United States v. State of South Carolina

ORDER This is a decision in a bifurcated trial before a three-judge court. Pursuant to a prior Order of the court, only issues of constitutional and federal law with respect to liability were considered at the first three-judge hearing. Procedural issues with respect to the alleged plaintiff class and all issues with respect to the availability or amount of damages were reserved for later determination. 1 The United States brought this action on September 15, 1975, against the State of South Carolina, the South Carolina State Board of Education, the South Carolina State Retirement System, the South Carolina Budget and Control Board, and three local school boards in their individual capacities and as representatives of a defendant class of all local school boards in the State. The defendants are charged with violations of the Fourteenth Amendment to the Constitution of the United States and Title VII of the Civil Rights Act of 1964, as amended, 42 U.S.C. § 2000e, et seq., (1970), through the use of minimum score requirements on the National Teacher Examinations (hereinafter “NTE”) to certify and determine the pay levels of teachers within the State. On September 17, 1975, the South Carolina Education Association (SCEA), the National Education Association (NEA), and nine named individuals as class plaintiffs brought another suit against the same defendants in their official and individual capacities, seeking substantially the same relief sought by the U. S. Attorney General’s suit, basing their actions also upon the Fourteenth Amendment and Title VII of the Civil Rights Act of 1964, as amended; and, in addition, alleging violations of Sections 1981 and 1983 of Title 42 of the United States Code. These plaintiffs moved to intervene in the Attorney General’s action. This motion was granted on January 6, 1976, and they became the plaintiff-intervenors, and their separate action was stayed. For over thirty years the State of South Carolina and its agencies have used scores on the NTE to make decisions with respect to the certification of teachers and the amount of state aid payable to local school districts. Local school boards within the State use scores on the NTE for selection and compensation of teachers. From 1969 to 1976, a minimum score of 975 was required by the State for its certification and state aid decisions. In June, 1976, after an *1098 exhaustive validation study by Educational Testing Service (ETS), and, after a critical review and evaluation of this study by the Board of Education’s Committee on Teacher Recruitment, Training and Compensation and the Department Staff, the State established new certification requirements involving different minimum scores in various areas of teaching specialization that range from 940 to 1198. The court holds that these regulations are properly before it since the plaintiffs 2 were undoubtedly aware that the validity study by ETS upon which these regulations were based was well under way when the suits were filed. The local boards are required by the State to hire only certified teachers, S.C. Code §§ 21-45, 21-371, 21-375, but there are no uniform standards with respect to test scores used by the local school boards in selecting from among the pool of certified applicants. Plaintiffs challenge each of the uses of the NTE. They contend that more blacks than whites historically have failed to achieve the required minimum score, and that this result creates a racial classification in violation of the constitutional and statutory provisions cited in their complaints. Each complaint seeks declaratory and injunctive relief with respect to the use of the minimum score requirement in certifying and determining the pay levels of teachers, injunctive relief to upgrade the certification levels of teachers adversely affected by the minimum score requirement, and monetary relief for alleged financial losses of teachers, together with costs. Plaintiff-intervenors also asked for attorneys’ fees. Preliminary matters: There are three preliminary matters pending before this court which should be decided at the outset. By the court’s order of June 17, 1976, rulings on the objections of all parties to the admissibility of certain evidence, and the determination of whether to certify the defendant class were reserved to the three-judge panel. At oral argument before the three-judge panel, plaintiffs moved for a review of the court’s prior order permitting the defendants to conduct redirect examination of one witness by written questions. Admissibility of evidence: We have considered all of the evidence offered by the parties and the objections thereto. In view of the fact that the court itself is the finder of fact, we admit and consider all of the evidence offered by all parties, and base our findings on the weight and probative value to be accorded the evidence in light of objections made by counsel. Certification of the defendant class : Plaintiffs sought to bring their actions against a defendant class of all local school boards within the State, and by timely motion sought to certify the class under Rule 23, Federal Rules of Civil Procedure. The motion to certify is denied. The defendant class sought by plaintiffs does not meet the requirements of the Rule since they have failed to establish that there are questions of law or fact common to the class. Each of the 92 school districts involved apparently has different policies with respect to the hiring, salary and tenure of teachers. It appears that some districts hire only teachers with professional certificates plus additional credentials; some hire on the basis of professional certificates alone; others are willing to consider warrant holders. Some districts provide a local salary supplement for nearly all positions; while others provide such supplements for a few classes of positions or none at all. Further, the plaintiffs have made no proper showing that the three school districts named as class representatives have defenses that are typical of the class, and that they will fairly and adequately protect the interests of the class. Thus another requirement of the Rule has not been met by plaintiffs. It is therefore concluded that plaintiffs have failed to carry their necessary burden to establish that their action falls within any of the provisions of Rule 23 in order to justify the court’s certifying it as a class action as sought by them. Review of the Order permitting examination by written questions : Plaintiffs asked *1099 the three-judge panel to review and overturn the order entered by the district judge (acting for the three-judge panel in all preliminary and prehearing matters) granting a motion by the State defendants to make certain written questions propounded during discovery and the answers thereto a part of the record. This dispute arose during the deposition of Dr. Winton H. Manning, a Vice President of ETS which administers the NTE. The plaintiffs deposed Dr. Manning for several days resulting in nearly 2000 pages of transcript. ETS being concerned with plaintiffs’ inability -to conclude the deposition of this senior officer offered to answer written questions in lieu of continuing the deposition by oral questioning. 3 Plaintiffs agreed and submitted 188 written questions on cross-examination. ETS answered these questions promptly. Thereafter, Dr. Manning submitted to another day of oral cross-examination. Defendants then submitted 174 written questions in lieu of redirect examination by oral questions. ETS answered those questions. Plaintiffs then submitted 54 written questions in lieu of oral re-cross examination which ETS answered. Plaintiffs objected to the admission into evidence of the answers to the questions on redirect examination. They urge that the procedure is not permitted by the Federal Rules, or authorized by any agreement of the parties. We hold that the agreement of the parties did not give plaintiffs the exclusive right to use written questions in this regard, and such agreed to procedure is not inconsistent with the Federal Rules, and did not adversely affect the plaintiffs. It would be unfair to deny the defendants the same opportunity to use written questions as was afforded to plaintiffs. Plaintiffs’ motion is therefore denied. I. CONSTITUTIONAL ISSUES We first consider whether the use by the State and its Board of Education of a minimum score requirement on the NTE violates the equal protection clause of the Fourteenth Amendment. 4 We conclude as a preliminary matter that the plaintiff-intervenors’ causes of action under the Fourteenth Amendment against the State and the three State agencies are barred by the Eleventh Amendment. 5 The Declaratory Judgment Act, 28 U.S.C. §§ 2201-02 (1970), does not explicitly authorize suits against states and therefore cannot be used to circumvent the Eleventh Amendment. Edelman v. Jordan, 415 U.S. 651, 94 S.Ct. 1347, 39 L.Ed.2d 662 (1974). The claims for monetary and injunctive relief under the Fourteenth Amendment are barred because a state is not divested of its immunity “on the mere ground that the case is one arising under the Constitution or laws of the United States.” Parden v. Ter *1100 minal R. Co., 377 U.S. 184, 186, 84 S.Ct. 1207, 1209, 12 L.Ed.2d 233 (1964), quoting Hans v. Louisiana, 134 U.S. 1, 10, 10 S.Ct. 504, 33 L.Ed. 842 (1890). The state agencies, by the nature of their responsibilities, function as an arm or “alter ego” of the State and are, therefore, protected by the State’s immunity. Brennan v. University of Kansas, 451 F.2d 1287, 1290 (10th Cir. 1971). We also dismiss the plaintiff-intervenors’ causes of action under Sections 1981 and 1983 against all defendants except the members and officers of the institutional defendants. 6 States are immune from suit by virtue of the Eleventh Amendment on the basis of 42 U.S.C. § 1983. Burton v. Waller, 502 F.2d 1261 (5th Cir. 1974), cert, den. 420 U.S. 964, 95 S.Ct. 1356, 43 L.Ed.2d 442 (1975); Percy v. Brennan, 384 F.Supp. 800, 809 (S.D.N.Y.1974); Francis v. Davidson, 340 F.Supp. 351, 370 (D.Md.1972), aff’d, 409 U.S. 904, 93 S.Ct. 223, 34 L.Ed.2d 168 (1972). None of these defendants are “persons” within the meaning of Section 1983. City of Kenosha v. Bruno, 412 U.S. 507, 93 5. Ct. 2222, 37 L.Ed.2d 109 (1973); Fitspatrick v. Bitzer, 427 U.S. 445, 96 S.Ct. 2666, 49 L.Ed.2d 614 (1976); Monroe v. Pape, 365 U.S. 167, 81 S.Ct. 473, 5 L.Ed.2d 492 (1961). Similarly, none of these defendants are subject to suit under Section 1981. Arunga v. Weldon, 469 F.2d 675 (9th Cir. 1972); Black Brothers Combined v. City of Richmond, 386 F.Supp. 147 (E.D.Va.1974); Bennett v. Gravelle, 323 F.Supp. 203, 216-17 (D.Md.), aff’d, 451 F.2d 1011 (4th Cir. 1971), cert. dismissed 407 U.S. 917, 92 S.Ct. 2451, 32 L.Ed.2d 692 (1972). In disposing of the remaining constitutional claims, separate consideration must be given to the State’s use of the test scores to certify teachers and to determine the amount of state aid for local school districts. Plaintiffs allege that the disparate racial impact of defendants’ certification and compensation systems creates a racial classification in violation of the Fourteenth Amendment. In order to sustain that allegation, the Supreme Court’s decision in Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976), requires plaintiffs to prove that the State intended to create and use a racial classification. If plaintiffs fail to prove intent (or defendants adequately rebut that proof), then we must evaluate this classification under the rational relationship standard required by the Fourteenth Amendment as to all such classifications. A. Discriminatory Intent Because of its paramount importance under Washington v. Davis, we look first at whether the plaintiffs have proved that any of the challenged decisions of defendants were motivated by an intent to discriminate. The purpose or intent that we must assess is the purpose or intent that underlies the particular act or acts under review. In its decision in Village of Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977), the Supreme Court suggested that evidence as to several factors might have probative value in proving intent; historical background, the sequence of events leading up to the challenged decision (including substantive and procedural departures from the norm), legislative his *1101 tory, and testimony from officials. 429 U.S. at 264-268, 97 S.Ct. 555. With these factors in mind, we look separately at the defendants’ actions with respect to the use of the NTE for certification purposes and those relating to its use for state aid purposes. 1. Certification South Carolina requires persons who teach in the public schools to hold a certificate issued by the State Board of Education. S.C.Code § 21-354. From 1945 to the present the State has had four certification systems, each requiring prospective teachers to take the NTE. Candidates are able to take the NTE an unlimited number of times. (The tests are given by the State three or four times each year). The record before us indicates that during this period, the racial composition of the South Carolina teacher force has closely paralleled the racial composition of the State’s population. The 1950 census estimated that 61.1% of the population was white and 38.8% was black. In 1953-54, the State employed a teacher force that was 58% white and 42% black. The 1960 census showed a population that was 65.1% white and 34.9% black. In 1966-67, the teacher force was 65.6% white and 34.4% black. The 1970 census showed a population that was 69.5% white and 30.5% black. In 1975-76, the teacher force was 70.1% white and 29.3% black. From 1945 through 1968, the State issued four grades of certificates: A, B, C and D. From 1945 through 1956, candidates were awarded certificates based on their relative standing with respect to test scores of all candidates in the State for the year: A certificates went to the top 25%; B certificates to the middle 50%; C certificates to the next 15%; and D certificates to the bottom 10%. Under this system, every candidate who met the other requirements was licensed. In 1957, a new system of absolute, rather than relative, requirements was instituted. Under this system, a score of 500 or more on the Common Examinations portion of the NTE 7 was required for an A certificate; a score of 425 to 499 for a B certificate; a score of 375 to 424 for a C certificate; and a score of 332 to 374 for a D certificate. Those with scores below 332 were not licensed. During the academic year 1967-68, that restriction eliminated less than 1% of the candidates from predominantly white colleges and approximately 3% of the candidates from predominantly black colleges. The reason for these low percentages is that the 332 minimum score, when placed on the NTE score scale of 300 to 900 points, is very near the lowest attainable score. In 1969, the certification system was further revised by replacing the four-tiered system with two types of certificates: the professional certificate and the warrant. The minimum score requirement was set at 975 for the professional certificate, including a score of at least 450 on the Common Examinations and at least 450 on the applicable Area Examination; and 850 for the warrant, including a score of at least 400 on the Common Examinations and 400 on the applicable Area Examination. Those who attained scores below 400 were not licensed. In the academic year 1969-70, the maximum score requirement of 400 on the Common Examinations eliminated approximately 41% of the graduates of predominantly black colleges and less than 1% of the graduates of predominantly white colleges. Similar results were obtained in succeeding years, despite the fact that a score of 400 is usually below the 11th percentile nationally, *1102 and almost 90% of the candidates who take these tests get a higher score. In 1976, the certification system was again revised, the two-tiered system being replaced with a single certificate with separate minimum score requirements in each of the 18 fields of teaching specialty replacing the single minimum score requirement. These combined scores on both Common Examinations and Area Examinations ranged from 940 in Agriculture to 1178 in Library and Media Specialties; and are set forth in detail hereinafter. There are no statistics in the record indicating the impact of the new score requirements because they will be applied first to the class of 1977; however, plaintiffs predict that, under these requirements, the disparate impact may be even greater. Plaintiffs have asserted four acts by the State indicating discriminatory intent: (1) The decision in 1945 to institute an NTE requirement; (2) The decision in 1956, effective in 1957, to institute an absolute rather than relative score requirement; (3) The decision in 1969 to revise the absolute score requirement to include the Area Examinations as well as the Common Examinations and raising the previous required scores; and (4) The decision in 1976 to change the single composite score requirement to separate composite score requirements for each teaching field, and in all but one case raising the required scores again. We conclude for the following reasons that such actions by the State were not motivated by an intent to discriminate because of race. The 1945 decision : In 1941, a committee was appointed by the General Assembly to review the certification and compensation of teachers. The committee recommended that a thorough study of the certification system be made and suggested that the use of an NTE score requirement be explored because of the great variation in teacher training institutions. The State Board undertook a two-year study that resulted in a four-volume report, made available in 1944. That report recommended four different teacher credentials based on relative NTE scores. On the basis of the report, the State Board adopted the four-tiered system described above under which no candidate was denied licensing. At that time, South Carolina maintained a dual public school system with segregated student bodies and faculties. After the decision to institute the new certification system, but before its effective date, the State Board tested 50% of the teachers in the State, which revealed that 90% of the white teachers, and only 27% of the black teachers, would qualify for A or B certificates (the two top grades). The remaining candidates — 10% of the whites and 73% of the blacks — would receive C and D certificates. After receiving these data, the State Board did not rescind its decision to institute the new certification system. We are unable to find any discriminatory intent from these facts. Although the historical background of segregated schools might provide some basis for the inference urged by plaintiffs, any such inference has been rebutted. The committee based its recommendation concerning the NTE in part on its conclusion that the tests “can be scored objectively and impartially and their use would not be subject to the accusation that they are used for purposes of discrimination.” 8 The Board’s extensive study is viewed as an earnest effort in its time, and provided reasonable support for its decision to institute an NTE requirement. The Board’s knowledge of differential impact, without more, does not support a finding of discriminatory intent. The 1956 decision: From 1945 through 1956, the'State Board received from ETS statistical summaries of the test scores of South Carolina candidates showing that blacks as a group had lower average scores than whites as a group. In 1954 and 1955, the Supreme Court handed down its historic decisions in Brown v. Board of Education, *1103 347 U.S. 483, 74 S.Ct. 686, 98 L.Ed. 873 (1954), 349 U.S. 294, 75 S.Ct. 753, 99 L.Ed. 1083 (1955), which announced the end of the “separate but equal” rationale for segregated school systems. South Carolina did not thereafter integrate its schools, but instead adopted policies which maintained its dual school system until well into the 1960’s. In 1956, the School Board adopted the absolute score system without a professional study relating the scores to effective teaching or academic achievement in the teacher preparation program. Again, we find nothing in the record to support a finding of discriminatory intent. Plaintiffs have not persuaded the court that the State’s reluctance to integrate its schools infected its decisions on teacher certification. The minimum score requirement of 332 was so low as to preclude us from presuming or even inferring that this was the case. Only 3% of the candidates from black colleges (and 1% of the candidates from white colleges) were denied certificates under this system during the last year it was in effect; and its effect in earlier years was not substantially different. The State Board did not move to change the system when it discovered that very few blacks were excluded, and significantly, this system was maintained for over 11 years. The 1969 decision: From 1957 through 1969, the State Board continued to receive from ETS statistical data indicating disparate impact by race. During this period, the State continued to maintain its essentially dual school system through a variety of administrative measures. 9 In 1968, the State Board received a report from a committee formed in 1967 to study the certification system. The committee included black and white members who were educators, teachers, state administrators and others. The committee had before it, inter alia, two documents, one which was critical of the NTE as a predictor of performance in student teaching, and the other which set out the ETS position that the test scores should not be used as a sole criterion for certification if other measures were available. The committee reviewed the academic programs and admission requirements at the State’s 25 teacher training institutions and determined that not all had the same resources and strengths. The committee recommended that the minimum score levels be raised to 975 for professional certificates and 850 for warrants. From the data supplied by ETS, the committee would have been able to predict the impact of its recommendations on black teacher candidates as a group and on white candidates as a group (although it is unclear that such a prediction was ever made). The State Board adopted the committee’s recommen1 dations. We are unable to find any intent to discriminate with respect to this decision. Plaintiffs offer no direct evidence of such an intent, even though one obvious source, the black members of the committee, was apparently available to them. The State’s authority to re-define minimal competence from time to time cannot reasonably be questioned. The 1976 decision: ETS urged that the State validate its cut-off score adopted in 1969. In 1971, ETS issued Guidelines that made its position clear. In the absence of any affirmative response by the State, in 1974 ETS announced its intention to cease the reporting of scores to South Carolina. In 1975, a three-judge court issued an interim decision in the ease styled United States v. North Carolina, 400 F.Supp. 343 (E.D.N. C.1975), vacated 425 F.Supp. 789 (E.D.N.C. 1977) which required objective proof by the State of a rational relationship between the minimum score requirement on the NTE and the State’s objective of certifying only minimally competent teachers. Shortly thereafter, the State Board authorized an extensive validity study which was conduct *1104 ed by ETS over a period of three months. The results of the study are set out in a two-volume, 300-page report, which was delivered to the State Board in January, 1976. The Board adopted the study report’s recommendation that there be separate minimum score requirements by teaching field, and set new higher score levels based on the data produced by the study. The 1976 decision carries a separate importance inasmuch as a finding with respect to it is the only basis for injunctive relief. Intent with respect to prior decisions is relevant only to damages. Plaintiffs offer no additional proof with respect to the defendants’ alleged intent to discriminate in establishing the new certification requirements in 1976. They apparently rely on the cumulative historical background and the imminence of litigation with regard to the prior system. Even if we had found intent to discriminate with respect to one of the State’s earlier decisions, this cumulative history would be without probative value as to the 1976 decision. While historical circumstances may illuminate the purpose of a particular State action, recent events are far more probative than distant events. With the exception of the 1969 adoption of minimum cutoff scores, the “history” relied on by plaintiffs occurred over 20 years ago. Plaintiffs’ elaborate web of historical circumstances, from which the court is being asked to infer discriminatory purpose, is noticeably silent about recent events. Significantly, the cases relied on by the Supreme Court in Village of Arlington Heights v. Metropolitan Housing Development Corp., supra, and by plaintiffs in this case (for the proposition that historical circumstances do bear upon invidious purpose) involved events occurring within four years of the state act or decision under review. With respect to the constitutional challenge to South Carolina’s use of the NTE for certification purposes, we conclude that the plaintiffs have not demonstrated the required discriminatory intent with respect to any of the specific decisions setting certification standards based on NTE scores. This is especially true in connection with the State’s 1976 change in requirements where there is no indication whatsoever that the State and its officers were motivated by anything more than a desire to use an accepted and racially neutral standardized test to evaluate the teacher applicants competing for relatively few jobs in South Carolina. The NTE are developed and administered by ETS, an independent non-profit organization of recognized professional reputation. ETS recommends that minimum score requirements not be used as a sole determinant of certification decisions where other appropriate information or criteria are available. In this case, the. plaintiffs have come forth with no other reasonably appropriate criteria upon which certification may be properly based. Neither have plaintiffs been able to establish any defect in the NTE indicating that the examinations themselves discriminate on the basis of race. The choices as to subject matter and question format are reasonable and well-documented on the record, and although other subject matters or other examination forms might be possible or even preferable, there is no proof of any inherent discrimination. The inference that plaintiffs would have us draw from the statistics which indicate that blacks as a group have lower average scores than whites is rebutted by the evidence with respect to the construction of the tests and their content validity. Since we find that the NTE create classifications only on permissible bases (presence or absence of knowledge or skill and ability in applying knowledge), and that they are not used pursuant to any intent to discriminate, their use in making certification decisions by the State is proper and legal. 2. Pay Scales Plaintiffs raise a separate constitutional challenge to South Carolina’s use of the NTE as a partial determinant of the salaries paid to public school teachers. The use *1105 of the NTE for salary purposes is distinct from, and, in significant ways, unrelated to the use of the NTE for certification purposes. Accordingly, we must again examine plaintiffs’ proof with respect to intent. Public school teachers are hired and paid by local school boards in each of the 92 school districts within the State. Each year since 1942 the Legislature has provided state funds for use by the local school districts in paying teacher salaries. This state aid is provided in accordance with a'schedule, enacted into law annually. The schedule is the formula by which the amount of state aid for each school district is calculated. In some districts, teachers are paid only the amount available through state aid. In most districts there is a local supplement. The formula for state aid with respect to each teacher’s salary takes into account the number of years of college or post-graduate training, the number of years of teaching experience, and the type of teaching credential held (professional certificate, warrant, or A, B, C, or D graded certificate). 10 Plaintiffs challenge the state-aid formula insofar as it uses as one criterion the type of teaching credential and thus, indirectly, NTE scores. 11 Plaintiffs claim that the formula creates a racial classification and was so intended by the State. In this instance we have approximately 34 separate annual decisions by the Legislature to enact the state aid schedule and to appropriate funds pursuant to that schedule. However, there are two turning points with respect to our inquiry into intent: the decision in 1945 to relate eligibility for state aid to certification requirements and thereby incorporate NTE scores into salary decisions; and the decision in 1969 to retain as a basis for the state aid formula the four-tiered certification system with its built-in NTE score requirements. With respect to each of these decisions, we find as follows: The 1945 decision: Before 1940, South Carolina had a dual pay system in which black teachers were paid less than white teachers; both having the same credentials and responsibilities. In 1940, a similar dual pay system maintained by the State of Virginia was held unconstitutional. Alston v. School Board, 112 F.2d 992 (4th Cir.), cert, denied, 311 U.S. 693, 61 S.Ct. 75, 85 L.Ed. 448 (1940). There was no consideration, in that case, of South Carolina’s pay system, but the Court’s view of the legality of any dual pay system was made clear. In 1941, a committee appointed by the General Assembly began to examine the salary system, and that work was carried on by the State Board which produced in 1944 a recommendation for a unitary system. The dual pay system thus survived until 1945 when the Legislature enacted a unitary system of state aid applicable to all teachers. That system was based in part on the certification grading system adopted by the Board of Education in that year. At the time the Legislature acted, it had access to the results of the November, 1944, administration of the NTE to about half of the teachers in the State. Those results indicated that 32% of the white candidates would be classified as A certificate holders while only 2% of the black candidates would be so classified. At the other end of the scale, 1% of the whites and 37% of the blacks would be classified as D certificate holders. In between, 67% of the whites and 61% of the blacks would be entitled to B and C certificates. This meant that although blacks and whites would have equal access to salary incentives for college and graduate work and salary rewards for longevity, the amount of those salary supple *1106 ments would be affected by the grade of the certificate held, with the highest supplements available to the A certificate holders and the lowest supplements available to the D certificate holders. The difference in supplements available to the holders of various grades of certificates was not uniform between each of the four tiers of the system. This differential was generally highest, in absolute terms, between D and C certificates, somewhat less between C and B certificates, and lowest between B and A certificates. Under this system, the Legislature provided to the holders of D certificates the greatest monetary incentive for improvement of the grade of the certificate. The other incentives to acquire additional education or rewards for longevity while holding the same certificate were to some extent independent of the incentives to improve the grade of the certificate. Soon after this salary system was put into effect, it was challenged on Fourteenth Amendment grounds. The federal district court (Judge Waring) found that, although the prior dual pay system was arbitrary and could not be sustained, the new system had “eliminate[d] entirely any danger of disparity or discrimination by reason of race or color.” Thompson v. Gibbes, 60 F.Supp. 872, 878 (E.D.S.C.1945). We are unable to find a discriminatory intent from these facts, even though the historical background of a dual pay system and a delay in implementing a unitary system after the Fourth Circuit struck down a similar system in another state provide some support for such an inference. Such inference is adequately rebutted by the evidence with respect to what the Legislature actually did. The unitary pay system was based in part on the amount of educational training and years of teaching experience possessed by each teacher. Plaintiffs make no claim that the use of either of these factors was motivated by discriminatory intent, and it is evident that the monetary rewards available through these avenues alone, without regard to the grade of certificate, were significant. The link between the new unitary pay system and the new certification system is not without a reasoned basis. It was important to the State to use its limited resources to improve the quality of the teacher force and to put whatever monetary incentives were available in the salary schedule to that task. As before stated, we have found that no discriminatory intent has been established with respect to the decision in 1945 to adopt a certification system based in part on NTE scores; and, therefore, without independent proof, there is no associated discriminatory intent in linking the certification and salary systems. The only independent proof offered by plaintiffs is the availability of statistics provided by ETS showing differential performance by race and the presumption that the Legislature knew of this impact when it enacted the salary system. Statistics indicating differential impact are without independent constitutional significance, Washington v. Davis, 426 U.S. 229, 242, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976), therefore knowledge of such statistics is similarly unpersuasive. The 1969 decision: In July 1969, the State Board replaced the old A-B-C-D certificates (prospectively only) with a professional certificate and a warrant. At the same time the Legislature retained the state-aid formula based on the A-B-C-D classifications. Holders of professional certificates and warrants were paid at the rates applicable to holders of A certificates. 12 Some changes tinkering with the *1107 system had been made over the years and was reflected in the system retained in 1969. From 1945 through 1969, statistical data was made available indicating that the percentage distribution of teachers among the four categories of teaching credentials was not uniform by race. Significant numbers of blacks qualified each year for A and B certificates, but a relatively higher percentage of blacks were in the C and D categories. Similarly, although it could be predicted that significant numbers of blacks would achieve professional certificates, a relatively higher percentage of blacks would hold warrants. Plaintiffs urge that the Legislature was motivated by an intent to discriminate because the state aid schedule provided fewer benefits and incentives for two classifications of teachers in which there were relatively more blacks than whites; the classifications continued to be based, if only in part, on examination scores; and the developer of the tests was later opposed to their use for this purpose. In the absence of any competent or persuasive direct evidence of intent, plaintiffs would have us draw an inference from the facts that are available. But such an inference can stand only if there is no equally persuasive explanation consistent with a legitimate intent. Here the reordering of priorities with respect to the limited resources available for teacher salaries appears entirely consistent with an intent to obtain the maximum incentive for improvement that could be accomplished with a fixed number of dollars. We are unable to find an intent to discriminate. B. Application of Rational Relationship Standard In the absence of discriminatory intent, the classifications of teachers for both certification and pay purposes may be assessed under the “rational relationship” standard required by the Fourteenth Amendment of all classifications. The Supreme Court has defined this standard in the following terms: Although no precise formula has been developed, the Court has held that the Fourteenth Amendment permits the States a wide scope of discretion in enacting laws which affect some groups of citizens differently than others. The constitutional safeguard is offended only if the classification rests on grounds wholly irrelevant to the achievement of the State’s objective. State legislatures are presumed to have acted within their constitutional power despite the fact that, in practice, their laws result in some inequality. A statutory discrimination will not be set aside if any state of facts reasonably may be conceived to justify it. McGowan v. Maryland, 366 U.S. 420, 425-26, 81 S.Ct. 1101,1105, 6 L.Ed.2d 393 (1961). We conclude that the State’s use of the NTE for both certification and pay purposes meets the “rational relationship” standard of McGowan v. Maryland, supra, and consequently does not violate the equal protection clause of the Fourteenth Amendment. No more rigorous constitutional standard need be applied to a case which does not involve express differentiation by race or other “suspect” classification, and does not involve discriminatory intent. We find however, that were an intermediate standard applied, the defendants’ use of the NTE bears a “fair and substantial relationship to the achievement of an important and constitutionally permissible governmental objective.” (See Wood v. Mills, 528 F.2d 321 (4th Cir. 1975), and Eslinger v. Thomas, 476 F.2d 225 (4th Cir. 1973)). Nevertheless, we find that the defendants have offered a legitimate and important governmental objective for their use of the NTE. The State has the right to adopt academic requirements and to use written achievement tests designed and validated to disclose the minimum amount of knowledge necessary to effective teaching. *1108 Dent v. West Virginia, 129 U.S. 114, 9 S.Ct. 231, 32 L.Ed. 623 (1889); United States v. North Carolina, 400 F.Supp. 343, 348-49 (E.D.N.C.1975) (three-judge court), vacated on other grounds, 425 F.Supp. 789 (E.D.N.C. 1977). The evidence in the record supports a finding that South Carolina officials were concerned with improving the quality of public school teaching, certifying only those applicants possessed of the minimum knowledge necessary to teach effectively, utilizing an objective measure of applicants coming from widely disparate teacher training programs, and providing appropriate financial incentives for teachers to improve their academic qualifications and thereby their ability to teach. We conclude that these are entirely legitimate and clearly important governmental objectives. In considering whether defendants’ use of the NTE bears a fair and substantial relationship to these governmental objectives, we conclude that it does. The record supports the conclusion that the NTE are professionally and carefully prepared to measure the critical mass of knowledge in academic subject matter. The NTE do not measure teaching skills, but do measure the content of the academic preparation of prospective teachers. United States v. North Carolina, 400 F.Supp. 343, 350 (E.D.N.C.1975), vacated on other grounds, 425 F.Supp. 789 (E.D.N.C.1977). Like the test at issue in Washington v. Davis, supra, the NTE program “is neutral on its face and rationally may be said to serve a purpose the Government is constitutionally empowered to pursue.” 426 U.S. at 246, 96 S.Ct. at 2051. Plaintiffs have not contended nor proved that the NTE are racially biased or otherwise deficient when measured against the applicable professional and legal standards. Furthermore, there is ample evidence in the record of the content validity of the NTE. The NTE have been demonstrated to provide a useful measure of the extent to which prospective teachers have, mastered the content of their teacher training programs. In a similar challenge to a bar examination the Fourth Circuit has held that proof of such content validity is persuasive evidence that the equal protection clause has not been violated. Richardson v. McFadden, 540 F.2d 744 (4th Cir. 1976). The Supreme Court has held that a substantial relationship between a test and a training program — such as is found here — is sufficient to withstand challenge on constitutional grounds. Washington v. Davis, 426 U.S. at 248-52, 96 S.Ct. 2040. 13 State officials surely have the right to require graduation from approved teacher training programs as a prerequisite to being certified to teach in South Carolina. Plaintiffs have acknowledged the substantial relationship between the academic training program and the job of teaching by advocating that a requirement of graduation from an approved program alone is sufficient to protect the public interest. Also, the minimum score of 975 selected by South Carolina in 1969, and the range of scores selected in 1976 were permissible exercises of judgment by the responsible officials. In fact the minimum scores adopted by the State in 1976 are for the most part substantially lower than those recommended for adoption by ETS. As in Washington v. Davis, “there is no evidence that the required passing grade was set at an arbitrarily high level.” Id. at 254, 96 S.Ct. at 2055 (concurring opinion of Mr. Justice Stevens). The evidence supports the conclusion that the 975 score was selected in large part because it enabled the State to deny certification to only those who scored in the lowest 10% of all candidates taking the NTE throughout the United States; and that score reflected the judgment as to *1109 minimal competence of a committee of independent responsible professionals. Such exercise of judgment, although subjective and imperfect, may nonetheless serve to support a minimum test score requirement challenged on equal protection grounds. Richardson v. McFadden, 540 F.2d 744 (4th Cir. 1976); Tyler v. Vickery, 517 F.2d 1089 (5th Cir. 1975), cert. denied, 426 U.S. 940, 96 S.Ct. 2660, 49 L.Ed.2d 393 (1976). Although we agree that a professionally designed and executed validity study is not necessarily required to demonstrate the relationship between a challenged use of a test and the governmental objective for which it is being used, Tyler v. Vickery, supra, we find support for our conclusions with respect to the NTE in the validity study conducted in this case and discussed below in connection with our analysis of plaintiffs’ contentions under Title VII. We also conclude that defendants’ use of the NTE for salary purposes bears the necessary relationship to South Carolina’s objectives with respect to its public school teaching force. 14 Although the NTE were not designed to evaluate experienced teachers, the State could reasonably conclude that the NTE provided a reliable and economical means for measuring one element of effective teaching — the degree of knowledge possessed by the teacher. Having so concluded, defendants could properly design a classification system relying on NTE scores for compensating teachers and providing incentives for teachers to improve their knowledge in the areas that they teach. The retention of this system for pay purposes after 1969, when the requirements for certification were changed, cannot be said to be arbitrary or completely without justification. The defendants were under no obligation to change all components of the program at the same time irrespective of the impact of such changes upon the State’s financial situation and its continued interest in providing incentives for teachers with low NTE scores. II. TITLE VII ISSUES We turn now to the question whether defendants’ uses of the NTE violate Title VII, 42 U.S.C. § 2000e, et seq. (1970). Defendants raise four legal challenges to plaintiffs’ Title VII claims which we must consider at the outset: (1) whether the Eleventh Amendment bars plaintiffs’ claims; (2) whether defendants are employers within the coverage of Title VII; (3) whether Title VII applies to certification activities; and (4) whether the Attorney General has the authority to bring this “pattern or practice” action. Neither the claims of the United States nor the plaintiff-intervenors under Title VII are barred by the Eleventh Amendment. In Fitzpatrick v. Bitzer, 427 U.S. 445, 96 S.Ct. 2666, 49 L.Ed.2d 614 (1976), the Supreme Court held: [T]he Eleventh Amendment, and the principle of state sovereignty which it embodies, . . . are necessarily limited by the enforcement provisions of § 5 of the Fourteenth Amendment. . When Congress acts pursuant to § 5, not only is it exercising legislative authority that is plenary within the terms of the constitutional Amendment whose other sections by their own terms embody limitations on state authority. We think that Congress may, in determining what is “appropriate legislation” for the purpose of enforcing the provisions of the Fourteenth Amendment, provide for private suits against States or state officials which are constitutionally impermissible in other contexts. (427 U.S. 445, at 456, 96 S.Ct. at 2671). Applying that test, the Court held that an action under Title VII cannot be dismissed on Eleventh Amendment grounds because Title VII was enacted pursuant to Section 5 of the Fourteenth Amendment. The State and its three agencies assert that they are not “employers” of teachers within the meaning of Title VII because all *1110 public school teachers are hired, paid and discharged by local school boards. These defendants further assert that their activities are only “certification” of teachers and not “selection,” therefore, they are not within the coverage of Title VII. The factual record on these issues was not well-developed because the parties concentrated their efforts on the disparate impact statistics and the validity study. All parties concede that the local school boards are “employers” within the meaning of Title VII and that their activities include “selection” practices that are covered by Title VII. Even though we have declined to certify the defendant class that would have put at issue the practices of all 92 local school boards within the State, the three individual school board defendants are properly before us. The local school boards defend against plaintiffs’ claims by asserting that their practices with respect to the NTE are governed for the most part by the State’s regulations. Because of our decision as to liability, we find it unnecessary to determine whether the State and its three agencies are employers of teachers within the meaning of Title VII. The defendants also assert that, after the 1972 amendments to Title VII, the Attorney General of the United States was without authority to bring this “pattern or practice” action against them. The 1972 Amendments broadened the scope of Title VII significantly by extending its coverage to state governments and governmental agencies, and augmented the powers of the Equal Employment Opportunity Commission (EEOC) by authorizing the Commission to bring civil actions where it is unable to obtain acceptable conciliation agreements from employers. Section 707 of Title VII was amended to transfer to the Commission the Attorney General’s authority to initiate “pattern or practice” actions. With regard to private employers, there can be no doubt that, since March 24, 1974, at which point the transfer of authority under Section 707 became complete, the Commission has had sole authority to bring “pattern or practice” actions, and that the Attorney General has no such authority. The Attorney General has not claimed otherwise in this case. It is also clear from the language of Section 706(f) 15 that individual complaints against public employers are to be brought to the Commission, or initiated by Commissioners, and that the Commission is to proceed with its “informal methods of conference, conciliation and persuasion.” 16 Only when such methods fail does authority shift to the Attorney General, who is then empowered to bring a civil action. The Attorney General has no authority to investigate such charges or to bring such actions on his own initiative, but can only step in to sue public employers with respect to individual complaints when a case is referred to him by the Commission following the procedures prescribed in Section 706. Similarly, with the transfer to the Commission of the Attorney General’s authority “to investigate and act on a charge of a pattern or practice of discrimination” in Section 707, 42 U.S.C. § 2000e-6 (1970), .the Attorney General is stripped of any authority to initiate any “pattern or practice” litigation, as in the present case. But Section 707 need not be read to preclude prosecution of “pattern or practice” actions against public employers nor to permit the Commission to bring such actions against public employers in the face of Congress’ *1111 explicit rejection of such authority in cases involving individual complaints of discrimination. Section 707(e), after granting the Commission authority to investigate and act on “pattern or practice” charges, mandates that “[a]ll such actions shall be conducted in accordance with the procedures set forth in Section 706 [42 U.S.C. § 2000e-5 (1970)] of this Act.” Those procedures, as discussed above, require, inter alia, that the Commission investigate the validity of any charges, and attempt to reach a conciliation agreement with employers before bringing any court actions. 17 The logical and consistent reading of Section 707 is that the Commission’s “pattern or practice” authority is identical to its authority to act upon charges of individual discriminatory actions. It initiates investigations into charges of both types, seeks conciliation agreements, and, failing in those attempts, goes to court. And, in both types of actions, where the defendant is a public employer, the Commission cannot sue but, following unsuccessful attempts at a conciliation agreement, must refer the matter to the Attorney General. The Commission is thus assigned no greater authority to act against public employers in “pattern or practice” cases than it has in other actions under Section 706, yet public employers are not immunized against “pattern or practice” charges and litigation. Section 707 does not preclude the Attorney General from bringing “pattern or practice” actions against public employers on referral from the Commission, and, in fact, requires that such actions be brought, if at all, by the Attorney General. But the Act vests no authority in the Attorney General to initiate Title VII actions — “pattern or practice” or otherwise — against employers except upon referral from the Commission following - the procedures of Section 706. In the present case, as in United States v. Fresno Unified School Dist., 412 F.Supp. 392 (E.D.Cal.1976), the Commission has had no involvement with this case from its inception and . the Commission has not complied with the express requirements of Title 42, United States Code, Sections 2000e-6(e) and 2000e-5. The Attorney General has not contended otherwise. He has based his standing solely upon the contention that the Justice Department’s former independent pattern or practice authority survived the 1972 Civil Rights Act Amendment. The Attorney General’s argument is, however, without merit. He presently has no authority to act on his own initiative in a pattern or practice case. (Id. at 393). 18 The claims of the United States under Title VII are dismissed. The remaining claims under Title VII must be tested under statutory standards. In Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040,48 L.Ed.2d 597 (1976), the Supreme Court summarized the order of proof: Under Title VII, Congress provided that when hiring and promotion practices disqualifying substantially disproportionate numbers of blacks are challenged, discriminatory purpose need not be proved, and that it is an insufficient response to demonstrate some rational basis for the challenged practices. It is necessary, in addition, that they be “validated” in terms of job performance in any one of several ways, perhaps by ascertaining the minimum skill, ability or potential necessary for the position at issue and determining whether the qualifying tests are appropriate for the selection of qualified applicants for the job in question. (Id. at 246, 96 S.Ct. at 2051). Thus, it was held not sufficient for the governmental entity to prove that the clas *1112 sification resulting from the test scores had a rational basis, that is, that it differentiated between persons who did and did not have some minimum verbal and communication skill. It was necessary, in addition, for the governmental entity to demonstrate that the minimum verbal and communication skill, in turn, had some rational relationship to the legitimate employment objectives of the employer. And Washington v. Davis left intact the holding in Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971), that the employment practice must be a “business necessity.” Id. at 431, 91 S.Ct. 849. A. Certification Plaintiffs have proved that the use of NTE scores by the State in its certification decisions disqualifies substantially disproportionate numbers of blacks. The burden of proof was thereby shifted to the defendants, and in an effort to meet this burden the State commissioned an extensive validity study by ETS. The design of this study is novel, but consistent with the basic requirements enunciated by the Supreme Court, and we accordingly hold such study sufficient to meet the burden placed on defendants under Title VII. The study seeks to demonstrate content validity by measuring the degree to which the content of the tests matches the content of the teacher training programs in South Carolina. It also seeks to establish a minimum score requirement by estimating the amount of knowledge (measured by the ability to answer correctly test questions that have been content validated) that a minimally qualified teacher candidate in South Carolina would have. To conduct the study, all 25 of the teacher training institutions in South Carolina were canvassed for experienced teacher educators in the various specialty fields tested in the NTE program. A group of 456 persons with the requisite professional credentials was assembled including representative numbers from each institution and both races. All were volunteers nominated by the colleges themselves. These 456 participants were divided into two general groups. One group was assigned the task of assessing the content validity of the NTE as compared to the curriculum in South Carolina institutions. The other was assigned the task of establishing the minimum score requirement. The two large groups were then each subdivided into panels of about 10 participants assigned to each test in the Common Examinations and each of the Area Examinations. The panelists were given the questions and answers on two current forms of the NTE and asked to record certain judgments about the tests. Each content review panel member was asked to decide whether each question on the tests involved subject matter that was a part of the curriculum at his or her teacher training institution, and therefore could be judged to be appropriate for use in South Carolina. Each minimum score panel member was asked to look at each of the questions on the test and estimate the percentage of minimally qualified students in teacher education programs in South Carolina who would know the correct answer. Each test was evaluated by teacher educators specializing in the field or one of the major fields covered by the test. Art education teachers evaluated the test in Art Education; French teachers evaluated the test in French, and so on. The content review panels determined that from 63% to 98% of the questions on the various tests were content valid for use in South Carolina. The panel members’ overview of the tests as a whole also found the NTE to be sufficiently closely related to the curriculum in South Carolina to be an appropriate measure of achievement with respect to that curriculum. The estimates made by the minimum score panel members (as to the percentage of minimally qualified students who would answer the question correctly) were combined statistically and analyzed to generate scaled scores that reflected, for each test, the level that would be achieved by the minimally knowledgeable candidate. Only test questions that had been determined by a majority of the content review panel members to be content appropriate for use *1113 in South Carolina were used in making the minimum score estimates. These scores were 548 on the Common Examinations and a range from 458 (Agriculture) to 690 (Media Specialist) on the Area Examinations. After receiving the recommended minimum scores the State Board considered a variety of statistical and human factors. Among these considerations were the standard error of measurement of the NTE, the possibility of sampling error in the study, the consistency of the results (internal comparisons of panel results), the supply of and demand for teachers in each specialty field, and the racial composition of the teacher force. It decided to lower the minimum scores by one, two, or three standard errors of measurement. Test State’s New Scores ETS Study Scores SEM * Units Agriculture 940 1006 -2 Art Education 1040 1104 -2 Biology and General Science 1115 1196 -3 Business Education 1071 1167 -3 Chemistry, Physics, and Gen. Science 1100 1158 -2 Early Childhood Education 1052 1124 -2 Education in the Elementary School 1132 1194 -2 Education of the Mentally Retarded 1120 1234 -3 English Language and Literature 1041 1069 -1 French 1144 1210 -2 Home Economics Education 1075 1143 -2 Industrial Arts Education 1100 1162 -2 Mathematics 1113 1203 -3 Librarian-Media Specialist 1178 1238 -2 Music Education 1020 1086 -2 Physical Education 1125 1195 -2 Social Studies 1086 1148 -2 Spanish 1132 1163 -1 Common Examinations 510 ** 548 -2 The design of the validity study is adequate for Title VII purposes. The Supreme Court made clear once again in Washington v. Davis that a content validity study that satisfies professional standards also satisfies Title VII. 426 U.S. at 247, n. 13, 96 S.Ct. 2040. The defendants called as an expert witness Dr. Robert M. Guión, the principal author of Standards for Educational and Psychological Tests published by the American Psychological Association and a nationally recognized authority in the field of testing and measurement who testified in,an unqualified fashion that in his expert opinion the ETS study design met all of the requirements of the APA Standards, the Division 14 Principles, 19 and the EEOC Guidelines. 20 Two other experts testified similarly, and ETS sought and obtained favorable opinions on the study design, before its implementation, from another two independent experts. The ETS decision to validate against the academic training program rather than job performance is specifically endorsed in principle in Davis, supra: [A] positive relationship between the test and training-course performance was sufficient to validate the former, wholly aside from its possible relationship to actual job] performance as a police officer. . Nor is the conclusion foreclosed by either Griggs or Albemarle Paper Co. v. Moody, 422 U.S. 405, 95 S.Ct. 2362, 45 L.Ed.2d 280 (1975); and it seems to us the much more sensible construction of the job-relatedness requirement. 426 U.S. at 250-251, 96 S.Ct. at 2053. (Emphasis added). The principal issue raised by plaintiffs in attacking the validity study is whether the execution of the design was such that the results can be trusted. Plaintiffs deposed 81 of the 456 panel members selected at random and, on the basis of those depositions, claim that the panel members did not *1114 prepare for the tasks they were asked to undertake, and did not understand or follow the instructions. Admittedly, there is a showing from these depositions which indicates some misunderstanding did exist; however, our review of these depositions shows that the misunderstandings were much less extensive than claimed by plaintiffs. When so many people are involved some misunderstanding and failure adequately to follow instructions are inevitable. The new cutoff scores adopted by the State which are substantially below the scores recommended by ETS after completion of its validation study provide a substantial margin of error. It appears that most of the deposed panel members recalled receiving the instructions sent in advance of the panel meetings, but did not recall doing any of the preparation suggested by the instructions. We are asked to find that this lack of preparation requires a finding that the results are untrustworthy. It is evident from the design of the study that the panel members were selected on the basis of their professional experience, and it was this experience that they were asked to apply to the task of content review and minimum score determination, not the results of current research in college catalogues or interviews of teaching colleagues. For many, the preparation suggested by the instructions may have been unnecessary. In order to support the allegation of untrustworthiness, plaintiffs would have to offer some evidence as to the necessity of preparation, the reasons why the panel members did not prepare and the effect of the lack of preparation on the panel members’ judgments. We find that the results of the validity study are sufficiently trustworthy to sustain defendants’ burden under Title VII. First, the possible error rate was not high. Second, the key question is not whether some of the panel members failed to understand the instructions or for other reasons failed to follow the instructions, but whether they would have reached any different result if they had understood and followed the instructions. Plaintiffs misconceive their burden once defendants have made a reasonable showing that the study was executed in a responsible, professional manner designed to produce trustworthy results. In order to rebut the presumption that trustworthy results were indeed produced, plaintiffs must not only show that the study was not executed as intended, but also that the results were adversely affected. Our findings with respect to the validity study end the inquiry for purposes of injunctive relief with respect to the certification system because the regulations now in effect are based on the validity study, and the actions of the State Board in lowering the scores recommended by the study to account for various statistical and human factors were reasonable, and, as before stated, provide and compensate for a substantial margin of error. However, an ancillary inquiry remains with respect to the 975 minimum score requirement in effect from 1969 through 1976 and not specifically validated. Plaintiff-intervenors’ claims for damages bring the legality of that requirement into question. There is nothing in the record to indicate that the requirements of the teaching profession changed substantially from 1969 to 1976 so that what was judged to be the minimum amount of knowledge to graduate from a teacher training program or to teach effectively would have been higher or lower in 1976 than it was in 1969. The State has offered some evidence to the contrary, indicating that the expectations with respect to teacher capability remained about the same during this period. The 975 minimum score requirement in effect during these years is substantially below the State’s new minimum score requirements in every field except in Agriculture, which has been reduced to 940. It is reasonable to conclude, based on the present validity study, that the 975 minimum score excluded only those who did not have the *1115 requisite minimum knowledge. The only discrimination question raised is with respect to the classification of those who do not have minimum knowledge. During the years 1969 through 1976 some of those who lacked minimum knowledge were denied a professional certificate because their scores were below 975, while others who also lacked minimum knowledge but whose scores were above 975 but below the 1976 minimum score requirements, were certified. In any decision-making process that relies on a standardized test, there is some risk of error. The risk of excluding a truly qualified candidate whose low test score does not reflect his or her real ability can be decreased by lowering the minimum score requirement. That also increases the risk of including an unqualified candidate whose low test score does reflect his or her real ability. The State must weigh many facets of the public interest in making such a decision. If there is a teacher shortage, a relatively high minimum score requirement may mean that some classrooms will be without teachers, and it may be better to provide a less than fully competent teacher than no teacher at all. But to the extent that children are exposed to incompetent teachers, education suffers. It may be that education suffers less than would be the case if classrooms were overcrowded due to lack of teachers. We think it is within the prerogative of the State to accept some unqualified teachers under circumstances where that is judged by the State to be on balance in the public interest, and that such an action by the State is not a violation of Title VII. There remains, however, the question whether the State has satisfied the “business necessity” requirement set out in Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971). This “business necessity” doctrine appears neither in the explicit language nor the legislative history of Title VII. The Court in Griggs and subsequent Title VII cases did not establish judicial standards for determining whether a particular practice is a business necessity. The EEOC Guidelines are of little assistance because they were published before Griggs and have not been updated since that time. 21 We think that Griggs did not import into Title VII law the concept of “compelling interest” developed as a part of the “strict scrutiny” standard for assessing certain classifications under the Fourteenth Amendment. Under this concept, the Court would balance the disparate impact on blacks against the business purpose of the employer and uphold the business practice only if it were sufficiently “compelling” to overcome the disparate impact. It is our view that the Supreme Court intended an examination of the alternatives available with respect to the legitimate employment objective identified by the employer to determine whether there is available to the employer an alternative practice that would achieve his business purpose equally well but with a lesser disparate impact by race. In examining alternatives, the risk and cost to the employer are relevant. Here, plaintiffs have suggested only one alternative to the use of the NTE for certification purposes. Plaintiffs contend that mere graduation from an approved program should be sufficient and would have a lesser disparate impact on blacks. We cannot find that this alternative will achieve the State’s purpose in cer *1116 tifying minimally competent persons equally well as the use of a content-validated standardized test. The record amply demonstrates that there are variations in admissions requirements, academic standards and grading practices at the various teacher training institutions within the State. The approval that the State gives to the teacher training program is to general subject matter areas covered by the program, not to the actual course content of the program, and not to the means used within the program to measure whether individual students have actually mastered the course content to which they have been exposed. The standardized test scores do reflect individual achievement with respect to specific subject matter content, which is directly relevant to (although not sufficient in itself to assure) competence to teach, and thus the use of these scores for certification purposes survives the business necessity test under Title VII. B. Pay scales There remains, finally, the question whether the uses of the NTE for salary purposes are a violation of Title VII. 22 Where the salary system was linked to the certification system, some salary benefits were available only by improving the grade of the certificate; and that, in turn, could be done only by achieving certain minimum NTE scores. That system continued in effect after March 24, 1972, when Title VII was made applicable to states, and in 1976, the distribution of teachers within the four grades of the pay scale system showed a substantial disparate impact by race. A higher proportion of whites (98%) were classified in the two higher grades (A and B) and a higher proportion of blacks (51%) were classified in the lower grades (C and D). By such showing plaintiffs have satisfied their burden of proof under Title VII to shift the burden of proof to the defendants to show a rational relationship to a legitimate employment objective and to show business necessity. The State identifies its legitimate employment objective as providing an incentive for improvement, so that teachers without adequate knowledge to teach effectively will upgrade their capability; and the State offers the same evidence of a rational relationship between its pay scales and this objective as it did with respect to the constitutional challenges. We think that evidence is sufficient to establish the relationship. We believe that a distinction for pay purposes between those who are qualified as well as between those who are not qualified survives the business necessity test. There appears to be no alternative available to the State, within reasonable limits of risk and cost, for providing the incentive necessary to motivate thousands of persons to acquire, generally on their own time and at their own expense, the necessary additional academic training so that they will be minimally competent teachers. Having made the investment of four years in an undergraduate education, it seems reasonable to try to upgrade the talent of unqualified teachers where possible, rather than rejecting them altogether. In accordance with the foregoing findings and conclusions, we conclude that plaintiff and plaintiff-intervenors have failed to establish their right to any of the relief sought in their respective complaints. It is, therefore, ORDERED that judgment be entered in favor of the defendants. Footnotes [1] In view of the court’s conclusions as hereinafter set forth this is a final Order that effectively disposes of all issues in this litigation, and no further proceedings or hearings are necessary. [2] When used in the plural, “plaintiffs” includes the plaintiff and the plaintiff-intervenors. [3] Because ETS is not a party, it could not be subjected to discovery through written interrogatories except by its consent. Rule 33, Federal Rules of Civil Procedure. [4] The United States appears to base its complaint exclusively on the equal protection clause of the Fourteenth Amendment. Plaintiff-intervenors expressly base their complaint on both the due process and equal protection clauses, but do not specify their due process grounds. To the extent that plaintiff-intervenors invoke the substantive guarantees of the due process clause, the standard applied is the same as under the equal protection clause, and we treat them on the same basis. [5] The comparable cause of action brought by the United States is unaffected by the Eleventh Amendment. United States v. California, 297 U.S. 175, 56 S.Ct. 421, 80 L.Ed. 567 (1936); United States v. Texas, 143 U.S. 621, 642^43, 12 S.Ct. 488, 36 L.Ed. 285 (1892). The plaintiff-intervenors’ causes of action for injunctive relief against the officials of the three state agencies are not barred by the Eleventh Amendment. When the State’s officers act illegally, they act “without the authority” of the State and thus outside the shield of sovereign immunity. Ex parte Young, 209 U.S. 123, 28 S.Ct. 441, 52 L.Ed. 714 (1908). The plaintiff-intervenors’ claims for relief against the local school boards are not barred by the Eleventh Amendment because the relative autonomy of these boards puts them outside the category of agencies that are “alter-egos” of the State. Burt v. Board of Trustees, 521 F.2d 1201 (4th Cir. 1975); Adams v. Rich-land School District # 1, 412 F.Supp. 647 (D.S. C.1976). [6] The plaintiff-intervenors named as defendants the members of the State Board of Education, the executive secretary and director of the State Retirement System, the members of the Budget and Control Board, and the members of the three local school boards, all in their official capacities only. Prospective injunctive relief against state agency officers is not precluded by the Eleventh Amendment. Ex Parte Young, 209 U.S. 123, 28 S.Ct. 441, 52 L.Ed. 714 (1908); Edelman v. Jordan, 415 U.S. 651, 97 S.Ct. 1347, 39 L.Ed.2d 662 (1974). Insofar as the plaintiffintervenors seek damages from state agency officers in their official capacities, the Eleventh Amendment requires dismissal because the agencies with which they are affiliated are “alter egos” of the State and the State’s funds are at stake. Edelman v. Jordan, 415 U.S. 651, 97 S.Ct. 1347, 39 L.Ed.2d 662 (1974); Ford Motor Co. v. Dept, of Treasury, 323 U.S. 459, 65 S.Ct. 347, 89 L.Ed. 389 (1945); Thonen v. Jenkins, 517 F.2d 3, 6 (4th Cir. 1975); Francis v. Davidson, 340 F.Supp. 351 (D.Md.), aff’d, 409 U.S. 904, 93 S.Ct. 223, 34 L.Ed.2d 168 (1972); Williams v. Eaton, 443 F.2d 422 (10th Cir. 1971). [7] The National Teacher Examinations are composed of two batteries of tests. The Common Examinations include tests in professional education (psychological and social foundations of education, and generally applicable methods and principles of teaching) and general education (English, social studies, literature, fine arts, science and mathematics). The Area Examinations are tests in the subject matter that a prospective teacher expects to teach. Art teachers take the Area Examination in Art Education; French teachers take the Area Examination in French, and so on. There was no minimum score requirement on the Area Examination portion of the NTE at this time. [8] Report of the Special Committee to Investigate Statutory Laws Dealing with Education, p. 9 (1941). [9] In 1966, the federal government began to take action to cut off federal funds because of lack of integration in the school system. In 1968, the Supreme Court held in Green v. County School Board, 391 U.S. 430, 88 S.Ct. 1689, 20 L.Ed.2d 716 (1968), that the “freedom of choice” system was unconstitutional indicating that unitary school systems could no longer be avoided. [10] In 1969 and 1976, when the State changed the certification system, all those certified under the previous system maintained their then-current classification unless they met the more stringent qualifications for the new professional certificate. At present there are still numerous holders of graded certificates and warrants. [11] Plaintiffs also challenge the formula for calculation of retirement benefits which are determined on the basis of salary earned during employment because salary is, in part, determined by the type of teaching credential held, and again the NTE scores are involved. There is no different question of law or fact with respect to the retirement system and our decision on the salary issue controls. [12] In 1969 the State ceased the practice of granting B, C or D certificates. Since these teachers lack the knowledge to teach effectively according to the ETS study, they probably should not have been hired in the first place. We are informed by counsel for the State that only a small percentage of teachers hired prior to 1969 are paid by the grade of their certificate, and that these teachers are rapidly decreasing in number (for instance, at last report there were only 31 D certificates, distributed almost equally between blacks and whites, and 399 C certificates held almost entirely by blacks). While we do not think this practice is unconstitutional, and we understand that the ETS objection is based upon testing or psycho *1107 metric considerations rather than legal ones, we do not think such a practice should be continued in the future. Unqualified teachers should no longer be hired, and certainly not retained. (Defendants’ Proposed Findings of Fact and Conclusions of Law). [13] We therefore reject as irrelevant to the equal protection challenge plaintiffs’ evidence directed at the relationship (or lack of it) between the NTE and the actual job of teaching. The statistical studies in the record do not prove that high NTE scores would correlate with high scores on measures of teaching effectiveness. However, all the experts who testified in this case, including those offered by plaintiffs, agreed that there is, as yet, no satisfactory measure of teaching effectiveness. For that reason, evidence of a lack of correlation is unpersuasive. [14] Surely children attending South Carolina’s public schools are entitled to the best qualified available teachers, regardless of whether they are white or black, and the State has a solemn duty to accomplish this goal in keeping with its financial ability to do so. [15] Section 706(f) provides that if, after a charge is filed by or on behalf of a person claiming to be aggrieved, or by a member of the Commission, the Commission has been unable to secure from the respondent a conciliation agreement acceptable to the Commission, the Commission may bring a civil action against any respondent not a government, governmental agency, or political subdivision named in the charge. In the case of a respondent which is a government, governmental agency, or political subdivision, if the Commission has been unable to secure from the respondent a conciliation agreement acceptable to the Commission, the Commission shall take no further action and shall refer the case to the Attorney General who may being a civil action against such respondent in the appropriate United States district court. (42 U.S.C. § 2000e-5). Emphasis added. [16] Section 706(a). [17] Such conciliation efforts may have been particularly useful in this case where defendants were planning an extensive validity study using a new model. If the government’s objections to the design, execution and data analysis of the State’s study had been made known in a timely fashion through conciliation, this litigation might have been avoided. [18] We are advised that the United States has appealed this decision to the Ninth Circuit. However, there is no indication when such appeal may be heard. [*] Standard error of measurement [**] The score from the Common Examinations alone is used to certify teachers in specialty fields for which there is no Area Examination under the NTE program. [19] Division of Industrial Organizational Psychology (Division 14), American Psychological Association, Principles for the Validation and Use of Personnel Selection Procedures (1975). [20] Guidelines for Employee Selection Procedures, 29 C.F.R. 1607 (1970). To the extent that the EEOC Guidelines conflict with well-grounded expert opinion and accepted professional standards, they need not be controlling. General Electric Co. v. Gilbert, 429 U.S. 125, 97 S.Ct. 401, 50 L.Ed.2d 343 (1976). [21] The EEOC re-published these regulations, without change, on November 24, 1976, 41 Fed. Reg. 51985 (1976). The EEOC evidently equates the concept of business necessity with the measurement concepts of statistical and practical significance. 29 C.F.R. § 1607.5(c) provides: In assessing the utility of a test the following considerations will be applicable: (1) The relationship between the test and at least one relevant criterion must be statistically significant . (2) In addition to statistical significance, the relationship between the test and criterion should have practical significance. By their terms, these regulations are applicable only to a'criterion-related validity study. The EEOC sets out no analogous requirements for a content validity study. [22] These uses, enacted by the Legislature, have remained the same since 1944 and were not changed in 1957, 1969, or 1976, when the certification system was changed by the State Board.