REPRODUCED WITH PERMISSION OF WEST PUBLISHING CO.                                                106 S.Ct. 1758 90 L.Ed.2d 137, 54 USLW 4449

A.L. LOCKHART, Director, Arkansas Department of Correction, Petitioner  v. Ardia V. McCREE.

REHNQUIST, J., delivered the opinion of the Court, in which BURGER, C.J., and WHITE, POWELL, and O'CONNOR, JJ., joined. BLACKMUN, J., concurred in the result. MARSHALL, J., filed a dissenting opinion, in which BRENNAN and STEVENS,  JJ., joined, post, p. 1760.
**1760 John Steven Clark, Little Rock, Ark., for petitioner.
*164 Samuel R. Gross, Stanford, Cal., for respondent.

In this case we address the question left open by our decision nearly 18 years  ago in Witherspoon v. Illinois, 391 U.S. 510, 88 S.Ct. 1770, 20 L.Ed.2d 776  (1968): Does the Constitution prohibit the removal for cause, prior to the  guilt phase of a bifurcated capital trial, of prospective jurors whose  opposition to the death penalty is so strong that it would prevent or  substantially impair the performance of their duties as jurors at the  sentencing phase of the trial? See id., at 520, n. 18, 88 S.Ct., at 1776,  n. 18; Bumper v. North Carolina, 391 U.S. 543, 545, 88 S.Ct. 1788, 1790, 20  L.Ed.2d 797 (1968). We hold that it does not.

The District Court held a hearing on the "death qualification" issue in July  1981, receiving in evidence numerous social science studies concerning the  attitudes and beliefs of "Witherspoon -excludables," along with the  potential effects of excluding them from the jury prior to the guilt phase of a  bifurcated capital trial. In August 1983, the court concluded, based on the  social science evidence, that "death qualification" produced juries that "were  more prone to convict" capital defendants than "non-death-qualified" juries.  Grigsby v. Mabry, 569 F.Supp., at 1323. The court ruled *168 that "death  qualification" thus violated both the fair-cross-section and impartiality  requirements of the Sixth and Fourteenth Amendments, and granted McCree  habeas relief. Id., at 1324. [FN2]

The Eighth Circuit found "substantial evidentiary support" for the District  Court's conclusion that the removal for cause of "Witherspoon -excludables"  resulted in "conviction-prone" juries, and affirmed the grant of habeas relief  on the ground that such removal for cause violated McCree's constitutional  right to a jury selected from a fair cross section of the community.  Grigsby v. Mabry, 758 F.2d, at 229. The Eighth Circuit did not address  McCree's impartiality claim. Ibid. The Eighth Circuit left it up to the  discretion of the State "to construct a fair process" for future capital trials  that would comply with the Sixth Amendment. Id., at 242-243. Four judges  dissented. Id., at 243-251.

Before turning to the legal issues in the case, we are constrained to point  out what we believe to be several serious flaws in the evidence upon which the  courts below reached the conclusion that "death qualification" produces  "conviction-prone" juries. [FN3] McCree introduced into evidence *169 some  15 social science studies in support of his constitutional claims, but only 6  of the studies even purported to measure the potential effects on the guilt-  innocence determination of the removal from the jury of "Witherspoon -  excludables." [FN4] Eight of the remaining **1763 nine studies dealt solely  with generalized attitudes and beliefs about the death penalty and other  aspects of the criminal justice system, and were thus, at best, only marginally  relevant to the constitutionality of McCree's conviction. [FN5] The 15th and  final study *170 dealt with the effects on prospective jurors of voir dire  questioning about their attitudes toward the death penalty, [FN6] an issue  McCree raised in his brief to this Court but that counsel for McCree admitted  at oral argument would not, standing alone, give rise to a constitutional  violation. [FN7]

FN3. McCree argues that the "factual" findings of the District Court and the Eighth Circuit on the effects of "death qualification" may be reviewed by this Court only under the "clearly erroneous" standard of Federal Rule of Civil Procedure 52(a). Because we do not ultimately base our decision today on the invalidity of the lower courts' "factual" findings, we need not decide the "standard of review" issue. We are far from persuaded, however, that the "clearly erroneous" standard of Rule 52(a) applies to the kind of "legislative" facts at issue here. See generally Dunagin v. City of Oxford, Mississippi, 718 F.2d 738, 748, n. 8 (CA5 1983) (en banc) (plurality opinion of Reavley, J.). The difficulty with applying such a standard to "legislative" facts is evidenced here by the fact that at least one other Court of Appeals, reviewing the same social science studies as introduced by McCree, has reached a conclusion contrary to that of the Eighth Circuit. See Keeten v. Garrison, 742 F.2d 129, 133, n. 7 (CA4 1984) (disagreeing that studies show relationship between generalized attitudes and behavior as jurors), cert. pending, No. 84-6187.

FN4. The Court of Appeals described the following studies as "conviction- proneness surveys": H. Zeisel, Some Data on Juror Attitudes Toward Capital Punishment (University of Chicago Monograph 1968) (Zeisel); W. Wilson, Belief in Capital Punishment and Jury Performance (unpublished manuscript, University of Texas, 1964) (Wilson); Goldberg, Toward Expansion of Witherspoon: Capital Scruples, Jury Bias, and Use of Psychological Data to Raise Presumptions in the Law, 5 Harv.Civ. Rights-Civ.Lib.L.Rev. 53 (1970) (Goldberg); Jurow, New Data on the Effect of a "Death Qualified" Jury on the Guilt Determination Process, 84 Harv.L.Rev. 567 (1971) (Jurow); and Cowan, Thompson, & Ellsworth, The Effects of Death Qualification on Jurors' Predisposition to Convict and on the Quality of Deliberation, 8 Law & Hum.Behav. 53 (1984) (Cowan-Deliberation). In addition, McCree introduced evidence on this issue from a Harris Survey conducted in 1971. Louis Harris & Associates, Inc., Study No. 2016 (1971) (Harris-1971).

FN5. The Court of Appeals described the following studies as "attitudinal and demographic surveys": Bronson, On the Conviction Proneness and Representativeness of the Death-Qualified Jury: An Empirical Study of Colorado Veniremen, 42 U.Colo.L.Rev. 1 (1970); Bronson, Does the Exclusion of Scrupled Jurors in Capital Cases Make the Jury More Likely to Convict? Some Evidence from California, 3 Woodrow Wilson L.J. 11 (1980); Fitzgerald & Ellsworth, Due Process vs. Crime Control: Death Qualification and Jury Attitudes, 8 Law & Hum.Behav. 31 (1984); and Precision Research, Inc., Survey No. 1286 (1981). In addition, McCree introduced evidence on these issues from Thompson, Cowan, Ellsworth, & Harrington, Death Penalty Attitudes and Conviction Proneness, 8 Law & Hum.Behav. 95 (1984); Ellsworth, Bukaty, Cowan, & Thompson, The Death-Qualified Jury and the Defense of Insanity, 8 Law & Hum.Behav. 81 (1984); A. Young,  Arkansas Archival Study (unpublished, 1981); and various Harris, Gallup,  and National Opinion Research Center polls conducted between 1953 and 1981.

FN6. McCree introduced evidence on this issue from Haney, On the Selection  of Capital Juries: The Biasing Effects of the Death-Qualification  Process, 8 Law & Hum.Behav. 121 (1984).

FN7. We would in any event reject the argument that the very process of  questioning prospective jurors at voir dire about their views of the death  penalty violates the Constitution. McCree concedes that the State may  challenge for cause prospective jurors whose opposition to the death  penalty is so strong that it would prevent them from impartially  determining a capital defendant's guilt or innocence. Ipso facto, the  State must be given the opportunity to identify such prospective jurors by  questioning them at voir dire about their views of the death penalty.

Of the six studies introduced by McCree that at least purported to deal with  the central issue in this case, namely, the potential effects on the  determination of guilt or innocence of excluding "Witherspoon -excludables"  from the jury, three were also before this Court when it decided  Witherspoon. [FN8] There, this Court reviewed the studies and concluded:

FN8. The petitioner in Witherspoon cited the Wilson and Goldberg  studies, and a prepublication draft of the Zeisel study. 391 U.S., at  517, n. 10, 88 S.Ct., at 1774, n. 10; see n. 4, supra.

"The data adduced by the petitioner ... are too tentative and fragmentary to  establish that jurors not opposed to the death penalty tend to favor the  prosecution in the determination of guilt. We simply cannot conclude, either  on the basis of the record now before us or as a matter of judicial notice,  that the exclusion of jurors *171 opposed to capital punishment results in  an unrepresentative jury on the issue of guilt or substantially increases the  risk of conviction. In light of the presently available information, we are  not prepared to announce a per se constitutional rule requiring the reversal of  every conviction returned by a jury selected as this one was." 391 U.S., at  517-518, 88 S.Ct., at 1774-75 (footnote omitted).

It goes almost without saying that if these studies were "too tentative and  fragmentary" to make out a claim of constitutional error in 1968, the same  studies, unchanged **1764 but for having aged some 18 years, are still  insufficient to make out such a claim in this case.

Nor do the three post-Witherspoon studies introduced by McCree on the  "death qualification" issue provide substantial support for the "per se  constitutional rule" McCree asks this Court to adopt. All three of the "new"  studies were based on the responses of individuals randomly selected from some  segment of the population, but who were not actual jurors sworn under oath to  apply the law to the facts of an actual case involving the fate of an actual  capital defendant. [FN9] We have serious doubts about the value of these  studies in predicting the behavior of actual jurors. See Grigsby v. Mabry,  758 F.2d, at 248, n. 7 (J. Gibson, J., dissenting). In addition, two of the  three "new" studies did not even attempt to simulate the process of jury  deliberation, [FN10] and none of the "new" studies was able to predict to what  extent, if any, the presence of one or more "Witherspoon -excludables" on a  *172 guilt-phase jury would have altered the outcome of the guilt  determination. [FN11]

FN9. The Harris-1971 study polled 2,068 adults from throughout the United States, the Cowan-Deliberation study involved 288 jury-eligible residents of San Mateo and Santa Clara Counties in California, and the Jurow study was based on the responses of 211 employees of the Sperry Rand Corporation in New York.

FN10. The Harris-1971 and Jurow studies did not allow for group deliberation, but rather measured only individual responses.

FN11. Justice MARSHALL's dissent refers to an "essential unanimity" of support among socialscience researchers and other academics for McCree's assertion that "death qualification" has a significant effect on the outcome of jury deliberations at the guilt phase of capital trials. See post, at 6. At least one of the articles relied upon by the dissent candidly acknowledges, however, that its conclusions ultimately must rest on "[a] certain amount of ... conjecture" and a willingness "to transform behavioral suspicions into doctrine." Finch & Ferraro, The Empirical Challenge to Death-Qualified Juries: On Further Examination, 65 Neb.L.Rev. 21, 67 (1986). As the authors of the article explain: "[U]ncertainty inheres in every aspect of the capital jury's operation, whether one focuses on the method of identifying excludable jurors or the deliberative process through which verdicts are reached. So it is that, some seventeen years after Witherspoon, no definitive conclusions can be stated as to the frequency or the magnitude of the effects of death qualification.

  * * *  "Nor is it likely that further empirical research can add significantly to  the current understanding of death qualification. The true magnitude of  the phenomenon of conviction proneness is probably unmeasurable, given the  complexity of capital cases and capital adjudication." Id., at 66-67  (footnote omitted).

Finally, and most importantly, only one of the six "death qualification"  studies introduced by McCree even attempted to identify and account for the  presence of so-called "nullifiers," or individuals who, because of their deep-  seated opposition to the death penalty, would be unable to decide a capital  defendant's guilt or innocence fairly and impartially. [FN12] McCree concedes,  as he must, that "nullifiers" may properly be excluded from the guilt-phase  jury, and studies that fail to take into account the presence of such  "nullifiers" thus are fatally flawed. [FN13] Surely a "per se constitutional  rule" as far *173 reaching as the one McCree proposes should not be based on  the results of the lone study that avoids this fundamental flaw.

FN12. Only the Cowan-Deliberation study attempted to take into account the presence of "nullifiers."  FN13. The effect of this flaw on the outcome of a particular study is likely to be significant. The Cowan-Deliberation study revealed that approximately 37% of the "Witherspoon -excludables" identified in the study were also "nullifiers."

[1] Having identified some of the more serious problems with McCree's  studies, however, we will assume for purposes of this opinion that the studies  are both methodologically valid and adequate to establish that "death  qualification" in fact produces juries somewhat more "conviction-prone" than  "non-death-qualified" juries. We hold, nonetheless, that the Constitution does  not prohibit the States from "death qualifying" juries in capital cases.

[2] **1765 The Eighth Circuit ruled that "death qualification" violated  McCree's right under the Sixth Amendment, as applied to the States via  incorporation through the Fourteenth Amendment, see Duncan v. Louisiana, 391  U.S. 145, 148-158, 88 S.Ct. 1444, 1446-52, 20 L.Ed.2d 491 (1968), to a jury  selected from a representative cross section of the community. But we do not  believe that the fair-cross-section requirement can, or should, be applied as  broadly as that court attempted to apply it. We have never invoked the fair-  cross-section principle to invalidate the use of either for-cause or peremptory  challenges to prospective jurors, or to require petit juries, as opposed to  jury panels or venires, to reflect the composition of the community at large.

[3] But even if we were willing to extend the fair-cross-section requirement  to petit juries, we would still reject the Eighth Circuit's conclusion that  "death qualification" violates that requirement. The essence of a "fair-cross-  section" claim is the systematic exclusion of "a 'distinctive' group in the  community." Duren, supra, 439 U.S., at 364, 99 S.Ct., at 668. In our view,  groups defined solely in terms of shared attitudes that would prevent or  substantially impair members of the group from performing one of their duties  as jurors, such as the "Witherspoon -excludables" at issue here, are not  "distinctive groups" for fair-cross-section purposes.

The group of "Witherspoon -excludables" involved in the case at bar differs  significantly from the groups we have previously recognized as "distinctive."  "Death qualification," unlike the wholesale exclusion of blacks, women, or  Mexican-Americans from jury service, is carefully designed to serve the State's  concededly legitimate interest in obtaining a single jury that can properly and  impartially apply the law to the facts of the case at both the guilt and  sentencing phases of *176 a capital trial. [FN15] There is very little  danger, therefore, and McCree does not even argue, that "death qualification"  was instituted as a means for the State to arbitrarily skew the composition of  capital-case juries.

Furthermore, unlike blacks, women, and Mexican-Americans, "Witherspoon -  excludables" are singled out for exclusion in capital cases on the basis of an  attribute that is within the individual's control. It 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. Because the group of "Witherspoon -excludables"  includes only those who cannot and will not conscientiously obey the law with  respect to one of the issues in a capital case, "death qualification" hardly  can be said to create an "appearance of unfairness."

Finally, the removal for cause of "Witherspoon -excludables" in capital  cases does not prevent them from serving as jurors in other criminal cases, and  thus leads to no substantial deprivation of their basic rights of citizenship.  They are treated no differently than any juror who expresses the view that he  would be unable to follow the law in a particular case.

**1767 In sum, "Witherspoon -excludables," or for that matter any other  group defined solely in terms of shared attitudes that render members of the  group unable to serve as jurors in a *177 particular case, may be excluded  from jury service without contravening any of the basic objectives of the fair-  cross-section requirement. See Lockett v. Ohio, 438 U.S. 586, 597, 98 S.Ct.  2954, 2961, 57 L.Ed.2d 973 (1978) ("Nothing in Taylor, however, suggests  that the right to a representative jury includes the right to be tried by  jurors who have explicitly indicated an inability to follow the law and  instructions of the trial judge"). It is for this reason that we conclude  that "Witherspoon -excludables" do not constitute a "distinctive group" for  fair-cross-section purposes, and hold that "death qualification" does not  violate the fair-cross-section requirement.

[4] McCree argues that, even if we reject the Eighth Circuit's fair-cross-  section holding, we should affirm the judgment below on the alternative ground,  adopted by the District Court, that "death qualification" violated his  constitutional right to an impartial jury. McCree concedes that the individual  jurors who served at his trial were impartial, as that term was defined by this  Court in cases such as Irvin v. Dowd, 366 U.S. 717, 723, 81 S.Ct. 1639,  1643, 6 L.Ed.2d 751 (1961) ("It is sufficient if the juror can lay aside his  impression or opinion and render a verdict based on the evidence presented in  court"), and Reynolds v. United States, 98 U.S. (8 Otto) 145, 25 L.Ed. 244  (1879). He does not claim that pretrial publicity, see Rideau v. Louisiana,  373 U.S. 723, 83 S.Ct. 1417, 10 L.Ed.2d 663 (1963), ex parte communications,  see Remmer v. United States, 347 U.S. 227, 74 S.Ct. 450, 98 L.Ed. 654  (1954), or other undue influence, see Estes v. Texas, 381 U.S. 532, 85 S.Ct.  1628, 14 L.Ed.2d 543 (1965), affected the jury's deliberations. In short,  McCree does not claim that his conviction was tainted by any of the kinds of  jury bias or partiality that we have previously recognized as violative of the  Constitution. Instead, McCree argues that his jury lacked impartiality because  the absence of "Witherspoon -excludables" "slanted" the jury in favor of  conviction.

We do not agree. McCree's "impartiality" argument apparently is based on the  theory that, because all individual jurors are to some extent predisposed  towards one result or another, a constitutionally impartial jury can be  constructed *178 only by "balancing" the various predispositions of the  individual jurors. Thus, according to McCree, when the State "tips the  scales" by excluding prospective jurors with a particular viewpoint, an  impermissibly partial jury results. We have consistently rejected this view of  jury impartiality, including as recently as last Term when we squarely held  that an impartial jury consists of nothing more than "jurors who will  conscientiously apply the law and find the facts." Wainwright v. Witt, 469  U.S. 412, 423, 105 S.Ct. 844, 852, 83 L.Ed.2d 841 (1985) (emphasis added); see  also Smith v. Phillips, 455 U.S. 209, 217, 102 S.Ct. 940, 946, 71 L.Ed.2d  78 (1982) ("Due process means a jury capable and willing to decide the case  solely on the evidence before it"); Irvin v. Dowd, supra, 366 U.S., at 722,  81 S.Ct., at 1642 ("In essence, the right to jury trial guarantees to the  criminally accused a fair trial by a panel of impartial, 'indifferent'  jurors").

The view of jury impartiality urged upon us by McCree is both illogical and  hopelessly impractical. McCree characterizes the jury that convicted him as  "slanted" by the process of "death qualification." But McCree admits that  exactly the same 12 individuals could have ended up on his jury through the  "luck of the draw," without in any way violating the constitutional guarantee  of impartiality. Even accepting McCree's position that we should focus on the  jury rather than the individual jurors, it is hard for us to understand the  logic of the argument that a given jury is unconstitutionally partial when it  results from a state-ordained process, yet impartial when exactly the same jury  results from mere **1768 chance. On a more practical level, if it were true  that the Constitution required a certain mix of individual viewpoints on the  jury, then trial judges would be required to undertake the Sisyphean task of  "balancing" juries, making sure that each contains the proper number of  Democrats and Republicans, young persons and old persons, white-collar  executives and blue-collar laborers, and so on. Adopting McCree's concept of  jury impartiality would also likely require the elimination of peremptory  challenges, which are commonly used by both *179 the State and the defendant  to attempt to produce a jury favorable to the challenger.

In our view, it is simply not possible to define jury impartiality, for  constitutional purposes, by reference to some hypothetical mix of individual  viewpoints. Prospective jurors *184 come from many different backgrounds,  and have many different attitudes and predispositions. But the Constitution  presupposes that a jury selected from a fair cross section of the community is  impartial, regardless of the mix of individual viewpoints actually represented  on the jury, so long as the jurors can conscientiously and properly carry out  their sworn duty to apply the law to the facts of the particular case. We hold  that McCree's jury satisfied both aspects of this constitutional standard. The  judgment of the Court of Appeals is therefore

Reversed. Justice BLACKMUN concurs in the result.

Justice MARSHALL, with whom Justice BRENNAN and Justice STEVENS join,  dissenting.

With a glib nonchalance ill suited to the gravity of the issue presented and  the power of respondent's claims, the Court upholds a practice that allows the  State a special advantage in those prosecutions where the charges are the most  serious and the possible punishments, the most severe. The State's mere  announcement that it intends to seek the death penalty if the defendant is  found guilty of a capital offense will, under today's decision, give the  prosecution license to empanel a jury especially likely to return that very  verdict. Because I believe that such a blatant disregard for the rights of a  capital defendant offends logic, fairness, and the Constitution, I dissent.

I

*187 In the wake of Witherspoon, a number of researchers set out to  supplement the data that the Court had found inadequate in that case. The  results of these studies were exhaustively analyzed by the District Court in  this case, see 569 F.Supp. 1273, 1291-1308 (ED Ark.1983), and can be only  briefly summarized here. [FN2] The data strongly suggest that death  qualification excludes a significantly large subset--at least 11% to 17%--of  potential jurors who could be impartial during the guilt phase of trial. [FN3]  Among the members of this excludable class are a disproportionate number of  blacks and women. See id., at 1283, 1293-1294.

FN2. Most of the studies presented here were also comprehensively summarized in Hovey v. Superior Court, supra. Because the California Supreme Court found the studies had not accounted for jurors who could be excluded because they would automatically vote for the death penalty where possible, that court ultimately rejected a defendant's constitutional challenge to death qualification. But see Kadane, After Hovey: A Note on Taking Account of the Automatic Death Penalty Jurors, 8 Law & Hum.Behav. 115 (1984).

FN3. Bronson, On the Conviction Proneness and Representativeness of the Death-Qualified Jury: An Empirical Study of Colorado Veniremen, 12 U.Colo.L.Rev. 1 (1970) (using classification only approximating Witherspoon standard, and finding 11% of subjects Witherspoon - excludable); Bronson, Does the Exclusion of Scrupled Jurors in Capital Cases Make the Jury More Likely to Convict? Some Evidence from California, 3 Woodrow Wilson L.J. 11 (1980) (using more appropriate Witherspoon question and finding 93% overlap of "strongly opposed" group in prior Bronson study with Witherspoon -excludables); Jurow, New Data on the Effect of a "Death Qualified Jury" on the Guilt Determination Process, 84 Harv.L.Rev. 567 (1971) (finding only 10% of sample excludable, but likely to have underestimated size of class in general population because sample 99% white and 80% male); Fitzgerald & Ellsworth, Due Process vs. Crime Control: Death Qualification and Jury Attitudes, 8 Law & Hum.Behav. 31 (1984) (random sample with nullifiers screened out finding 17% still excludable under Witherspoon ); A. Young, Arkansas Archival Study (unpublished, 1981) (14% of jurors questioned in voir dire transcripts excludable); Precision Research, Inc., Survey No. 1286 (1981) (11% excludable, not counting nullifiers); see 569 F.Supp., at 1285; Grigsby v. Mabry, 758 F.2d 226, 231 (CA8 1985).

*188 The perspectives on the criminal justice system of jurors who survive  death qualification are systematically different from those of the excluded  jurors. Death-qualified jurors are, for example, more likely to believe that a  defendant's failure to testify is indicative of his guilt, more hostile to the  insanity defense, more mistrustful of defense attorneys, and less concerned  about the danger of erroneous convictions. Id., at 1283, 1293, 1304. This  proprosecution bias is reflected in the greater readiness of death-qualified  jurors to convict or to convict on more serious charges. Id., at 1294-1302;  Grigsby v. Mabry, 758 F.2d 226, 233-236 (CA8 1985). **1773 And, finally,  the very process of death qualification--which focuses attention on the death  penalty before the trial has even begun--has been found to predispose the  jurors that survive it to believe that the defendant is guilty. 569  F.Supp., at 1302-1305; 758 F.2d, at 234.

The evidence thus confirms, and is itself corroborated by, the more intuitive  judgments of scholars and of so many of the participants in capital trials--  judges, defense attorneys, and prosecutors. See 569 F.Supp., at 1322. [FN4]  *189

II
A

Respondent's case would of course be even stronger were he able to produce  data showing the prejudicial effects of death qualification upon actual  trials. Yet, until a State permits two separate juries to deliberate on the  same capital case and return simultaneous verdicts, defendants claiming  prejudice from death qualification should not be denied recourse to the only  available means of proving their case, recreations of the voir dire and trial  processes. See Grigsby v. Mabry, supra, at 237 ("[I]t is the courts who  have often stood in the way of surveys involving real jurors and we should not  now reject a study because of this deficiency").

The chief strength of respondent's evidence lies in the essential unanimity of  the results obtained by researchers using diverse subjects and varied  methodologies. Even the Court's haphazard jabs cannot obscure the power of the  array. Where studies have identified and corrected apparent flaws in prior  investigations, the results of the subsequent work have only corroborated the  conclusions drawn in the earlier efforts. Thus, for example, some studies  might be faulted for failing to distinguish within the class of  Witherspoon -excludables, between nullifiers (whom respondent concedes may  be excluded from the guilt phase) and those who could assess guilt  impartially. Yet their results are entirely consistent with those obtained  after nullifiers had indeed been excluded. See, e.g., Cowan, Thompson, &  Ellsworth, The Effects of Death Qualification on Jurors' Predisposition to  Convict and on the Quality of Deliberation, 8 Law & Hum. Behav. 53 (1984).  And despite the failure of certain studies to "allow for group deliberations,"  ante, *190 at 1762, n. 10, the value of their results is underscored by the  discovery that initial verdict preferences, made prior to group deliberations,  are a fair predictor of how a juror will vote when faced with opposition in the  jury room. See Cowan, Thompson, & Ellsworth, supra, at 68-69; see also R.  Hastie, S. Penrod, & N. Pennington, Inside the Jury 66 (1983); H. Kalven & H.  Zeisel, The American Jury 488 (1966).

The evidence adduced by respondent is quite different from the "tentative and  **1774 fragmentary" presentation that failed to move this Court in  Witherspoon. 391 U.S., at 517, 88 S.Ct., at 1774. Moreover, in contrast  to Witherspoon, the record in this case shows respondent's case to have  been 'subjected to the traditional testing mechanisms of the adversary process, '  Ballew v. Georgia, 435 U.S. 223, 246, 98 S.Ct. 1029, 1042, 55 L.Ed.2d  334 (1978) (POWELL, J., concurring in judgment). At trial, respondent  presented three expert witnesses and one lay witness in his rebuttal. Testimony  by these witnesses permitted the District Court, and allows this Court, better  to understand the methodologies used here and their limitations. Further  testing of respondent's impirical case came at the hands of the State's own  expert witnesses. Yet even after considering the evidence aduced by the State,  the Court of Appeals properly noted: 'there are no studies which contradict  the studies submitted [respondent]; in other words, all of the documented  studies support the district court's findings.' 758 F.2d, at 238.

Faced with the near unanimity of authority supporting respondent's claim that  death qualification gives the prosecution a particular advantage in the guilt  phase of capital trials, the majority here makes but a weak effort to contest  that proposition. Instead, it merely assumes for the purposes of this opinion  "that 'death qualification' in fact produces juries somewhat more 'conviction-  prone' than 'non-death-qualified' juries," ante, at 1764, and then holds  that this result does not offend the Constitution. This disregard for the  clear import of the evidence tragically misconstrues the settled constitutional  principles that guarantee a defendant the right to to a fair trial and an  impartial jury whose composition is not biased toward the prosecution.  In Witherspoon, the Court observed that a defendant convicted by a jury  from which those unalterably opposed to the death penalty had been excluded  "might still attempt to establish *193 that the jury was less than neutral  with respect to guilt." 391 U.S., at 520, n. 18, 88 S.Ct., at 1776, n. 18.  Respondent has done just that. And I believe he has succeeded in proving that  his trial by a jury so constituted violated his right to an impartial jury,  guaranteed by both the Sixth Amendment and principles of due process, see  Ristaino v. Ross, 424 U.S. 589, 595, n. 6, 96 S.Ct. 1017, 1020, n. 6, 47  L.Ed.2d 258 (1976). We therefore need not rely on respondent's alternative  argument that death qualification deprived him of a jury representing a fair  cross section of the community. [FN6]

VI

On occasion, this Court has declared what I believe should be obvious--that  when a State seeks to convict a defendant of the most serious and severely  punished offenses in its criminal code, any procedure that "diminish[es] the  reliability of the guilt determination" must be struck down. Beck v.  Alabama, 447 U.S., at 638, 100 S.Ct., at 2390. But in spite of such  declarations, I cannot help thinking that respondent here would have stood a  far better chance of prevailing on his constitutional claims had he not been  challenging a procedure peculiar to the administration of the death penalty.  For in no other context would a majority of this Court refuse to find any  constitutional violation in a state practice that systematically operates to  render juries more likely to convict, and to convict on the more serious  charges. I dissent.