399 F. Supp. 683 | D. Colo. | 1975
UNITED STATES of America, Plaintiff,
v.
John E. TEST, Defendants.[*]
United States District Court, D. Colorado.
*684 *685 James L. Treece, U. S. Atty., by J. Terry Wiggins and Richard Slivka, Asst. U. S. Attys., Denver, Colo., for plaintiff.
Walter Gerash, Denver, Colo., for defendants Test, Tosti, and Marquez.
Robert C. Floyd, Denver, Colo., for defendants Hermanson and Chavez.
Leonard E. Davies and Michael G. Sabbeth, Denver, Colo., for defendants Gutierrez, O'Malley, Jackson, and Hudson.
Forrest C. O'Dell, Denver, Colo., for defendant Moorer.
Joseph Saint-Veltri, Denver, Colo., for defendants Small and Allen.
MEMORANDUM AND ORDER
ARRAJ, Chief Judge.
The above-captioned criminal actions have been consolidated for hearing on defendants' individual motions challenging the Jury Selection Plan of this judicial district. The motions are predicated on allegations that both the grand and petit jurors selected under the Plan are not drawn from a "fair cross section of the community" as required by the Jury Selection and Service Act of 1968 [28 U. S.C. § 1861 et seq., as amended]. It is further alleged that the selection procedures violate the Fifth and Sixth Amendments to the United States Constitution.
As originally filed, defendants' motions cite four grounds for attacking the Jury Selection Plan: (1) that the use of voter registration lists as the source of prospective jurors results in significant underrepresentation of Blacks, Chicanos, American Indians, students, and young persons in the Master Jury Wheel;[1] (2) that the Plan does *686 not assure "random selection" of jurors from the master wheel; (3) that excuses and/or exemptions from jury service granted to certain categories of persons tend to systematically exclude "poor women" and "poor people" from the pool of available jurors; and (4) that the Plan fails to supplement voter registration lists with other sources of prospective jurors so as to alleviate the aforementioned underrepresentation of minority groups.
Under the mandate of Test v. United States, 420 U.S. 28, 95 S. Ct. 749, 42 L. Ed. 786 (1975), we granted the motions of all defendants to inspect the Master and Qualified Jury Wheels used in this district since January 1972. Upon completion of their inspection, defendants were ordered to submit "statements of intended proof" and memoranda of law in support of their positions. On June 19, 1975, we conducted a hearing during which defendants offered documentary evidence and testimony concerning the alleged defects in the Jury Selection Plan. After considering this evidence and the arguments of counsel, we are now prepared to rule on defendants' motions.
I. Defendants' Evidence
During the course of the hearing, counsel for defendants conceded that they had no evidence of systematic exclusion or underrepresentation of American Indians, students, young persons, poor women, or poor people. They also reported that in comparing the racial and ethnic make-up of the Master Jury Wheel with that of the Qualified Jury Wheel, no "serial dilution" of cognizable groups was discovered. This indicates that any failings of the Jury Selection Plan do not stem from the process of qualifying jurors for service once their names are drawn from the original source list. Consequently, all parties agree that defendants' challenge to the Plan is limited to the alleged underrepresentation of Blacks and Chicanos in the voter registration list used to fill the Master Jury Wheel.[2]
For purposes of making the relevant analysis, defendants employed Dr. George E. Bardwell, a professor of mathematics and statistics, who was qualified as an expert witness at the evidentiary hearing. Starting with demographic information contained in the 1970 United States census, Dr. Bardwell determined the percentage of voting-aged Chicanos and Blacks in the population of each of the three jury divisions in this district.[3] These are referred to as the Denver, Grand Junction, and Pueblo divisions. He then analyzed five random samples drawn from the Master and Qualified Jury Wheels for the various divisions. The percentage comparisons and the statistical conclusion drawn therefrom for each data set are as follows:
*687
A. Denver Division From January 1972 through June 1973 (sample size, 723)
Percentage in Percentage in Statistical
Race Random Sample Voting-Aged Population Conclusion
------------------------------------------------------------------------
| | | | Statistically |
|Chicano | 4.88% | 8.93% | Significant |
| | | | Difference |
|---------|----------------|---------------------------|-----------------|
| | | | Not |
|Black | 1.94% | 3.00% | Statistically |
| | | | Significant |
========================================================================
B. Denver Division from January 1973 through December 1974 (sample size, 2, 020)
Percentage in Percentage in Statistical
Race Random Sample Voting-Aged Population Conclusion
------------------------------------------------------------------------
| | | | Statistically |
|Chicano | 6.73% | 8.93% | Significant |
| | | | Difference |
|---------|----------------|---------------------------|-----------------|
| | | | Not |
|Black | 2.48% | 3.00% | Statistically |
| | | | Significant |
========================================================================
C. Grand Junction Division on July 2, 1973 (sample size, 832)
Percentage in Percentage in Statistical
Race Random Sample Voting-Aged Population Conclusion
------------------------------------------------------------------------
| | | | Statistically |
|Chicano | 4.81% | 8.89% | Significant |
| | | | Difference |
|---------|----------------|---------------------------|-----------------|
| | | | Not |
|Black | 0.00% | 0.21% | Statistically |
| | | | Significant |
=========================================================================
D. Pueblo Division on July 6, 1973 (sample size, 872)
Percentage in Percentage in Statistical
Race Random Sample Voting-Aged Population Conclusion
------------------------------------------------------------------------
| | | | Statistically |
|Chicano | 12.84% | 16.29% | Significant |
| | | | Difference |
|---------|----------------|---------------------------|-----------------|
| | | | Statistically |
|Black | 1.49% | 2.90% | Significant |
| | | | Difference |
========================================================================
E. Denver Division on July 30, 1973 and May 29, 1974 (composite sample size, 2111)
Percentage in Percentage in Statistical
Race Random Sample Voting-Aged Population Conclusion
------------------------------------------------------------------------
| | | | Statistically |
|Chicano | 6.20% | 8.93% | Significant |
| | | | Difference |
|---------|----------------|---------------------------|-----------------|
| | | | Statistically |
|Black | 1.94% | 3.00% | Significant |
| | | | Difference[4] |
========================================================================
*688 On the basis of this evidence, and particularly the statistical conclusions of Dr. Bardwell, defendants contend that the Jury Selection Plan of this district is both unconstitutional and violative of the Jury Selection and Service Act of 1968. For the reasons set forth below, we disagree.
II. The Constitutional Claim
In a line of decisions spanning nearly a century, the United States Supreme Court has fashioned certain basic principles for judging the constitutionality of jury selection systems. [See, e. g., Taylor v. Louisiana, 419 U.S. 522, 95 S. Ct. 692, 42 L. Ed. 2d 690 (1975); Alexander v. Louisiana, 405 U.S. 625, 92 S. Ct. 1221, 31 L. Ed. 2d 536 (1972); Turner v. Fouche, 396 U.S. 346, 90 S. Ct. 532, 24 L. Ed. 2d 567 (1970); Sims v. Georgia, 389 U.S. 404, 88 S. Ct. 523, 19 L. Ed. 2d 634 (1967); Whitus v. Georgia, 385 U.S. 545, 87 S. Ct. 643, 17 L. Ed. 2d 599 (1967); Swain v. Alabama, 380 U.S. 202, 85 S. Ct. 824, 13 L. Ed. 2d 759 (1965); Eubanks v. Louisiana, 356 U.S. 584, 78 S. Ct. 970, 2 L. Ed. 2d 991 (1958); Avery v. Georgia, 345 U.S. 559, 73 S. Ct. 891, 97 L. Ed. 1244 (1953); Cassell v. Texas, 339 U.S. 282, 70 S. Ct. 629, 94 L. Ed. 839 (1950); Patton v. Mississippi, 332 U.S. 463, 68 S. Ct. 184, 92 L. Ed. 76 (1947); Akins v. Texas, 325 U.S. 398, 65 S. Ct. 1276, 89 L. Ed. 1692 (1945); Smith v. Texas, 311 U.S. 128, 61 S. Ct. 164, 85 L. Ed. 84 (1940); Norris v. Alabama, 294 U.S. 587, 55 S. Ct. 579, 79 L. Ed. 1074 (1935); Carter v. Texas, 177 U.S. 442, 20 S. Ct. 687, 44 L. Ed. 839 (1900); Strauder v. West Virginia, 100 U.S. 303, 25 L. Ed. 664 (1879)] From these precedents, it is clear that a criminal defendant has no constitutional right to a grand or petit jury which proportionally reflects the various "cognizable" groups within the community. Nor is he entitled to proportional representation of such groups "on the venire or jury roll from which petit [or grand] jurors are drawn." [Swain v. Alabama, supra, 380 U.S. at 208, 85 S.Ct. at 829] What is guaranteed by the Sixth Amendment and the Due Process Clauses of the Fifth and Fourteenth Amendments are grand and petit juries selected from source lists, and by procedures, which are free from discrimination. As the Court stated in Apodaca v. Oregon, 406 U.S. 404, 92 S. Ct. 1628, 32 L. Ed. 2d 184 (1972):
All that the Constitution forbids . . . is systematic exclusion of identifiable segments of the community from jury panels and from the juries ultimately drawn from those panels . . .. [Id. at 413, 92 S. Ct. at 1634 (emphasis added)]
In applying these principles, courts have uniformly held that the challenger of a jury selection system must prove "that a recognizable, identifiable class of persons, otherwise entitled to be jurymembers, has been purposefully and systematically excluded from jury service." [Leggroan v. Smith, 498 F.2d 168, 170 (10th Cir. 1974) (emphasis added), citing, Brown v. Allen, 344 U.S. 443, 73 S. Ct. 397, 97 L. Ed. 469 (1953)] Consequently, a successful challenge to a jury plan depends upon proof of a prima facie case, comprised of two basic elements: (1) evidence that some "distinctive" or "cognizable" group or groups, (2) have been "purposefully and systematically excluded" from jury service.
The requirement of "cognizability" derives primarily from Hernandez v. Texas, 347 U.S. 475, 74 S. Ct. 667, 98 L. Ed. 866 (1954), wherein the Supreme Court held:
Throughout our history differences in race and color have defined easily identifiable groups which have at times required the aid of the courts in securing equal treatment under the laws. But community prejudices are not static, and from time to time other differences from the community norm may define other groups which need the same protection. Whether such a group exists within a community is a question of fact. When the existence of a distinct class is demonstrated, and it is further shown that the laws, as written or as applied, single out *689 that class for different treatment not based on some reasonable classification, the guarantees of the Constitution have been violated.
. . . . . .
The petitioner's initial burden in substantiating his charge of group discrimination was to prove that persons of Mexican descent constitute a separate class in Jackson County, distinct from "whites." [Id. at 478, 479, 74 S.Ct. at 670 (footnote omitted)]
The burden of proving cognizability is upon the challenger of the jury plan. [See, e. g., United States v. De Alba-Conrado, 481 F.2d 1266, 1270 (5th Cir. 1973)] In United States v. Guzman, 337 F. Supp. 140 (S.D.N.Y.), aff'd, 468 F.2d 1245 (2d Cir. 1972), cert. denied, 410 U.S. 937, 93 S. Ct. 1397, 35 L. Ed. 2d 602 (1973), the trial court set forth the following factors to be considered in determining cognizability: (1) the presence of some quality or attribute which "defines and limits" the group; (2) a cohensiveness of "attitudes or ideas or experience" which distinguishes the group from the general social milieu; and (3) a "community of interest" which may not be represented by other segments of society. [Id. at 143-44]
In the case of Blacks, these factors are often presumed to exist [see, e. g., Whitcomb v. Chavis, 403 U.S. 124, 91 S. Ct. 1858, 29 L. Ed. 2d 363 (1971); Cassell v. Texas, supra], and in the case of Chicanos, there are "accepted" compilations of Spanish surnames which "define" a cognizable group [see, e. g., Hernandez v. Texas, supra, 347 U.S. at 480, n. 12, 74 S. Ct. 667] The cognizability of "other groups" may be more difficult to demonstrate, but defendants here have limited the basis of their challenge to underrepresentation of Blacks and Chicanos. As to the former, we presume cognizability. As to the latter, we are satisfied that defendants have used an "accepted" list of Spanish surnames in collecting their data,[5] and that the evidence before us does relate to a cognizable group.
The second element of a prima facie case namely proof of purposeful and systematic exclusion is a more formidable obstacle for defendants. The precise meaning of the term, "purposeful and systematic exclusion," has been the subject of some controversy. [See, United States v. Butera, 420 F.2d 564, 568-69 (1st Cir. 1970)] But a majority of courts now recognize that a prima facie case may be established without proof of a specific intent or willful scheme to discriminate. In the words of Judge Gewin in An Analysis of Jury Selection Decisions, appendix to Foster v. Sparks, 506 F.2d 805, 811 (5th Cir. 1975):
The Supreme Court has long acknowledged the hazards involved in testing constitutionality by legislative motive and has been equally sensitive to charges of sophisticated as well as naive modes of discrimination. Moreover, it has repeatedly entertained contentions that a legislative or administrative scheme operates to discriminate without requiring that evidence of purposeful discrimination be adduced. . . . [P]roof which evidences neglect or the arbitrary quality of thoughtlessness involving clear overtones of racial discrimination and resulting in the same evils which characterize intentional or purposeful discrimination should suffice. Ignorance of the inadequacy of selection procedures would be as reprehensible as knowledge of such inadequacy and unwillingness to rectify.
The meaning of discriminatory purpose could thence be explained in two ways. One such explanation would be that the term connotes malevolent motive. Since this is irreconcilable with decisions which decry well-intentioned but nevertheless deliberate exclusions, the second and preferable explanation is that purpose connotes deliberate, *690 intentional or knowing as opposed to inadvertent interference with the selection system, irrespective of the design with which such deliberate interference is undertaken. [Id. at 822-23 (emphasis added) (footnotes omitted)]
We agree with this reasoning. In Alexander v. Louisiana, supra, the Court struck down a jury system notwithstanding the absence of any evidence of specific discriminatory intent. Moreover, in its most recent jury selection decision the Supreme Court seemingly abandoned the rubric of "purposeful discrimination," stating simply that:
[T]he jury wheels, pools of names, panels or venires from which juries are drawn must not systematically exclude distinctive groups in the community and thereby fail to be reasonably representative thereof. [Taylor v. Louisiana, supra, 419 U.S. at 538, 95 S.Ct. at 702]
But there is another implication to this language which is far more critical to defendants' position. In holding that the Constitution prohibits jury selection systems which "systematically exclude distinctive groups . . . and thereby fail to be reasonably representative . . .." [Id. (emphasis added)], the Court acknowledged a causal link between the "deliberate interference" to which Judge Gewin makes reference and the "underrepresentation" upon which defendants here rely. As observed earlier, the Constitution does not require "mathematical representativeness;" nor does it prohibit "mathematical underrepresentativeness," as such. What is constitutionally proscribed is that underrepresentatives of distinct groups which is caused by systematic exclusion, or stated alternatively, by "deliberate interference" with the natural laws of probability and sampling.
There are then two ways in which defendants might sustain their burden. First, if they prove that there are affirmative barriers to voter registration by Blacks or Chicanos in Colorado, then the use of Colorado voter registration lists by this Jury Plan would be unconstitutional, even if the district did not adopt those lists with the specific intent to discriminate. Second, if during the process of qualifying jurors drawn from the original source lists, there is a "progressive decimation" or "serial dilution" of the proportion of Blacks or Chicanos available for jury service, and if there is some "opportunity" for "deliberate interference" with the selection process, then again a prima facie case of unconstitutionality is established. [See, e. g., Alexander v. Louisiana, supra But even assuming arguendo that the underrepresentation of Chicanos and Blacks in evidence here constitute a disparity of constitutionally significant proportions, defendants' claims must still fail.[6] There is simply no evidence here of "systematic exclusion." As previously *691 noted, defendants' own investigations disclosed that no "serial dilution" of racial or ethnic groups occurred during the process of qualifying jurors drawn from the original source list. And, although defendants allege that, for various reasons, many Chicanos and Blacks are not registering to vote in Colorado, we have been offered no evidence of "deliberate interference" with voter registration in this state.[7] [See also, United States v. Whitley, 491 F.2d 1248 (8th Cir.), cert. denied, 416 U.S. 990, 94 S. Ct. 2399, 40 L. Ed. 2d 769 (1974)]
Upon the foregoing, we hold that defendants have failed to make a prima facie showing that the Jury Selection Plan of this district violates the guarantees of the Fifth and Sixth Amendments.
III. The Statutory Claim
Under the overwhelming majority of decisions which address the issue, a failure to establish systematic exclusion is fatal not only to a constitutional challenge to a jury plan but also to any challenge based on the Jury Selection Act. [See, e. g., Government of the Virgin Islands v. Navarro, 513 F.2d 11, 19 (3d Cir. 1975); Hallman v. United States, 490 F.2d 1088, 1091-92 (8th Cir. 1973); United States v. Greene, 160 U.S.App. D.C. 21, 489 F.2d 1145, 1149-50 (1973) (dictum), cert. denied, 419 U.S. 977, 95 S. Ct. 239, 42 L. Ed. 2d 190 (1974); United States v. Hamling, 481 F.2d 307, 313-14 (9th Cir. 1973), aff'd, 418 U.S. 87, 94 S. Ct. 2887, 41 L. Ed. 2d 590 (1974); United States v. Ware, 473 F.2d 530, 537 (9th Cir. 1973); United States v. Lewis, 472 F.2d 252, 256 (3d Cir. 1973); United States v. Guzman, 468 F.2d 1245, 1247-48 (2d Cir. 1972), cert. denied, 410 U.S. 937, 93 S. Ct. 1397, 35 L. Ed. 2d 602 (1973); United States v. Ross, 468 F.2d 1213, 1217 (9th Cir. 1972), cert. denied, 410 U.S. 989, 93 S. Ct. 1500, 36 L. Ed. 2d 188 (1973); United States v. James, 453 F.2d 27, 29 (9th Cir. 1971); United States v. Parker, 428 F.2d 488, 489 (9th Cir.), cert. denied, 400 U.S. 910, 91 S. Ct. 155, 27 L. Ed. 2d 150 (1970); Camp v. United States, 413 F.2d 419, 421 (5th Cir.), cert. denied, 396 U.S. 968, 90 S. Ct. 451, 24 L. Ed. 2d 434 (1969); United States v. DiTommaso, 405 F.2d 385, 390-91 (4th Cir. 1968), cert. denied, 394 U.S. 934, 89 S. Ct. 1209, 22 L. Ed. 2d 465 (1969); United States v. Johnson, 386 F. Supp. 1034 (W.D.Pa.1974)] A case in point is Hamling v. United States, supra, wherein the defendants moved for a continuance of their trial pursuant to 28 U.S.C.A. § 1867(a). As grounds for their motion, they cited the fact that the master jury wheel for the Southern District of California had not been refilled for approximately four years. Thus, persons between the ages of 21 and 24 could not possibly be included in the voter registration list used to fill the *692 wheel, and "young persons" as a group were, according to defendants, grossly underrepresented in the pool of prospective jurors. The trial court denied the motion, and the court of appeals affirmed. The issue was presented to the United States Supreme Court which held:
The Court of Appeals assumed, without deciding, that the young do constitute a cognizable group or class, but concluded that petitioners had "failed to show, let alone establish, a purposeful systematic exclusion of the members of that class whose names, but for such systematic exclusion would otherwise be selected for the master jury wheel," and therefore that the District Court's refusal to grant a continuance was not an abuse of discretion. 481 F.2d, at 314. We agree with the Court of Appeals.
. . . . . .
Since petitioners failed to establish a discriminatory exclusion of the young from their jury, the District Court properly exercised its discretion in refusing to grant petitioners' motion for a continuance. [418 U.S. at 137, 138, 94 S.Ct. at 2917]
This language clearly suggests to us that the "rule of exclusion" also controls in jury challenges brought under the Act. Furthermore, we believe it is important that the underrepresentation of Chicanos and Blacks in the instant case has no proven cause other than the failure of members of these groups to register to vote in the same proportions as other groups in the community. Courts have uniformly held that: "[A] group of persons who choose not to vote do not constitute a `cognizable group.'" [United States v. Lewis, supra at 256 (footnote omitted); see also, United States v. Freeman, 514 F.2d 171, 173 (8th Cir. 1975)] Following these precedents, we would deny both the constitutional and statutory claims of defendants without further inquiry.
However, it is defendants' position that in enacting the Jury Selection and Service Act, Congress established statutory standards for judging the legality of jury plans which are more rigorous than the constitutional standard of "systematic exclusion." Specifically, they contend that under 28 U.S.C.A. §§ 1861 and 1867(a), any jury plan which "substantially fails" to draw jurors from a "fair cross section of the community" violates the Act. Further, they argue that a "statistically significant underrepresentation" of cognizable groups constitutes a "substantial failure to comply" with the "fair cross section" requirement, irrespective of the cause of such underrepresentation. They maintain that whenever such underrepresentation occurs, the court is under an affirmative duty to supplement the voter registration lists so as to obtain a better cross section.
In making this argument, defendants rely, in part, on the following language from the Government's Brief in Test v. United States, supra:
We believe . . . that the Act created broader rights than those established under Sixth Amendment principles, and, specifically, that substantial underrepresentation of a cognizable group (such as blacks or Mexican-Americans) would necessitate resort to supplementary sources under Section 1863 (if it has a substantial impact on selection of jurors . . .) even though no history of voter discrimination against the group is proved. [Brief for Appellee at 13-14 (footnote omitted)]
At least three reported decisions seem to support this construction of the Act. [United States v. Jenkins, 496 F.2d 57, 64-66 (2d Cir. 1974), cert. denied, 420 U.S. 925, 95 S. Ct. 1119, 43 L. Ed. 2d 394 (1975); United States v. Fernandez, 480 F.2d 726, 733 (2d Cir. 1973); United States v. McDaniels, 370 F. Supp. 298, 301-02 (E.D.La.1973), aff'd sub nom, *693 United States v. Goff, 509 F.2d 825 (5th Cir. 1975)] We find it unnecessary to add our opinion to this debate, for even if we adopt defendants' conception of the Act, their evidence falls short of establishing a "substantial failure to comply" with the "fair cross section" requirement.
The pivotal issues here are two-fold: (1) What is meant by a "fair cross section of the community?" and (2) What is a "substantial" deviation from that cross section? The term, "fair cross section" is not defined by the Act, and legislative history reflects a Congressional intent to "leave the definition of `substantial' to the process of judicial decision." [H.R.Report No. 1076, U.S.Code Cong. & Adm.News at 1794 (1968)]
Several possible definitions of these terms have been suggested, the most prominent of which are summarized by Judge Gewin in his article, An Analysis of Jury Selection Decisions, supra. One proposal is the "percentage deficiency test," which is described as follows:
If a 20% "deficiency" is to be allowed, this means that where nonwhites comprise 30% of the eligible population, they must comprise no less than 24% of the names in the jury wheel. The 20% deficiency standard is computed as the difference between 30% and 24% of 6%, expressed as a proportion of 30%. Thus, 6/30 = 20%. In other words, the disparity of 6% between nonwhite share in voting age population and nonwhite share in the jury wheel is 20% of the nonwhite percentage of the population. (6/30 = 20%) [506 F.2d at 818 (footnote omitted)]
Of course, if a larger percentage deficiency, say 30%, is to be permitted, then greater deviations from the population percentage will be tolerated.
A second formulation is one suggested by Professor Henry B. Moore of the University of Alabama. In contrast to the "percentage deficiency" test, which seeks to define "substantial deviation," the Moore test emphasizes the concept of "fair cross section:"
"Fair cross-sectional representation" is defined as the range of ten percentage points above and below the relevant percentage of voter age population. For example, if nonwhites represent 40% of the voter age population of a district, it is assumed for present purposes that nonwhites would be cross-sectionally represented in the master jury wheel if nonwhites represented any percentage from 30 to 50% of the names in the wheel. [506 F.2d at 818 (footnote omitted)]
If we apply these two tests to defendants' evidence, we reach the following conclusions: Accepting the twenty percent standard suggested in the example above, every sample for both Chicanos and Blacks reported in Tables A through E, supra, fails the percentage deficiency test. For example, if we take the data for Chicanos in the Denver division from January 1972 through June 1973 [Table A], we find that a twenty percent deficiency from the 8.93% population figure equals 7.14%. Since Chicanos constituted only 4.88% of the random sample drawn from the corresponding jury wheel, the wheel fails to adequately represent this group. But if we adopt the Moore formulation, we obtain an opposite result for every sample reported by defendants. Again taking the data for Chicanos in Table A, any source list which contains 0 to 18.93% Chicanos would fall within the Moore definition of a "fair cross section." Thus, a list containing 4.88% Chicanos would comply with the statutory requirement.
Both the percentage deficiency and Moore tests utilize pre-selected numerical criteria for defining "substantial deviation from a fair cross section of the community." We believe it is unrealistic, if not impossible, to a priori fix such numerical standards. If either of these approaches were intended by the Act, Congress could, and no doubt would, have selected the criteria for us, rather than leaving the matter to "judicial decision." Moreover, the two tests lead to diametrically opposite conclusions when applied to *694 the evidence in this case. We are unable to ascertain any sound basis for favoring one formulation over the other, and the parties have suggested none. We, therefore, reject both proposals.
A third possibility is that employed by Dr. Bardwell in reaching the "statistical conclusions" reported in Tables A through E, supra. During the evidentiary hearing, Dr. Bardwell explained the theoretical basis of the statistical tools he used and the premises or assumptions which underlie his conclusions. As we understand his testimony, Dr. Bardwell applied a statistical test which measures the probability or "chances" that the samples drawn for each of the five data sets could have come from a source list which contains the percentage of Chicanos and Blacks indicated by the 1970 census figures. To illustrate, Dr. Bardwell concluded that there is a "statistically significant difference" between the percentages of Chicanos in the sample taken from the Denver division for January 1972 through June 1973 [Table A] and the corresponding population figures of the 1970 census. These values are 4.88% and 8.93%, respectively. As a matter of "probability theory," Dr. Bardwell found that the chances of drawing a sample which contains 4.88% Chicanos from a jury wheel which actually contained 8.93% Chicanos is approximately seven chances in ten million. Although this figure is, to say the least, impressive, we believe it has little if any relevance to the legal issues now before us.
It is clear from his testimony that Dr. Bardwell's statistical conclusions are based on a comparison between an "observed value" (i. e., the percentage of Chicanos or Blacks in a sample drawn from a particular jury wheel) and a discrete, ideally "expected value" (i. e., the corresponding census figure). The statistics applied here measure the "mathematical significance" of this observed difference. In essence, then, Dr. Bardwell's testimony tells us that jurors in this district are selected from a source list which contains "statistically significant" departures from a perfect cross section of the community. However, under the Act we are not concerned with deviations from a perfect cross section. By the plain meaning of the term, a "fair cross section" is something other than a perfect cross section, defined by discrete numerical values. Instead, the question that is legally relevant here is whether the percentage of Chicanos and Blacks in the jury wheels is "substantially" outside of some acceptable range of values. This is a question, which according to his own testimony, Dr. Bardwell cannot answer with the statistical techniques which he employed. In United States v. McDaniels, supra, a case which accepts the interpretation of the Act urged by defendants here, the court observed:
[The Jury Selection Act] does not contemplate that the jury be selected from a list of names that is the quintessence of the community, nor that the list be a statistical apotheosis. [370 F.Supp. at 301]
We conclude, therefore, that the notion of a "statistically significant underrepresentation," as used by defendants here, cannot be equated with the legal standard of a "substantial deviation from a fair cross section of the community."
The percentage deficiency and Moore formulations were rejected, in part because they required fixed, numerical definitions of "a fair cross section" and of "a substantial deviation." Defendants' "probability theory" approach avoided a priori standards, but failed to measure what the Act requires us to measure.[8] Perhaps it is problems such *695 as these which explain why courts have not fashioned any generally applicable and objective rules for applying the Jury Selection Act, and why they have, instead, relied upon the familiar standard of "systematic exclusion." In wrestling with this issue, the court in McDaniels stated:
By the very nature of the tests adopted by Congress, there can be no absolute measure of what is substantial; the term must be defined in its relationship to the situation in a particular federal judicial district, considering the size of each racial group, the trends in voting, the actual results in jury selection, the overall impact of the selection process, and any other relevant evidence. [370 F.Supp. at 303-04]
We are left then with only one basis for decision in this controversy, namely a comparison of the extent of underrepresentation here with the extent of underrepresentation found in other cases which interpret the Act in a manner consistent with defendants' position.
In United States v. Jenkins, supra, the evidence indicated that the jury wheel for the New Haven division of the District of Connecticut contained only 3.3% Blacks, while the corresponding population was 5.45% Black. In affirming the trial court's judgment that such underrepresentation was not "substantial," within the meaning of the Act, the appellate court pursued the following reasoning:
There remains the question of whether the substantiality of the disparity, judged by the "local community" standard, should be determined by (1) the ratio of the Negro percentage of the adult population to the Negro percentage of those adults available for jury service under the district court's plan or (2) by the difference that would result in the absolute racial composition of the venire as a result of the under-registration of blacks as voters. Appellants contend that the former standard should be used, pointing to the 5 to 3 ratio, which by itself appears substantial indeed. Judge Newman, on the other hand, adopted the latter test, noting that a 6% to 3% disparity between the census figures and the voter registration figures for the community would add to an average array of 50 to 60 veniremen only one (1) additional Negro, which he found not to be substantial enough to deprive appellants of their constitutional or statutory rights. We agree.
The Act was not intended to require precise proportional representation of minority groups on grand or petit jury panels. See United States v. Fernandez, supra, 480 F.2d p. 733. Nor does the Constitution. . . . (proportional representation "has never been a part of the Anglo-American jury system and, indeed, is repugnant to that system.") Carried to its logical conclusion appellants' argument would require that if the adult Negroes in a community constituted 1% of the population, but only ¼ of 1% were registered voters, the 4 to 1 disparity would deny a Negro a fair trial. This we must reject. Besides color, Congress, in enacting the 1968 Jury Selection and Service Act, designated race, religion, sex, national origin and economic status as "suspect categories." See 28 U.S.C. § 1862. Since disparities similar to the one presented here undoubtedly exist for at least one or more of the multitude of groups within these categories in virtually every federal district in the country, adoption of appellants' position would necessarily lead to total abandonment of *696 voter registration lists as the primary tool for the selection of jurors. This would be contrary to both the language of the statute and its legislative history, which makes clear that Congress intended the voter registration lists to serve an important screening function by eliminating those individuals "who are either unqualified to vote or insufficiently interested in the world about them to do so." 1968 U.S.Code Cong. and Admin.News, pp. 1795-1796.
The test of fairness intended by Congress is the more practical one of the difference in absolute numbers rather than a difference in percentages. Judged by this standard, a difference of one (1) Negro in a panel of 60 jurors is not substantial. [496 F.2d at 65-66 (citations omitted)]
A similar approach was taken by the court in United States v. Goff, supra. There the court scrutinized the degree of underrepresentation of Blacks and "food stamp recipients" in the jury pool for the Eastern District of Louisiana:
Appellants' statistics showed that in 1970 (the date of grand jury selection) blacks comprised 26.33% of the voting age population and 21.06% of the registered voters in the Eastern District of Louisiana. Thus there was a 5.27% absolute differential between blacks in the voting age population and registered black voters, and a 20.02 percentage underrepresentation of blacks on the jury list. The court below found that this underrepresentation was not substantial within the meaning of the Jury Selection Act after assessing the impact of this underrepresentation on a grand jury of twenty-three persons. 370 F.Supp. at 302-04. A grand jury that statistically mirrored the voter registration list would contain 4.6 black persons, while one statistically mirroring the voting age population would contain 6.0 black persons. The court concluded that this amount of underrepresentation was not sufficiently substantial to require the Eastern District of Louisiana to supplement its voter registration list. 28 U.S.C. § 1863. We hold that the district court properly disposed of that issue. See United States v. McDaniels, 370 F. Supp. 298, 302-04 (E.D.La.1973).
. . . . . .
The district court also examined the underrepresentation of the class of food stamp recipients in the Eastern District of Louisiana on the jury list. . . . The figures showed that food stamp recipients constitute 10.51% of the voting age population, and 4.34% of those on the jury list. Thus a twenty-three person grand jury that mirrored the jury list would contain 1.0 food stamp recipients, while one that mirrored the voting age population would contain 2.4 food stamp recipients. The registration rate differential has the same impact as in the case of blacks. We find this underrepresentation is not so substantial as to require supplementation of the voter registration list. [509 F.2d at 826, 827 (footnote omitted)]
There are three modes of comparison utilized by the courts in Jenkins and Goff: (1) absolute percentage difference; (2) comparative percentage underrepresentation; and (3) actual impact on jury panels, grand juries, and petit juries.[9] The following table summarizes *697 the evidence in Jenkins and Goff, and compares this evidence with the data reported by defendants in the present case.
TABLE F
Mode of Comparison Jenkins[10] Goff[11] Test[12]
------------------------------------------------------------------------------------
|Absolute Percentage | 2.15% | 6.17% | 4.08% |
|Difference | (min.) | (max.) | (max.) |
|-------------------------|-----------------|---------------------|------------------|
|Comparative Percentage | 39% | 59% | 46% |
|Underrepresentation | (min.) | (max.) | (max.) |
|-------------------------|-----------------|---------------------|------------------|
|Impact on Panel of 50 | 1.1 persons | 3.1 persons | 2.0 persons |
|Jurors | (min.) | (max.) | (max.) |
|-------------------------|-----------------|---------------------|------------------|
|Impact on Grand Jury | 0.5 persons | 1.4 persons | 0.9 persons |
|of 23 | (min.) | (max.) | (max.) |
|-------------------------|-----------------|---------------------|------------------|
|Impact on Petit Jury | 0.26 persons | 0.74 persons | 0.49 persons |
|of 12 | (min.) | (max.) | (max.) |
-------------------------------------------------------------------------------------
As indicated in the above Table, the maximum underrepresentation of either Chicanos or Blacks in the jury wheels of this district falls approximately halfway between the minimum underrepresentation found in Jenkins and the maximum underrepresentation permitted in Goff. Indeed, the underrepresentation reflected in the evidence before us is well within the "limits" set by Goff. Finally, the maximum absolute percentage difference established by defendants here is 4.08%. This measure is more meaningful than the comparative percentage underrepresentation measure, particularly where, as in the present case, the general population precentage for the group in question is relatively small. [See, Gewin, supra at 834-35 and cases cited therein] In our view, an absolute percentage difference of 4.08% is not "substantial" within the meaning of the Act, and certainly the impact of this difference on jury panels of fifty, grand juries of twenty-three, and petit juries of twelve is not so "substantial" as to require supplementation of voter lists, pursuant to 28 U.S.C. § 1863(b)(2). For the foregoing reasons, we hold that the underrepresentation of Chicanos and Blacks in this case does not constitute a "substantial failure to comply" with the "fair cross section" *698 requirement of 28 U.S.C. § 1861. It is, therefore,
ORDERED as follows:
(1) that defendants' challenges to the Jury Selection Plan of this district are without merit and that defendants' respective motions directed thereto be, and the same hereby are, denied; and
(2) that the findings and conclusions contained in the foregoing memorandum shall constitute the court's report in conformance with the mandates of the Tenth Circuit Court of Appeals in Criminal Nos. 72-CR-352, 74-CR-136, and 74-CR-336.
APPENDIX TO MEMORANDUM AND ORDER
The dates of Indictment and trial of the respective
defendants were:
72-CR-352 John E. Test
Indictment 10-26-72
Trial 2-27-73
74-CR-421 Thomas K. Hudson
John B. O'Malley, Jr.
Indictment 10-10-74
Trial 2- 3-75
74-CR-422 John B. O'Malley, Jr.
Indictment 10-10-74
Trial Not yet held
74-CR-423 John B. O'Malley, Jr.
Indictment 10-10-74
Trial Not yet held
74-CR-242 Clemente Marquez
Indictment 6-13-74
Trial 2- 6-75
74-CR-495 Dean Martin Small
Indictment 12- 5-74
Trial 5-12-75
75-CR-58 John Brett Allen
Indictment 2-14-75
Trial 6-23-75
75-CR-58 Frank Lewis Jackson
Indictment 2-14-75
Trial 6-23-75
75-CR-127 Enrique Sandoval Chavez
Indictment 5-15-75
Trial Not yet held
74-CR-336 George Moorer
Indictment 8- 7-74
Trial 10-15-74
74-CR-244 Joe Gutierrez
Indictment 6-13-74
Trial 8-19-74
74-CR-136 Irving Joseph Tosti
Indictment 4- 9-74
Trial 7-29-74
74-CR-136 Phillip Timothy Hermanson
Indictment 4- 9-74
Trial 7-29-74
NOTES
[*] Consolidated with:
U. S. v. Tosti, 74-CR-136; Hermanson, 74-CR-136; Marquez, 74-CR-242; Gutierrez, 74-CR-244; Moorer, 74-CR-336; Hudson, 74-CR-421; O'Malley, 74-CR-422 & 423; Small, 74-CR-495; Jackson, 75-CR-58; Allen, 75-CR-58; and Chavez, 75-CR-127.
[1] Even though many of the defendants herein are not members of the allegedly excluded classes, they still have standing to assert this claim. [See, Peters v. Kiff, 407 U.S. 493, 92 S. Ct. 2163, 33 L. Ed. 2d 83 (1972)]
[2] Defendants also contend that the Jury Plan fails to employ true "random sampling" in that every seventy-fifth name is selected from the source lists for the Denver and Pueblo divisions, while every fiftieth name is drawn from the list for the Grand Junction division. Counsel suggests that this "over-represents" the latter divsion; however, the evidence does not link this fact to any systematic inclusion or exclusion of a cognizable group; nor does the evidence indicate how this "defect" detracts from the cross sectional quality of the composite source lists. Mere geographical imbalance, standing alone, does not render a jury plan illegal. [See, e. g., United States v. Lewis, 504 F.2d 92, 99 (6th Cir. 1974)] Moreover, defendants' own expert testified that the selection of names at fixed intervals from an alphabetical list is mathematically random, so long as the source list contains no periodicity of minority names. Defendants have offered no evidence of such periodicity. We conclude, therefore, that defendants' attack upon the sampling procedures of the Plan is without merit. [See also discussion of standards for judging legality of procedures under the Jury Selection and Service Act, text infra at 692.]
[3] According to Dr. Bardwell's testimony, the documents utilized for this purpose are as follows: Final Report PC1C7, Tables 119 and 129 (Spanish surnamed) and Final Report C1B7, Table 35 (Black), United States Dep't of Commerce, Bureau of Census (1970).
[4] The apparent anamoly between this conclusion and the conclusion reached for Blacks in Table A is, according to Dr. Bardwell's testimony, explained by the larger sample taken for data set E.
[5] See, U.S. Census, General Coding Procedures Manual (Attachment J2), entitled, "Detailed List of Spanish Surnames" (1970).
[6] Defendants apparently seek to circumvent this "causal link" requirement with the argument that a statistical disparity alone raises an inference of "purposeful discrimination." Such an inference is possible under certain circumstances. For example, the court in United States v. Hyde, 448 F.2d 815 (5th Cir. 1971), cert. denied, 404 U.S. 1058, 92 S. Ct. 736, 30 L. Ed. 2d 745 (1972), observed:
When there is a complete absence or "spectacular" underrepresentation of Negroes on juries, the courts have read the figures as establishing a prima facie case of purposeful discrimination on the theory that discrimination against Negroes is the most reasonable explanation for the composition of the jury system under scrutiny. [Id. at 825 (footnote omitted); see also, Whitus v. Georgia, supra]
But the evidence before us does not indicate a "complete absence" or such "spectacular underrepresentation" of minority groups to permit this inference. More importantly, defendants' own evidence that the relative percentages of Blacks and Chicanos in the pool of prospective jurors remained stable during the qualifying process overcomes any possible inference of deliberate interference or systematic exclusion. If there were evidenc of "tampering," then a smaller statistical disparity would suffice for the prima facie showing required. But where, as in this case, there is no evidence of "tampering," the statistical disparity must be indeed "spectacular." We hold that it is not. [Compare, Blackwell v. Thomas, 476 F.2d 443 (4th Cir. 1973); Quadra v. Superior Court of City & County of San Francisco, 378 F. Supp. 605 (N.D.Cal.1974)]
[7] The only "evidence" which defendants have offered relevant to this issue comes from legislative hearings conducted by Congress in connection with a proposed expansion of the Voting Rights Act of 1965. [H.R. 6219, 94th Cong., 1st Sess. (1975)] One provision of the bill is designed to facilitate voter registration by "persons of Spanish origin." Defendants place particular emphasis on the "legislative fact" that registration procedures and elections which are conducted only in English have tended to exclude Chicanos from the electoral process. If we were to carry defendants' argument to its logical conclusion, they would have us decide the constitutionality of 28 U.S.C.A. § 1865(b) (2), which disqualifies any prospective juror who "is unable to read, write, and understand the English language with a degree of proficiency sufficient to fill out satisfactorily the juror qualification form[.]" [See also, United States v. Tijerina, 446 F.2d 675 (10th Cir. 1971) (disqualification of prospective jurors because of their lack of proficiency in English did not constitute "substantial failure to comply" with Jury Selection Act)] In our view, the mere existence of a "language barrier" is not sufficient for a factual finding of "deliberate interference" or "systematic exclusion" within the meaning of the precedents cited in text.
[8] Under a fourth proposal made by the United States Commission on Civil Rights, "any disparity of 20% or more between the proportion of eligible whites selected for [the] master jury wheel and [the] proportion of eligible minority persons [should] be remedied by supplementation." [Gewin, supra at 818] This test cannot be applied in the present controversy, because the parties have not submitted evidence on the proportion of Chicanos, Blacks, and Whites who are registered to vote in Colorado.
[9] We acknowledge that the approaches taken by the courts in Jenkins and Goff also measure the difference between an "observed value" and a "discrete, ideally expected value." For this reason, we do not believe these "tests" are consistent with the intent of the Act [see, discussion in text, pp. 694-695, supra]; however since Jenkins and Goff are the only precedents which apply defendants' construction of the Act to actual data, we follow their reasoning.
[10] Apparently, the trial court in Jenkins found that there were 6% Blacks in the eligible population and only 3% Blacks in the jury pool. [496 F.2d at 65] But the numbers appearing in this column are based on a 5.45% to 3.3% discrepancy, and, therefore, reflect the minimum underrepresentation of Blacks in the Jenkins case. Calculations for each entry in the Jenkins column were performed as follows:
(a) Absolute Percentage Difference: (5.45%-3.3%) = 2.15%
(b) Comparative Percentage Difference: (5.45%-3.3%)/5.45% = 39%
(c) Impact on Panel of 50 Jurors: (2.15% × 50 persons) = 1.1 persons
(d) Impact on Grand Jury of 23: (2.15% × 23 persons) = 0.5 persons
(e) Impact on Petit Jury of 12: (2.15% × 12 persons) = 0.26 persons
[11] The values appearing in this column are based on the evidence in Goff for "food stamp recipients." Since the percentage of the eligible population in this group (10.51%) more closely approximates the corresponding figures for Chicanos and Blacks in the instant case (8.93% and 3.00%, respectively), we have used the figures for the "food stamp recipients" group instead of Blacks in Goff (26.33% of the eligible population). For the method of calculation, see note 10, supra.
[12] The values reported in this column are based on defendants' data for Chicanos in the Grand Junction Division [Table C, supra]. Because this data reflects the greatest absolute percentage difference in any of the five data sets (8.89%-4.81%) = 4.08%, it is directly comparable to the evidence in Goff. For method of calculation, see note 10, supra.