioner is not injured by the failure to investigate.

To determine whether Horsley -has met this burden we look to all the circumstances of the case and consider all the evidence presented. See Strickland, 466 U.S. at 695-97, 104 S.Ct. at 2069. Based on a review of the record, we conclude that Horsley has made no showing that it was reasonably probable that an ordinary, reasonable lawyer given the constraints of. time and money Horsley’s counsel faced and using reasonable diligence would have discovered mental health experts who would have testified as did Dr. Phillips and Dr. Lyman.¹⁸ That experts were found who would testify favorably almost twenty years later is irrelevant. The record in this case simply does not demonstrate that either of Horsley’s experts would have come to Monroe County, Alabama to testify in 1977. The record also does not show that other experts who would testify favorably to the plaintiff would have been available at that time.¹⁹ The record fails to demonstrate what kind and how much investigation a reasonable lawyer would have made in the circumstances of this case. As in Elledge, we make no comment on whether similar experts were reasonably discoverable or whether a source of funds would have made their testimony possible., We merely hold that the record reveals too little to demonstrate the likelihood of such an occurrence. Accordingly, Horsley has failed to demonstrate that he was prejudiced by counsels’ alleged failure to investigate his mental condition and failure to producé a favorable expert witness.²⁰

CONCLUSION

The district court’s denial of the petition is AFFIRMED.

1 . Under Alabama's 1975 Death Penally Act, once a defendant was convicted of any of the specified offenses, see Ala.Code § 13 — 11—2(a) (1975), and the jury returned the mandatory death sentence, the trial court was required to hold a sentencing hearing, § 13-11-3. Thus, the jury's sentence was not dispositive. Baldwin v. Alabama, 472 U.S. 372, 373-75, 105 S.Ct. 2727, 2729, 86 L.Ed.2d 300 (1985). At the sentencing hearing, evidence “may be presented as to any matter that the court deems relevant to sentence and shall include any matters relating to any of the aggravating or mitigating circumstances enumerated in sections 13-11-6 and 13-11-7.” § 13-11-3 (emphasis added). The court was then required to sentence the defendant to death or to life imprisonment without parole. § 13-11-4. If the court imposed a sentence of death, it was required to set forth in writing findings of fact from the trial and the sentencing hearing, including at least one or more of the aggravating circumstances enumerated in section 13-11-6 and any of the statutory mitigating circumstances enumerated in section 13-11-7 which it found insufficient to outweigh the aggravating circumstances. Id.

After sentencing, the Court of Criminal Appeals was required to review the decision of the trial court, and if that court affirmed, certiorari by the state supreme court was automatic. §§ 13-11-5, 12-22-150; see Baldwin, 472 U.S. at 379 n. 5, 105 S.Ct. at 2731 n. 5. "Both appellate courts 'review ... the aggravating and mitigating circumstances found in the case by the trial judge' and independently weigh those circumstances to determine whether the imposition of a death sentence is appropriate." Id. (emphasis added) (citations omitted).

The 1975 death penalty statute was repealed in its entirety in 1981. It remains effective, however, for crimes committed while it was in force. See 1981 Ala.Acts § 20, codified as Ala.Code, § 13A-5-57 (1982).

2 . The trial court stated:

The Court now considers mitigating circumstances as described in Title 15, Section 342(9) of the 1940 Code of Alabama, as amended:

(a) The court finds the only mitigating circumstance would be the age of the defendant at the time of the commission of the crime, to-wit, nineteen.

Horsley v. State, 374 So.2d 363, 375 (Ala.Cr.App.1978), aff'd, Ex parte Horsley, 374 So.2d 375 (Ala.1979) (emphasis added).

3 . Horsley’s conviction and sentence were affirmed on direct appeal. Horsley v. State, 374 So.2d 363 (Ala.Cr.App.1978), aff'd, Ex parte Horsley, 374 So.2d 375 (Ala.1979). The Supreme Court vacated the judgment and remanded in the light of Beck v. Alabama, 447 U.S. 625, 100 S.Ct. 2382, 65 L.fed.2d 392 (1980). Horsley v. Alabama, 448 U.S. 903, 100 S.Ct. 3043, 65 L.Ed.2d 1133 (1980). The Alabama Supreme Court remanded to the court of criminal appeals which reversed. Horsley v. State, 409 So.2d 1347 (Ala.Crim.App.1981). The U.S. Supreme Court vacated and remanded on authority of Hopper v. Evans, 456 U.S. 605, 102 S.Ct. 2049, 72 L.Ed.2d 367 (1982). Alabama v. Horsley, 457 U.S. 1114, 102 S.Ct. 2921, 73 L.Ed.2d 1326 (1982). On remand, the conviction and sentence were once again upheld, Horsley v. State, 476 So.2d 623 (Ala.Cr.App.1983), aff'd, Ex parte Horsley, 476 So.2d 626 (Ala.1985), cert. denied, Horsley v. Alabama, 475 U.S. 1031, 106 S.Ct. 1239, 89 L.Ed.2d 347 (1986). In 1986, Horsley filed a state coram nobis petition which was denied. Horsley v. State, 527 So.2d 1355 (Ala.Crim.App.1988), cert. denied, 527 So.2d 1355 (Ala.1988), cert. denied, 489 U.S. 1059, 109 S.Ct. 1328, 103 L.Ed.2d 596 (1989).

4 . The district court granted an evidentiary hearing on the following ineffective-assistance-of-counsel claims: 1) failure to object to an alleged Swain violation; 2) failure to object to excessive security at trial; 3) failure to present psychiatric evidence at sentencing; and 4) failure to prepare Horsley as a witness.

5 . Horsley raises other claims in this appeal: 1) ineffective assistance of counsel for failure to object to an alleged violation of Swain v. Alabama, 380 U.S. 202, 85 S.Ct. 824, 13 L.Ed.2d 759 (1965); 2) ineffective assistance of counsel for failure to object to an alleged violation of Witherspoon v. Illinois, 391 U.S. 510, 88 S.Ct. 1770, 20 L.Ed.2d 776 (1968); 3) ineffective assistance of counsel for failure to challenge alleged under-representation of blacks in the grand and petit jury selection processes in violation Duren v. Missouri, 439 U.S. 357, 99 S.Ct. 664, 58 L.Ed.2d 579 (1979); and 4) a claim that Horsley was entitled to funds for mental health experts. We affirm the district court's rulings: the claims are proeedurally barred or wholly without merit.

6 . In Hitchcock, the Court repeated its earlier holding that "in capital cases, the sentencer may not refuse to consider or be precluded from considering any relevant mitigating evidence.” Hitchcock, 481 U.S. at 394, 107 S.Ct. at 1822 (citing Skipper v. South Carolina, 476 U.S. 1, 106 S.Ct. 1669, 90 L.Ed.2d 1 (1986); Eddings v. Oklahoma, 455 U.S. 104, 102 S.Ct. 869, 71 L.Ed.2d 1 (1982); Lockett v. Ohio, 438 U.S. 586, 98 S.Ct. 2954, 57 L.Ed.2d 973 (1978)). A. Hitchcock violation is based upon a Lockett violation, and we have held that Lockett is to be applied retroactively. Hargrave v. Dugger, 832 F.2d 1528, 1533 (11th Cir.1987) (in banc).

7 . Horsley contends that the clearly erroneous standard of review applies to the district court's factual findings on the Hitchcock violation. But for Hitchcock purposes, the district court, as noted above, held no evidentiary hearing, made no factual findings, and never resolved the Hitchcock issue on the merits. Accordingly, our review is plenary. See Abdi v. Georgia, 744 F.2d 1500, 1503 n. 5 (11th Cir.1984).

8 . Because the Alabama sentencing scheme provides no role for the jury in weighing the aggravating and mitigating circumstances, this factor is not relevant to our inquiry.

9 . We note that the Alabama appellate courts independently reviewed the record and seem to have weighed the aggravating and mitigating factors for this case. See Horsley, 476 So.2d at 624. As a result, the appellate courts may have cured weighing errors that may have resulted from the trial court's failure, if any, to weigh all mitigating circumstances. See Clemons v. Mississippi, 494 U.S. 738, 739-42, 110 S.Ct. 1441, 1444, 108 L.Ed.2d 725 (1990) (state appellate courts in weighing states may independently weigh aggravating and mitigating circumstances and thereby cure certain sentencing phase errors). Because the State did not raise this argument in its original brief, we decline to address this issue.

10 . We reject Horsley’ claim that the State has abandoned any harmless error argument. And, even if the State had abandoned this argument we have the discretion to overlook a failure to argue harmlessness and to undertake sua sponte the task of considering harmlessness. E.g. United States v. Giovannetti, 928 F.2d 225, 227 (7th Cir.1991); United States v. Cortes, 949 F.2d 532, 542-43 (1st Cir.1991).

11 . We know that one circuit has limited Brecht’s application considerably and in such a way that Brecht might not apply to this case. See Omdorff v. Lockhart, 998 F.2d 1426 (8th Cir.1993). We think the Supreme Court’s rationale (advancing comity, federalism, finality, and the importance of the trial) for the Brecht rule reaches to almost all federal habeas cases, including this one.

12 . We note that the circuits are split on the question of who bears the burden of proof on collateral review under the Brecht standard, see Ayala v. Leonardo, 20 F.3d 83, 91 (2d Cir.1994) (conflicting authorities noted), and that panels of this court seem to disagree, compare Bolender, 16 F.3d at 1567 (burden on petitioner); Cumbie v. Singletary, 991 F.2d 715, 724 (11th Cir.), cert. denied, - U.S. -, 114 S.Ct. 650, 126 L.Ed.2d 608 (1993) (same); with Bonner v. Holt, 26 F.3d 1081, 1083 (11th Cir.1994) (burden on state). The Supreme Court has recently granted certio-rari on this issue. See O’Neal v. Morris, 3 F.3d 143 (6th Cir.1993), cert. granted, - U.S. -, 114 S.Ct. 1396, 128 L.Ed.2d 70 (1994).

We need not decide this question in this case because even in applying the Kotteakos-Brecht standard and reviewing the error' de novo to determine "whether the error 'had a substantial and injurious effect or influence in determining the [court's] verdict,’ " Brecht, - U.S. at -, 113 S.Ct. at 1722, we conclude the error is harmless. Put differently, given the facts of this case, and the balance of aggravating and mitigating circumstances, any failure of the sentencing court to consider the nonstatutoiy mitigating evidence before it would have been harmless error no matter which party bears the burden of proof in federal court.

13 . We note that Horsley seeks to expand the scope of his claim in the habeas petition by alleging for the first time on appeal that trial counsel were ineffective for not presenting certain non-expert testimony at sentencing. This testimony consists of evidence from school teachers and friends of petitioner; and, according to petitioner, the testimony would have shown that he suffered from fetal exposure to alcohol, a learning disability, a dysfunctional family, exposure to violence, and an early history of drug and alcohol abuse.

In our order for supplemental briefing, we asked Horsley to cite to the record to show where this issue was preserved. No citation was provided. A review of the record shows that the issue of non-expert mental health evidence was raised neither in the state coram nobis proceedings nor in the habeas proceedings below. Hors-ley's habeas petition alleged ineffective assistance because of "[failure to present mental health evidence concerning domination as mitigating evidence." Horsley’s brief in the district court addresses only the absence of expert mental health testimony.

Now Horsley claims that this issue was effectively tried by the implied consent of the parties under Federal Rules of Civil Procedure 15(b). Once again, the record does not support this argument. The district court granted an eviden-tiary hearing limited to four issues. The issue of non-expert testimony in mitigation was not one of those issues. Although the court permitted lay witnesses to testify on Horsley's background, that evidence merely provided the factual background upon which the expert's opinions were based. That neither the court nor the state believed a new claim of ineffective assistance was being raised, or tried, is clear from the record and from the district court's order. Because this issue was raised for the first time on appeal, we decline to consider it. See Lightbourne v. Dugger, 829 F.2d 1012, 1017 n. 5 (11th Cir.1987).

14 . In sum, their testimony reflected that Horsley had fallen as a four year old; was unconscious for a month, was in the hospital for about a month, and that after the fall he suffered from spasms, was unable to remember as his siblings did, and was easy to persuade.

, Horsley's mental health experts testified at the evidentiary hearing in federal court. The expert testimony indicates that Horsley was neither unconscious for a month nor hospitalized for a month: Horsley was in and out of consciousness, was in the hospital for approximately six days, and seemingly had one seizure while in the hospital.

15 . We have held that when the state coram nobis record is insufficient to permit a determination of whether counsel's decision was strategic or negligent it was proper to hold an evidentiary hearing. Thomas v. Kemp, 796 F.2d 1322, 1324 (11th Cir.1986).

16 . Dr. Robert D. Lyman, a professor of neuro-psychology in Tuscaloosa, Alabama, testified on behalf of Horsley. He found that Horsley had an IQ of 92, normal to low normal. About Hors-ley's neurological functioning Dr. Lyman stated, "I think I would conclude that Mr. Horsley probably does have a mild degree of neurological impairment.” He also said that he believed that there were multiple causal factors such as "possible fetal alcohol effects," head injury, and possibly a genetic predisposition. He concluded that all these factors manifested themselves in "attention deficit disorder," and probably in a learning disorder. Dr. Lyman described Hors-ley's mental state at the time of the crime as, "under time pressure and emotional pressure, I think [Horsley] has tended to act impulsively, without considering consequences, without attending to all the events around him." Thus, Dr. Lyman concluded that, at the time of the crime, Horsley acted under the substantial domination of Brian Baldwin, and that Horsley’s capacity to conform his behavior to the requirements of the law or to appreciate the wrongfulness of his conduct was substantially impaired. But, when asked on cross-examination whether a causal connection could be established between Hors-ley's impairment and the murder, Dr. Lyman stated that he did not think so, at least not definitively.

The second expert testifying for Horsley was Dr. Robert Phillips, a psychiatrist. He stated that Horsley had an average IQ. He found that many factors had affected Horsley’s judgment. He concluded that possible fetal exposure to alcohol, a dysfunctional family, head injuiy, organic personality disorder, and constant abuse of drugs and alcohol left Horsley lacking the capacity to conform his behavior because of vulnerability to domination. But, he also described Horsley's participation in the crime as “halfhearted” but "clearly participatory.” Further, Dr. Phillips stated that he believed that Horsley's choice was not "totally free and not totally volitional” because of the underlying set of circumstances which predisposed Horsley to faulty judgments. From his examination he concluded that Horsley appreciated the criminalily of his conduct but lacked the capacity to conform his conduct. Under cross-examination he admitted that none of the factors described above were causally related to the murder.

17 .Because ineffectiveness of counsel presents a mixed question of law and fact we exercise plenary review. The underlying factual findings of the district court are presumptively correct unless clearly erroneous. Bush v. Singletary, 988 F.2d 1082, 1089 (11th Cir.1993).

18 . We have said that to make such a showing a petitioner could present testimony from:

(a) members of the bar relating to the amount of investigation that is reasonable in such a situation and the ease or difficulty in finding such experts at that time, (b) psychiatrists, or other experts ... relating to how widely the proposed theory was accepted at the time ... and the ease an attorney would have had in getting such experts, and (c) any other relevant testimony that would tend to demonstrate it was reasonably probable that reasonable diligence would uncover an expert similar to the one eventually located.

Elledge, 823 F.2d at 1447 n. 17. The record reveals no such showing in this case.

19 . Horsley claims that the availability requirement was met by. the responses of trial counsel at both state and federal hearings. But, the record reflects that the relevant testimony was equivocal. For example, counsel stated that experts “probably” would have been available from the state. More important, however, despite Hors-ley's argument to the contrary, the record is completely silent about whether any experts “would have presented testimony similar to that which was eventually adduced.” Elledge, 823 F.2d at 1447 n. 15. Thus, Horsley’s claim must fail.

20 .And, perhaps even more important, even if Horsley had provided these expert witnesses at the 1977 sentencing hearing, we nonetheless would still affirm the district court. Horsley has not established that it is reasonably probable that having heard this testimony "the sentencer ... would have concluded that the balance of aggravating and mitigating circumstances did not warrant death." Strickland, 466 U.S. at 695, 104 S.Ct. at 2069.

---

Dissenting Opinion

HATCHETT, Circuit Judge,

dissenting:

The majority holds that if the trial court committed a Hitchcock error in sentencing Horsley to death, it was harmless.¹ The majority also holds that Horsley’s counsel was not ineffective for failing to present expert psychiatric testimony at his sentencing hearing because Horsley has not shown that such testimony was available in 1977, at the time of his sentencing. I disagree with both of these holdings and dissent.

I. THE HITCHCOCK ERROR

The district court held that Horsley procedurally defaulted his Hitchcock claim. I agree with the majority’s conclusion that the district court erred in this holding. This, however, is where my agreement with the majority ends.

The majority “suggests” that the trial court did not commit a Hitchcock error, but states “for the purpose of deciding this appeal, we will assume that nonstatutory mitigating evidence was not considered by the sentencing judge.” Thus, the majority does not hold that Horsley’s sentencing court complied with Hitchcock. Nonetheless, I am compelled to respond to the “suggestion” that a Hitchcock error was not committed.

In its order granting an evidentiary hearing, the district court found that the trial court committed a Hitchcock error:

Likewise, it is apparent from the record that the trial judge in Horsley’s case considered himself bound by the enumerated statutory mitigating factors:

THE COURT: The court now considers mitigating circumstances as described in Title IS, § 342(9) of the 1940 Code of Alabama, as amended:

(a) The Court finds the only mitigating circumstance would be the age of the defendant at the time of the commission of the crime, to-wit, nineteen.

The court having considered the aggravating circumstances and the mitigating circumstances and after weighing the aggravating and mitigating circumstances, it is the judgment of the Court that the aggravating circumstances far outweigh the mitigating circumstances and that the death penalty as fixed by the jury should be and hereby is accepted.

This procedure does not appear to satisfy Hitchcock or the eases cited therein.

However, no evidentiary hearing is required on this claim as the merits can be considered without further evidence. Whether this claim is procedurally barred is a matter that will be addressed in the Court’s final Order following the evidentia-ry hearing.

(Citations omitted). In footnote 7, the majority declares that we are not obliged to defer to this finding under the clearly erroneous standard because the district court did not hold an evidentiary hearing and, therefore, did not make findings of fact on this issue. This reasoning does not comport with this circuit’s most recent precedent.

In Spaziano v. Singletary, 36 F.3d 1028, 1032 (11th Cir.1994), cert. denied, — U.S. -, 115 S.Ct. 911, 130 L.Ed.2d 793 (1995), we recently stated: “While the issue of whether Hitchcock error occurred is a legal one, it is almost entirely dependent upon the answer to a question of fact: Did the sentencing judge consider any and all nonstatu-tory mitigating circumstance evidence that was presented to him?” In Spaziano, the district court found no Hitchcock error because “the trial judge knew that he was bound to consider, and did actually consider, the non-statutory mitigating circumstance evidence.” Spaziano, 36 F.3d at 1032. In reviewing this determination, we held:

Because that is a finding of historical fact — what the judge knew and what he did — we review the finding under the deferential clearly erroneous standard, which the Supreme Court defined as follows: [We then quoted a passage from Anderson v. City of Bessemer City, N.C., 470 U.S. 564, 573-74, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985).] In the present case, the district court did not hold an evidentia-ry hearing, but instead based its fact findings on the state record, documentary evidence, or inferences from other facts. However, the Supreme Court has held that the clearly erroneous standard, as defined in Anderson, is applicable.to fact findings drawn solely from documents, records, or inferences from other facts. Anderson, 470 U.S. at 574, 105 S.Ct. at 1511. Therefore, it applies with full force here.

Spaziano, 36 F.3d at 1032 (parallel citations omitted). In this case, the district court “did not hold an evidentiary hearing, but instead based its fact findings on the state record,” Spaziano, 36 F.3d at 1032, and, therefore, the clearly erroneous standard of review applies. Thus, under Spaziano, the majority’s reasoning in footnote 7 is simply wrong.

The majority also states that because the district court determined that Horsley’s Hitchcock claim was procedurally barred, it could never have reached the merits of the claim. Consequently, the majority reasons that the district court’s comments were merely preliminary in nature and the clearly erroneous standard does not apply. This reasoning is also flawed.

Although the majority suggests otherwise, it is common practice for federal courts, including this court, to hold that a habeas corpus claim is procedurally defaulted and then to also address the merits of the claim. See, e.g., Roberts v. Singletary, 29 F.3d 1474, 1478 n. 2 (11th Cir.1994) (“Although we find the Ritchie claim to be procedurally defaulted, our review of the record persuades us that even if we addressed the issue on the merits, Roberts would be entitled to no relief.”). More importantly, however, the majority’s reasoning once again manifests a disregard for this circuit’s most recent precedent. In Alderman v. Zant, 22 F.3d 1541, 1554 (11th Cir.), cert. denied, — U.S. -, 115 S.Ct. 673, 130 L.Ed.2d 606 (1994), the district court considered a habeas corpus claim and “found that the claim was procedurally defaulted, an abuse of the writ, or in the alternative without merit.” When reviewing this decision, we stated that although we found the claim “to be procedurally barred, we feel it is necessary to briefly discuss the merits of the claim and the district court’s factual findings.” Alderman, 22 F.3d at 1553. Even though the district court held that the claim was procedurally defaulted, we went on to specifically state that the district court made “factual findings which are entitled to be reviewed under the clearly erroneous standard.” Alderman, 22 F.3d at 1554. Essentially, the district court in Alderman did exactly what the district court did in this case (both courts found the petitioner’s claim to be procedurally barred, but made factual findings on the merits), and we still determined that the district court’s findings were subject to the clearly erroneous standard.

I believe the district court did not clearly err in finding that the trial court committed a Hitchcock error. Even if the correct standard of review were de novo, however, I would still conclude that the trial court committed a Hitchcock error because of the similarity between the trial court’s statements and the sentencing judge’s statements in Hitchcock. Since this is not the holding of the majority, however, I will not discuss this issue further.

The majority holds that any Hitchcock error in Horsley’s case was harmless. Even though Horsley addressed harmless error in his initial brief, the state did not mention it once in its response brief: it did not list harmless error in its statement of the issues; it did not mention harmless error in the summary of its argument; it did not list harmless error as a point heading in its argument; and it did not mention harmless error anywhere in the body of its argument. This panel, however, stepping out of, its traditional role, raised the issue and asked the parties to provide supplemental briefing. The state’s first mention of harmless error came as a response to this request. The state waived the harmless error argument.

In Hitchcock, the Supreme Court indicated that the harmless error issue should not be raised sua sponte in this context. The Supreme Court stated: “Respondent has made no attempt to argue that this error was harmless.... In the absence of such a showing, our cases hold that the exclusion of mitigating evidence of the sort at issue here renders the death sentence invalid.” Hitchcock, 481 U.S. at 399, 107 S.Ct. at 1824.

Instead of following the dictates of the Supreme Court, which this court is required to do, the majority cites a First Circuit case and a Seventh Circuit case for the proposition that the court can raise the harmless error issue sua sponte in this instance. Obviously, these cases are not controlling; moreover, they are not on point, for they do not address Hitchcock errors. Nonetheless, even this non-binding, inapplicable authority reveals the impropriety of raising the harmless error issue sua sponte in this instance.

These cases declare that when deciding whether to raise harmless error sua sponte, “the controlling considerations are the length and complexity of the record, whether the harmlessness of the error or errors found is certain or debatable, and whether reversal will result in protracted, costly, and ultimately futile proceedings in the district court.” United States v. Giovannetti, 928 F.2d 225, 227 (7th Cir.1991).² The majority chooses not to analyze these factors. Indeed, if it had, it would have recognized that none of the factors weigh in favor of invoking harmless error sua sponte in this instance.

The analyses of the first and third factors are quite straightforward. When considering the first factor, we are faced with the voluminous records of a complex death penalty case. After a full-blown murder trial and capital sentencing hearing, this case has traveled the typical circuitous route of death penalty cases. Footnote 3 of the majority opinion summarizes the length and complexity of the post-trial proceedings in this case. These proceedings have included several trips through the state courts, a trip to the United States Supreme Court, and, of course, this federal habeas corpus petition. Clearly this is not the type of record suited for a sua sponte review. When considering the third factor, a reversal of Horsley’s sentence will not result in protracted, costly, and ultimately futile proceedings. To the contrary, an Alabama trial court will simply have to re-sentence Horsley. Thus, the third factor also weighs against the majority’s sua sponte discussion.

The analysis of the second factor is a bit more complex. Whether the harmlessness is certain or debatable necessarily depends upon the standard of harmless error utilized.³ Under the Chapman standard, the harmlessness of the Hitchcock error in Horsley’s ease is far from certain. Even the majority states that “we think we could conclude” (emphasis added) that the error was harmless under the Chapman standard, thereby conceding that the issue is debatable.⁴ Of course, the majority rejects Chapman and implements the Brecht standard. Still, I do not believe that the harmlessness of the trial court’s Hitchcock error is certain under Brecht. Indeed, in applying the Brecht standard, the majority does not use decisive language, as it hesitantly states that it can “say with greater reason” that the error was harmless. Finally, even if the harmlessness of the trial court’s error were certain under the Brecht standard, the other two factors clearly weigh against conducting a sua sponte harmless error review. Thus, even the majority’s authority indicates that a sua sponte harmless error analysis was ill-advised in this instance.

In sum, I agree with the district court’s finding that the Alabama trial court committed a Hitchcock error and its finding should be deferred to under a clearly erroneous standard. I also conclude that Hitchcock specifically prohibits the majority’s sua sponte harmless error holding. Additionally, the majority’s non-binding, inapplicable authority indicates that a sua sponte harmless error analysis is inappropriate in this instance. Consequently, I disagree with the majority’s decision to deny Horsley relief on his Hitchcock claim.

II. INEFFECTIVE ASSISTANCE OF COUNSEL

The majority correctly states that in Elledge v. Dugger, 823 F.2d 1439 (11th Cir.1987), cert. denied, 485 U.S. 1014, 108 S.Ct. 1487, 99 L.Ed.2d 715 (1988), we explained that the two-pronged Strickland test becomes a four-pronged inquiry when we are faced with a claim of ineffective assistance of counsel for failure to present expert testimony at sentencing.⁵ The majority holds that Horsley has failed to satisfy the third prong of this inquiry, which requires a petitioner to show that “it is reasonably probable that a reasonable investigation would have turned up an expert who would have presented testimony similar to that which was eventually adduced.” Elledge, 823 F.2d at 1447 n. 15.

In deciding whether Horsley has satisfied this third prong, we “must look to all the circumstances of the ease and consider all the evidence presented.” Elledge, 823 F.2d at 1446 (citing Strickland, 466 U.S. at 695, 104 S.Ct. at 2068). In holding that Horsley’s trial counsel was not ineffective at sentencing, the district court committed an error of law because it failed to recognize this requirement. Instead, the district court was under the erroneous impression that it was only to consider the evidence presented at the federal evidentiary hearing. This error becomes apparent when comparing the district court’s order granting an evidentiary hearing on the ineffective assistance of counsel claim with its final order denying the claim.

In its initial order, the district court recognized that “Horsley’s counsel recalled that psychiatric experts probably would have been available, had such assistance been requested.” In support of this determination, the district court quoted testimony that Horsley’s trial counsel gave at his coram nobis proceeding:

Q: At the time Edward Horsley was tried, was a psychiatrist available from the State to assist indigent defendants in the handling of their cases?

A: Whether or not this person was a psychologist or a psychiatrist I don’t remember. My best recollection is that there probably would have been somebody available at that time.

Conversely, after conducting an evidentiary hearing, the district court denied Horsley’s ineffective assistance claim, explaining: “Petitioner presented no evidence that it was reasonably probable that experts sueh as those who testified at the evidentiary hearing were available to Petitioner’s attorneys in 1977.” Apparently, the district court incorrectly believed that it was confined to the evidence presented at the evidentiary hearing, for it had previously stated, when viewing the evidence as a whole, that psychiatric experts “probably would have been available.”

Unlike the district court, the majority recognizes that we are obliged to consider all of the evidence in the record. It states, however, that “the record reveals too little” evidence to satisfy the third-prong of the El-ledge test. I disagree. In addition to the testimony quoted in the district court’s opinion, three other portions of the record indicate the availability of expert testimony in 1977.

First, immediately before Horsley’s trial counsel testified that experts were probably available, he testified as follows:

Q. What was your understanding as to the availability of expert assistance from the state?

A. Expert assistance in what regard?

Q. Well, for example — first of all, is there any psychiatrist in Monroeville?

A. No, sir, but there is one that from time to time has been made available through the Department of Health, I believe.

Later on in the coram nobis proceeding, Horsley’s counsel testified:

Q. Did you know of any assistance that might be available from the State in assisting you to analyze this case from the psychological or psychiatric point of view?

A. I think upon petition to the Court,probably he could have been examined if it was shown that he had something like that, but I don’t recall specifically. I don’t know.

Moreover, at the evidentiary hearing before the district court, Horsley’s counsel testified:

THE COURT: At that time back in 1977 was there a psychiatrist practicing in Monroeville?

THE WITNESS: No, sir, there wasn’t. At that time I believe the practice would have been to have made a showing before the trial judge about the need for a psychiatric evaluation, and then I suppose he could order one based on that. I mean-the practice has changed.

THE COURT: Where would have been the nearest psychiatrist back in 1977?

THE WITNESS: Well, it would have been the state institution, Mt. Vernon or Searcy or something of that sort, I believe, Your Honor.

THE COURT: Would that be true for a psychologist as well as for a psychiatrist?

THE WITNESS: Well, the mental health center had a psychologist there part-time.

THE COURT: In Monroeville?

THE WITNESS: Yes, sir. Who worked there part-time anyway.

In reviewing the evidence as a whole, I believe that these portions of the record are enough to satisfy the availability prong of the Elledge test. The majority emphasizes that in Elledge we stated that a petitioner must show that experts were available at the time of sentencing “who would have presented testimony similar to that which was eventually adduced.” Elledge, 823 F.2d at 1447 n. 15. The experts that Horsley produced at the evidentiary hearing, however, did not present “cutting edge,” newly discovered, or radical psychiatric theories. In this respect, this ease is distinguishable from Elledge, where the testimony that the expert gave at the evidentiary hearing was unacceptable at the time of Elledge’s sentencing. Elledge, 823 F.2d at 1446-47. Put simply, the evidence as a whole shows that psychiatric experts from Alabama were available to Horsley’s counsel in 1977, and these experts could have testified about the basic concepts that the experts described in the evidentiary hearing before the district court.

Finally, because I also believe that Horsley has satisfied the other prongs of the Elledge test, I respectfully dissent from the majority’s decision to deny him relief on his ineffective assistance of counsel claim.⁶

1 . Hitchcock v. Dugger, 481 U.S. 393, 107 S.Ct. 1821, 95 L.Ed.2d 347 (1987) (in capital cases, the sentencer may not refuse to consider or be precluded from considering relevant mitigating evidence).

2 . Admittedly, we have relied on the reasoning in Giovannetti before, emphasizing that we may address the harmless error issue sua sponte only where the harmlessness “is patently obvious." United States v. Adams, 1 F.3d 1566, 1576 (11th Cir.1993), cert. denied, - U.S. -, 114 S.Ct. 1310, 127 L.Ed.2d 660 (1994).

3 . As the majority points out, in Brecht v. Abrahamson, - U.S. -, 113 S.Ct. 1710, 123 L.Ed.2d 353 (1993), the Supreme Court replaced the traditional harmless error standard, as articulated in Chapman v. California, 386 U.S. 18, 87 S.Ct. 824, 17 L.Ed.2d 705 (1967), with the less onerous standard, described in Kotteakos v. United States, 328 U.S. 750, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946), for cases on collateral review. I disagree with the majority's decision to apply the Brecht harmless error standard in this case. Like the Eighth Circuit, I believe that the Chapman harmless error standard should apply when a federal habeas corpus court is the first court to conduct a harmless error analysis. See Orndorff v. Lockhart, 998 F.2d 1426, 1430 (8th Cir.1993) (reasoning that "the Brecht rule is based largely on the notion that because the state courts can properly apply the Chapman harmless error standard on direct review, the federal habeas courts need only review those decisions under the Kot-teakos harmless error standard"), cert. denied, - U.S. -, 114 S.Ct. 1631, 128 L.Ed.2d 354 (1994).

4 .While the question here is only whether the harmlessness of the Hitchcock error was certain or debatable, it should be noted that this circuit has frequently concluded that Hitchcock errors were not harmless under the Chapman standard. See Booker v. Dugger, 922 F.2d 633, 634 (11th Cir.), cert. denied, 502 U.S. 900, 112 S.Ct. 277, 116 L.Ed.2d 228 (1991); Jackson v. Dugger, 931 F.2d 712, 716 (11th Cir.), cert. denied, 502 U.S. 973, 112 S.Ct. 452, 116 L.Ed.2d 470 (1991); Delap v. Dugger, 890 F.2d 285, 304-06 (11th Cir.1989), cert. denied, 496 U.S. 929, 110 S.Ct. 2628, 110 L.Ed.2d 648 (1990); Jones v. Dugger, 867 F.2d 1277, 1279-80 (11th Cir.1989); Magill v. Dugger, 824 F.2d 879, 893-95 (11th Cir.1987).

5 . Strickland v. Washington, 466 U.S. 668, 687, 104 S.Ct. 2052, 2064, 80 L.Ed.2d 674 (1984) ("First, the defendant must show that counsel's performance was deficient.... Second, the defendant must show that the deficient performance prejudiced the defense.”).

6 . As an afterthought, the majority states, in footnote 20, that Horsley also failed to satisfy the fourth prong of Elledge: that the expert "testimony would have affected the sentence eventually imposed.” Elledge, 823 F.2d at 1447 n. 15. Since the majority did not engage in any analysis in arriving at this conclusion, it is enough to say that we have often held that a failure to present similar psychiatric testimony has affected the sentence eventually imposed. See, e.g., Middleton v. Dugger, 849 F.2d 491, 495 (11th Cir.1988).