with whom Justice Marshall joins, dissenting.
The Court holds that the Due Process Clause of the Fourteenth Amendment did not require that the trial judge in each of the instant cases hold a “fair hearing,” Jackson v. Denno, 378 U. S. 368, 377 (1964), to decide the admissibility of eyewitness identification evidence, and that a remand is not now required to accord such a hearing. While freely conceding that a “judicial determination outside the presence of the *350jury of the admissibility of identification evidence may often be advisable [and i]n some circumstances . . . constitutionally necessary,” ante, at 349, the Court holds that the Constitution does not require “a per se rule compelling such a procedure in every case,” ibid. I dissent. In my view, the Due Process Clause mandates such a hearing whenever a defendant, as both petitioners did at their respective trials below, has proffered some evidence that pretrial police procedures directed at identificatiop were impermissibly suggestive. The flaw in the Court’s reasoning lies in its statement that identification evidence does not implicate the “special considerations” on which Jackson v. Denno relied to “justify a departure from the presumption that juries will follow instructions.” Ante, at 347. Surely jury instructions can ordinarily no more cure the erroneous admission of powerful identification evidence than they can cure the erroneous admission of a confession. Accordingly, the separate judicial determination of admissibility required by Jackson for confessions is equally applicable for eyewitness identification evidence. Because the record before us is inadequate to conclude that in each case the identification evidence was properly admitted, see Jackson v. Denno, supra, at 376-377, I would remand these cases for further proceedings.
At least since United States v. Wade, 388 U. S. 218 (1967), the Court has recognized the inherently suspect qualities of eyewitness identification evidence.1 Two particular attributes of such evidence have significance for' the instant cases. First, eyewitness identification evidence is notoriously unreliable:
“The vagaries of eyewitness identification are well-known ; the annals of criminal law are rife with instances *351of mistaken identification. Mr. Justice Frankfurter once said: ‘What is the worth of identification testimony even when uncontradicted? The identification of strangers is proverbially untrustworthy. The hazards of such testimony are established by a formidable number of instances in the records of English and American trials. These instances are recent — not due to the brutalities of ancient criminal procedure.’ The Case of Sacco and Vanzetti 30 (1927).” Id., at 228 (footnote omitted).
Manson v. Brathwaite, 432 U. S. 98, 111-112 (1977), emphasized this troublesome characteristic of such evidence:
“The driving force behind United States v. Wade, 388 U. S. 218 (1967), Gilbert v. California, 388 U. S. 263 (1967) (right to counsel at a post-indictment lineup), and Stovall, all decided on the same day, was the Court’s concern with the problems of eyewitness identification. Usually the witness must testify about an encounter with a total stranger under circumstances of emergency or emotional stress. The witness’ recollection of the stranger can be distorted easily by the circumstances or by later actions of the police.”
Accordingly, to guard against the “dangers inherent in eyewitness identification,” United States v. Wade, supra, at 235, the Court has required the presence of counsel at postindietment lineups, 388 U. S., at 236-237,2 and has held inadmissible identification evidence tainted by suggestive confrontation procedures and lacking adequate indicia of reliability, *352Manson v. Brathwaite, supra, at 114. “Thus, Wade and its companion cases reflect the concern that the jury not hear eyewitness testimony unless that evidence has aspects of reliability.” 432 U. S., at 112. An important thrust of our eyewitness identification evidence cases from Wade to Manson, therefore, has been to prevent impairment of the jury’s decisionmaking process by the introduction of unreliable identification evidence.
Second, despite its inherent unreliability, much eyewitness identification evidence has a powerful impact on juries. Juries seem most receptive to, and not inclined to discredit, testimony of a witness who states that he saw the defendant commit the crime.3
“[E]yewitness testimony is likely to be believed by jurors, especially when it is offered with a high level of confidence, even though the accuracy of an eyewitness and the confidence of that witness may not be related to one another at all. All the evidence points rather strikingly to the conclusion that there is almost nothing more convincing than a live human being who takes the stand, points a finger at the defendant, and says ‘That’s the one!’ ” 4
The powerful impact that much eyewitness identification evidence has on juries, regardless of its reliability,5 virtually *353mandates that, when such evidence is inadmissible, the jury should know nothing about the evidence. See Manson v. Brathwaite, supra, at 112. For certainly the resulting prejudice to the defendant cannot be erased by jury instructions. See generally E. Loftus, Eyewitness Testimony 189-190 (1979); P. Devlin, Report to the Secretary of State for the Home Department of the Departmental Committee on Evidence of Identification in Criminal Cases 149-150 (1976). The Court’s contrary conclusion cavalierly dismisses the inherent unreliability of identification evidence and its effect on juries — two attributes of confession evidence that led the Court to mandate a “fair hearing” safeguard in Jackson v. Denno.
Any purported distinction between the instant cases and Jackson is plainly specious. In Jackson, this Court invalidated a New York State procedure whereby the jury was instructed first to determine the voluntariness of a defendant’s confession6 and then to disregard the confession if it concluded that the confession was involuntary. Jackson struck down this practice and required first that the voluntariness *354of a confession be determined by the judge before its admission in evidence, and second that the jury not be allowed to consider an inadmissible confession. Jackson refused to rely on the curative effect of jury instructions where the trial judge had not applied “ ‘the exclusionary rules before permitting evidence to be submitted to the jury.’ ” 378 U. S., at 382, n. 10, quoting Meltzer, Involuntary Confessions: The Allocation of Responsibility Between Judge and Jury, 21 U. Chi. L. Rev. 317, 327 (1954).7
For purposes of the instant cases, three factors central to our decision in Jackson are apposite. First, Jackson stated, as the Court today notes, ante, at 347, “that the Fourteenth Amendment forbids the use of involuntary confessions . . . because of the probable unreliability of confessions that are obtained in a manner deemed coercive.” 378 U. S., at 385-386. Second, Jackson stated, as the Court today further notes, ante, at 347, that involuntary confessions are inadmissible “because of the ‘strongly felt attitude of our society that important human values are sacrificed where an agency of the government, in the course of securing a conviction, wrings a confession out of an accused against his will.’ ” 378 U. S., at 386.8 Third, because of the sensitive nature of confession *355evidence, Jackson found that instructions were not adequate to assure that the jury would ignore involuntary confession evidence:
“Under the New York procedure, the fact of a defendant’s confession is solidly implanted in the jury’s mind, for it has not only heard the confession, but it has also been instructed to consider and judge its voluntariness and is in position to assess whether it is true or false.[9] If it finds the confession involuntary, does the jury — indeed, can it — then disregard the confession in accordance with its instructions? If there are lingering doubts about the sufficiency of the other evidence, does the jury unconsciously lay them to rest by resort to the confession? Will uncertainty about the sufficiency of the other evidence to prove guilt beyond a reasonable doubt actually result in acquittal when the jury knows the defendant has given a truthful confession.” Id., at 388 (footnote omitted).
Similar considerations plainly require a hearing in the case of identification evidence. First, there can be little doubt that identification evidence is as potentially unreliable as confession evidence. See supra, at 350-352. Second, suggestive confrontation procedures which, in the totality of the circumstances, create “ ‘a very substantial likelihood of irreparable misidentification,’ ” Manson v. Brathwaite, 432 U. S., at 116, quoting Simmons v. United States, 390 U. S. 377, 384 (1968), are as impermissible a police practice as the securing of a custodial confession determined, in the totality of the circumstances, to be involuntary, see United States v. Washington, 431 U. S. 181, 188 (1977); cf. North Carolina v. Butler, *356441 U. S. 369, 374-375 (1979) (waiver). See also Manson v. Brathwaite, supra, at 112; Foster v. California, 394 U. S. 440, 442-443 (1969); United States v. Wade, 388 U. S., at 228-229, 232-235; Stovall v. Denno, 388 U. S. 293, 302 (1967). And third, because of the extraordinary impact of much eyewitness identification evidence, juries hearing such evidence will be no more able fully to ignore it upon instruction of the trial judge than will juries hearing confession evidence.10 To expect a jury to engage in the collective mental gymnastic of segregating and ignoring such testimony upon instruction is utterly unrealistic. The Court’s bald assertion, therefore, that jury instructions are adequate to protect the accused, is as untrue for identification evidence as it is for involuntary confessions.
Nor can it be assumed, as the Court has, that cross-examination will protect the accused in this circumstance. That is no more true here than it was in Jackson, where the defendant was also allowed to cross-examine on the question of voluntariness. Cross-examination, of course, affects the weight and credibility given by the jury to evidence,11 but cross-examination is both an ineffective and a wrong tool for purging inadmissible identification evidence from the jurors’ minds. It is an ineffective tool because all of the scientific *357evidence suggests that much eyewitness identification testimony has an unduly powerful effect on jurors. Thus, the jury is likely to give the erroneously admitted evidence substantial weight, however skillful the cross-examination. See generally E. Loftus, Eyewitness Testimony 9 (1979). Cross-examination is also a wrong tool in the sense that jury instructions are the means normally employed to cure the erroneous introduction of evidence. At best, cross-examination might diminish the weight the jury accords to the admissible evidence. The likelihood is, however, that the jury would continue to give the improperly admitted evidence substantial weight, even if properly instructed to disregard it.
It is clear beyond peradventure, I submit, that because of the dangers to a just result inherent in identification evidence — its unreliability and its unusual impact on the jury— a “fair hearing and a reliable determination” of admissibility, Jackson v. Denno, 378 U. S., at 377, are constitutionally mandated. The Due Process Clause obviously precludes the jury from convicting on unreliable identification evidence. Manson v. Brathwaite, supra.12 But the only way to be sure that the jury will not rest its verdict on improper identification evidence, as a practical matter, is by not permitting the jury to hear it in the first place. A Jackson v. Denno hearing would expediently accomplish that purpose.13 I believe that the Due Process Clause requires no less.
*358A large and distinguished group shares my view. The lower federal courts with virtual unanimity have encouraged the type of hearing sought by petitioners.14 As already noted, *359the Court too states that “[a] judicial determination outside the presence of the jury of the admissibility of identification evidence may often be advisable [and i]n some circumstances . . . constitutionally necessary.” Ante, at 349. I should think it follows from this congruence of opinion on the desirability of such a judicial hearing that evolving standards of justice15 mandate such a hearing whenever a defendant proffers sufficient evidence to raise a colorable claim that police confrontation procedures were impermissibly suggestive. See, e. g., United States ex rel. Fisher v. Driber, 546 F. 2d 18, 22 (CA3 1976).
In the instant cases, the suggestiveness of the confrontation procedures was clearly shown, and equally clearly cross-examination in front of the jury was inadequate to test the reliability of the evidence because of the undoubted inhibiting effect on cross-examination from fear that rigorous questioning of hostile witnesses would strengthen the eyewitnesses’ testimony and impress it upon the jury. See United States v. Wade, 388 U. S., at 240-241.16 In any event, the record *360is inadequate to decide that petitioners could not have succeeded in foreclosing admission of the evidence if they had been afforded a hearing out of the jury’s presence in the first place. Accordingly, I would remand for such further proceedings as are necessary to give these petitioners “a fair hearing and a reliable determination,” Jackson v. Denno, 378 U. S., at 377, that the identification evidence in each trial was not erroneously admitted.
The special nature of eyewitness identification evidence has produced an enormous reservoir of scholarly writings, many based on solid empirical research. For a bibliography of that literature, see E. Loftus, Eyewitness Testimony 237-247 (1979).
“ [Suggestibility inherent in the context of the pretrial identification” is a factor that has led the Court to require the presence of counsel at postindictment lineups. United States v. Wade, 388 U. S., at 235. If counsel is not present at such a lineup, the identification may not be introduced into evidence at trial and an in-court identification may be made only if the prosecutor establishes “by clear and convincing evidence that the in-court identification [was] based upon observatio[n] ... of the suspect other than the lineup identification.” Id., at 240.
“[JJuries unfortunately are often unduly receptive to [identification] evidence . . . Manson v. Brathwaite, 432 U. S. 98, 120 (1977) (Marshall, J., dissenting) (footnote omitted). See Loftus, supra, at 8-19; P. Wall, Eyewitness Identification in Criminal Cases 19-23 (1965) ; Hammelmann & Williams, Identification Parades — II, 1963 Crim. L. Rev. 545, 550. See generally A. Yarmey, The Psychology of Eyewitness Testimony (1979).
Loftus, supra, at 19 (emphasis supplied). Professor Loftus exhaustively canvasses statistical and psychological evidence which persuasively supports her conclusion that eyewitness identification evidence is “overwhelmingly influential.” Id., at 9.
Professor Loftus, ibid, (emphasis in original), observes that “[j]urors *353have been known to accept eyewitness testimony pointing to guilt even when it is far outweighed by evidence of innocence.”
Wall, supra, at 19 (footnotes omitted) (emphasis supplied), concludes: “[Tjuries are unduly receptive to identification evidence and are not sufficiently aware of its dangers. It has been said that ‘positive recognition by well intended uninterested persons is commonly accepted unless the alibi is convincing/ and that evidence of identification, however untrustworthy, is ‘taken by the average juryman as absolute proof.’ ”
Distinguishing Jackson from the instant cases on the basis that the jury there was first instructed to determine voluntariness is not persuasive. That consideration goes to the weight given the evidence by the jury. Jackson itself recognized that the lingering effect of the involuntary confession might be decisive in the jury’s deliberations. Such an effect is no less likely to be decisive in the case of powerful eyewitness identification evidence that a jury has been instructed to ignore. In both instances, peculiarly powerful evidence must leave an indelible impact on a juror’s mind. See n. 7, infra.
The Court in Jackson noted:
“ ‘Due Process of law requires that a coerced confession be excluded from consideration by the jury. It also requires that the issue of coercion be tried by an unprejudiced trier, and, regardless of the pious fictions indulged by the courts, it is useless to contend that a juror who has heard the confession can be uninfluenced by his opinion as to the truth or falsity of it. . . . And the rule of exclusion ought not to be emasculated by admitting the evidence and giving to the jury an instruction which, as every judge and lawyer knows, cannot be obeyed.’ ” 378 U. S., at 382-383, n. 10, quoting E. Morgan, Some Problems of Proof Under the Anglo-American System of Litigation 104-105 (1956).
Of course, police misbehavior is not always so lacking in subtlety that involuntary confessions are invariably wrenched from an accused by force. Thus, indirect methods of interrogation which seek to elicit a statement *355from a custodial suspect may also warrant a conclusion of involuntariness. See Rhode Island v. Innis, 446 U. S. 291, 301 (1980) (interrogation includes actions which “the police should know are reasonably likely to elicit an incriminating response”); cf. Brewer v. Williams, 430 U. S. 387 (1977) (Sixth Amendment violation).
See n. 6, supra.
In both of these cases, the • eyewitnesses were also the victims of the crimes. Not only does that dual status affect the reliability of the identification, but it also is likely to make the testimony more powerful and thus less curable by jury instructions. Clearly, this is not a case where 14 reliable identifications were properly received in evidence, but a 15th by a nonvictim witness was subject to suggestive confrontation procedures and was unreliable, thereby raising the possibility that the error was harmless beyond a reasonable doubt.
In Manson v. Brathwaite, 432 IT. S., at 116, the Court stated:
“We are content to rely upon the good sense and judgment of American juries, for evidence with some element of untrustworthiness is customary grist for the jury mill. Juries are not so susceptible that they cannot measure intelligently the weight of identification testimony that has some questionable feature.”
In Jackson v. Denno, the Court was concerned that the jury not hear a defendant’s confession until a trial judge had made a preliminary determination of voluntariness. The Court assumed that were this not done, a deleterious impact on the jury’s deliberations would operate:
“[I]t is only a reliable determination on the voluntariness issue which satisfies the constitutional rights of the defendant and which would permit the jury to consider the confession in adjudicating guilt or innocence.” 378 U. S., at 387.
The Court errs in any event in deciding these eases on the premise that petitioners request a per se rule requiring a hearing out of the jury’s presence in every case. In the first place, petitioners rely substantially *358on authority which does not go that far. Brief for Petitioners 43-45. Clearly, they have sought reversal of their convictions on the basis that they were entitled to such a hearing. Moreover, there is no question here that they raised a colorable claim that the confrontation procedures were impermissibly suggestive. See, e. g., United States ex rel. Fisher v. Driber, 546 F. 2d 18, 22 (CA3 1976); United States v. Cranson, 453 F. 2d 123, 127 (CA4 1971), cert. denied, 406 U. S. 909 (1972).
If the Court’s result is out of concern for not adding another layer of complexity to criminal litigation, that is understandable, but not sufficient to supplant an accused’s constitutional right. Moreover, a rule requiring the defendant to proffer some minimum quantum of evidence showing the suggestiveness of the confrontation procedures would eliminate frivolous requests. See, e. g., United States ex rel. Fisher v. Driber, supra, at 22.
United States ex rel. Fisher v. Driber, supra, at 22 (requiring hearing outside presence of jury where motion for such hearing is not frivolous); United States v. Smith, 546 F. 2d 1275, 1279 (CA5 1977) (evi-dentiary hearing not required where no critical facts in dispute); United States v. Mitchell, 540 F. 2d 1163, 1166 (CA3 1976) (defendant could have “requested a hearing outside the presence of the jury in accordance with Neil v. Biggers”), cert. denied, 429 U. S. 1099 (1977); Nassar v. Vinzant, 519 F. 2d 798, 802, n. 4 (CA1) (commending hearing out of jury’s presence), cert, denied, 423 U. S. 898 (1975); United States v. Cranson, supra, at 125-126 (“evidentiary hearing outside the jury’s presence is required” upon motion to suppress); Haskins v. United States, 433 F. 2d 836, 838 (CA10 1970) (requiring hearing outside of jury’s presence); United States v. Ranciglio, 429 F. 2d 228, 230 (CA8) (“trial court, out of the hearing and presence of the jury, conducted a hearing as required in Wade”), cert. denied, 400 U. S. 959 (1970); United States ex rel. Phipps v. Follette, 428 F. 2d 912, 913, n. 1 (CA2) (“commend[ing] . . . practice” of hearing out of jury’s presence), cert. denied, 400 U. S. 908 (1970); United States v. Allison, 414 F. 2d 407, 410 (CA9) (requiring hearing outside of jury’s presence), cert. denied, 396 U. S. 968 (1969); United States v. Broadhead, 413 F. 2d 1351, 1359 (CA7 1969) (pretrial hearing approved), cert. denied, 396 U. S. 1017 (1970); Clemons v. United States, 133 U. S. App. D. C. 27, 34, 408 F. 2d 1230, 1237 (1968) (en banc) (requiring hearing outside of jury’s presence or disclosure of prosecutor’s evidence), cert. denied, 394 U. S. 964 (1969). Even the Court of Appeals deciding these eases stated that it had “no *359doubt that” a hearing out of the jury’s presence “is the preferable procedure.” Summitt v. Bordenkircher, 608 F. 2d 247, 250 (CA6 1979).
In addition, the Commonwealth of Kentucky, where petitioners were tried and convicted, appears to require a hearing out of the presence of the jury, upon defendant’s motion, for confession and for search evidence. See Ky. Rule Crim. Proc. 9.78. In addition, Moore v. Commonwealth, 569 S. W. 2d 150, 153 (Ky. 1978), decided after petitioners were convicted, held that the trial court’s refusal to hold a suppression hearing to determine the admissibility of identification evidence constituted error. Previous Kentucky appellate decisions had reached a similar conclusion. E. g., Francis v. Commonwealth, 468 S. W. 2d 287 (App. 1971).
See, e. g., Harper v. Virginia Board of Elections, 383 U. S. 663, 669 (1966) (equal protection); Trap v. Dulles, 356 U. S. 86, 100-101 (1958) (plurality opinion of Warren, C. J.) (Eighth Amendment).
It is no answer to say, as the Court does, that the record does not reflect that petitioners’ respective counsel were deterred by the presence of the jury, for the simple reason that a cold record cannot reflect questions not asked.