MR. JUSTICE BLACKMUN announced the judgment of the Court and delivered an opinion in which MR. JUSTICE STEVENS joined.

This case presents the issue whether a state criminal trial to a jury of only five persons deprives the accused of the right to trial by jury guaranteed to him by the Sixth and Fourteenth Amendments. Our resolution of the issue requires an application of principles enunciated in Williams v. Florida, 399 U.S. 78 (1970), where the use of a six-person jury in a state criminal trial was upheld against similar constitutional attack.

I

In November 1973 petitioner Claude Davis Ballew was the manager of the Paris Adult Theatre at 320 Peachtree Street, Atlanta, Ga. On November 9 two investigators from the Fulton County Solicitor General's office viewed at the theater a motion picture film entitled "Behind the Green Door." Record 46-48, 90. After they had seen the film, they obtained a warrant for its seizure, returned to the theater, viewed the film once again, and seized it. Id., at 48-50, 91. Petitioner and a cashier were arrested. Investigators returned to the theater on November 26, viewed the film in its entirety, secured still another warrant, and on November 27 once again viewed the motion picture and seized a second copy of the film. Id., at 53-55.

On September 14, 1974, petitioner was charged in a two-count misdemeanor accusation with

"distributing obscene materials in violation of Georgia Code Section 26-2101 in that the said accused did, knowing the obscene nature thereof, exhibit a motion picture film entitled 'Behind the Green Door' that contained obscene and indecent scenes . . . ." App. 4-6.

Petitioner was brought to trial in the Criminal Court of Fulton County. After a jury of 5 persons had been selected and sworn, petitioner moved that the court impanel a jury of 12 persons. Record 37-38. That court, however, tried its misdemeanor cases before juries of five persons pursuant to Ga. Const., Art. 6. § 16, ¶ 1, codified as Ga. Code § 2-5101 (1975), and to 1890-1891 Ga. Laws, No. 278, pp. 937-938, and 1935 Ga. Laws, No. 38, p. 498. Petitioner contended that for an obscenity trial, a jury of only five was constitutionally inadequate to assess the contemporary standards of the community. Record 13, 38. He also argued that the Sixth and Fourteenth Amendments required a jury of at least six members in criminal cases. Id., at 38.

The motion for a 12-person jury was overruled, and the trial went on to its conclusion before the 5-person jury that had been impaneled. At the conclusion of the trial, the jury deliberated for 38 minutes and returned a verdict of guilty on both counts of the accusation. Id., at 205-208. The court imposed a sentence of one year and a $1,000 fine on each count, the periods of incarceration to run concurrently and to be suspended upon payment of the fines. Id., at 16-17, 209. After a subsequent hearing, the court denied an amended motion for a new trial.

Petitioner took an appeal to the Court of Appeals of the State of Georgia. There he argued: First, the evidence was insufficient. Second, the trial court committed several First Amendment errors, namely, that the film as a matter of law was not obscene, and that the jury instructions incorrectly explained the standard of scienter, the definition of obscenity, and the scope of community standards. Third, the seizures of the films were illegal. Fourth, the convictions on both counts had placed petitioner in double jeopardy because he had shown only one motion picture. Fifth, the use of the five-member jury deprived him of his Sixth and Fourteenth Amendment right to a trial by jury. Id., at 222-224.

The Court of Appeals rejected petitioner's contentions. 138 Ga. App. 530, 227 S.E.2d 65 (1976). The court independently reviewed the film in its entirety and held it to be "hard core pornography" and "obscene as a matter of constitutional law and fact." Id., at 532-533, 227 S.E.2d, at 67-68. The evidence was sufficient to support the jury's conclusion that petitioner possessed the requisite scienter. As manager of the theater, petitioner had advertised the movie, had sold tickets, was present when the films were exhibited, had pressed the button that allowed entrance to the seating area, and had locked the door after each arrest. This evidence, according to the court, met the constructive-knowledge standard of § 26-2101. The court found no errors in the instructions, in the issuance of the warrants, or in the presence of the two convictions. In its consideration of the five-person-jury issue, the court noted that Williams v. Florida had not established a constitutional minimum number of jurors. Absent a holding by this Court that a five-person jury was constitutionally inadequate, the Court of Appeals considered itself bound by Sanders v. State, 234 Ga. 586, 216 S.E.2d 838 (1975), cert. denied, 424 U.S. 931 (1976), where the constitutionality of the five-person jury had been upheld. The court also cited the earlier case of McIntyre v. State, 190 Ga. 872, 11 S.E.2d 5 (1940), a holding to the same general effect but without elaboration.

The Supreme Court of Georgia denied certiorari. App. 26.

In his petition for certiorari here, petitioner raised three issues: the unconstitutionality of the five-person jury; the constitutional sufficiency of the jury instructions on scienter and constructive, rather than actual, knowledge of the contents of the film; and obscenity vel non. We granted certiorari. 429 U.S. 1071 (1977). Because we now hold that the five-member jury does not satisfy the jury trial guarantee of the Sixth Amendment, as applied to the States through the Fourteenth, we do not reach the other issues.

II

The Fourteenth Amendment guarantees the right of trial by jury in all state nonpetty criminal cases. Duncan v. Louisiana, 391 U.S. 145, 159-162 (1968). The Court in Duncan applied this Sixth Amendment right to the States because "trial by jury in criminal cases is fundamental to the American scheme of justice." Id., at 149. The right attaches in the present case because the maximum penalty for violating § 26-2101, as it existed at the time of the alleged offenses, exceeded six months' imprisonment. See Baldwin v. New York, 399 U.S. 66, 68-69 (1970) (opinion of WHITE, J.).

In Williams v. Florida, 399 U.S., at 100, the Court reaffirmed that the "purpose of the jury trial, as we noted in Duncan, is to prevent oppression by the Government. 'Providing an accused with the right to be tried by a jury of his peers gave him an inestimable safeguard against the corrupt or overzealous prosecutor and against the compliant, biased, or eccentric judge.' Duncan v. Louisiana, [391 U.S.,] at 156." See Apodaca v. Oregon, 406 U.S. 404, 410 (1972) (opinion of WHITE, J.). This purpose is attained by the participation of the community in determinations of guilt and by the application of the common sense of laymen who, as jurors, consider the case. Williams v. Florida, 399 U.S., at 100.

Williams held that these functions and this purpose could be fulfilled by a jury of six members. As the Court's opinion in that case explained at some length, id., at 86-90, common-law juries included 12 members by historical accident, "unrelated to the great purposes which gave rise to the jury in the first place." Id., at 89-90. The Court's earlier cases that had assumed the number 12 to be constitutionally compelled were set to one side because they had not considered history and the function of the jury. Id., at 90-92. Rather than requiring 12 members, then, the Sixth Amendment mandated a jury only of sufficient size to promote group deliberation, to insulate members from outside intimidation, and to provide a representative cross-section of the community. Id., at 100. Although recognizing that by 1970 little empirical research had evaluated jury performance, the Court found no evidence that the reliability of jury verdicts diminished with six-member panels. Nor did the Court anticipate significant differences in result, including the frequency of "hung" juries. Id., at 101-102, and nn. 47 and 48. Because the reduction in size did not threaten exclusion of any particular class from jury roles, concern that the representative or cross-section character of the jury would suffer with a decrease to six members seemed "an unrealistic one." Id., at 102. As a consequence, the six-person jury was held not to violate the Sixth and Fourteenth Amendments.

III

When the Court in Williams permitted the reduction in jury size — or, to put it another way, when it held that a jury of six was not unconstitutional — it expressly reserved ruling on the issue whether a number smaller than six passed constitutional scrutiny. Id., at 91 n. 28. See Johnson v. Louisiana, 406 U.S. 356, 365-366 (1972) (concurring opinion). The Court refused to speculate when this so-called "slippery slope" would become too steep. We face now, however, the two-fold question whether a further reduction in the size of the state criminal trial jury does make the grade too dangerous, that is, whether it inhibits the functioning of the jury as an institution to a significant degree, and, if so, whether any state interest counterbalances and justifies the disruption so as to preserve its constitutionality.

Williams v. Florida and Colgrove v. Battin, 413 U.S. 149 (1973) (where the Court held that a jury of six members did not violate the Seventh Amendment right to a jury trial in a civil case), generated a quantity of scholarly work on jury size. These writings do not draw or identify a bright line below which the number of jurors would not be able to function as required by the standards enunciated in Williams. On the other hand, they raise significant questions about the wisdom and constitutionality of a reduction below six. We examine these concerns:

First, recent empirical data suggest that progressively smaller juries are less likely to foster effective group deliberation. At some point, this decline leads to inaccurate fact-finding and incorrect application of the common sense of the community to the facts. Generally, a positive correlation exists between group size and the quality of both group performance and group productivity. A variety of explanations have been offered for this conclusion. Several are particularly applicable in the jury setting. The smaller the group, the less likely are members to make critical contributions necessary for the solution of a given problem. Because most juries are not permitted to take notes, see Forston, Sense and Non-Sense: Jury Trial Communication, 1975 B. Y. U. L. Rev. 601, 631-633, memory is important for accurate jury deliberations. As juries decrease in size, then, they are less likely to have members who remember each of the important pieces of evidence or argument. Furthermore, the smaller the group, the less likely it is to overcome the biases of its members to obtain an accurate result. When individual and group decisionmaking were compared, it was seen that groups performed better because prejudices of individuals were frequently counterbalanced, and objectivity resulted. Groups also exhibited increased motivation and self-criticism. All these advantages, except, perhaps, self-motivation, tend to diminish as the size of the group diminishes. Because juries frequently face complex problems laden with value choices, the benefits are important and should be retained. In particular, the counterbalancing of various biases is critical to the accurate application of the common sense of the community to the facts of any given case.

Second, the data now raise doubts about the accuracy of the results achieved by smaller and smaller panels. Statistical studies suggest that the risk of convicting an innocent person (Type I error) rises as the size of the jury diminishes. Because the risk of not convicting a guilty person (Type II error) increases with the size of the panel, an optimal jury size can be selected as a function of the interaction between the two risks. Nagel and Neef concluded that the optimal size, for the purpose of minimizing errors, should vary with the importance attached to the two types of mistakes. After weighting Type I error as 10 times more significant than Type II, perhaps not an unreasonable assumption, they concluded that the optimal jury size was between six and eight. As the size diminished to five and below, the weighted sum of errors increased because of the enlarging risk of the conviction of innocent defendants.

Another doubt about progressively smaller juries arises from the increasing inconsistency that results from the decreases. Saks argued that the "more a jury type fosters consistency, the greater will be the proportion of juries which select the correct ( i. e., the same) verdict and the fewer 'errors' will be made." Saks 86-87. From his mock trials held before under-graduates and former jurors, he computed the percentage of "correct" decisions rendered by 12-person and 6-person panels. In the student experiment, 12-person groups reached correct verdicts 83% of the time; 6-person panels reached correct verdicts 69% of the time. The results for the former-juror study were 71% for the 12-person groups and 57% for the 6-person groups. Ibid. Working with statistics described in H. Kalven H. Zeisel, The American Jury 460 (1966), Nagel and Neef tested the average conviction propensity of juries, that is, the likelihood that any given jury of a set would convict the defendant. They found that half of all 12-person juries would have average conviction propensities that varied by no more than 20 points. Half of all six-person juries, on the other hand, had average conviction propensities varying by 30 points, a difference they found significant in both real and percentage terms. Lempert reached similar results when he considered the likelihood of juries to compromise over the various views of their members, an important phenomenon for the fulfillment of the commonsense function. In civil trials averaging occurs with respect to damages amounts. In criminal trials it relates to numbers of counts and lesser included offenses. And he predicted that compromises would be more consistent when larger juries were employed. For example, 12-person juries could be expected to reach extreme compromises in 4% of the cases, while 6-person panels would reach extreme results in 16%. All three of these post-Williams studies, therefore, raise significant doubts about the consistency and reliability of the decisions of smaller juries.

Third, the data suggest that the verdicts of jury deliberation in criminal cases will vary as juries become smaller, and that the variance amounts to an imbalance to the detriment of one side, the defense. Both Lempert and Zeisel found that the number of hung juries would diminish as the panels decreased in size. Zeisel said that the number would be cut in half — from 5% to 2.4% with a decrease from 12 to 6 members. Both studies emphasized that juries in criminal cases generally hang with only one, or more likely two, jurors remaining unconvinced of guilt. Also group theory suggests that a person in the minority will adhere to his position more frequently when he has at least one other person supporting his argument. In the jury setting the significance of this tendency is demonstrated by the following figures: If a minority viewpoint is shared by 10% of the community, 28.2% of 12-member juries may be expected to have no minority representation, but 53.1% of 6-member juries would have none. Thirty-four percent of 12-member panels could be expected to have two minority members, while only 11% of 6-member panels would have two. As the numbers diminish below six, even fewer panels would have one member with the minority viewpoint and still fewer would have two. The chance for hung juries would decline accordingly.

Fourth, what has just been said about the presence of minority viewpoint as juries decrease in size foretells problems not only for jury decisionmaking, but also for the representation of minority groups in the community. The Court repeatedly has held that meaningful community participation cannot be attained with the exclusion of minorities or other identifiable groups from jury service. "It is part of the established tradition in the use of juries as instruments of public justice that the jury be a body truly representative of the community." Smith v. Texas, 311 U.S. 128, 130 (1940). The exclusion of elements of the community from participation "contravenes the very idea of a jury . . . composed of 'the peers or equals of the person whose rights it is selected or summoned to determine.'" Carter v. Jury Comm'n, 396 U.S. 320, 330 (1970), quoting Strauder v. West Virginia, 100 U.S. 303, 308 (1880). Although the Court in Williams concluded that the six-person jury did not fail to represent adequately a cross-section of the community, the opportunity for meaningful and appropriate representation does decrease with the size of the panels. Thus, if a minority group constitutes 10% of the community, 53.1% of randomly selected six-member juries could be expected to have no minority representative among their members, and 89% not to have two. Further reduction in size will erect additional barriers to representation.

Fifth, several authors have identified in jury research methodological problems tending to mask differences in the operation of smaller and larger juries. For example, because the judicial system handles so many clear cases, decisionmakers will reach similar results through similar analyses most of the time. One study concluded that smaller and larger juries could disagree in their verdicts in no more than 14% of the cases. Disparities, therefore, appear in only small percentages. Nationwide, however, these small percentages will represent a large number of cases. And it is with respect to those cases that the jury trial right has its greatest value. When the case is close, and the guilt or innocence of the defendant is not readily apparent, a properly functioning jury system will insure evaluation by the sense of the community and will also tend to insure accurate factfinding.

Studies that aggregate data also risk masking case-by-case differences in jury deliberations. The authors, H. Kalven and H. Zeisel, of The American Jury (1966), examined the judge-jury disagreement. They found that judges held for plaintiffs 57% of the time and that juries held for plaintiffs 59%, an insignificant difference. Yet case-by-case comparison revealed judge-jury disagreement in 22% of the cases. Id., at 63, cited in Lempert 656. This casts doubt on the conclusion of another study that compared the aggregate results of civil cases tried before 6-member juries with those of 12-member jury trials. The investigator in that study had claimed support for his hypothesis that damages awards did not vary with the reduction in jury size. Although some might say that figures in the aggregate may have supported this conclusion, a closer view of the cases reveals greater variation in the results of the smaller panels, i. e., a standard deviation of $58,335 for the 6-member juries, and of $24,834 for the 12-member juries. Again, the averages masked significant case-by-case differences that must be considered when evaluating jury function and performance.

IV

While we adhere to, and reaffirm our holding in Williams v. Florida, these studies, most of which have been made since Williams was decided in 1970, lead us to conclude that the purpose and functioning of the jury in a criminal trial is seriously impaired, and to a constitutional degree, by a reduction in size to below six members. We readily admit that we do not pretend to discern a clear line between six members and five. But the assembled data raise substantial doubt about the reliability and appropriate representation of panels smaller than six. Because of the fundamental importance of the jury trial to the American system of criminal justice, any further reduction that promotes inaccurate and possibly biased decisionmaking, that causes untoward differences in verdicts, and that prevents juries from truly representing their communities, attains constitutional significance.

Georgia here presents no persuasive argument that a reduction to five does not offend important Sixth Amendment interests. First, its reliance on Johnson v. Louisiana, 406 U.S. 356 (1972), for the proposition that the Court previously has approved the five-person jury is misplaced. In Johnson the petitioner challenged the Louisiana statute that permitted felony convictions on less-than-unanimous verdicts. The prosecution had to garner only nine votes of the 12-member jury to convict in a felony trial. The Court held that the statute did not violate the due process guarantee by diluting the reasonable-doubt standard. Id., at 363. The only discussion of the five-person panels, which heard less serious offenses, was with respect to the petitioner's equal protection challenge. He contended that requiring only nine members of a 12-person panel to convict in a felony case was a deprival of equal protection when a unanimous verdict was required from the 5-member panel used in a misdemeanor trial. The Court held merely that the classification was not invidious. Id., at 364. Because the issue of the constitutionality of the five-member jury was not then before the Court, it did not rule upon it.

Second, Georgia argues that its use of five-member juries does not violate the Sixth and Fourteenth Amendments because they are used only in misdemeanor cases. If six persons may constitutionally assess the felony charge in Williams, the State reasons, five persons should be a constitutionally adequate number for a misdemeanor trial. The problem with this argument is that the purpose and functions of the jury do not vary significantly with the importance of the crime. In Baldwin v. New York, 399 U.S. 66 (1970), the Court held that the right to a jury trial attached in both felony and misdemeanor cases. Only in cases concerning truly petty crimes, where the deprivation of liberty was minimal, did the defendant have no constitutional right to trial by jury. In the present case the possible deprivation of liberty is substantial. The State charged petitioner with misdemeanors under Ga. Code Ann. § 26-2101 (1972), and he has been given concurrent sentences of imprisonment, each for one year, and fines totaling $2,000 have been imposed. We cannot conclude that there is less need for the imposition and the direction of the sense of the community in this case than when the State has chosen to label an offense a felony. The need for an effective jury here must be judged by the same standards announced and applied in Williams v. Florida.

Third, the retention by Georgia of the unanimity requirement does not solve the Sixth and Fourteenth Amendment problem. Our concern has to do with the ability of the smaller group to perform the functions mandated by the Amendments. That a five-person jury may return a unanimous decision does not speak to the questions whether the group engaged in meaningful deliberation, could remember all the important facts and arguments, and truly represented the sense of the entire community. Despite the presence of the unanimity requirement, then, we cannot conclude that "the interest of the defendant in having the judgment of his peers interposed between himself and the officers of the State who prosecute and judge him is equally well served" by the five-person panel. Apodaca v. Oregon, 406 U.S., at 411 (opinion of WHITE, J.).

Fourth, Georgia submits that the five-person jury adequately represents the community because there is no arbitrary exclusion of any particular class. We agree that it has not been demonstrated that the Georgia system violates the Equal Protection Clause by discriminating on the basis of race or some other improper classification. See Carter v. Jury Comm'n, 396 U.S. 320 (1970); Smith v. Texas, 311 U.S. 128 (1940). But the data outlined above raise substantial doubt about the ability of juries truly to represent the community as membership decreases below six. If the smaller and smaller juries will lack consistency, as the cited studies suggest, then the sense of the community will not be applied equally in like cases. Not only is the representation of racial minorities threatened in such circumstances, but also majority attitude or various minority positions may be misconstrued or misapplied by the smaller groups. Even though the facts of this case would not establish a jury discrimination claim under the Equal Protection Clause, the question of representation does constitute one factor of several that, when combined, create a problem of constitutional significance under the Sixth and Fourteenth Amendments.

Fifth, the empirical data cited by Georgia do not relieve our doubts. The State relies on the Saks study for the proposition that a decline in the number of jurors will not affect the aggregate number of convictions or hung juries. Tr. of Oral Arg. 27. This conclusion, however, is only one of several in the Saks study; that study eventually concludes:

"Larger juries (size twelve) are preferable to smaller juries (six). They produce longer deliberations, more communication, far better community representation, and, possibly, greater verdict reliability (consistency)." Saks 107.

Far from relieving our concerns, then, the Saks study supports the conclusion that further reduction in jury size threatens Sixth and Fourteenth Amendment interests.

Methodological problems prevent reliance on the three studies that do purport to bolster Georgia's position. The reliability of the two Michigan studies cited by the State has been criticized elsewhere. The critical problem with the Michigan laboratory experiment, which used a mock civil trial, was the apparent clarity of the case. Not one of the juries found for the plaintiff in the tort suit; this masked any potential difference in the decisionmaking of larger and smaller panels. The results also have been doubted because in the experiment only students composed the juries, only 16 juries were tested, and only a video tape of the mock trial was presented. The statistical review of the results of actual jury trials in Michigan erroneously aggregated outcomes. It is also said that it failed to take account of important changes of court procedure initiated at the time of the reduction in size from 12 to 6 members. The Davis study, which employed a mock criminal trial for rape, also presented an extreme set of facts so that none of the panels rendered a guilty verdict. None of these three reports, therefore, convinces us that a reduction in the number of jurors below six will not affect to a constitutional degree the functioning of juries in criminal trials.

V

With the reduction in the number of jurors below six creating a substantial threat to Sixth and Fourteenth Amendment guarantees, we must consider whether any interest of the State justifies the reduction. We find no significant state advantage in reducing the number of jurors from six to five.

The States utilize juries of less than 12 primarily for administrative reasons. Savings in court time and in financial costs are claimed to justify the reductions. The financial benefits of the reduction from 12 to 6 are substantial; this is mainly because fewer jurors draw daily allowances as they hear cases. On the other hand, the asserted saving in judicial time is not so clear. Pabst in his study found little reduction in the time for voir dire with the six-person jury because many questions were directed at the veniremen as a group. Total trial time did not diminish, and court delays and backlogs improved very little. The point that is to be made, of course, is that a reduction in size from six to five or four or even three would save the States little. They could reduce slightly the daily allowances, but with a reduction from six to five the saving would be minimal. If little time is gained by the reduction from 12 to 6, less will be gained with a reduction from 6 to 5. Perhaps this explains why only two States, Georgia and Virginia, have reduced the size of juries in certain nonpetty criminal cases to five. Other States appear content with six members or more. In short, the State has offered little or no justification for its reduction to five members.

Petitioner, therefore, has established that his trial on criminal charges before a five-member jury deprived him of the right to trial by jury guaranteed by the Sixth and Fourteenth Amendments.

VI

The judgment of the Court of Appeals is reversed, and the case is remanded for further proceedings not inconsistent with this opinion.

It is so ordered.