Opinion
Plaintiffs who brought school desegregation action against school board and state officials filed motion to determine sufficiency of state officials' response to request to admit or deny validity of statistics which concerned racial composition of schools and which were arrived at through random sampling techniques or to have matter deemed admitted. The District Court, Blumenfeld, J., held that request to admit or deny accuracy of the statistics was not objectionable on theory that request called for an ‘ opinion’ or on theory that rule only required state officials to admit or deny verifiable facts; that officials' response that, in the ‘ abstract,’ sampling techniques and techniques used by plaintiffs' expert in particular were valid methods for arriving at an accurate approximation of racial imbalance of school system during years in question constituted an admission that sampling techniques used were valid; and that officials could with reasonable inquiry, answer request to admit or deny accuracy of the ststistics.
Order accordingly.
Douglas M. Crockett, John A. Dziamba, Tolland-Windham Legal Assistance, Willimantic, Conn., Igor I. Sikorsky, Jr., Hartford, Conn., Raymond Marcin, Asst. Professor, Catholic University of America, Washington, D. C., John E. Tener, Hartford, Conn., Louis R. Lucas, Memphis, Tenn., for plaintiffs.
Robert K. Killian, Atty. Gen., F. Michael Ahern, David B. Beizer, David J. Della-Bitta, Asst. Attys. Gen., Hartford, Conn., Louis Parley, Univ. of Conn. Legal Clinic, W. Hartford, Conn., for defendants.
MEMORANDUM OF DECISION ON PLAINTIFFS' MOTION TO DETERMINE THE SUFFICIENCY OF THE DEFENDANT STATE OFFICIALS' RESPONSE AND/OR MOTION TO DEEM THE MATTER ADMITTED
BLUMENFELD, District Judge.
This motion presents a rather narrow and technical issue within the context of this complex school desegregation case instituted in 1970. Although narrow, the question at issue does have substantial bearing on the speed with which this prolonged litigation can finally reach the state of adjudication and the length of the trial which must ultimately be conducted before a three-judge court of this district. Germane to the issues in this case is the racial balance of the Hartford school system from approximately 1950 to the present. Before 1964 the school system did not maintain statistics regarding the racial composition of either individual schools or the system as a whole. Thus, for the 1950-1964 period plaintiffs have been forced to compile these statistics on their own. Apparently the school system maintained individual records on all students throughout this period which indicated the child's race and such other data (pertinent for the purpose of determining whether a child was a Latino) as the child's place of birth and the language spoken in his home.
A three-judge panel was convened in this case pursuant to 28 U.S.C. § 2281 (1970) because plaintiffs seek injunctive relief against the state-wide enforcement of Conn. Gen.Stat.Ann. §§ 10-220 (Supp.1974) and 10-240 (1958) which prescribe school district and attendance zones. They contend that this Court can only grant the necessary interdistrict relief by enjoining the enforcement of those statutes.
In compiling their statistics, plaintiffs were faced with the choice of either reading through the files of all 321,000 students who attended Hartford elementary schools during the period in question or else employing a random sampling technique to arrive at an accurate approximation of the racial composition of the schools. Not unreasonably, they chose the latter alternative which, according to their statistical expert, Dr. Paul Hadden of the University of Connecticut, required the examination of only 11,000 files to reach an accuracy level of 5% with a confidence level of 95%. It is clear that an actual reading of all 321,000 files would have involved incalculable time and expense.
According to Dr. Hadden's affidavit, this means that one can be 95% confident that the statistics are accurate within a 5% range.
Having completed their statistical survey, the plaintiffs then submitted the results in tabular form to the defendant school board and the state defendant on March 4, 1974, with a request to admit their accuracy pursuant to Fed.R.Civ.P. 36(a). On March 19, 1974, the defendant school board admitted the accuracy of the results. The state defendant, on the other hand, has refused to admit or deny the accuracy of the results, relying upon a rather narrow reading of the requirements of Rule 36(a). Within the 30 days required by the rule, the defendant did submit a response in which it stated that it could not ‘ truthfully admit or deny the accuracy of the contents of the table because it is based on a statistical formulation employing random sampling to obtain estimates of the racial composition of the schools so that the conclusions reached are not verifiable facts.’
Upon receipt of this response, plaintiffs conferred with the defendant and agreed to break down their request for admission into three parts. As so drafted, they now read:
‘ 1. Where the Hartford Board of Education maintains that actual figures for the racial composition of the Hartford Public Schools were not kept, a formulation based on statistically valid random sampling techniques is a reliable means of approximating the racial composition of those individual schools. ‘ 2. The methodology employed by Dr. Kenneth Paul Hadden as set out in the ‘ Affidavit of Kenneth Paul Hadden in Support of the Attached Document ‘ Racial Composition of Hartford Public Schools, Determined by Sample Count of Student Population’ ' is a statistically valid random sampling technique. ‘ 3. The results of the methodology referred to in item #2 above and depicted in the table entitled, ‘ Racial Composition of Hartford Public Schools, Determined by Sample Count of Student Population’ are accurate as a reliable means of approximating the racial composition of the schools listed in that table.'
The defendant responded to these revised requests by admitting that the table of results was ‘ obtained in accordance with the methodology described in the affidavit of Kenneth Paul Hadden . . ..’ However, again it stated that it could not ‘ truthfully admit or deny the accuracy of the contents of the table.’ Presumably, its difficulty was still chiefly based upon the notion that Rule 36(a) only requires a response to a request to admit ‘ verifiable facts.’ This assumption is borne out by the argument of defendant's counsel in his brief and in the hearing held before this Court on this matter.
Initially, then, in determining the adequacy of the defendant's response, an inquiry must be made into the scope of Rule 36(a). The defendant refuses to answer on the grounds that the rule only requires it to admit or deny ‘ verifiable facts.’ It argues that the plaintiffs are asking it to express an opinion as to the validity of sampling techniques in general, the validity of the particular technique used in this case and the accuracy of the results obtained. It firmly maintains that Rule 36(a) does not contemplate that kind of request.
In light of the recent amendment to Rule 36(a), this Court is somewhat perplexed by the defendant's argument. Prior to 1970, the rule read, in relevant part:
‘ After commencement of an action a party may serve upon any other party a written request for the admission by the latter of the genuinenss of any relevant documents described in and exhibited with the request or of the truth of any relevant matters of fact set forth in the request. . . .’ (Emphasis added.)
It is true that under this language there was some confusion and division among the courts as to the type of material which could be the subject matter of an admission. Some courts, viewing the scope of the rule broadly, held that requests in the shadowy area between fact and opinion were proper. Photon, Inc. v. Harris Intertype, 28 F.R.D. 327 (D.Mass.1961), aff'd on other grounds 349 F.2d 856 (1st Cir. 1965), cert. denied sub nom. Tansel v. Photon, Inc., 382 U.S. 1011, 86 S.Ct. 621, 15 L.Ed.2d 527 (1966); Jones v. Boyd Truck Lines, Inc., 11 F.R.D. 67 (W.D.Mo.1951). Others, narrowly restricting the rule, held that requests for admission of matters of opinion or conclusion were not proper. Trabon Engineering Corp. v. Eaton Mfg. Co., 37 F.R.D. 51 (N.D.Ohio 1964); Kasar v. Miller Printing Machinery Co., 36 F.R.D. 200 (W.D.Pa.1964).
In revising Rule 36(a), the Advisory Committee on Rules was well aware of this problem area. As revised, the rule now reads in pertinent part:
‘ A party may serve upon any other party a written request for the admission, for purposes of the pending action only, of the truth of any matters within the scope of Rule 26(b) set forth in the request that relate to statements or opinions of fact or of the application of law to fact, including the genuineness of any documents described in the request. . . .’ (Emphasis added.)
Thus, the revised rule on its face embraces requests for admissions of opinions of fact. As the parties admit, there is virtually no case law on the impact of the revision upon practice under Rule 36(a). However, the Advisory Committee on Rules itself discussed the intent of the change as follows:
‘ Subdivision (a) . As revised, the subdivision provides that a request may be made to admit any matters within the scope of Rule 26(b) that relate to statements or opinions of fact or of the application of law to fact. It thereby eliminates the requirement that the matters be ‘ of fact.’ This change resolves conflicts in the court decisions as to whether a request to admit matters of ‘ opinion’ and matters involving ‘ mixed law and fact’ is proper under the rule. . . . ‘ Not only is it difficult as a practical matter to separate ‘ fact’ from ‘ opinion,’ see 4 Moore's Federal Practice ¶ 36.04 (2d ed. 1966); cf. 2A Barron & Holtzoff, Federal Practice and Procedure 317 (Wright ed. 1961), but an admission on a matter of opinion may facilitate proof or narrow the issues or both. An admission of a matter involving the application of law to fact may, in a given case, even more clearly narrow the issues.' 48 F.R.D. 487, 532 (1970).
In view of this clear language, it becomes rather apparent why virtually no other court has been forced to resolve the chimerical problem presented here.
I have chosen to discuss this matter at some length because of the need to emphasize the usefulness of Rule 36(a) in verifying statistical data at a pretrial stage in this piece of litigation and others of like complexity. It is well established that Rule 36 ‘ offers great possibilities for facilitating the proof at the trial by weeding out facts and items of proof over which there is no dispute, but which are often difficult and expensive to prove.’ 4A J. Moore, Federal Practice ¶ 36.02 at 36-15 (2d ed. 1948). In this case, the racial distribution in the Hartford school system over a period of years is clearly an underlying factual issue of some importance, yet one which ideally should not involve serious dispute among the parties. Plaintiffs make the reasonable assertion that if they are required to prove the racial composition at trial that such proof would require ‘ four to five full days of trial time including the testimony of Dr. Hadden and upwards of twenty-five witnesses from the Hartford School system who are familiar with the racial composition of the Hartford Schools.’ Plaintiffs' Brief (on the instant motion) 10. The avoidance of such potentially unnecessary use of court time, particularly when three judges are involved, is a matter of importance to this court. The use of a Rule 36 request for admission seems like a procedure well-suited to achieving this result.
The defendant objects to being forced to admit or deny the statistics which are based upon a random sampling technique. As discussed, any objection to such a request on the grounds that it calls for an ‘ opinion’ is clearly without foundation. Moreover, on a policy level, there seems to be no good reason why such statistics would not be an appropriate subject for such a Rule 36 request.
As early as 1953, Judge Wyzanski in United States v. United Shoe Machinery Corp., 110 F.Supp. 295 (D.Mass.1953), aff'd 347 U.S. 521, 74 S.Ct. 699, 98 L.Ed. 910 (1954), observed, when faced with the admissibility of a statistical sample in a complex antitrust case, that ‘ [i]f antitrust trials are to be kept manageable, samples must be used, and a sample which is in general reasonable should not be rejected in the absence of the offer of a better sample.’ Id. at 305-306. Clearly, such an observation is equally applicable to complex civil rights cases.
More recently, Judge Weinstein in Rosado v. Wyman, 322 F.Supp. 1173 (E.D.N.Y.1970), aff'd 437 F.2d 619 (2d Cir. 1970), aff'd 402 U.S. 991, 91 S.Ct. 2169, 29 L.Ed.2d 157 (1971), devoted considerable attention to the role of sampling in litigation. Beginning with the observation that ‘ [s]ampling has long been considered an acceptable method of determining the characteristics of a large universe,’ he cited numerous instances in which courts have accepted such methods as being ‘ reliable and acceptable in determining adjudicative facts.’ Id. at 1180.
In Rosado itself a sampling techinque was used by stipulation of the parties to determine the average value per recipient per month of ‘ special need’ items included in the need standard under the state welfare program over a period of a year. See Johnson v. White, 353 F.Supp. 69, 75 (D.Conn.1972); New Jersey Welfare Rights Organization v. Cahill, 349 F.Supp. 501, 508 (D.N.J.1972), aff'd 483 F.2d 723 (3d Cir. 1973). The court's discussion as to the propriety and need for sampling data under the facts of that case are peculiarly relevant to the instant case:
‘ In the instant case where the question of the value of special needs arose after the fact, the State did not have adequate retained information or data at hand from which it could determine the amount of special need grants made during the relevant period, nor could any such information be readily produced. The only source of such information was a review of the individual case records of AFDC families. It was entirely impracticable to review all the case records. The monthly average number of cases in New York City from July 1968 through June 1969 was 186,629. Accordingly, a sample of such cases was the only feasible technique.’ 322 F.Supp. at 1181.
Although the numbers involved in this case are perhaps smaller than those in Rosado, it is still obvious that a sampling of the records of the Hartford school children in the years in question is the only feasible method of obtaining the necessary statistics for use at trial. In other words, this Court will ultimately have to rely upon some set of statistics derived through the use of some sampling techinque. See Manual for Complex Litigation, 2.7 12-. 713 (1973); H. Zeisel, The Uniqueness of Survey Evidence, 45 Cornell L. Q. 322 (1960).
Having established the need for such sampling data, it is equally apparent that there is no substantial reason why the defendant should not be required to make ‘ reasonable inquiry,’ Fed.R.Civ.P. 36(a), to determine whether it can admit or deny the accuracy of the particular statistics obtained through the use of Dr. Hadden's particular sampling technique. If it can admit the accuracy of the statistics, the fact-finding process in this case will be greatly facilitated.
As discussed earlier, plaintiffs' actual request for admissions was divided into three separate parts. The first addressed the validity of sampling techniques in the abstract; the second dealt with the validity of Dr. Hadden's particular technique; and the third with the accuracy of the statistics derived through the use of the technique. At the hearing on this motion, the Court asked defendant's counsel what its answer would be to the substance of the first two requests if it were forced to answer. The defendant responded that its opinion was that in the ‘ abstract’ such sampling techniques, and Dr. Hadden's in particular, were valid methods for arriving at an accurate approximation of the racial balance in the school system during the years in question. Thus, for the purposes of this litigation, the Court will deem the first two requests admitted.
‘ THE COURT: Right
At the hearing, however, the defendant was not willing to admit the accuracy of the statistics, even assuming its construction of Rule 36 to be wrong. It cited a number of valid concerns with the way in which the research may have been conducted. For example, it did not understand why Dr. Hadden had sampled only 17 of the 30 schools in Hartford or know the degree of Dr. Hadden's supervision over the students who actually did the field research. Valid as these concerns are, there seems to be no apparent reason why through ‘ reasonable inquiry’ the defendant would not be able to obtain those facts which it contends are necessary to put it in a position to make a judgment as to the accuracy of plaintiffs' statistics. Plaintiffs point out that they have offered to make Dr. Hadden available to the defendant for the purpose of a deposition. In addition, the defendant could inspect the field notes and working papers of Dr. Hadden. In this way, it could obtain information regarding the way in which the study was conducted. Its expert(s) could then advise it whether plaintiffs' statistics are reliable within the range indicated by Dr. Hadden's affidavit.
‘ MR. AHERN: And it's as simple as that. We just don't know how it was arrived at.
Hans Zeisel, Professor emeritus of Law and Sociology at the University of Chicago Law School, in his article The Uniqueness of Survey Evidence, supra, points out that there are essentially two ways relevant to this case in which a sampling operation can be invalid: the selection of an improper ‘ universe’ or an inadequate sample. Defendant's concerns seem to center around the latter. However, as outlined in the text, the defendant should be able to make judgments about the adequacy of the universe and the sample at this stage in the proceedings and thus respond to a Rule 36(a) request to admit the accuracy of the statistics.
The burden which this places upon the State of Connecticut is not unduly heavy. The defendant argues that the ‘ reasonable inquiry’ requirement of Rule 36 does not contemplate the kind of inquiry which would be required of it in this case. Again, this objection is answered by the Report of the Advisory Committee on Rules in reference to this requirement in the amended Rule 36:
‘ The rule as revised adopts the majority view, as in keeping with a basic principle of the discovery rules that a reasonable burden may be imposed on the parties when its discharge will facilitate preparation for trial and ease the trial process. It has been argued against this view that one side should not have the burden of ‘ proving’ the other side's case. The revised rule requires only that the answering party make reasonable inquiry and secure such knowledge and information as are readily obtainable by him. In most instances, the investigation will be necessary either to his own case or to preparation for rebuttal. Even when it is not, the information may be close enough at hand to be ‘ readily obtainable.’ Rule 36 requires only that the party state that he has taken these steps.' 48 F.R.D. 487, 533 (1970).
In the instant case, the defendant is only being asked to ‘ secure such knowledge and information as are readily obtainable by [it].’ In addition, it is clear that if the defendant finds that it cannot admit the accuracy of the statistics, the research required of it here will not have been wasted. It will be ‘ necessary either to [its] own case or to preparation for rebuttal.’ Indeed, it does not even seem unreasonable to require the defendant to conduct independent research if necessary to verify the accuracy of the raw data obtained by Dr. Hadden's field workers. Accordingly, this Court will require the defendant to make reasonable inquiry, as set out in this opinion, to acquire the facts necessary to admit or deny the request submitted by the plaintiffs. It is directed to respond within 30 days of the date of this order to plaintiffs' request for admission unless good cause be shown why time to respond should be enlarged. If no response is made within that time period, the matter shall be deemed admitted. Fed.R.Civ.P. 36(a).
The Court understands that such verification might require taking a sample of 10% of the 11,000 folders inspected by plaintiffs' field workers to insure that they correctly read the folders or made reasonable judgments in doubtful cases regarding whether particular students were Spanish-speaking. In fact, the plaintiffs themselves conducted such an accuracy test which is described in an affidavit by Dr. Hadden which they submitted to this Court and to the defendant. The defendant may elect to rely upon the results of that check or undertake its own.
So ordered.
‘ Now suppose the Court were to say, well, you have to respond by stating what your opinion is as to the validity of that method as a statistical method? ‘ MR. AHERN: In the abstract we can say it is a valid method.’ Transcript of Hearing, May 28, 1974, at 25.
‘ For example, in determining that the cost is prohibitive, Professor Hadden mentions the fact that to get the race of every student attending every Hartford school during the twenty-three year period would be prohibitive. However, he only did the random sampling evidently for twelve years, not 23 years. ‘ Furthermore, each school in Hartford wasn't sampled: They only did 17 out of 30 schools. We don't know why. ‘ We don't know the determinations that were made by the students. We don't know how many students worked on this. We don't know what supervision Professor Hadden gave to them. We don't know whether Professor Hadden is an expert. We don't know whether Professor Hadden ever performed or ran a random sampling before. He doesn't mention it in his affidavit. ‘ THE COURT: All right. ‘ MR. AHERN: So for all of these reasons we feel that the request for admitting the accuracy of these figures cannot be admitted with a yes or no answer.’ Transcript of Hearing, May 28, 1974, at 26-27.