There is no federal constitutional requirement that members of a grand jury be selected in any particular manner. The constitution demands only that the selection process employed produce an array from which a cognizable group or class of citizens has not been systematically excluded. A grand jury need not, however, be a statistical mirror of the community. Under the rule of exclusion, a defendant may establish a prima facie case of discrimination in the selection of a grand jury by showing a disparity between the percentage which his race or class constitutes of the group from which the array is selected and the percentage which his race or class constitutes of the panel selected. In situations where a group claiming to be victimized by discrimination makes up a relatively small percentage of the total electorate eligible for service as grand jurors, statistical analysis to prove that the selection process was discriminatory is an appropriate, although not an exclusive approach. The defendant, of Puerto Rican parentage, claimed that the grand jury which indicted him for the crime of first-degree murder was illegally impaneled in that in its selection a systematic exclusion of persons of his racial, ancestral or cultural background was practiced. The trial court, although it permissibly relied on Spanish surnames as a sufficiently accurate and reliable guide in determining the percentage of Puerto Rican electors eligible for grand jury duty in Fairfield County, incorrectly concluded, from an overestimate of the percentage of Spanish-surnamed electors in that county, that the defendant had established a prima facie case of systematic discrimination and quashed the indictment against him. Since it erred in doing so, this court reinstated the indictment.
Argued February 7, 1973
Decided May 2, 1973
Indictment charging the defendant with the crime of murder in the first degree, brought to the Superior Court in Fairfield County; the defendant's plea in abatement and motion to quash sustained, Saden, J., and, on permission being granted, the state appealed to this court. Error; further proceedings.
Donald A. Browne, assistant state's attorney, with whom, on the brief, was Joseph T. Gormley, Jr., state's attorney, for the appellant (state).
L. Scott Melville, for the appellee (defendant).
The defendant was indicted for the crime of first-degree murder by an eighteen-member grand jury on December 6, 1971. On December 10, 1971, the defendant entered a plea of not guilty to the murder indictment and on February 29, 1972, filed in one pleading a combined plea in abatement and motion to quash. The defendant's primary claim in the plea in abatement and motion to quash was that the grand jury which indicted him was illegally impaneled, in that a systematic exclusion of persons of the same racial, ancestral or cultural background as that of the defendant, who is of Puerto Rican parentage, was practiced in the selection of the grand jury.
Following a hearing on the pleadings, the court sustained the defendant's plea in abatement and motion to quash. In accordance with General Statutes 54-96 the court granted the state permission to appeal. The state has assigned error in several of the court's conclusions. The trial court's conclusions are tested by the finding. Sea Beach Assn. v. Water Resources Commission, 164 Conn. 90, 318 A.2d 115; Brauer v. Freccia, 159 Conn. 289, 293, 268 A.2d 645.
The state assigns error in the finding and specifically attacks the court's failure to find material facts as set forth in twenty-seven paragraphs of the draft finding. Seven, of these paragraphs, which the state seeks to include in the finding, were in fact found or were implicit in the finding. A finding need not be in language identical with the draft finding. Walsh v. Turlick, 164 Conn. 75, 316 A.2d 759. Of the remaining paragraphs, eight were not mentioned in the state's brief or appendix. To secure an addition to the finding the state must point to some part of the appendix, the pleading or an exhibit properly before us which discloses that the defendant admitted the truth of the fact or that its validity was conceded to be undisputed. Gordon v. Indusco Management Corporation, 164 Conn. 262, 320 A.2d 811; Gary Excavating, Inc. v. North Haven, 164 Conn. 119, 318 A.2d 84. A fact is not admitted or undisputed merely because it has not been contradicted. Malarney v. Peterson, 159 Conn. 342, 344, 269 A.2d 274. While there is evidence in the appendix to the state's brief to establish that the court heard testimony supporting the remaining requested additions, these paragraphs were in dispute. The state has failed to show that facts were admitted or undisputed and, therefore, no additions to the finding are warranted.
The finding reveals the following facts which are relevant to the claims pressed by the state. During the nine years from 1963 to 1972, John P. Previdi, as sheriff of Fairfield County, selected individuals to serve on grand juries. Sheriff Previdi maintains a list or panel of individuals who are residents of all the towns in Fairfield County with the exception of the communities of New Fairfield, Sherman, Redding and Wilton. This panel of potential grand jurors includes individuals with different occupations and with a wide range of ages. The list has been altered by approximately twenty-five names during the previous nine years for a variety of reasons. Sheriff Previdi's selections for the various grand juries during the period from 1963 through 1971 were restricted to the list of 138 names maintained in his office. The list was composed largely of friends and acquaintances of the sheriff. In an attempt to obtain more Puerto Rican individuals for his panel of potential grand jurors, Sheriff Previdi requested the assistance of the registrar of voters in Bridgeport, A. Edward Sandula. To determine whether Puerto Rican individuals were systematically excluded from grand jury participation, the defendant recruited a committee headed by Kathleen Earley to conduct a survey. This committee, which included three persons who had received training in the Spanish language, counted the number of Spanish-sounding surnames appearing on voting lists for the years 1963, 1966, 1969, 1970 and 1971 for Bridgeport, Norwalk, Stratford, Danbury, Fairfield, Stamford and Trumbull. The committee could ascertain only if a name generally appeared to be of Spanish origin and would count as Spanish-surnamed anyone with a name of Hispanic origin. It was assumed that no Spanish-surnamed electors lived in the remaining towns in the county. The results of the seven-town survey indicated that in 1971 approximately 1.35 percent of the total of 373,575 electors in Fairfield County were Spanish-surnamed. The survey showed that in 1970 1.11 percent of the electors were Spanish-surnamed. The percentage of Spanish-surnamed electors in the county for the following years was not computed, but a simple analysis of the raw data in the finding reveals that in 1969 .9 percent, in 1966 .8 percent and in 1963 .5 percent of the population were Spanish-surnamed. The nine-year average is therefore approximately .93 percent.
A publication of the greater Bridgeport regional planning agency dated August, 1971, shows Spanish-speaking persons as one of three racial categories considered, and certain sections of Bridgeport and Danbury are predominantly Puerto Rican in character. The court found further that a total of 738 selections for grand jury service were made during the nine-year period between 1963 and 1971 and that only two persons of the 738 selected were Puerto Rican. After analyzing the data of the defendant's exhibit, and the testimony of an expert witness, the court found that based on established statistical methods the chance of selecting only two Spanish-surnamed electors out of 738 selections was 2000 to 1, and that such a result was not consistent with a random selection. In addition, the court found that the selection of one Spanish-surnamed grand juror over the period 1969-71 was not consistent with a random selection of grand juries and, similarly, the selection of no Spanish-surnamed grand jurors in 1971 was not a result of a random selection. In determining the existence or nonexistence of randomness in the selection process, the percentage of Spanish-surnamed electors in the seven towns surveyed was an essential factor. The contents of the defendant's exhibits, made part of the findings, show that a nine-year average of 1.6 percent was employed by the defendant's expert in arriving at her conclusions which were accepted in the finding.
While the finding is more extensive than the brief narrative presented here, the foregoing statement is sufficient to review the conclusions attacked by the state.
The conclusions reached by the trial court are tested by the finding and must stand unless they are legally or logically inconsistent with the facts found or unless they involve the application of some erroneous rule of law material to the case. Johnston Jewels, Ltd. v. Leonard, 156 Conn. 75, 79, 239 A.2d 500; Lewis v. Lewis, 162 Conn. 476, 480, 294 A.2d 637; Connecticut Bank Trust Co. v. Bovey, 162 Conn. 201, 205-6, 292 A.2d 899.
The state during oral argument abandoned its claim that the defendant's motions were untimely. This claim will not be considered.
It must be noted at the outset that both parties agree that the issue here involves the claimed systematic exclusion of Puerto Rican electors from grand jury service. The court's references to Spanish-speaking electors are simply an alternative identification of the class of Puerto Rican electors involved in this case.
The state has assigned error in the court's conclusion that Spanish surnames serve as a sufficiently accurate and reliable guide in determining the number of Puerto Rican electors eligible for grand jury duty in Fairfield County. The state produced no evidence to demonstrate that this methodology was inaccurate or would introduce measurable error into the calculations. Instead, the state relied on repeated assertions that this method of determining the number of Puerto Rican electors was inherently inaccurate and imprecise.
The court's conclusion is logically consistent with its finding. The survey in one instance was cross-checked by two different groups with the result that only a two-name difference out of a total count of 370 names was discovered. An estimate of Puerto Rican electors in Bridgeport made by the registrar of voters for Bridgeport, relying on actual place of birth of the elector, found 4 1/2 to 5 percent of the total number of electors in Bridgeport were Puerto Rican. This result closely paralleled the data in the defendant's survey. It cannot therefore be said that the court's conclusion that Spanish surnames were a reasonably acceptable and reliable method of determining the number of Puerto Rican electors was logically inconsistent with the finding.
Furthermore, the use of, Spanish surnames as a measure of the electoral strength of Mexican-Americans has been viewed favorably by the United States Supreme Court. In Hernandez v. Texas, 347 U.S. 475, 74 S.Ct. 667, 98 L.Ed. 866, the state of Texas challenged reliance on surnames as a device for showing the descent of persons. The Supreme Court there held (p. 480 n. 12) that, "just as persons of a different race are distinguishable by color, these Spanish names provide ready identification of the members of this class." The use of Spanish surnames as a measure of the size of the class of Mexican-American voters was similarly accepted in Montoya v. People, 141 Colo. 9, 345 P.2d 1062; see also State v. Lopez, 182 Kan. 46, 318 P.2d 662. The survey technique employed here thus has been held to be permissible as a measure of the Mexican-American population in the southwest. The finding indicates that the procedure employed achieved an acceptable level of reliability in measuring the Puerto Rican electoral population of Fairfield County. Puerto Rican electors make up a significant minority of voters in Fairfield County and are easily recognizable and identifiable by their Spanish surnames. The court was not in error in relying on Spanish surnames as a device for identifying the percentage of Puerto Rican electors in Fairfield County.
The state has also assigned error in the court's conclusion that it could not determine whether the names Douglas Bora, Amadeo Morrello and Ferdinand Equi were of Hispanic origin. The state claims that these names are Spanish surnames and that these individuals have served on various grand juries in Fairfield County. There is no evidence in the finding which would establish that Bora, Morrello and Equi are Spanish surnames. The court's conclusion is thus not legally or logically inconsistent with the finding and must be accepted.
It is perhaps appropriate to peruse the Supplement to Manual of Immigration, Spanish — Spanish Personal Names (1936), selected by George Lockwood, United States Immigration and Naturalization Service, as updated by the census department in General Coding Procedures Manual, Attachment J2, Detailed List of Spanish Surnames, for use in the 1970 census. This list of over 8000 Spanish surnames does not include the surnames Bora or Equi. See also Montez v. Superior Court, 10 Cal.App.3d 343, 352, 88 Cal.Rptr. 736.
The remaining assignments of error attack the court's conclusions that the defendant had established a prima facie case of systematic exclusion of Puerto Rican individuals in the selection of grand jurors and that the method of selection used violated the equal protection clause of the fourteenth amendment to the United States Constitution and was unconstitutional because it failed to expose Puerto Rican individuals to a fair chance of selection for grand jury service. The main thrust of the court's memorandum of decision was that only a random method of selection was constitutionally acceptable. The conclusions in the finding imply that random selection is required to satisfy constitutional standards for grand jury selections.
This court has recently reviewed the standards for grand jury selection. State v. Cobbs, 164 Conn. 402, 324 A.2d 234. To reiterate, there is no constitutional requirement that members of a grand jury be chosen in any particular manner. The common-law selection method followed by the sheriff was not inherently unconstitutional. Grand jurors need not be picked at random in order to comply with constitutional requirements forbidding intentional discrimination against race or class. State v. Cobbs, supra. The constitution demands only that the system employed by the state to select grand jurors produce an array from which a cognizable group or class of citizens has not been systematically excluded. United States ex rel. Chestnut v. Criminal Court, 442 F.2d 611 (2d Cir.); United States v. Butera, 420 F.2d 564 (1st Cir.); State v. Cobbs, supra, 4. The court's conclusion that a prima facie case of systematic exclusion of Puerto Rican individuals had been proven was not based only on the absence of a random selection method as claimed by the state, but was also supported by the conclusion that the selection method as used failed to expose Puerto Rican individuals to a fair chance of selection. The court's conclusion that a prima facie case of intentional exclusion was established on the latter basis will be reviewed to determine if it is legally and logically inconsistent with the facts found or involves application of an erroneous rule of law material to the case. Barnini v. Sun Oil Co., 161 Conn. 59, 64, 283 A.2d 217; Johnston Jewels, Ltd. v. Leonard, 156 Conn. 75, 79, 239 A.2d 500.
Differences in race, color or ethnic background define easily identifiable groups which exist within our communities. Whether a particular group is a recognizable class subject to discrimination in grand jury selection is an issue of fact. Hernandez v. Texas, supra, 478. The trial court concluded that Spanish-speaking persons (Puerto Ricans) represented a distinct class in Fairfield County and this conclusion was not attacked by the state. It appears from the finding that this group does constitute a cognizable class. Persons of Puerto Rican heritage share a similarity of attitudes, ideas and experience which cannot be adequately represented if they are excluded from grand jury selection. The protection of the fourteenth amendment is not limited to whites and blacks; it applies to any identifiable group which may be the subject of community prejudice. Hernandez v. Texas, supra; State v. Plenty Horse, 85 S.D. 401, 405, 184 N.W.2d 654; see also United States ex rel. Leguillou v. Davis, 115 F. Sup. 392, 398 (D.V.I.) (Puerto Rican electors in the Virgin Islands). The defendant's proof that Puerto Ricans are a cognizable group was more than adequate and never seriously disputed by the state. The constitution does not require, however, that the grand jury be a statistical mirror of the community. Hoyt v. Florida, 368 U.S. 57, 82 S.Ct. 159, 7 L.Ed.2d 118. It stands to reason that in a body limited in number, such as a grand jury, accurate and complete representation of all of the groups and categories that comprise a modern heterogeneous community is difficult to achieve even when measured over a substantial period. United States v. Ditommaso, 405 F.2d 385, 390 (4th Cir.).
In attempting to meet his burden of proof to establish a prima facie case of discrimination, the defendant relied on the "rule of exclusion" which was first enunciated in Norris v. Alabama, 294 U.S. 587, 55 S.Ct. 579, 79 L.Ed. 1074.
Under this rule as it has developed since the Norris decision, a defendant may establish a prima facie case of discrimination by showing a disparity between (1) the percentage which his race or class constitutes of the group from which a grand jury panel is selected, here the electors in Fairfield County, and (2) the percentage which his race or class constitutes on the jury panels selected. Sims v. Georgia, 389 U.S. 404, 407-8, 88 S.Ct. 523, 19 L.Ed.2d 634; Muniz v. Beto, 434 F.2d 697, 700 (5th Cir.).
The defendant must also show an opportunity on the part of the official selecting the panel to discriminate. Alexander v. Louisiana, 405 U.S. 625, 630, 92 S.Ct. 1221, 31 L.Ed.2d 536. The state never disputed the claim here that the sheriff had an opportunity to discriminate. Since he selected each grand jury from a limited private list, the sheriff had an opportunity to discriminate and could have discriminated in his selections against Puerto Ricans as identified by Spanish-sounding surnames. The opportunity to discriminate alone is not fatal to the selection method. The defendant must still demonstrate significant underrepresentation of Puerto Ricans to prove a prima facie case of discrimination.
In order to show a disparity between the percentage of Puerto Rican electors qualified for selection as grand jurors of the total qualified and the percentage which Puerto Ricans constituted of persons actually selected as grand jury members, it was first necessary for the defendant to show the percentage of the total electors qualified who were members of the class. While the use of statistics in equal protection cases can be misleading, statistics do serve a useful purpose in pointing up potential problems that require explanation. Smith v. Yeager, 465 F.2d 272, 279 n. 18 (3d Cir.). Comparative and absolute percentage statistics may not reveal a true picture and thus the presentation of some probability analysis along with actual percentages is appropriate.
Purposeful discrimination exists whenever significant unexplained disparities exist alongside an opportunity to discriminate. United States v. Butera, 420 F.2d 564 (1st Cir.). Obviously, proof of such underrepresentation is more readily available in situations where intuitively the disparity can be recognized as exclusionary. See Hairston v. Cox, 459 F.2d 1382 (4th Cir.) (no blacks served while 20 percent of the population was black); Black v. Curb, 464 F.2d 165 (5th Cir.) (60 percent of the population was black, 35 percent on selection list); Carter v. Jury Commission, 396 U.S. 320, 90 S.Ct. 518, 24 L.Ed.2d 549 (65 percent of the population was black, 32 percent on grand jury list); Hernandez v. Texas, 347 U.S. 475, 74 S.Ct. 667, 98 L.Ed. 866 (no Mexican-Americans ever served on grand jury).
In the present case such striking underrepresentation was not argued and the court properly heard evidence by a mathematician who attempted to prove that the selection process was discriminatory on the basis of a methodology known as statistical decision theory. This approach evaluates results expected to be produced by a random selection and compares these expected results with the observed data of actual selections to determine whether such actual results are consistent with randomness. Since circumstances or chance may well dictate that no person in a certain class will serve on a particular jury or during some period, it is appropriate to determine whether mere chance could explain the results actually observed. See Hernandez v. Texas, supra, 480. The defendant attempted to demonstrate that as a mathematical probability it was impossible for so few Puerto Rican individuals to be selected for grand jury service absent discrimination.
Finkelstein, "The Application of Statistical Decision Theory to the Jury Discrimination Cases," 80 Harv. L. Rev. 338.
The United States Supreme Court in Alexander v. Louisiana, supra, commented on the use of statistical decision theory and this statistical device was employed in Witcher v. Peyton, 405 F.2d 725 (4th Cir.). We believe that in situations where the group claimed to be victimized by discrimination makes up a relatively small percentage of the total electorate eligible for service as grand jurors, the statistical approach adopted by the trial court in this case is appropriate although we hasten to say that it is not an exclusive approach.
In any event, the critical starting point for the analysis is a determination of what percentage of the total eligible electors are members of the class. In the finding, the court found that the percentage of Spanish-surnamed electors for the entire nine-year period was an essential factor for determining whether randomness existed. This figure taken from the corrected finding appears to be .93 percent, or less than 1 percent. The defendant's exhibit, relied on by the expert witness, uses the figure of 1.6 percent for the Puerto Rican population in determining that the selection was not at random. To resolve this conflict in the finding we have examined the transcript. The transcript reveals that the defense statistician used the base percentage of 1.6 percent in all calculations. This nine-year average figure is greater than the largest percentage, 1.35 percent, found for any year in the finding. The 1.6 percent figure was obtained by the defendant's statistician who determined the percentage of Spanish surnames by weighing the percent of Spanish surnames from each of the towns by the respective percentage of grand jurors selected from those towns. The total figure of 1.6 percent is thus a weighted average which was explained as the percentage of Spanish-surnamed individuals in a town multiplied by the likelihood of a grand juror coming from that town. This likelihood of selection was based on the frequency of actual selections from that town over the nine-year period.
This .93 percent figure is the average of the percentages for the five years tabulated.
While it is clear that statistical proofs are valuable tools for the court in evaluating claims of discrimination, the use of such analysis must be restricted to a showing of results which might have been expected from mere chance. To determine this expected result the court must insist that the data offered comprise theoretical results. Although weighted averages are frequently utilized by statisticians, the court must determine results expected by random chance assuming an equiprobable selection process in which, under conditions of true randomness, all citizen electors of all towns had an equal chance to be selected for grand jury service. In this manner the court will have as a touchstone an expected result from random selection and can then test the actual observed results by this theoretically random result.
The defendant here made no claim that discrimination resulted from overrepresentation of some parts of the county at the expense of others and therefore the use of weighted averages based on the historical frequency of selection from certain towns in the county was inappropriate. Since the court's ultimate conclusion that the defendant had established a prima facie case was based on these impermissible statistical proofs and since the average of 1.6 percent used in the calculations cannot be supported by the finding which indicates the greatest percentage of Spanish-surnamed electors in the county at any time was 1.35 percent, we must hold that the court erred in quashing the indictment and concluding that Spanish-surnamed persons were not exposed to a fair chance of selection for grand jury service. We recognize that this case, was complicated by the difficulty in gathering statistical data and in interpreting the statistical theories presented by the defendant, but we are constrained to hold that a grand jury selection method may not be declared unconstitutional in the absence of even a prima facie case of discrimination. The defendant's offer in this case overestimated the percentage of Spanish-surnamed electors and thus rendered any conclusion based on it erroneous. It is inappropriate to speculate on the results which might have been found from an evaluation of the data assuming equiprobable selection of citizens regardless of residence and a nine-year average of .93 percent Spanish-surnamed electors, or any other shorter time span with a verified percentage of Spanish-surnamed electors. Such issues would require a full exploration at the trial level.
It is unnecessary to review the state's claim that the court erred in concluding that the state failed to rebut the prima facie case of systematic exclusion. It should be noted, however, that mere protestations of good faith are not sufficient to rebut a prima facie case. Alexander v. Louisiana, 405 U.S. 625, 92 S.Ct. 1221, 31 L.Ed.2d 536.