dissenting.
Unpersuaded by petitioner’s claims, the state trial court found that “there was no purposeful discrimination by the prosecution] in the use of... peremptory strikes,” App. 878. This finding established that petitioner had failed to carry his burden at step three of the inquiry set out in Batson v. Kentucky, 476 U. S. 79 (1986). Title 28 U. S. C. § 2254(e)(1) requires that a federal habeas court “presum[e]” the state court’s findings of fact “to be correct” unless petitioner can rebut the presumption “by clear and convincing evidence.” The majority decides, without explanation, to ignore §2254(e)(1)’s explicit command. I cannot. Because petitioner has not shown, by clear and convincing evidence, that any peremptory strikes of black veniremen were exercised because of race, he does not merit a certificate of appealability (COA). I respectfully dissent.
I — I
A
The Court agrees, ante, at 342, that the state court’s finding at step three of Batson is a finding of fact ordinarily subject to § 2254(e)(1)’s presumption of correctness:
“In a proceeding instituted by an application for a writ of habeas corpus by a person in custody pursuant to the judgment of a State court, a determination of a factual *355issue made by a State court shall be presumed to be correct. The applicant shall have the burden of rebutting the presumption of correctness by clear and convincing evidence.”
However, the Court implicitly rejects the obvious conclusion that the COA determination under § 2253(c) is part of a “proceeding instituted by an application for a writ of habeas corpus.” Instead of presuming the state court’s factfindings to be correct, as § 2254(e)(1) requires, the Court holds that petitioner need only show that reasonable jurists could disagree as to whether he can provide clear and convincing evidence that the finding was erroneous. Ante, at 341.
The Court’s main justification for this conclusion is supposed fidelity to Slack v. McDaniel, 529 U. S. 473 (2000). See ante, at 338 (“‘The petitioner must demonstrate that reasonable jurists would find the district court’s assessment of the constitutional claims debatable or wrong’” (quoting Slack, supra, at 484)). But neither Slack nor any other decision of this Court addressing the COA procedure has ever considered a “constitutional claim” that turns entirely on issues of fact. In these circumstances, it is the text of § 2254(e)(1) that governs.
Unlike the majority, I begin with the plain text of the statute that instructs federal courts how to treat state-court findings of fact. At issue is what constitutes a “proceeding” for purposes of § 2254(e)(1). The word, “proceeding,” means “[t]he regular and orderly progression of a lawsuit, including all acts and events between the time of commencement and the entry of judgment.” Black’s Law Dictionary 1221 (7th ed. 1999) (emphasis added). The COA, “standing alone,. .. does not assert a grievance against anyone, does not seek remedy or redress for any legal injury, and does not even require a ‘party’ on the other side. It is nothing more than a request for permission to seek review.” Hohn v. United States, 524 U. S. 236, 256 (1998) (Scalia, J., dissenting).
*356I agree with the majority that the existence of a COA is a jurisdictional prerequisite to the merits appeal. Ante, at 336. However, the Court takes a wrong turn when it implies that the merits appeal is part of the habeas process (or “proceeding”) but the COA determination somehow is not. Overwhelming authority (including the majority opinion) confirms that § 2254(e)(1) applies to the merits appeal. See ante, at 342; Weaver v. Bowersox, 241 F. 3d 1024, 1030 (CA8 2001); Putman v. Head, 268 F. 3d 1223, 1241 (CA11 2001); Johnson v. Gibson, 254 F. 3d 1155, 1160 (CA10 2001); Francis S. v. Stone, 221 F. 3d 100, 114-115 (CA2 2000); Weeks v. Snyder, 219 F. 3d 245, 258 (CA3 2000); Mueller v. Angelone, 181 F. 3d 557, 575 (CA4 1999); Ashford v. Gilmore, 167 F. 3d 1130, 1131 (CA7 1999); cf. Sumner v. Mata, 449 U. S. 539, 546-547 (1981) (pre-Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA) factual deference provision with virtually identical language applies to merits appeal). The COA determination should be treated no differently, because § 2254(e)(1) draws no distinction between the merits appeal and the COA. The Court’s silent conclusion to the contrary is simply illogical. The COA’s status as the jurisdictional prerequisite for the merits appeal requires that both the COA determination and the merits appeal be considered a part of the same “proceeding.”
The Court’s rejection of this conclusion also conflicts with pre-AEDPA practice. Prior to AEDPA, access to a merits appeal in federal habeas corpus proceedings was governed by a mechanism similar to the COA, known as a certificate of probable cause, or CPC. See Slack, supra, at 480. There was also a standard of factual deference similar to, though weaker than, the standard in § 2254(e)(1). See 28 U. S. C. § 2254(d) (1994 ed.).1 Under these provisions (indis*357tinguishable from AEDPA’s for these purposes), courts concluded that §2254(e)(1)’s predecessor applied directly to the CPC proceeding, without any filtering through the “debatability” standard the Court has used in both the CPC and COA contexts. See, e. g., Barnard v. Collins, 13 F. 3d 871, 876-877 (CA5 1994); Cordova v. Collins, 953 F. 2d 167, 169 (CA5 1992). These cases support the straightforward notion that § 2254(e)(1), like its predecessor did with respect to CPC proceedings, applies directly to the COA proceeding.
The Court’s decision in Hohn, supra, which holds that the COA determination constitutes a “case” in the court of appeals for purposes of this Court’s jurisdiction under 28 U. S. C. § 1254, is not to the contrary. Hohn does not hold, nor does its logic require, that the COA determination be regarded as separate from the rest of the habeas proceeding. In fact, Hohn rejected the proposition that “a request to proceed before a court of appeals should be regarded as a threshold inquiry separate from the merits ....” 524 U. S., at 246 (emphasis added). Indeed, Hohn analogized the COA to the filing of a notice of appeal, id., at 247, which in the civil context all would consider to be part of the same “proceeding” (“instituted by” a complaint) as the trial and merits appeal.
B
The Court also errs, albeit in dicta, when it implies that delayed state factfinding — here the two years between voir dire and the post-trial Batson hearing2 — is an excuse for *358weakened factual deference. Ante, at 342-343. Even putting aside the fact that an appellate court on direct review should (and would) still give heavy deference to 2-year-old credibility findings,3 this reasoning is in tension with the plain text of § 2254(e)(1) and ignores changes wrought by AEDPA to the role of federal courts on collateral review.
Unlike an appellate court’s review of district court findings of fact for clear error, § 2254(e)(1) establishes a presumption of correctness. It requires that the federal habeas court assume the state court that entered the findings was the best placed factfinder with the most complete record and only then ask whether the pétitioner can refute that factual finding by clear and convincing evidence. Procedural imperfections ordinarily will not affect this presumption; thus, it does not matter whether the state judge made his decision two years late or with a less-than-perfect record. Admittedly these conditions might increase the odds that a habeas applicant could locate helpful evidence, but to “presume” facts “correct” means a court cannot allow a habeas applicant to evade § 2254(e)(1) by attacking the process employed by the state factfinder rather than the actual factfindings.
This reading is confirmed by the changes worked by AEDPA. Section 2254(e)(1) does not, as its predecessor did, create exceptions to factual deference for procedural infirmities. For example, prior to AEDPA, a federal habeas court *359would not defer to state-court determinations of fact if “the factfinding procedure employed by the State court was not adequate to afford a full and fair hearing,” 28 U. S. C. § 2254(d)(2) (1994 ed.), “the material facts were not adequately developed at the State court hearing,” § 2254(d)(3), or “the applicant did not receive a full, fair, and adequate hearing,” § 2254(d)(6). The removal of these exceptions forecloses the use of marginal procedural complaints — such as a delay between voir dire and a Batson hearing — to determine whether or “how much” a federal habeas court will defer to state-court factfinding.
Section 2254(e)(1) simply cannot be read to contain an implied sliding scale of deference. I do not understand the Court to disagree with this view, however, as its dicta does not actually purport to interpret the text of § 2254(e)(1).4
II
Because § 2254(e)(1) supplies the governing legal standard, petitioner must provide “clear and convincing” evidence of purposeful discrimination in order to obtain a COA. Petitioner’s constitutional claim under Batson turns on this fact and “reasonable jurists could debate,” ante, at 336 (internal *360quotation marks omitted), whether a Batson violation occurred only if petitioner first meets his burden under § 2254(e)(1). And the simple truth is that petitioner has not presented anything remotely resembling “clear and convincing” evidence of purposeful discrimination.
A
The evidence amassed by petitioner can be grouped into four categories: (1) evidence of historical discrimination by the Dallas District Attorney’s office in the selection of juries; (2) the use of the “jury shuffle” tactic by the prosecution; (3) the alleged similarity between white veniremen who were not struck by the prosecution and six blacks who were: Edwin Rand, Wayman Kennedy, Roderick Bozeman, Billy Jean Fields, Joe Warren, and Carroll Boggess; and (4) evidence of so-called disparate questioning with respect to veniremen’s views on the death penalty and their ability to impose the minimum punishment.
The “historical” evidence is entirely circumstantial, so much so that the majority can only bring itself to say it “casts doubt on the State’s claim that [discriminatory] practices had been discontinued before petitioner’s trial.” Ante, at 335. And the evidence that the prosecution used jury shuffles no more proves intentional discrimination than it forces petitioner to admit that he sought to eliminate whites from the jury, given that he employed the tactic even more than the prosecution did.5 Ultimately, these two categories of evidence do very little for petitioner, because they do not address the genuineness of prosecutors’ proffered race-neutral reasons for making the peremptory strikes of these particular jurors.
In short, the reasons that Justice Scalia finds this to be a “close case,” ante, at 348 (concurring opinion), are reasons that, under the correct reading of § 2254(e)(1), it is a losing *361case. I write further to explore two arguments advanced by petitioner that the Court deemed helpful in establishing petitioner’s “debatable” entitlement to relief, apparently because the majority’s “debatability” inquiry requires a less-thorough review of the record and a more permissive attitude toward a COA movant’s representations.
B
As noted, petitioner argues the prosecution struck six blacks — Rand, Kennedy, Bozeman, Fields, Warren, and Boggess — who were similarly situated to unstruck whites. I see no need to repeat Justice Scalia’s dissection of petitioner’s tales of white veniremen as ambivalent about the death penalty as Kennedy, Bozeman, Warren, and Boggess. Ante, at 350-353 (concurring opinion). However, the majority’s cursory remark that “three of the State’s proffered race-neutral rationales for striking [black] jurors pertained just as well to some white jurors who were not challenged and who did serve on the jury,” ante, at 343 (emphasis added), is flatly incorrect and deserves some discussion.
For the three challenged peremptory strikes used on Fields, Warren, and Boggess, petitioner has not even correctly alleged the existence of “similarly situated” white veniremen. The majority’s discussion of this subject is misleading, stating that “prosecutors explained that their peremptory challenges against six [black] potential jurors were based on ambivalence about the death penalty; hesitancy to vote to execute defendants capable of being rehabilitated; and the [veniremens’] own family history of criminality.” Ibid. The implication is that for each of the six challenged veniremen, the prosecution gave all three reasons as justifications for the use of a peremptory strike. To clarify: Rand, Kennedy, Bozeman, Warren, and Boggess were struck for ambivalence about the death penalty. Fields, Warren, and Boggess were struck for having family mem*362bers with criminal histories. Bozeman and Fields were struck for making prodefense remarks about rehabilitation.
Simple deduction, and an analysis of petitioner’s contentions that includes the names of these allegedly similar white veniremen, cf. ibid., reveals that petitioner has unearthed no white venireman who, like Warren and Boggess, was both ambivalent about the death penalty and related to individuals who had previous brushes with the law.6 Petitioner also produces no white venireman who, like Fields, expressed prodefense views on rehabilitation and had a family member with a criminal history.7 “Similarly situated” *363does not mean matching any one of several reasons the prosecution gave for striking a potential juror — it means matching all of them.
This leaves Rand, Kennedy, and Bozeman.8 Petitioner alleges that white jurors Hearn and Mazza were as ambivalent about the death penalty as these three struck black veniremen. Justice Scalia has adequately demonstrated that this is absurd with respect to Kennedy and Bozeman, but I agree that petitioner makes a slightly better case with Rand. Ante, at 352-353 (concurring opinion). However, since the burden is on petitioner to show, by clear and convincing evidence, that Rand was struck because of his race, I find this sliver of evidence, even when combined with petitioner’s circumstantial evidence, insufficient to rebut §2254(e)(1)’s presumption.
C
Petitioner's accounts of "disparate questioning” also amount to little of substance. Petitioner argues that the prosecution posed different questions at voir dire depending on the race of the venireman on two subjects: the death penalty and the minimum punishment allowed under law. Neither accusation can withstand a careful examination of the full record or help petitioner assemble the requisite clear and convincing evidence.
*3641
Respondent counters petitioner’s complaints about the so-called “graphic formulation” or “script” by arguing that this depiction was used only with those potential jurors who “expressed reservations about the death penalty in their juror questionnaires.” Brief for Respondent 17. The majority discounts this explanation, stating that “[t]his cannot be accepted without further inquiry.” Ante, at 344. Under my view, however, petitioner bears the burden of showing purposeful discrimination by clear and convincing evidence.
The Court’s treatment of this issue focuses on the apparent disparity in treatment of 10 black veniremen and 10 white veniremen who were supposedly similar in their opposition to the death penalty. The majority notes that only 2 out of these 10 whites got the graphic description while 7 out of 10 blacks did. Ante, at 344-345. But the Court neglects to mention that the eight white veniremen who petitioner thinks should have received the graphic formulation, Reply Brief for Petitioner 15, n. 19, were so emphatically opposed to the death penalty that such a description would have served no purpose in clarifying their position on the issue. No trial lawyer would willingly antagonize a potential juror ardently opposed to the death penalty with an extreme portrait of its implementation. The strategy pursued by the prosecution makes perfect sense: When it was necessary to draw out a venireman’s feelings about the death penalty they would use the graphic script, but when it was overkill they would not.
The record demonstrates that six of these eight white veniremen were so opposed to the death penalty that they were stricken for cause without the need for the prosecution to spend a peremptory challenge. For example, John Nelson wrote on his questionnaire, “I believe that the State does not have the right to take anyone’s life,” Tr. of Voir Dire in No. F85-78668-NL (5th Crim. Dist. Ct., Dallas County, Tex.), p. 625 (hereinafter VDR) (internal quotation marks omitted), and testified flatly, “I would not be able to vote for the death *365penalty.”9 Id., at 614. Nelson was struck for cause. Id., at 662-663. Linda Berk was “always” opposed to the death penalty, id., at 1449, and felt so strongly on the subject that the prosecutor remarked upon her discomfort, after which she stated, “[y]ou’re going to have to excuse me because I’m getting a little emotional, okay?” Id., at 1445. Later, after she had begun crying, Berk was struck for cause. Id., at 1478. Gene Hinson stated curtly, “I put on the form there that I didn’t agree with it,” id., at 1648, and was struck for cause. Sheila White said, “I have always been against . . . the death penalty,” id., at 2056, and was struck for cause.
Even those two not struek for cause had firm views. Margaret Gibson said: “I don’t believe in the death penalty. I don’t know why it was started. I don’t think it solves anything,” id., at 485, and was struck by the prosecution with a peremptory strike. And James Holtz thought the death penalty appropriate only if a policeman or fireman was murdered. Id., at 1021. I can apprehend simply no reason to fault the prosecution for failing to give a more graphic description of lethal injection to prospective jurors with such firm views against capital punishment.
I recognize that these voir dire statements only indirectly support respondent’s explanation because the graphic script was typically given at the outset of voir dire — before the above quoted veniremen had the chance to give their stark answers. Nevertheless, all available evidence supports respondent’s view that those who were unclear in their views on the death penalty in their juror questionnaires received the graphic formulation — and that those who were adamantly for or against the death penalty in their questionnaires did not.
The jury forms at issue asked two questions directly relevant to the death penalty. Question 56 asked, “Do you believe in the death penalty?,” offered potential jurors the *366chance to circle “yes” or “no,” and then asked them to “[pllease explain your answer.” See, e. g., Joint Lodging 44 (Boggess questionnaire). Question 58 allowed potential jurors to circle “yes” or “no” in answering the following question: “Do you have any moral, religious, or personal beliefs that would prevent you from returning a verdict which would ultimately result in the execution of another human being?” Ibid.
First, as already noted, the deeper and clearer opposition to the death penalty on the part of the eight whites who did not receive the graphic script (but petitioner thinks should have) indirectly supports respondent’s contention that this opposition came out in their questionnaires (presumably by an answer of “no” to question 56 and an answer of “yes” to question 58). But this is not the only evidence supporting respondent’s view. Hinson, a white venireman who did not receive the graphic formulation, stated during voir dire that he “put on the form there that [he] didn’t agree with [the death penalty] for both moral and religious reasons.” VDR 1648. Similarly, Nelson, a white venireman not receiving the graphic formulation, stated on his questionnaire, “I believe that the State does not have the right to take anyone’s life.” Id., at 625 (internal quotation marks omitted). Fernando Gutierrez, a juror who received the graphic formulation, answered “yes” to question 56, but also “yes” to question 58, indicating he had “moral, religious, or personal beliefs” that would obstruct his voting for the death penalty despite the fact that he believed in it. Joint Lodging 205.
The prosecution treated the black veniremen no differently. The blacks who did not receive the graphic formulation (whose questionnaires are contained in the record) all answered “yes” to question 56, stating they believed in the death penalty, and “no” to question 58, indicating that their beliefs wouldn’t prevent them from imposing a death sentence. See id., at 12 (Bozeman), 20 (Fields), 28 (Warren), 36 *367(Rand). The black veniremen who were given the graphic formulation, by contrast, gave ambiguous answers on their juror questionnaires expressing hesitation, rather than philosophical opposition, to the death penalty. Boggess answered “yes” to question 56 but also “yes” to question 58. Id., at 44. Kennedy answered “yes” to question 56 but indicated that he believed in the death penalty “[o]nly in extreme cases, such as multiple murders.” Id., at 51. Troy Woods answered “no” to question 56, but also “no” to question 58, indicating he did not believe in the death penalty but would have no personal objection to imposing it. Id., at 180. He wrote “that [sic] not punishment,” in the space provided for question 56. Ibid. It happened that, while not completely clear about it in the questionnaire (and hence receiving the graphic formulation), Woods was an enthusiastic supporter of the death penalty, and he was, in fact, seated on petitioner’s jury. Further confirming respondent’s explanation, black veniremen Linda Baker, Janice Mackey, Paul Bailey, and Anna Keaton all gave unclear responses to questions 56 and 58 and all received the graphic formulation. See Tr. of Pretrial Hearings in No. F85-78660-NL (5th Crim. Dist. Ct., Dallas County, Tex.) (Def. Exh. 7).10
*368To sum up, the correlation between questionnaire answers and the use of the graphic script is far stronger than any correlation with race. Sixteen veniremen clearly indicated on the questionnaires their feelings on the death penalty,11 and 15 of them did not receive the graphic script.12 Eight veniremen gave unclear answers and those eight veniremen got the graphic script.13 In other words, for 23 out of 24, or 96%, of the veniremen for whom questionnaire information is available, the answers given accurately predict whether they got the graphic script.14 Petitioner’s theory that race determined whether a venireman got the graphic script produces a race-to-script correlation of only 74% — far worse.15
2
Petitioner fares no better with his allegation that the prosecution employed two different scripts on the basis of *369race when asking questions about imposition of the minimum sentence. Indeed, this disparate questioning argument is as flawed as the last one. Respondent admits that the different questioning on minimum sentences was used as an effort to get veniremen the prosecution felt to be ambivalent about the death penalty dismissed for cause. In making the decision whether to employ the “manipulative” minimum punishment script, prosecutors could rely on both the questionnaires and substantial voir dire testimony, as the minimum punishment questioning occurred much later in voir dire than the graphic formulation.
Seven black veniremen were given the allegedly “manipulative” minimum punishment script, all of whom were opposed to the death penalty in varying degrees. Rand, Kennedy, Bozeman, Warren, and Boggess’ views on the death penalty have all been exhaustively discussed. This leaves Baker and Fields. Baker’s views on the death penalty were so clearly ambivalent that she is not even the subject of petitioner’s Batson challenge. And Fields’ family history of criminality and views on rehabilitation, as earlier discussed, supra, at 362, and n. 7, convinced the prosecution to use a peremptory strike.16 Finally, petitioner’s objection to the prosecution’s decision not to use the “manipulative” punishment script on Woods, Reply Brief for Petitioner 17, n. 23, makes no sense. Woods gave answers indicating he would be an excellent State’s juror — why would the prosecution have tried to eliminate him? Of course, if petitioner were correct that the prosecution sought to eliminate blacks then one might expect that all methods, including the use of the “manipulative” script, would have been deployed against Woods, who happened to also be black.
As with graphic questioning, respondent’s explanation goes unrebutted by petitioner. Unless a venireman indicated he would be a poor State’s juror (using the criteria that *370respondent has identified here) and would not otherwise be struck for cause or by agreement, there was no reason to use the “manipulative” script. Thus, when petitioner points to the “State’s failure to use its manipulative method with the vast majority of white veniremembers who expressed reservations about the death penalty,” ibid., he ignores the fact that of the 10 whites who expressed opposition to the death penalty, 8 were struck for cause or by agreement, meaning no “manipulative” script was necessary to get them removed. The other two whites were both given the “manipulative” script and peremptorily struck,17 just like Rand, Kennedy, Bozeman, Fields, Warren, Boggess, and Baker.
* * *
Quite simply, petitioner’s arguments rest on circumstantial evidence and speculation that does not hold up to a thorough review of the record. Far from rebutting § 2254(e)(1)’s presumption, petitioner has perhaps not even demonstrated that reasonable jurists could debate whether he has provided the requisite evidence of purposeful discrimination — but that is the majority’s inquiry, not mine. Because petitioner has not demonstrated by clear and convincing evidence that even one of the peremptory strikes at issue was the result of racial discrimination, I would affirm the denial of a COA.
The pre-AEDPA standard of factual deference provided:
“In any proceeding instituted in a Federal court by an application for a writ of habeas corpus by a person in custody pursuant to the judgment of a State court, a determination after a hearing on the merits of a factual *357issue, made by a State court of competent jurisdiction . . . shall be presumed to be correct, unless the applicant shall establish or it shall otherwise appear, or the respondent shall admit [enumerated exceptions omitted]. . . . And in an evidentiary hearing ... the burden shall rest upon the applicant to establish by convincing evidence that the factual determination by the State court was erroneous.”
Not all the factfinding was so hindered. Prosecutors gave reasons for 2 of the 10 strikes of black veniremen at the post-trial Batson hearing. One of those, Joe Warren, is at issue here. App. 856-860.
I am puzzled by the majority’s willingness to hold against respondent the failure of prosecutors to testify at the post-trial Batson hearing. Petitioner could easily have requested that the reasons for the allegedly unconstitutional peremptory strikes be given again, and did not. The attorney representing the State at the post-trial Batson hearing made certain that both trial prosecutors were present to reiterate the reasons they gave in the record for striking the challenged black veniremen. App. 865. Petitioner’s counsel explicitly refused the opportunity to do so when it was offered. Ibid. Furthermore, I fail to understand why a move that resulted in a more efficient hearing without redundant testimony should redound to the benefit of petitioner, who bears the burden of proof in this federal habeas corpus proceeding.
I do, however, agree with the majority that the Court’s decisions in Hernandez v. New York, 500 U. S. 352 (1991), and Purkett v. Elem, 514 U. S. 765 (1995) (per curiam), can be helpful in guiding a federal habeas court deciding a claim under Batson v. Kentucky, 476 U. S. 79 (1986). For instance, both cases confirm that Batson step three turns on an evaluation of the prosecutor’s proffered race-neutral justifications for the peremptory challenges at issue. Purkett, supra, at 768-769; Hernandez, 500 U. S., at 364-365 (plurality opinion); id., at 372 (O’Connor, J., concurring in judgment); see also Batson, supra, at 98, n. 21. Additionally, because Hernandez's, dear-error standard is less demanding of a criminal defendant than § 2254(e)(1) is of a habeas applicant, a federal habeas court can deny relief on § 2254(e)(1) grounds if it determines it would do so when reviewing the same facts for dear error. Cf. Marshall v. Lonberger, 459 U. S. 422, 434-435 (1983) (“We greatly doubt that Congress . . . intended to authorize broader federal review of state court credibility determinations than are authorized in appeals within the federal system itself”).
Petitioner shuffled the jury five times; the prosecution did so only three times. Brief for Respondent 21.
Petitioner directs the Court to white veniremen Noad Vickery, Cheryl Davis, Chatta Nix, and Joan Weiner as having family members with criminal histories, but points to white veniremen Sandra Hearn and Marie Mazza as equally ambivalent about the death penalty. Brief for Petitioner 22. Of course, as Justice Scalia demonstrates, Hearn and Mazza were not ambivalent about the death penalty. Ante, at 351-352 (concurring opinion).
Again petitioner points to Vickery, Davis, Nix, and Weiner for similar family histories. Justice Scalia has shown that none of these four were in fact similarly situated to Fields with respect to this justification. Ante, at 353-354 (concurring opinion). Petitioner also alleges that Hearn made prodefense remarks about rehabilitation similar to those made by Felds. Again, no white venireman even allegedly fits both reasons given for striking Felds. Furthermore, even if Felds had only been struck for his views on rehabilitation, those views were in no way equivalent to those expressed by Hearn. Felds answered “yes” to the question whether he believed that “everyone can be rehabilitated.” App. 118. Felds went on to say that “[i]t may be far-fetched, but I feel like, if a person has the opportunity to really be talked about God and he commits himself, whereas he has committed this offense, then if he turns his life around, that is rehabilitation.” Ibid. In contrast, Hearn stated that she “believe[d] in the death penalty if a criminal cannot be rehabilitated.” Id., at 694.
Petitioner tries to muddy the waters by pointing out that Felds was, in other respects, a good State’s juror because he supported the death penalty. Brief for Petitioner 24-25. However, that does not change the fact that Felds said that everyone could be rehabilitated (and thus might have been swayed by a penitent defendant’s testimony) and Hearn insisted that *363some people could not be rehabilitated. In analyzing Batson claims the focus should not be on the “reasonableness of the asserted nonracial motive . . . [but] rather [on] the genuineness of the motive.” Purkett, 514 U. S., at 769 (emphasis in original).
The prosecution’s stated reasons for striking Bozeman were that he was ambivalent about the death penalty and that he made prodefense remarks about rehabilitation. This is one case where the prosecution gave multiple reasons for a strike and petitioner actually correctly alleged the existence of a similarly situated white venireman, Hearn. Petitioner believes, albeit erroneously, see ante, at 351-353 (Scalia, J., concurring), that Hearn expressed similar ambivalence about the death penalty and made prodefense remarks about rehabilitation.
Nelson was also a doctor and presumably did not need to have the lethal injection process described to him.
Questions 56 and 58, and the responses thereto, are found on page 6 of each questionnaire. Baker did not circle “yes” or “no” in answering question 56, but wrote “[m]y strongest feeling is against the death penalty; however, being aware of the overcrowding in jails and the number of murders],] I would have to know the facts to make a decision . . . .” (Emphasis added.) Baker also did not answer question 58, writing “undecided” instead. Mackey answered question 56 “no,” indicating she did not believe in the death penalty, and wrote “Thou Shall Not Kill” in the explanation space. She then proceeded to answer question 58 “no” as welL Bailey circled “yes” in answering question 56, but wrote in “NO” with a circle around it, along with such explanations as “yes for a major crime” and “[n]o one have [sic] the right to take anothe [sic] ones [sic] life.” (Emphases in original.) He then circled “no” in answering question 58. Keaton circled “no,” indicating she did not believe in the death penalty, when she answered question 56, writing “It’s not for me to pun*368ished [sic] anyone.” However she then circled “no” in answering question 58, indicating that she did not have any objection to imposing the death penalty.
See VDR 1648 (Hinson), 625 (Nelson); Joint Lodging 12 (Bozeman), 20 (Melds), 28 (Warren), 36 (Rand), 125 (Mary Sumrow), 132 (Ronnie Long), 140 (Weiner), 148 (Mazza), 156 (Vivian Sztybel), 164 (Debra McDowell), 172 (Kevin Duke), 189 (Brenda Walsh), 197 (Filemon Zablan), 213 (Hearn).
Sztybel received the graphic script. VDR 2828.
Boggess, Kennedy, Baker, Mackey, Bailey, Keaton, Gutierrez, and Woods.
This analysis considers Hinson and Nelson as being clearly opposed to the death penalty in their questionnaires (answering question 56 “no” and question 58 “yes”) and Kennedy as being ambiguous (though in fact he answered question 56 “yes” and 58 “no”). Even without these assumptions, 13 out of 15 veniremen who answered “yes” to question 56 and “no” to question 58 — indicating clear support for the death penalty — did not receive the graphic script. And seven out of seven of those answering “no” and “no” or “yes” and “yes” — indicating ambiguous or mixed feelings about the death penalty — or not answering clearly at all received the graphic script. This yields an accuracy rate of 20 out of 22, or 91%.
For whites, 10 out of 12 did not get the graphic script. For blacks, 7 out of 11 did get the graphic script. This means race predicted use of the graphic script only 74% of the time.
The prosecution in fact used peremptory strikes on all seven of these black veniremen.
See Joint Lodging 110; VDR 502-511 (Gibson), 1046-1050 (Holtz).