OPINION
VAN ANTWERPEN, Circuit Judge.
In this appeal and cross-appeal in a capital case, the Commonwealth of Pennsylvania and Andre Stevens ask us to review the District Court’s partial grant of Stevens’s petition for writ of habeas corpus. We will affirm the judgment of the District Court in all respects.
I.
Stevens was convicted at a 1993 bench trial on two counts of first-degree murder.
At trial, Stevens did not dispute that he shot his estranged wife, Brenda Jo Stevens, and her acquaintance, Michael Love, at a Beaver County, Pennsylvania, bar. A jury was empaneled for the capital sentencing phase. It unanimously found the existence of aggravating factors, and it unanimously found that those aggravating factors outweighed the mitigating factors. Stevens was sentenced to death. On direct appeal, the Supreme Court of Pennsylvania affirmed.
Commonwealth v. Stevens,
543 Pa. 204, 670 A.2d 623, 627-28 (1996).
Shortly thereafter, Stevens filed a petition pursuant to the Pennsylvania PosfiConviction Relief Act (PCRA). The PCRA court denied relief. In affirming, the Pennsylvania Supreme Court split, 5-2, on the merits of Stevens’s claim that trial counsel rendered ineffective assistance at the guilt phase by failing to investigate, develop, and present a diminished capacity defense.
See Commonwealth v. Stevens,
559 Pa. 171, 739 A.2d 507, 511-16 (1999). Unanimously, the court rejected Stevens’s argument that counsel ineffectively failed to argue that a potential juror was improperly excluded under
Witherspoon v. Illinois,
391 U.S. 510, 522, 88 S.Ct. 1770, 20 L.Ed.2d 776 (1968).
Id.
at 520-21, 88 S.Ct. 1770.
Several months later, Stevens filed his federal habeas petition. In the petition, he raised fourteen claims, including,
inter alia,
that (a) counsel ineffectively failed to formulate and press a diminished capacity defense and (b) the potential juror’s exclusion constituted a
Witherspoon
violation. The District Court denied relief on the ineffectiveness claim, which was Stevens’s only guilt-phase claim. The District Court granted relief on Stevens’s
Witherspoon
claim.
Stevens v. Horn,
319 F.Supp.2d 592, 617 (W.D.Pa.2004). It did not reach Stevens’s remaining sentencing claims. Both Stevens and the Commonwealth timely appealed.
II.
A.
Stevens argued in Ms habeas petition that his trial attorney, Wayne Lipeeky, Esq., provided constitutionally inadequate representation by failing to investigate, develop, and present a diminished capacity defense. To establish Meffectiveness on the part of Lipeeky, Stevens was required to show both that (a) the attorney performed unreasonably under prevailing professional norms and (b) there exists a “reasonable probability” that, but for counsel’s unprofessional errors, he would not have been convicted.
See Strickland v. Washington,
466 U.S. 668, 688-89, 694, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984).
Stevens originally retained Wendell Freeland, Esq., to represent him. Free-land subsequently withdrew because of the cost of hiring mental health experts, and the trial court appointed the Beaver County Public Defender’s Office to represent Stevens. Lipeeky, representing that office, ultimately took over the primary responsibility for the case. Stevens contends that Lipeeky failed to forward certain categories of evidence — a journal, Stevens’s letters to Freeland, his complete military and employment records, and information from family members about his childhood — to psychiatric experts. During the state postconviction proceedings, several experts opined that, had Lipeeky provided this information to an expert prior to trial, he would have been able to secure and present expert testimony at the guilt phase that Stevens suffered — at the time of the offense — from profound neuropsychological ailments. Among other things, these experts testified that Stevens suffered from organic brain damage, psychosis, profoundly disorgamzed thinMng, paranoia, and extreme depression.
The Pennsylvania Supreme Court rejected the ineffectiveness claim, holding that Stevens had “not proved, by a preponderance of the evidence, that there was a reasonable probability that presentation of expert testimony concerning [his] psychosis would have successfully prevented the Commonwealth from proving ... premeditation beyond a reasonable doubt.”
Stevens,
739 A.2d at 516. The court relied, in particular, on the trial judge’s assertion— when sitting as the PCRA court — that even if he had had the benefit of additional testimony concerning Stevens’s neuropsychological deficiencies, he would nevertheless have found that Stevens acted with the specific intent to kill.
Id.
The PCRA court emphasized, m particular, the evidence indicating that Stevens had purposefully retrieved his weapon from Ms vehicle before returning to the bar to attack Brenda Jo and Love.
Id.
On habeas review, the District Court demed relief on the ineffectiveness claim. Recognizing that the state courts had not decided whether Lipeeky’s performance was unreasonably deficient, the District Court conducted a
de novo
review of that issue.
See Stevens,
319 F.Supp.2d at 608 (citing
Wiggins v. Smith,
539 U.S. 510, 534, 123 S.Ct. 2527, 156 L.Ed.2d 471 (2003)). The District Court generally found that Lipeeky’s performance was reasonable, although it opted not to decide whether Lipeeky performed unreasonably in failmg to obtain and forward information from Stevens’s family about his life
and difficult childhood.
Stevens,
319 F.Supp.2d at 613.
Turning to the issue of prejudice, the District Court recognized that it would normally owe deference to the Pennsylvania state courts’ resolution of that issue.
Stevens,
319 F.Supp.2d at 613-14. Nevertheless, it noted Stevens’s argument that the Pennsylvania Supreme Court’s resolution of the prejudice issue — which seemed to rely on the PCRA judge’s subjective observations about how he personally would have viewed the evidence of neuropsychological ailments — was “contrary to” the Supreme Court’s binding precedent on ineffectiveness.
See id.
at 614;
see also
28 U.S.C. § 2254(d) (requiring deference by habeas courts to state courts’ legal conclusions unless they constituted an “unreasonable application of’ or were “contrary to” Supreme Court precedent). The District Court concluded that it need not decide whether the analysis of prejudice conducted by the state courts was “contrary to”
Strickland
because no prejudice could be shown even under
de novo
review.
Stevens,
319 F.Supp.2d at 614.
B.
Although the substance of our analysis is somewhat different than the District Court’s, we agree that Stevens cannot prevail on his ineffectiveness claim. In reaching this conclusion, we exercise our discretion to decide the claim solely on the basis of the Strickland prejudice prong.
See Strickland,
466 U.S. at 697, 104 S.Ct. 2052 (“[i]f it is easier to dispose of an ineffectiveness claim on the ground of lack of sufficient prejudice, which we expect will often be so, that course should be followed”). This approach is particularly appropriate in this case because (a) the Pennsylvania Supreme Court did not pass on the reasonableness of Lipecky’s performance and (b) the District Court itself opted not to decide some aspects of Stevens’s claim that Lipecky unreasonably failed to forward evidence to the psychiatric experts.
In finding that Stevens had not demonstrated a reasonable probability of prejudice under
Strickland,
the District Court relied on two themes. First, the District Court concluded that Stevens had not sufficiently pointed to the existence of any admissible evidence that would have been developed if counsel had provided better information to the psychiatric experts. Second, the District Court concluded that there was such “overwhelming evidence of specific intent” that Stevens simply could
not demonstrate a reasonable probability that Lipecky’s failures had altered the outcome of the guilt phase. Of these two grounds, the second is more persuasive.
“In Pennsylvania, diminished capacity 'is an extremely limited defense, which requires extensive psychiatric testimony establishing a defendant suffered from one or more mental disorders which prevented him from formulating the specific intent to Mil.’ ”
Jacobs v. Horn,
395 F.3d 92, 105 (3d Cir.2005) (quoting
Commonwealth v. Cuevas,
574 Pa. 409, 832 A.2d 388, 393 (2003)). In the District Court’s view, Stevens pointed on postconviction review to too little evidence suggesting he suffered from psychological disorders that might have prevented the formulation of the specific intent to Mil.
Stevens,
319 F.Supp.2d at 614-16 & n. 18. Specifically, the District Court believed the postconviction testimony of Stevens’s experts had been too focused on what they might have been able to tell the jury at the sentencing phase (and not focused enough on what they could have said at the guilt phase).
See id.
In this regard, the District Court stressed that Dr. Rodney Altman “never stated (and in fact was never asked by Stevens’ [s] PCRA counsel) that ... he would have been able to testify at the guilt phase that Stevens’[s] impairments affected his ability to formulate the specific intent to Mil.”
Id.
at 615.
It is true that Dr. Altman’s testimony focused largely on what evidence he would have been able to present at the penalty phase if fuller neuropsychological testing had been conducted prior to trial. That said, Dr. Altman expressly testified, with the benefit of the information available on postconviction review, that — at the time of the offense — Stevens was likely psychotic, paranoid, schizophrenic, and exhibited a “thought process disorder.” Dr. Ralph E. Landefeld’s testimony was similar. Notably, Dr. Christine Martone specifically opined that she believed Stevens’s psychosis was serious enough that he was unable to form the specific intent to kill.
Indeed, as Stevens stresses, each of his three primary experts agreed that he suffered from profound neuropsychological ailments at the time of the crime. Although much of the experts’ postconviction testimony was focused on what they might have offered at the penalty phase as mitigation evidence, this same testimony was seemingly of the type that Pennsylvania courts have found to be relevant and admissible as to the question of diminished capacity.
See, e.g., Jacobs, 395 F.3d
at
105; see also Commonwealth v. Legg,
551 Pa. 437, 711 A.2d 430, 433 (1998) (stressing that evidence of cognitive disorders affecting the defendant’s ability to act deliberately or permissively was admissible).
More convincing is the District Court’s alternative prejudice-prong conclusion, that the evidence presented at trial was so overwhelming that Stevens could not establish a reasonable probability that an objective trier-of-fact, even one presented with a full diminished capacity defense supported by expert testimony, would have found him not guilty.
See Stevens,
319 F.Supp.2d at 616-17. In
Zettlemoyer v. Fulcomer,
923 F.2d 284, 297 (3d Cir.1991), we concluded that — even if it were assumed,
dubitante,
that the attorney in that
capital case performed unreasonably by failing to obtain and present evidence in support of a diminished capacity defense— the habeas petitioner could not establish
Strickland
prejudice in the face of “compelling uncontroverted evidence” establishing an intent to kill.
In the District Court’s view, the evidence against Stevens was analogous:
Stevens admitted during the guilt phase of his trial that he left Armando’s bar the night of the murders, retrieved his gun from his car, cocked it three times, and then reentered the bar. He stood near the dance floor and waited for Love and his wife to finish dancing. When they walked away from each other, he approached his wife, pressed the gun to the back of her head and shot her. He then turned to Love — who had raised his arms in the air — and shot him twice in rapid succession. He shot Love several more times while Love lay on the floor, including an intentional shot to his scrotum area. After shooting Love, he then refocused on his wife, who lay lifeless on the floor. He shot her again in the head and he admitted that he had intentionally aimed at her back.
Stevens,
319 F.Supp.2d at 616 (internal citations omitted).
The Pennsylvania Supreme Court provided a similar summary of the testimony:
[Tjhere was testimony that, on the evening of the murders, Appellant became very angry when he saw Brenda Jo dancing with Love. Appellant slammed his forearms on the bar and stood and stared at Love and Brenda Jo and then left the bar. He retrieved his gun from beneath the seat of his vehicle and returned to the bar. Appellant walked toward the table area, pulled the gun from underneath his jacket, raised it, but then lowered it to his side and waited until Love and Brenda Jo were finished dancing. When they left the dance floor, Appellant cocked the gun and without hesitation walked up behind Brenda Jo and shot her in the back of the head at close range. He then turned to Love and shot him twice rapidly in succession, and fired at Love several times while he lay on the floor, including an intentional shot through the scrotum area. Appellant fired a last shot at Brenda Jo, hitting her again in the head. Appellant had fired a total of ten shots.
Stevens,
739 A.2d at 516.
The Supreme Court offered this recounting in the course of expressly agreeing with the PCRA court when it held that, even in the face of a diminished capacity defense fully supplemented with appropriate expert testimony, it “would still have concluded that [Stevens] had the capacity to form the specific intent to kill.”
Id.
Stevens argues that (i) the trial (and PCRA) judge erred when he
relied on what his own personal view of a fully-supported diminished capacity defense would have been and (ii) the Pennsylvania Supreme Court exacerbated this error by deferring to the trial judge’s subjective views. This is a fair point. Courts have generally indicated that a determination of
Strickland
prejudice should be the result of an objective analysis (i.e., how an objective decisionmaker would have acted if counsel’s unprofessional errors had not occurred).
See Strickland,
466 U.S. at 695, 104 S.Ct. 2052 (stating that the inquiry into prejudice does “not depend on the idiosyncraeies of the particular decision-maker”);
see also United States v. Moran,
393 F.3d 1, 11 n. 6 (1st Cir.2004);
Miller v. Angliker,
848 F.2d 1312, 1323 (2d Cir.1988).
However, even if the approach taken by the state courts was “contrary to” — or, at a minimum, an “unreasonable application”
of
— Strickland,
see
28 U.S.C. § 2254(d), making deference to the state courts inappropriate, Stevens still cannot prevail. As the District Court concluded, even under a
de novo
review, Stevens cannot establish a reasonable probability that an objective trier-of-fact would have found him not guilty.
Stevens,
319 F.Supp.2d at 614. As in
Zettlemoyer,
the evidence that Stevens formed, and acted upon, the specific intent to kill was too strong. Stevens did not simply erupt into violence when he saw Brenda Jo and Love dancing at a bar. Instead, he went to his vehicle, located his weapon underneath a seat, and cocked it three times. He reentered the bar, pulled the gun from underneath his coat, pointed the weapon at the ceiling and then lowered it. He waited until Brenda Jo and Love finished dancing. These events evince a real deliberateness, and — significantly— they indicate that Stevens acted after having a moment to reflect.
When Brenda Jo and Love finished their dance, Stevens again cocked the gun; he then quickly approached Brenda Jo and shot her in close range in the back of the head. Stevens turned to Love, whose arms were in the air. He shot Love two or three times. He then fired several more times at Love, with the last shot being an intentional shot at the man’s groin.
Again, these events evince specific intent. Notably, Stevens had to cock the gun before firing it. He fired first on Brenda Jo, then turned to Love. The placement of his final shot at Love was obviously purposeful. After Stevens shot Love, he focused attention again on Brenda Jo, who was lifeless on the floor. Stevens shot her again, having — according to his own testimony — intentionally aimed at her head. As with the prior events, these final events, too, suggest real intent. Given this overwhelming evidence of deliberateness and purposefulness, the District Court correctly concluded that Stevens did not establish a reasonable probability that he would have been found not guilty if he had been able to present better testimony from psychiatric experts. The evidence of his own crime, including the evidence he gave at the guilt phase, betray his claim that he acted without any specific intent to kill.
Zettlemoyer,
923 F.2d at 297.
For the foregoing reasons, Stevens has not established a reasonable probability that, but for counsel’s possible errors in faffing to forward evidence to the psychiatric experts, he would not have been convicted. Even a diminished capacity defense that was fully supplemented by expert testimony would have been fatally undermined by the evidence, inherent in
the events of the murders, that Stevens acted with the specific intent to kill.
III.
During voir dire prior to the sentencing phase, the following colloquy occurred between potential juror Nancy Hartling and the judge:
Judge:
Do you have an opinion about the death penalty which would prevent you from following the Court’s instructions as to what penalty should be imposed?
Nancy Hartling:
I don’t believe in the death penalty.
Judge:
Very well.
Prosecutor:
Challenge for cause, Your Honor.
Judge:
Very well. You are excused, thank you.
Stevens’s attorney did not object to the exclusion or seek in any way to rehabilitate Hartling. Stevens argued in his habeas petition that the removal of Hartling violated
Witherspoon.
In
Witherspoon,
the Supreme Court held that a member of a jury panel may not be excused for cause simply for “voicing] general objections to the death penalty or expressing] conscientious or religious scruples against its infliction.” 391 U.S. at 522, 88 S.Ct. 1770. Otherwise, a defendant’s right to a fair and impartial jury would be violated.
See Szuchon v. Lehman,
273 F.3d 299, 327 (3d Cir.2001). In
Wainwright v. Witt,
469 U.S. 412, 424, 105 S.Ct. 844, 83 L.Ed.2d 841 (1985) (internal quotation and citation omitted), the Court clarified that the proper standard for determining a
Witherspoon
violation “is whether the juror’s views would prevent or substantially impair the performance of his duties as a juror in accordance with his instructions and his oath.” This standard does not require, as footnotes in
Witherspoon
itself intimated, that a juror’s bias against the death penalty be shown with “unmistakable clarity.”
Id.
This is so because “many veniremen simply cannot be asked enough questions to reach the point where their bias has been made ‘unmistakably clear.’ ”
Id.
at 424-25, 105 S.Ct. 844. It is enough to justify an excusal that the trial judge be “left with the definite impression that a prospective juror would be unable to faithfully and impartially apply the law.”
Id.
at 426, 105 S.Ct. 844. A trial judge’s finding under this standard is a factual determination, and it is therefore entitled to a presumption of correctness under 28 U.S.C. § 2254(e)(1).
Martini v. Hendricks,
348 F.3d 360, 363 (3d Cir.2003),
cert. denied,
543 U.S. 1025, 125 S.Ct. 662, 160 L.Ed.2d 503 (2004).
This recognizes that a trial judge is in the best position to judge the credibility and demeanor of po
tential jurors during voir dire.
Witt,
469 U.S. at 428, 105 S.Ct. 844;
Szuchon,
273 F.3d at 328.
When the issue was raised on state post-conviction review, the trial judge said he had “no specific recollection of Mrs. Hartling.” Nevertheless, he thought there must have been something about Hartling’s demeanor that was “emphatic and clearly conveyed her unwillingness to comply with the court’s instructions.” He noted that he had asked other potential jurors whether they would be able to set aside their opposition to the death penalty. In affirming the denial of postconviction relief, the Pennsylvania Supreme Court also noted that the trial judge had asked “most other potential jurors” whether they would be able to set aside their personal opinions about the death penalty.
Stevens,
739 A.2d at 521. It therefore “accept[ed] the statement of the trial court ... that its decision not to conduct further inquiry with respect to [Hartling] was based on its assessment of her demeanor.”
Id.
The District Court, however, found a straightforward violation of
Witherspoon. Stevens,
319 F.Supp.2d at 604.
As detailed above, the trial judge — acting on the prosecutor’s objection — excused Hartling for cause after she simply expressed her personal opposition to the death penalty. No follow-up questioning ensued to determine whether Hartling could, despite her qualms about the death penalty, set aside her views and follow the trial court’s instructions and the law. As the Commonwealth appears to recognize, if we were conducting
de novo
review, we would have to find a
Witherspoon
violation. Indeed, in
Szuchon,
conducting
de novo
review, we found a
Witherspoon
violation in an interchange that was, if anything, slightly more expansive than the colloquy here.
See Szuchon,
273 F.3d at 329.
Szuchon
was governed by a less deferential standard of review. Under the present, more deferential regime, the state courts’ finding that Hartling was biased against the death penalty is presumed to be correct, and Stevens bears “the burden of rebutting the presumption ... by clear and convincing evidence.” 28 U.S.C. § 2254(e)(1).
We agree with the District Court that Stevens met that burden here. As noted, and as in
Szuchon,
the prospective juror — here, Hartling — indicated no more than that she opposed the death penalty. Because the trial judge did not follow up Hartling’s single statement with any follow-up questioning, there is simply
no tangible evidence in the record to support the finding of the state courts. In the absence of
any
evidence in the record to support the state courts’ finding, it must be true that Stevens has established, by the requisite clear and convincing evidence, that Hartling was removed for cause on a “broader basis than inability to follow the law or abide by [her] oath [as a juror].”
Adams v. Tex.,
448 U.S. 38, 48, 100 S.Ct. 2521, 65 L.Ed.2d 581 (1980);
see also Gray v. Miss.,
481 U.S. 648, 652 n. 3, 107 S.Ct. 2045, 95 L.Ed.2d 622 (1987) (“[a] motion to excuse a venire member for cause of course must be supported by specified causes or reasons that demonstrate, as a matter of law, the venire member is not qualified to serve”). Indeed, the record supports but a single conclusion— that Hartling was removed merely because she expressed opposition to the death penalty. This violated Stevens’s right to a fair and impartial jury.
Witherspoon,
391 U.S. at 522, 88 S.Ct. 1770.
In rejecting Stevens’s
Witherspoon
claim, the state courts relied on the trial judge’s statement that, although he could not remember the interchange with Hartling, he must have been relying on something emphatic in her demeanor.
Stevens,
739 A.2d at 521. A trial judge is certainly entitled to rely on a potential juror’s demeanor.
See, e.g., Witt,
469 U.S. at 434, 105 S.Ct. 844. However, the dispositive question under
Witherspoon
and its progeny is “not whether a reviewing court might disagree with the trial court’s findings, but whether those findings are fairly supported by the record.”
Id.
As we indicated in
Szuchon,
the trial judge — when questioning a potential juror — has an obligation “to make a record of [any] bias” shown by the juror. 273 F.3d at 328 (citing
Gray).
By relying on supposition that there must have been something “emphatic” about Hartling’s demeanor, the state courts failed to heed the Supreme Court’s rule that findings of bias must be “fairly supported by the record.”
Witt,
469 U.S. at 434, 105 S.Ct. 844.
For that matter, even if we accepted the trial judge’s supposition that he relied on something emphatic in Hartling’s demean- or, the result would be no different. Even
if
Hartling
emphatically
stated that she did not “believe in” the death penalty, all that could reasonably be inferred is that her moral opposition to capital punishment was strongly felt. Supreme Court precedent makes clear, however, that the true question was whether Hartling would have been able to set aside her “conscientious or religious scruples against” the death penalty,
Witherspoon,
391 U.S. at 522, 88 S.Ct. 1770, and “faithfully and impartially apply the law.”
Witt,
469 U.S. at 426, 105 S.Ct. 844. That Hartling’s “conscientious or religious scruples” were strongly felt says remarkably little about whether she could follow the trial court’s instructions.
See Lockhart v. McCree,
476 U.S. 162, 176, 106 S.Ct. 1758, 90 L.Ed.2d 137 (1986) (“[i]t is important to remember that not all who oppose the death penalty are subject to removal for cause in capital cases; those who firmly believe that the death penalty is unjust may nevertheless serve as jurors in capital cases so long as they state clearly that they are willing to temporarily set aside their own beliefs in deference to the rule of law”);
see also Szuchon,
273 F.3d at 331 (holding that the fact that a potential juror did not “believe” in the death penalty was “by no means the equivalent of being unwilling to impose it”). The state courts had no basis for concluding that Hartling’s “views would [have] prevented] or substantially impair[ed] the performance of [her] duties as a juror in accordance with [her] instructions and [her] oath.”
Witt,
469 U.S. at 424, 105 S.Ct. 844.
In rejecting Stevens’s
Witherspoon
claim, the Pennsylvania Supreme Court relied on the fact that the trial judge asked “most other potential jurors” whether they “would be able to set aside ... personal opinion[s] of the death penalty and follow the instructions of the court.”
Stevens,
739 A.2d at 521. The Commonwealth does not press this argument in its brief to us. For that matter, the parties have not submitted a complete transcript of the voir dire. Our review of the partial record of voir dire provided to the District Court indicates, however, that only a few potential jurors were actually similarly situated to Hartling. Of those few potential jurors, perhaps seven, who responded to the judge’s opening question about the death penalty with an answer suggesting opposition to the death penalty, the trial judge failed twice — in the case of both Hartling and a prospective juror named James L. Perrott
— to conduct any colloquy whatsoever to determine whether the venire members would be able to set aside their moral difficulties with capital punishment. The fact that the trial judge conducted a proper colloquy in the remaining handful of similar instances, then, is of little persuasive value. The remaining instances do not comprise a significant number — or, for that matter, a large enough percentage of the similar situations. Under these circumstances, we cannot easily attribute the absence of the proper colloquy in Hartling’s case to something unremembered about her demeanor. Indeed, it may be noteworthy that Hartling was the first potential juror, and the only juror to do so on the first day of voir dire, who told the trial judge that she had moral qualms about the death penalty. This may suggest, if anything, that the trial judge’s failure to conduct proper follow-up questioning was an oversight, although one of constitutional proportions. The trial judge may simply have not yet well-thought out how he would handle a potential juror who expressed moral opposition to capital punishment.
In its brief, the Commonwealth focuses our attention on attorney Lipecky’s failure to object to Hartling’s removal. According to this argument, counsel’s failure to object suggests that Hartling was so obviously adamant about her opposition to the death penalty that everyone — prosecution, defense, and trial judge — could see there was no point in asking her any further questions.
See id.
In support, the Commonwealth cites this passage from
Witt:
Thus, whatever ambiguity [may be found] in this record, we think that the trial court, aided as it undoubtedly was by its assessment of [the prospective juror’s] demeanor, was entitled to resolve it in favor of the State. We note in addition that respondent’s counsel chose not to question [the prospective juror] himself, or to object to the trial court’s excusing her for cause. This questioning might have resolved any perceived ambiguities in the questions; its absence is all the more conspicuous because counsel did object to the trial court’s excusing other veniremen later on during the voir dire. Indeed, from what appears on the record it seems
that at the time [the prospective juror] was excused no one in the courtroom questioned the fact that her beliefs prevented her from sitting. The reasons for this, although not crystal clear from the printed record, may well have been readily apparent to those viewing [the prospective juror] as she answered the questions.
Witt,
469 U.S. at 434-35, 105 S.Ct. 844. In the Commonwealth’s view, the fact that Lipecky did not object should similarly be read as support for the trial judge’s supposition that he must have relied on something in Hartling’s demeanor. We cannot agree.
Witt
and Stevens’s case are quite different. First and foremost, in
Witt,
the trial judge conducted some follow-up questioning of the prospective juror who indicated that she had personal reservations about the death penalty.
Id.
at 415-16, 105 S.Ct. 844. Indeed, the trial judge went on to specifically ask the prospective juror
both
whether her views would “interfere with [her] sitting as a juror”
and
whether those views would impair her ability to “judg[e] the guilt or innocence of the Defendant.”
Id.
at 416, 105 S.Ct. 844. Remarkably, she answered yes to both questions.
Id.
Accordingly, when the Supreme Court wrote in
Witt
that the trial judge’s assessment of the potential juror’s demeanor could be used to resolve “ambiguity” in the record, the Court was referring to possible ambiguity in the juror’s statements about whether she would be able, notwithstanding her views on the death penalty, to follow the trial court’s instructions. In other words, the trial judge was entitled to use the potential juror’s demeanor to resolve the relevant question — whether she could follow the law.
Here, by contrast, the trial judge never followed up on Hartling’s
initial
statement that she opposed capital punishment. Under
Witherspoon
and its progeny, including
Witt,
that was error.
See Witherspoon,
391 U.S. at 522, 88 S.Ct. 1770;
Witt,
469 U.S. at 422, 105 S.Ct. 844 (indicating that the key question is whether the venire member would “refuse[] to follow the statutory scheme” controlling when the death penalty should be given). As noted above, the fact that no follow-up questioning was utilized means that the trial judge was able to compile no record concerning whether Hartling was fatally biased against imposition of the death penalty.
See Szuchon,
273 F.3d at 328;
see also Witt,
469 U.S. at 423, 105 S.Ct. 844 (noting that it is the adversary seeking exclusion [i.e., here, the Commonwealth] who “must demonstrate ... that the potential juror lacks impartiality”). Moreover, and significantly for present purposes, the fact that no follow-up questioning occurred also means that no ambiguity existed in Hartling’s answers as to the relevant question — whether she could follow the trial judge’s instructions. After all, as noted, even if Hartling emphatically stated that she was opposed to capital punishment, whether she could nevertheless follow the law posed a separate question.
See Lockhart,
476 U.S. at 176, 106 S.Ct. 1758. Accordingly, because Hartling was never asked about her willingness and ability to follow the law, the trial judge in Stevens’s case — unlike the trial judge in
Witt
— was not in a position to evaluate and factor in her demeanor when answering the relevant question. The quotation from Witt is not a satisfactory rebuttal to Stevens’s
Witherspoon
claim.
Because Hartling was removed from the venire on the sole basis of her statement that she opposed capital punishment, Stevens was denied his constitutional right to a fair and impartial jury.
Witherspoon,
391 U.S. at 522, 88 S.Ct. 1770. The trial judge did not have enough information be
fore him to conclude that she “would be unable to faithfully and impartially apply the law.”
Witt,
469 U.S. at 426, 105 S.Ct. 844. The state courts’ finding of bias was not supported by anything more than speculation and, accordingly, Stevens met his burden of showing the constitutional violation by the requisite clear and convincing evidence. 28 U.S.C. § 2254(e)(1).
IV.
For the reasons detañed, the District Court properly refused to grant habeas relief on Stevens’s claim that his attorney rendered ineffective assistance at the guüt phase (C.A. No. 04-9013). The District Court’s correctly awarded relief on Stevens’s
Witherspoon
claim, however (C.A. No. 04-9011). We wñl affirm the District Court’s judgment in all respects.