United States v. Gary W. Jackman ( 1995 )

JON 0. NEWMAN, Chief Judge:

This appeal concerns a procedure adopted by clerical staff in the District Court for the District of Connecticut to remedy a jury selection process that had inadvertently, but systematically, excluded from petit jury veni-res all residents of Hartford and New Britain, communities with large minority populations. The adequacy of this procedure, no longer in use, is challenged by Gary W. Jackman on his appeal of judgment of the District Court (Alfred V. Covello, Jr., Judge), convicting him of bank robbery, in violation of 18 U.S.C. § 2113(a).¹ Because the procedure used to select the venire from which Jackman’s jury was selected violates the Sixth Amendment’s fair cross-section guarantee, we reverse.

Background

A. Prior History of Jury Selection Process

In United States v. Osorio, 801 F.Supp. 966 (D.Conn.1992), Judge T.F. Gilroy Daly examined the jury selection procedures for the Hartford Jury Division of the District and found that all residents of the cities of Hartford and New Britain had been excluded from the Division’s jury pool. Judge Daly ruled that the exclusion violated the Sixth Amendment guarantee of a fair cross-section of the community. At the time of the decision in Osorio, the Division’s venires were, at least in theory, composed by following the Plan for the Random Selection of Grand and Petit Jurors, adopted December 4, 1989. See Osorio, 801 F.Supp. at 969-71.

The selection process under this Plan began with the construction of a “Master Wheel.” This wheel was composed of ten percent of the names randomly chosen from the voter registration lists of each of the 84 towns and cities within the Hartford Division. From this Master Wheel of approximately 68,000 names, 1,500 names were randomly drawn twice each year, and jury questionnaires were sent to each of these names. When the questionnaires were returned, the jury clerk would divide the jurors into three categories: “qualified,” “unqualified,” and “uncertain.” A district judge would review the clerk’s assignments and make a final decision as to those in the “uncertain” category.

The names of those determined to be qualified were placed in the “Qualified Wheel,” the source of prospective jurors for all grand and petit jury venires called for service in Hartford. When jurors were needed, names from the Qualified Wheel, which contained in excess of a thousand questionnaires, were selected at random, and jury summonses were mailed to those individuals.

The Hartford Qualified Wheel, however, contained what Judge Daly aptly characterized as “startling anomalies.” Id. at 972. While the 1990 census indicated that 6.34% of the voting-age population in the Division is Black and 5.07% is Hispanic, Blacks accounted for only 3.08% of the individuals in the Qualified Wheel, and Hispanics accounted for only 0.77%. Id. at 972. Although at the time of Osorio, 4,631 juror questionnaires had been mailed by the clerk to voting-age residents whose names appeared on the Master Wheel, no questionnaire had ever been sent to a resident of Hartford or New Britain, the two largest cities in the Division. Id. at 972-73.

As a result, no resident of Hartford and New Britain was included in the Qualified Wheel, and none was ever summoned to serve in a venire. Because Hartford and New Britain contain 62.93% of the voting-age Black population and 68.09% of the voting-age Hispanic population in the Division, id. at 972, exclusion of residents of these cities from the Qualified Wheel resulted in a proportion of Blacks and Hispanics in the Qualified Wheel (the jury pool from which venires were selected) significantly lower than their proportion of the voting-age population.

At the time of the decision in Osorio, there was no explanation for the omission of Hartford and New Britain residents from the jury pool. Subsequently, it was discovered that no jury questionnaires were sent to Hartford residents because a computer programming error had caused the letter “d” in “Hartford” to communicate to the computer that all potential jurors from Hartford were deceased and thus unavailable for jury service. No explanation has ever been provided as to why the names of New Britain residents were never even entered into the computer.

To remedy this “systematic exclusion” of two-thirds of the Blacks and Hispanics in the Division, id. at 973-80, as well as to improve the overall fairness of the system, the District of Connecticut adopted a new jury selection plan. See Second Restated Plan for Random Selection of Grand and Petit Jurors Pursuant to Jury Selection and Service Act of 1968 (as amended) (D.Conn. Nov. 23,1992) (requiring, among other things, that Master Wheels be drawn from both voter registration and motor vehicle operator lists).

In addition, the Qualified Wheel condemned in Osorio was abandoned and, as an interim measure until the Second Restated Plan could be implemented, the Clerk’s Office directed the Yale University computer programmer responsible for the administration of the jury system to make sure that all names of New Britain residents drawn from voter registration lists were placed into the Master Wheel, and to correct the computer error that had led to the exclusion of Hartford residents. One thousand names were then drawn at random from the revised Master Wheel that included Hartford and New Britain residents, a jury questionnaire was sent to each name drawn, and, after elimination of unqualified and exempt persons, the remaining names constituted a new, representative Qualified Wheel. This interim measure, as such, is not challenged, but the Clerk’s procedure for drawing names for ve-nires in the aftermath of Osorio has precipitated the pending appeal. That procedure was ejqdained at a hearing in appellant’s case, to which we now turn.

B. Selection of Appellant’s Venire

On September 14, 1993, the jury was scheduled to be selected for appellant’s trial. Prior to the venire’s entry into the courtroom, appellant objected to being required to select a jury from a venire that contained no Blacks and only a single Hispanic. Judge Covello instructed Jackman to file a written motion, but jury selection began, and a jury was selected from the venire as constituted.

On September 21,1993, prior to the swearing of the jury, the Court conducted an evi-dentiary hearing on appellant’s motion. Maria Carpenter, the jury administrator (“jury clerk”) for the Hartford seat of court, testified that she maintained in her computer a pool of names of persons whom she contacted when a jury venire was required (“the Jury Clerk’s Pool”). These names, which came from the Qualified Wheel, first entered the Jury Clerk’s Pool from the summonses provided by the Yale computer programmer. Prior to the decision in Osorio, the clerk never received from the programmer a summons for anyone residing in Hartford or New Britain.

Upon the initial service of a summons upon a juror, the clerk would enter the juror’s name into the Jury Clerk’s Pool. Thereafter, she would notify the jurors in the Jury Clerk’s Pool to appear for venire panels that were convened during the course of the two-year period following the juror’s initial summons. The names of jurors remained in the Jury Clerk’s Pool for two years or thirty days of service, after which the juror was dismissed. When the Jury Clerk’s Pool required replenishing, the clerk obtained more summonses from the Yale programmer, who in turn obtained the names for these summonses from the Qualified Wheel.

When she was notified that a venire was required, the jury clerk would enter certain selection criteria into her computer in order to cull from the Jury Clerk’s Pool a “picking list.” This list, for example, omitted the names of jurors who had served the previous month. Before sending out letters notifying jurors to report, the clerk would remove from the “picking list” jurors whose two-year term of service was about to expire, jurors who had written to indicate that they would be out of state for business or a vacation, and jurors who were pregnant or who had just delivered a child.

The clerk reported that, after the Osorio decision and the abandonment of the old Qualified Wheel, Hartford and New Britain names were entered into the Master Wheel, from which a new Qualified Wheel was drawn. However, this new Qualified Wheel did not become the exclusive or even the primary source of names for jurors, a circumstance that is the basis of appellant’s Sixth Amendment challenge. The jury clerk did not create a new Jury Clerk’s Pool from the new Qualified Wheel. Instead, when a venire was required, the clerk would first create a “picking list” from those names in her computer, all of which had come from the pre-Osorio Qualified Wheel. If there were not enough names on the “picking list” for a venire, the clerk would supplement the list with some additional names from the new Qualified Wheel that included Hartford and New Britain.

For appellant’s jury selection, the clerk generated a “picking list” of 78 names from the Jury Clerk’s Pool — a pool that excluded Hartford and New Britain residents. Needing to summon 100 persons in order to have a venire of adequate size show up for jury selection, she then added 22 names drawn from the new Qualified Wheel and sent out letters to all 100 individuals. Some of these individuals were excused, reducing the list of those summoned to 80 persons, among whom there was one Hispanic and no Blacks. Of the 74 potential jurors appearing in court, 62 (84%) were from the Jury Clerk’s Pool (ie., drawn from the pre-Osorio list), and 12 (16%) were from the new Qualified Wheel. Two Hartford residents and four New Britain residents were in the venire. Two other Hartford residents were summoned, but did not appear in court.

At the September 21, 1993, hearing, the Government presented no evidence, but argued that the minority composition of appellant’s venire was “an isolated, one time incident” and that the presence of six individuals from Hartford and New Britain sufficiently met appellant’s concern. The District Court agreed, rejecting appellant’s fair cross-section challenge on the ground that the presence of Hartford and New Britain residents on his venire demonstrated that “there was no systematic exclusion,” and thus that the problem identified in Osorio had been solved. Following the jury’s guilty verdict on September 24, 1993, appellant raised his objections to the jury selection process again in a motion for a new trial, which the District Court denied.

On October 1, 1993, new Qualified Wheels constructed in accordance with the Second Restated Plan became available, and the Clerk’s Office ceased to draw jurors from any pool that existed prior to that date.²

Discussion

The Sixth Amendment requires that jury panels be drawn from a source representing a “fair cross section” of the community in which the defendant is tried. Taylor v. Louisiana, 419 U.S. 522, 536, 95 S.Ct. 692, 700-01, 42 L.Ed.2d 690 (1975); United States v. LaChance, 788 F.2d 856, 864 (2d Cir.), cert. denied, 479 U.S. 883, 107 S.Ct. 271, 93 L.Ed.2d 248 (1986). This fair cross-section requirement applies only to the larger pool serving as the source of names and not to the petit jury. See Taylor, 419 U.S. at 538, 95 S.Ct. at 701-02 (“Defendants are not entitled to a jury of any particular composition.”). In other words, the Sixth Amendment guarantees the opportunity for a representative jury venire, not a representative venire itself. Roman v. Abrams, 822 F.2d 214, 229 (2d Cir.1987), cert. denied, 489 U.S. 1052, 109 S.Ct. 1311, 103 L.Ed.2d 580 (1989).

On its face, the procedure adopted by the jury clerk to remedy the problem disclosed in Osorio appears to constitute systematic — though perhaps unintentional — exclusion. Osorio alerted the clerk to the fact that the old Qualified Wheel seriously underrepresented residents of Hartford and New Britain and consequently underrepresented Blacks and Hispanics. It was perhaps a defensible decision not to discard the names already drawn from that list and entered in the jury clerk’s computer, names that had taken considerable citizen and staff time to assemble. But if the names drawn from the pre-Osorio list were not to be discarded, it should have been apparent that these names would have to be used with caution and that some adjustment in the procedure for selecting venires would have to be made to remedy the underrepresentation that Osorio had disclosed. Once the new Qualified Wheel was assembled, one way of continuing to use the names already drawn from the old Qualified Wheel and placed in the Jury Clerk’s Pool would have been to select names at random from the new Qualified Wheel, and then add to the Jury Clerk’s Pool Hartford and New Britain residents drawn from the new Qualified Wheel in sufficient numbers so that the Jury Clerk’s Pool would reflect the percentages of Hartford and New Britain residents in the population of the Hartford Division.³ Whatever procedure was used, effective steps needed to be taken to assure that the pool from which names for venires would be drawn would no longer reflect the underrep-resentation that Osorio had disclosed.

Unfortunately, an adequate procedure was not adopted. Instead, the jury clerk used the Jury Clerk’s Pool, derived from the old Qualified Wheel that had totally excluded Hartford and New Britain residents, as a source of names for her “picking list” and supplemented her “picking list” in an entirely unacceptable manner. She drew from the new Qualified Wheel only the number of additional names that she needed in order to have a sufficient total number of persons to be summoned. Her remedy was bound to fail. Since the new Qualified Wheel was presumably representative and the old Qualified Wheel was definitely not, any random selection of names derived from both the old wheel and the new wheel, without some careful adjustment, would inevitably produce a “picking list” that permitted the underrepre-sentational effect of the old list to continue to taint the “picking list.”

But the jury clerk’s remedy was deficient beyond the mere use of the old unmodified list. Her approach made that list her primary source of names and supplemented it only to a slight extent. If the new Qualified Wheel had contributed most of the names to the “picking list,” the underrepresentational effect of using the names derived from the old list might have been so diluted that the resulting degree of underrepresentation in a typical “picking list” might have been negligible. But by making minimal use of the new list, the clerk assured that the underrepre-sentativeness of the old list would significantly taint each “picking list.”

In sum, the jury clerk started out with an unrepresentative pool of names and continued to use it so extensively, without sufficient adjustment, as to assure that her interim “system” would remain substantially unrepresentative. In focussing on the clerk’s actions, we do not mean to fault her individually or to fix responsibility for what occurred. Her actions are necessarily what we must review. But responsibility is another matter, one that is understandably not clear on this record, but is also of no relevance to the validity of appellant’s contention. Since the inadequacy of the procedure used was plainly accomplished by action within the Court, what occurred presents a cognizable constitutional claim. The responsibility might not be the clerk’s at all, it might be someone else’s, or, as often happens in overburdened courts (like other institutions), the failure to adopt a proper procedure might have resulted simply from the unwarranted assumptions by all concerned — in this instance, judges and staff — that the problem was being solved.

The facial inadequacy of the interim procedure adopted in the Hartford Division seems sufficient to sustain appellant’s claim. But, since the District Judge has ruled that the procedure, as applied in appellant’s case, yielded an adequate result, we proceed to analyze appellant’s claim in terms of the specific result. As Judge Covello recognized, to establish a prima facie violation of the Sixth Amendment, a defendant must prove that: (1) the group claimed to be excluded is distinctive in the community, (2) the representation of the group in the jury pool is not fair and reasonable in relation to the number of members of the group in the community, and (3) the underrepresentation is the result of systematic exclusion of the group in the jury selection process. Duren v. Missouri, 439 U.S. 357, 364, 99 S.Ct. 664, 668-69, 58 L.Ed.2d 579 (1979). The defendant need not prove discriminatory intent on the part of those constructing or administering the jury selection process. See id. at 368 n. 26, 99 S.Ct. at 670; Alston v. Manson, 791 F.2d 255, 258 (2d Cir.1986), cert. denied, 479 U.S. 1084, 107 S.Ct. 1285, 94 L.Ed.2d 143 (1987).

There is little question that both Blacks and Hispanies are “distinctive” groups in the community for purposes of this test. United States v. Biaggi, 680 F.Supp. 641, 648 (S.D.N.Y.1988), aff'd, 909 F.2d 662 (2d Cir.1990), cert. denied, 499 U.S. 904, 111 S.Ct. 1102, 113 L.Ed.2d 213 (1991).

The second element requires the Court to determine whether either or both of these two “distinctive” groups are “significant[ly] underrepresent[ed]” in the jury selection process. Biaggi, 909 F.2d at 677. The relevant comparison, for purposes of assessing the representativeness of the system, would normally be between the number of minority persons in the population and the number of persons belonging to the class found in the jury pool. See Duren, 439 U.S. at 365-66, 99 S.Ct. at 669-70; LaChance, 788 F.2d at 868. This case is unusual, however, because unlike Osorio, where the exclusive source of potential jurors was the old Qualified Wheel (which in turn generated the Jury Clerk’s Pool), here there was no single “pool” from which the jury clerk drew names to create a final “picking list.” Instead, as names were needed to summon each venire, the clerk would take names from the Jury Clerk’s Pool, which still excluded Hartford and New Britain residents, and add a few names from the new Qualified Wheel, which included Hartford and New Britain residents, to make up for the shortfall.

Here, for example, the clerk took 78 names from the Jury Clerk’s Pool and then added 22 names from the post-Osorio wheel. Apart from those 100 names, generated solely to create a “picking list” for summoning appellant’s venire, there is no larger pool or pre-established source list that can be examined to determine the overall representativeness of the system with respect to either Hartford and New Britain residents or Blacks and Hispanies.

To assess the representativeness of the “picking list” procedure, the amici curiae suggest that we should examine what they call a “functional wheel,” a construct reflecting the wheel that would be assembled by drawing names from the old Qualified Wheel and the new Qualified Wflieel in the same proportion as those wheels served as the source for the 100 names on appellant’s “picking list,” i.e., 78% from the old Qualified Wheel and 22% from the new Qualified Wheel. In this “functional wheel,” the amici calculate, Blacks would be 3.8% and Hispanics would be 1.72%. These figures yield slightly smaller variations from the corresponding proportions of adults in the relevant population than occurred in Osorio, reflecting the 22% contribution of the new Qualified Wheel. Apprehending that under the so-called “absolute numbers” ⁴ or “absolute impact” approach, see United States v. Biaggi, 909 F.2d at 678; United States v. Jenkins, 496 F.2d 57, 65 (2d Cir.1974), cert. denied, 420 U.S. 925, 95 S.Ct. 1119, 43 L.Ed.2d 394 (1975), the underrepresentation in the “functional wheel” might not be regarded as substantial, amici urge us not to use the “absolute numbers” approach in this case. We agree that this approach, even as applied to the construct of the “functional wheel,” is inappropriate.

First, in Biaggi we recognized that the absolute numbers approach is of questionable validity when applied to an underrepresented group that is a small percentage of the total population, because an underrepresentation of such a group that can be “remedied” by adding only one or two members to a typical venire can lead to the selection of a large number of venires in which members of the group are substantially underrepresented or even totally absent. See Biaggi, 909 F.2d at 678. This problem is particularly acute here because of the small percentage of voting-age Blacks and Hispanies residing within the Hartford Division (6.34% and 5.07% respectively, as compared with 19.9% and 15.7% within the Manhattan Master Wheel of the Southern District of New York in Biaggi ).⁵ See Osorio, 801 F.Supp. at 978-79.

Second, although we tolerated the disparities in Biaggi, we noted that the degree of underrepresentation there “press[ed] the Jenkins ‘absolute numbers’ approach to its limit” and that we “would find the Sixth Amendment issue extremely close if the un-derrepresentations had resulted from any circumstance less benign than use of voter registration lists.” Biaggi, 909 F.2d at 679. The facts in this case reveal circumstances far less benign than those in Biaggi or, indeed, even those in Osorio, where the apparently inadvertent exclusion of Hartford and New Britain residents had not previously been discovered. The underrepresentation of Hartford and New Britain residents continued for more than a year after disclosure of constitutional infirmities in the selection process. The procedure adopted in the aftermath of Osorio was not designed to assure fair representation of either the populations of those cities or the significant minority populations living there. Names from the pre-Osorio wheel remained the principal source of potential jurors and were supplemented only to the limited extent necessary to satisfy the need for enough jurors to fill out a venire, rather than in sufficient numbers to assure that the selection system would adequately reflect a cross-section of the community.

The more basic flaw in trying to justify what occurred in appellant’s case by an absolute numbers approach is that, as stated earlier, there does not exist a pool, larger than the various “picking lists,” that was uniformly used as the source for venires. And if we regard appellant’s “picking list” as the source, we lack the data to show the demographic breakdown of that list. What is left is appellant’s venire, on which we know there were no Blacks and only one Hispanic. The absolute numbers approach, even if otherwise applicable, cannot save underrepre-sentation to that degree, at least where no broad source of names exists to support the argument that the representations in the ve-nire selected in any one ease were only slight departures from the representations in average venires that would regularly be selected.

For all of these reasons, we conclude that appellant has shown a significant level of underrepresentation of Blacks and Hispanics for purposes of the Sixth Amendment.

We also conclude that appellant has established that this underrepresentation was the result of a “systematic exclusion in the jury-selection process,” Duren, 439 U.S. at 366, 99 S.Ct. at 669, specifically the clerk’s decision to select most potential jurors from the pre-Osorio Jury Clerk’s Pool. Like the prior total exclusion of Hartford and New Britain residents from the jury pool, the underrepre-sentation here “was quite obviously due to the system by which juries were selected.” Id. at 367, 99 S.Ct. at 670. The fact that appellant’s venire included some residents of Hartford and New Britain (6 out of 74 or only 8.1%, compared to the 16.6% in the relevant population) does not defeat appellant’s challenge, since the existence of systematic underrepresentation turns on the process of selecting venires, not the outcome of that process in a particular case.

In sum, we conclude that appellant has established a prima facie fair cross-section violation. Though the Government is entitled to attempt to rebut appellant’s claim by demonstrating a significant interest in the procedure used, see Duren, 439 U.S. at 367, 99 S.Ct. at 670, no such interest has been shown here. The Government has provided no justification for failing to make the Jury Clerk’s Pool more representative of the community, or for continuing to rely on the pre-Osorio old Qualified Wheel as the primary source of names, with only a sporadic supplementation unrelated to achieving proper representation. Since the Government has failed to rebut appellant’s prima facie showing, the underrepresentation of Blacks and Hispanics as a result of the underrepresentation of residents from both Hartford and New Britain in the jury selection system violated appellant’s Sixth Amendment right to a fair cross-section of the community on his jury venire.

We note, however, that our ruling will not necessarily jeopardize judgments from other cases that were tried during this same period using the old Jury Clerk’s Pool. This is not a case “where the public’s interest in enforcing judgments outweighs its interest in broad participation in the administration of justice.” Osorio, 801 F.Supp. at 976. Thus, in the absence of a timely objection to the jury selection process, courts will retain the discretion to uphold convictions.

Conclusion

We reverse appellant’s conviction and order a new trial.

1 . Appellant also asserts a violation of the Jury Selection and Service Act, 28 U.S.C. § 1861 et seq., and challenges both the District Court’s failure to order disclosure of certain evidence and its decision to admit an out-of-court statement pursuant to the residual hearsay exception of Fed.R.Evid. 804(b)(5). In addition, amici curiae urge that we consider an equal protection claim, which appellant raised unsuccessfully below in objecting to the jury selection process but which he appears to have abandoned on appeal. In light of our ruling on his fair cross-section challenge, however, we need not consider these other arguments for reversal. Since we are not considering Jackman’s claim under the Jury Act, we express no view as to whether this claim was timely asserted under the Act or whether the District Court abused its discretion in denying a continuance to permit a full exploration of the claim before jury selection.

2 . Sometime after the September 21, 1993, evi-dentiary hearing in this case, then-Chief Judge Cabranes appointed a committee to examine and report upon the District’s jury selection process. The findings of the committee's report confirmed that after Osorio, the clerk continued to draw names for each venire primarily from the pre-Osorio Jury Clerk’s Pool, but supplemented those names when necessary by drawing from the new Qualified Wheel.

3 . Perhaps some adjustment of this solution, such as entering names in the Jury Clerk's Pool and also returning them to the new Qualified Wheel, might have been necessary to avoid creating un-derrepresentation in the new Qualified Wheel.

4 . The absolute numbers approach is a measure of the degree of underrepresentation accomplished by determining how many individuals from the distinctive group would have to be added to a randomly drawn venire to make that venire representative with respect to the excluded or underrepresented group. See Jenkins, 496 F.2d at 65. The analysis involves two steps. First, the Court must calculate "the difference between a group's proportion of the population and its proportion in the relevant jury pool.” United States v. Gerena, 677 F.Supp. 1266, 1271 (D.Conn.1987), aff'd, United States v. Maldonado-Rivera, 922 F.2d 934 (2d Cir.1990), cert. denied, 501 U.S. 1211, 111 S.Ct. 2811, 115 L.Ed.2d 984 (1991). Next, that percentage difference is multiplied by the size of a typical venire in order to determine the number of individuals from the distinctive group that have to be added to the venire, on average, to redress the group’s under-representation. Id.

5 . A brief review of some pertinent numbers makes the point more precisely.

If we assume a 75-person venire, in Biaggi, a perfectly representative wheel for jury selection would have yielded only a .0003% chance that a particular venire would have no Hispanies, while the actual wheel created a .016% chance of no Hispanies in any given venire. In other words, with either wheel in Biaggi, the likelihood of a venire with no Hispanies was minimal. In this case, however, if we use for purposes of argument the numbers derived from the “functional wheel” suggested by the amici, a perfectly representative wheel would have yielded only a 2% chance that a particular venire would have no Hispanies, while the functional wheel created a 27.2% chance of no Hispanies in any given veni-re. Thus, while a perfectly representative wheel would exclude Hispanies in only one out of every fifty venires, the technique used in this case would do so in more than one out of every four venires. Moreover, the wheel used in Biaggi created only a 2.88% chance that any given veni-re would have less than one-fourth of the number of Hispanies that would appear in a representa-five venire, while the functional wheel, appropriate for analysis in this case, created a 62.94% chance of such a substantially underrepresenta-tive venire.

The corresponding statistics for Blacks are not quite as dramatic, but they reflect a similar pattern in the likelihood of substantial underrepre-sentation, and reveal similar differences between this case and Biaggi. These numbers demonstrate that, due to the relatively small size of the relevant minority populations here, appellant's opportunity for a representative jury venire in this case was dramatically worse than the defendant’s opportunity for a representative jury veni-re in Biaggi —a case that we said was at the very edge of acceptability.

For a review of the relevant formulas used to derive these figures (all of them mathematical measures commonly employed in statistical analysis), see generally Michael O. Finkelstein & Bruce Levin, Statistics For Lawyers (1990), Peter A. Detre, Note, A Proposal for Measuring Underrepresentation in the Composition of the Jury Wheel, 103 Yale L.J. 1913 (1994), and D.H. Kaye, Statistical Analysis in Jury Discrimination Cases, 25 Jurimetrics J. 274 (1985).