TABLE OF CONTENTS

SUMMARY 932 OPINION 933 I. The Parties 933 II. Procedural Background 933 III. History of the I.Q. Test and of Special Education in California 935 A. The Early History 935 B. State Funded Programs for the Educable Mentally Retarded:' Statutory History 937 C. Special Classes for the Educable Mentally Retarded 941 D. The Racial Composition of E.M.R. Classes . 942 E. The Use of I.Q. Tests in E.M.R. Placement 945 1. The State Requirement of an I.Q. Test 945 2. The I.Q. Score and the Other Psychological Data for E.M.R. Placement' 948 3. The Moratorium on I.Q. Testing 951 IV. Intelligence Tests 952 A. Introduction: The Impossibility of Measuring Intelligence 952 B. Scores of Black and White Children on I.Q. Tests 954 C. Reasons for the Disparity in I.Q. Scores 954 1. The Genetic Argument 955 2. The Socio-Economic Argument 956 3. The Cultural Bias of Standardized I.Q. Tests 956 D. The Relevance of Cultural Bias and the Disparity in I.Q. Scores 959 V. Legal Analysis 960 A. Introduction 960 B. Federal Statutory Claims 961 1. Private Right of Action and Exhaustion of Administrative Remedies 961 . 2. Title VI of the Civil Rights Act of 1964: One Application of an “Effects Test” 964 3. The Rehabilitation Act of 1973 and the Education for All Handicapped Children Act of 1975 — Another “Effects Test” 966 4. The Question of Validation D,espite Discriminatory Effects 968 5. Alternatives to I.Q. Testing for E.M.R. Placement 973 C. The Equal Protection Clause of the Fourteenth Amendment 974 1. Intentional Discrimination 975 a. Determining the Standard 975 b. Applying the Law 979 2. An “Intermediate” Standard of Review 985

D. State Claims E. Adequacy of Representation F. Legal Conclusions VI. Remedy A. Introduction B. Permanent Injunctive Relief 1. Intelligence Tests 2. Disproportionate Placement 3. Reevaluation and Other Class Relief C. The Necessity of Limiting the Scope of the Remedy VII. Conclusion 986 987 988 989 989 989 989 990 990 990 991

Plaintiffs, representing the class of black children in California who have been or in the future will be wrongly placed and maintained in special classes for the “educable mentally retarded” (“E.M.R.”) challenge the placement process for those classes and particularly certain uses of standardized individual intelligence (“I.Q.”) tests in California. They contend that the I.Q. tests in their present form are biased and that defendants have discriminated against black children by using those tests. The tests allegedly result in the misplacement of black children in special classes that doom them to stigma, inadequate education, and failure to develop the skills necessary to productive success in our society. Black children represent only 10 percent of the present general student population in California, but provide some 25 percent of the population enrolled in E.M.R. classes.

These testing and placement problems arise in a setting of educational failure. California’s schools have been unable to meet the educational needs of disadvantaged minorities such as the black children who brought this case. As a result, poor minority children tend not only to start out behind their white, middle-class counterparts, but also tend to fall increasingly farther behind after exposure to the public school system. Defendant Wilson Riles stated in 1969, before he became Superintendent of Public Instruction, that disadvantaged black children, on the average, learn only about .7 as much as middle-class white children in any given year. More recent data confirmed that analysis. Black children fall increasingly behind to the point that it is not unusual for high school students to be reading at the third grade level and performing at only the fourth grade level in mathematics. ® The disproportionate placement of black children into E.M.R. classes is but one aspect of this troublesome situation.

Courts cannot solve our educational problems, but they played a part in the incremental effort to improve those aspects of our educational systems that effectively deny minorities an equal opportunity to succeed. In particular, the phenomenon of special education such as that for the “mentally retarded” has not yet been subjected to much judicial scrutiny. We have been forced in this case to enter that complicated area, and it raises special problems for court intervention.

A principal focus of this litigation is on testing — on the use of individual I.Q. tests — to classify black children and assign them to E.M.R. classes. Much of the more then 10,000-page transcript of the trial represents detailed expert testimony about these tests. The court has necessarily been drawn into the emotionally charged debate about the nature of “intelligence” and its basis in “genes” or the “environment.” This debate, which finds reknowned experts disagreeing sharply, obviously cannot be resolved by judicial decree. Despite ‘these problems, however, court intervention has been necessary. The history of this litigation has demonstrated the failure of legislators and administrative agencies to confront problems that clearly had to be faced, and it has revealed an all too typical willingness either to do nothing or to pass on issues to the courts.

Fortunately, the “scientific controversy” surrounding the I.Q. tests has not materialized to the extent that might have been expected. The experts have tended to agree about what I.Q. tests can and cannot do, even if they disagree about the utility of I.Q. testing for E.M.R. placement. Our decision, therefore, rests more on a consensus than on the testimony of any one line of experts. Given that consensus, coupled with the other factors present in this case, there is no choice but to invalidate California’s present system of classification of black children for E.M.R. classes. The bases for this ruling, both statutory and constitutional, will be explained in detail below, but it may be helpful to summarize them briefly at the outset before proceeding to the main body of the opinion.

SUMMARY

This court finds in favor of plaintiffs, the class of black children who have been or in the future will be wrongly placed or main- tamed in special classes for the educable mentally retarded, on plaintiffs’ statutory and state and federal constitutional claims. In violation of Title VI of the Civil Rights Act of 1964, the Rehabilitation Act of 1973, and the Education for All Handicapped Children Act of 1975, defendants have utilized standardized intelligence tests that are racially and culturally biased, have a discriminatory impact against black children, and have not been validated for the purpose of essentially permanent placements of black children into educationally dead-end, isolated, and stigmatizing classes for the so-called educable mentally retarded. Further, these federal laws have been violated by defendants’ general use of placement mechanisms that, taken together, have not been validated and result in a large over-representation of black children in the special E.M.R. classes.

Defendants’ conduct additionally has violated both state and federal constitutional guarantees of the equal protection of the laws. The unjustified toleration of disproportionate enrollments of black children in E.M.R. classes, and the use of placement mechanisms, particularly the I.Q. tests, that perpetuate those disproportions, provide a sufficient basis for relief under the California Constitution. And under the federal Constitution, especially as interpreted by the Ninth Circuit Court of Appeals, it appears that the same result is dictated.

Moreover, there is another basis for the federal constitutional ruling. Defendants’ conduct, in connection with the history of I.Q. testing and special education in California, reveals an unlawful segregative intent. This intent was not necessarily to hurt black children, but it was an intent to assign a grossly disproportionate number of black children to the special E.M.R. classes, and it was manifested, inter alia, in the use of unvalidated and racially and culturally biased placement criteria. This intent, consistent only with an impermissible and unsupportable assumption of a higher incidence of mental retardation among blacks, cannot be allowed in the face of the constitutional prohibition of racial discrimination.

The precise reasons for these findings, expressed here in simplified form, and the type of remedy that will be necessary, including a continuation of the present injunction against I.Q. testing for the placement of black children into E.M.R. classes, are discussed in the body of this opinion.

OPINION

I. THE PARTIES

The named plaintiffs, including Larry P., are black children who attended elementary schools in the San Francisco Unified School District and were placed in special classes for the educable mentally retarded. Their scores on individual standardized I.Q. tests contributed to the placement decisions, which plaintiffs contend were erroneous.

Defendant Wilson Riles has been the Superintendent of Public Instruction for the State of California since the inception of this lawsuit. He has overall responsibility for administering California’s public educational system, including classes for the mentally retarded. Defendants Marian W. Drinker, Michael W. Kirst, James W. Dent, Jonn R. Ford, Louis Honig, Jr., Patricia D. Ingoglia, Virla R. Krotz, Lorenza C. Schmidt, and Tony N. Sierra are members of the California State Board of Education. The Board is empowered to set the policies that the Superintendent executes.

Defendant Robert Alioto is the Superintendent of Schools for the San Francisco Unified School District, and defendant members of the Board of Education are Zuretti Goosby, Ben Tom, Peter Mezey, William Maher, Lucille Abrahamson, Eugene S. Hopp, and Rosario Anaya.

The United States is participating in this lawsuit as amicus curiae pursuant to a motion granted in August 1977. As amicus curiae the United States was given the right to present expert witnesses, file briefs, and make oral arguments to the court.

II. PROCEDURAL BACKGROUND

The initial complaint for declaratory and injunctive relief was filed against the San Francisco and state defendants on November 23,1971, challenging as unconstitutional the use of standardized intelligence tests for black E.M.R. placement in San Francisco. By memorandum and order filed on June 20, 1972, this court determined that the action could proceed as a class action on behalf “of all black San Francisco school children who have been classified as mentally retarded on the bases of I.Q. test results.” Larry P. v. Riles, 343 F.Supp. 1306, 1315 (N.D.Cal.1972). The order also granted a preliminary injunction in favor of plaintiffs. Defendants were

343 F.Supp. at 1315 .

Defendants appealed the preliminary injunction to the Court of Appeals for the Ninth Circuit. On August 16, 1974, by a per curiam decision, the circuit affirmed the 1972 decision. Larry P. v. Riles, 502 F.2d 963 (9th Cir. 1974).

After the rendering of the Ninth Circuit’s decision, plaintiffs moved to modify the class and the terms of the preliminary injunction. The court agreed and filed, an order on December 13, 1974, expanding the class to include “all Black California school children who have been or may in the future be classified as mentally retarded on the basis of I.Q. tests.” Order of December 13, 1974, at 1. The terms of the preliminary injunction were correspondingly expanded. The order, at page 2, provided as follows:

The statewide preliminary injunction, affecting only black children, has been in effect since the date of that order. In January of 1975, defendants voluntarily imposed a moratorium on I.Q. testing for E.M.R. placement of all children, regardless of race, in California. For a period of time it was thought that the Master Plan for special education in California, enacted in experimental form in 1974, would address and perhaps remedy the problems raised by this case, but that hope never materialized. The case had to be brought to trial on the merits.

This case has developed from a 1971 private San Francisco class action alleging only federal constitutional claims to a statewide action supported by the Department of Justice and involving numerous federal and state laws. The essence of the challenge is still the same, but with the recent proliferation of legislation and regulations touching on once neglected problems, plaintiffs have bolstered substantially their attack.

On January 18, 1977, pursuant to the court’s order of December 6,1976, plaintiffs filed their first amended complaint. Reflecting recent legislative reforms, the amended complaint alleged several statutory bases for the claims, including Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d, and the Emergency School Aid Act of 1972 and 1974, 20 U.S.C. § 3191 et seq. The amended complaint also alleged violations of the Constitution of the State of California, as well as of several sections of the California Education Code.

On August 8, 1977, the United States Department of Justice moved to participate as amicus curiae. The motion was granted on August 19, 1977. The position of the United States is that the use of I.Q. test scores to effect the placement of black children in special education classes violates federal law, including the provisions mentioned before, as well as the Education for All Handicapped Children Act, 20 U.S.C. § 1401 et seq., and section 504 of the Rehabilitation Act of 1973, 29 U.S.C. § 794 , if the tests have a disproportionate impact on black children or were not validated for the purpose of diagnosing mental retardation. Plaintiffs later were granted leave to file a second amended complaint, alleging in particular the violations of the Education for All Handicapped Act.

The court denied various motions to dismiss and for summary judgment, thus necessitating the trial that began on October II, 1977.

III. HISTORY OF THE I.Q. TEST AND OF SPECIAL EDUCATION IN CALIFORNIA

A. The Early History

We must recognize at the outset that the history of the I.Q. test, and of special education classes built on I.Q. testing, is not the history of neutral scientific discoveries translated into educational reform. It is, at least in the early years, a history of racial prejudice, of Social Darwinism, and of the use of the scientific “mystique” to legitimate such prejudices.

The first usable I.Q. tests were developed in France in 1905 by Alfred Binet, who sought to distinguish between “backward” and normal children in Paris. He had no illusions that his test could measure innate traits. He sought to develop diagnostic tools to help those who needed to improve their intellectual skills. Binet insisted that “it was necessary to react against and protest the brutal pessimism of those who regarded the test as measuring some fixed and unchanging quantity.” Kamin 849. This admonition, however, was not heeded when Binet’s test was transplanted in American soil.

The early leaders of the I.Q. testing movement in the United States were quick to assume that the Binet tests measured an innate capacity fixed in the genes. Typical and of particular interest for the California story were the views of Professor Lewis Terman of Stanford, the well-known developer of the Stanford-Binet I.Q. test in 1916. According to Terman, his test could be employed as an objective tool to learn the identity of the “feeble-minded” and “borderline feeble-minded” and hopefully to discourage them from breeding. He predicted great accomplishments for the testing movement:

He concluded further that, since the tests measured a fixed, innate ability, it would be useless to keep the feebleminded and borderline feebleminded in class with other, normal children.

Finally, it was Terman’s opinion, as well as that of others in the testing movement, that different racial and ethnic groups had different proportions of intellectually subnormal individuals. He declared in 1916 that the borderline feebleminded “represent the level of intelligence which is very, very common among Spanish-Indian and Mexican families of the Southwest and also among negroes. Their dullness seems to be racial or at least inherent in the family stocks from which they come.”

These opinions were widely accepted and influential in the programs for the “feeble-minded” set up during and after World War I. Indeed, in 1921, the year that the California state legislature first provided a legal basis for classes for the mentally retarded, 1921 Cal.Stats. ch. 685, the State Superintendent of Public Instruction made the following statement:

Special education and the I.Q testing that justified its existence thus found an enduring place in California in the 1920’s, and the assumptions brought to the special classes survived long past the presumed “scientific basis” for them. By 1927, according to a state survey, there were some 10,000 school children with I.Q. scores below 80, and 5,322 were in special schools or classes.

The growth of these special classes and schools stagnated during the depression years, and efforts to obtain state- funding were unsuccessful. The situation changed dramatically after the Second World War. The California State Bureau of Special Education was created, and, in the words of Dr. Simmons of the State Department of Education, the principal historian of these developments, “After many previous efforts, a state program for the educable mentally retarded was finally achieved for the first time in 1947.”

B. State Funded Programs for the Educable Mentally Retarded: Statutory History

The 1947 legislation substantially increased the state’s role in special education classes, which earlier had been state-sanctioned but not closely regulated. The new law defined “mentally retarded minors” as:

1947 Cal.Stats. ch. 1475. The “special training schools or classes” mandated by this law ratified the earlier assumptions behind these classes and set the pattern for education of the educable mentally retarded in California. State funding was equal to the excess cost of educating mentally retarded minors multiplied by the number of such persons who were in average daily attendance. Reference to the special classes was only to be made after a “careful individual examination by a competent psychologist”; an intelligence test was not specifically required by statute, but it is undisputed that such tests were given as a matter of course.

In order to understand the distinctive qualities of the E.M.R. program, it is first necessary to distinguish it from other state categorical programs. The “trainable mentally retarded” (T.M.R.) category was created in 1951 for school-age children considered one step below the category of educable mentally retarded. These T.M.R. classes no longer have the same title, but they continue to exist to enable severely retarded children to become socially adjusted and useful “in their homes and within sheltered environments.” Initially, such classes were at the option of each district. Classes for these children were finally mandated in 1963. 1963 Cal.Stats. ch. 2105.

See Cal.Educ.Code § 56515.

In 1963, in addition, the legislature created programs for “culturally disadvantaged minors,” 1963 Cal.Stats. ch. 98, and for “educationally handicapped minors.” 1963 Cal. Stats, ch. 2165. See Cal.Educ.Code § 56600. The former category is particularly significant since it specifically addresses the problems that underlie the challenge to I.Q. testing in this case. “Culturally disadvantaged minors” were defined as children “potentially capable of successfully completing a regular educational program” but who probably would be unable to do so because of “cultural, economic, and like disadvantages.” The 1963 legislation set up a program for “supplementary education” in deprived areas to help this group. See Cal. Educ.Code § 56600.

“Educationally handicapped minors” (E.H.) also were thought capable of returning to a regular school program. They are assigned to special classes because they “cannot benefit from the regular educational program” due to “marked learning or behavioral disorders, or both.” Cal.Educ. Code § 56600. Special E.H. classes were to emphasize returning the children — principally those considered “neurologically handicapped” or “emotionally disturbed” — back to the classroom.

In the mid-1960’s, therefore, California had programs for the “culturally deprived,” the “educationally handicapped,” the “educable mentally retarded,” and the “trainable mentally retarded.” The funding criteria described above encouraged placement in these classes, and the E.M.R. enrollment peaked at 58,000 in 1968-69. About 27 percent of the E.M.R. children were black, even though black children represented only 9 percent of the California school population. The swelling enrollments, particularly of minorities, led to criticisms of the program and several statutory changes. The first was the revision of the mechanism of funding to reduce the incentive to fill the classes. 1969 Cal.Stats. ch. 784.

In 1969 the legislature clarified the definitions of the categories of special classes and specified that admission to a special educational program for the mentally retarded “shall be made only on the basis of an individual evaluation according to standards established by the State Board of Education and upon individual recommendation of a local admission committee which shall include a teacher, a school nurse or social worker, a school psychologist or other pupil personnel worker authorized to serve as a school psychologist who has individually examined the minor, a principal or supervisor, and a licensed physician.” 1969 Cal. Stats, ch. 784. The law also required an nual reviews of E.M.R. placements and consultations with parents prior to enrollment.

Around this time, the legislature began to address tentatively the complaints that had been surfacing since 1967 about misplacement of minorities, especially Spanish-speaking children, in the E.M.R. classes. Indeed, as early as 1968, the State Division of Special Education held a two-day workshop on the problems of minorities. House Resolution 444, passed in 1969, then became the first example of legislative concern in this area.

House Resolution 444 stated that the category of educable mentally retarded was under “mounting criticism from representatives of certain minority groups, most particularly culturally bilingual groups, to the effect that a disproportionate number of children from such groups are assigned to classes for the mentally retarded.” This resolution then tossed the problem to the State Board of Education and through the latter to the State Department of Education. The legislature simply directed the Board to give attention to this problem and make suggestions for legislation.

Precisely what the Board and Department of Education were doing at this time will be examined more closely below, but it is important to note here that they prepared a report to comply with the legislative request The report’s recognition of the problem resulted in another House Resolution, numbered 262. This resolution proclaimed that “There does exist a disproportionate enrollment of Spanish surname and Negro pupils in classes for the educable mentally retarded.” Again, however, the legislature merely called for a further report on efforts to correct this over-enrollment.

Following a 1969 regulation adopted by the State Board of Education, the legislature in 1970 for the first time required by explicit statutory provision that an I.Q. test selected from a list approved by the State Board of Education be given for E.M.R. placement. The law, Education Code § 6902.06 (now § 56505), provided for the testing of children in their primary home language, and it prohibited placement into an E.M.R. class if a child scores better than two standard deviations below the norm on the test. Cal.Educ.Code § 56506. Written parental consent, obtained after an explanation of the E.M.R. program, was required before placement, Cal.Educ.Code § 56506, and the Department of Education was required to prepare an annual report for the state legislature on placements into E.M.R. classes. Cal.Educ.Code §56509. Another new enactment of 1970 required that all children presently in E.M.R. classes be retested according to the new statutory cutoff. Further, annual reports were mandated by all school superintendents in districts where there was a significant disproportion in the enrollment of minorities in the E.M.R. program. Cal.Educ.Code § 56508. Some provisions were made also for remedial attention to those found through retesting to have been qualified for regular school instruction. 1970 Cal.Stats. ch. 1543.

Significant new legislation also came in 1971, when the I.Q. test was singled out specifically for attention. 1971 Cal.Stats. ch. 78. The legislature made the following clarion declaration:

Cal.Educ.Code § 56504. No teeth were given to this declaration, however, and no mention was made of which of the state-mandated I.Q. tests might be discriminatory. Rather, the legislature added the requirement that “retarded intellectual development indicated by the individual test scores” be substantiated by “a complete psychological examination by a credentialed school psychologist investigating such factors as developmental history, cultural background, and school achievement.” Cal. Educ.Code § 56506. Studies of “adaptive behavior” — the ability to engage in social activities and perform everyday tasks— were also called for as well as home visits, with the consent of the parent or guardian. Cal.Educ.Code § 56506. Finally, the legislature sought to strengthen the requirement of parental consent by requiring that the E.M.R. program be explained to parents. Cal.Educ.Code § 56506.

The statutory picture created by the 1970 and 1971 legislation has remained essentially the same since that date, although there have been some signs of change. The legislature in 1974 established a pilot program, entitled the Master Plan for Special Education, designed to cover the education of all “individuals with exceptional needs.” Both the . original 1974 legislation and amendments made in 1977 would provide further protections to students in the same position as those termed E.M.R. by the Education Code sections now in effect. At the present time, while the legislature has indicated its intention to phase in the Master Plan for the entire state, it is still unclear whether the funds will be made available and, if they are, whether there would be any practical effect on the existence of special classes made up of those now called E.M.R. The Master Plan thus indicates some movement by the state in the direction sought by plaintiffs in this case; nevertheless, that movement has been slow and ambiguous, and this court must address itself to the situation as it is rather than what it might become.

C. Special Classes for the Educable Mentally Retarded

In contrast to the classes for the culturally deprived or the educationally handicapped, E.M.R. classes are designed to separate out children who are incapable of learning in the regular classes. According to a 1974 state handbook entitled “Programs for the Educable Mentally Retarded in California Public Schools” and “written in an attempt to provide information to those who are planning, managing, or evaluating programs for the educable mentally retarded,”

There was testimony at the trial indicating that at times these classes have been utilized to give special remedial attention to some students, e. g., Meyers 8711-13, but the administrators in the Department of Education indicated clearly that this use of the classes was discouraged. See, e. g., Simmons 3328-30; Brinegar 4620-26. In the words of Fred Hanson, a special consultant and one of the key state administrators, “slow learning must be caused by limited intellectual capacity.” Hanson 3565. These classes are not meant for remedial instruction.

Further, the curriculum was not and is not designed to help students learn the skills necessary to return to the regular instructional program. To quote again from the state handbook,

Finally, consistent with the first two aspects of E.M.R. classes, the classes are conceived of as “dead-end classes.” Children are placed there, generally at about eight to ten years of age, because they are thought to be incapable of learning the skills inculcated by the regular curriculum. They are provided with instruction that deemphasizes academic skills in favor of adjustment, and naturally they will tend to fall farther and farther behind the children in the regular classes. As stated by Mr. Hanson, “as they get older, the gap tends to widen between the two programs [regular and E.M.R.].

E.M.R. classes therefore tend to have a very low turnover rate. As Mr. Dean, Assistant Superintendent of Special Education Services in San Francisco, informed members of the San Francisco School Board in 1970, and reaffirmed on the witness stand, “the object is for the children to remain in the program from time of placement until graduation from senior high school.” Dean 4311-12. Independent empirical studies substantiate this observation. . Professor Mercer’s investigation of schools in Riverside, for example, found that less than 20 percent of those placed in E.M.R. classes ever returned to the regular classroom. Mercer 1433-34. A second survey, the State’s Uslan Report, found that the average student in an E.M.R. class had been there for four years and five months. These classes were designed and operate only to teach certain minimal skills and habits to those found mentally incapable of mastering the skills necessary to follow a regular classroom curriculum.

Since California has carved out a special “track” for children unable to progress beyond a minimal education, it is clear that the decision to place children in these classes is a crucial one. Children wrongly placed in these classes are unlikely to escape as they inevitably lag farther and farther behind the children-in regular classes. Consequently, as stated by Dr. Riles, it would “absolutely not” be appropriate to put a non-mentally retarded child in a class for the educable mentally retarded. Riles 2705.

D. The Racial Composition of E.M.R. Classes

There is abundant evidence that black children are substantially overrepresented in E.M.R. classes. The following table gives the aggregate data for the relevant periods of time.

TABLE I STATE DATA

School Year Total E.M.R. Enrollment Black % in Student Population Black % in E.M.R. Classes

1968- 69 57,148 not reported 25.50%

1969- 70 54,078 not reported 27.06%

1970- 71 47,864 not reported 25.73%

1971- 72 38,208 not reported 26.70%

1972- 73 33,091 9.3% 25.00%

1973- 74 29,609 9.3% 24.56%

1974- 75 28,136 not reported 22.61%

1976-77 19,289 10% 25.43%

Sources: SDE * Reports, in Exhibits P-40, P — 41, P-42(b), P-131

U. S. DATA

1973- 74 32,500 12% 26%

1974- 75 26,494 12% 27%

1976-77 26,224 10% 17%

Sources: U. S. Exhibits 5a, 5b, 5c, 7, 8

This table provides two sets,of data which clearly indicate that, while the total enrollment in E.M.R. classes has declined drastically since the academic year 1969-70, the overrepresentation of black children in those classes has remained essentially constant. In the academic year 1974-75, for example, the lowest percentage differential according to state data was reached. Black children provided approximately 10 percent of the student population but 22.6 percent of the E.M.R. population. As illustrated by Table 2, the practical effect of these overall disproportions is even more dramatic.

TABLE 2

1976-1977 ENROLLMENT DATA ON TWENTY SCHOOL DISTRICTS WHICH ACCOUNT FOR 80% OF CALIFORNIA’S BLACK POPULATION *

District EMR Enrollmt. Black EMR Total District Enrollmt. Total Black Enrollmt. Black District % Black EMR % Non-Black EMR Enrollmt. Rate Black EMREnrollmt. Rate

L.A.U.S.D. 3155 1786 592,931 142,778 24.1% 56.6% .30% 1.25%

•Oakland 751 663 52,275 35,086 67.1% 88.3% .51% 1.89%

Compton 461 442 31,054 26,356 84.9% 95.9% .40% 1.68%

S.F.U.S.D. 446 240 69,982 21,730 31.1% 53.8% .43% 1.10%

San Diego 592 202 121,233 17,637 14.6% 34.1% .38% 1.14%

Richmond 109 67 34,880 12,278 35.2% 61.5% .19% .54%

Pasadena 38 29 25,409 10,872 42.8% 76.3% .06% .27%

Inglewood 62 51 13,068 10,428 79.8% 82.3% .42% .49%

Sacramento City _ No Data----------No Data •

Long Beach 458 75 60,187 9,058 15.1% 16.4% .75% .83%

Pomona 139 77 20,863 6,195 29.7% 55.4% .42% 1.24%

Fresno City 399 110 53,643 5,550 10.4% 27.6% .60% 1.98%

Berkeley 25 22 12,142 5,221 43.0% .04% .42%

San Bernardino City 265 79 30,970 4,712 15.2% 29.8% .02% 1.68%

Lynwood 94 56 9,450 4,560 48.3% 59.6% .78% 1.23%

Stockton City 260 84 26,868 4,293 16.0% 32.3% .78% 1.96%

Vallejo 81 39 14,357 4,030 28.1% 48.2% .41% .97%

Ravenswood 10 9 3,443 2,913 84.6% 90.0% .19% .31%

Bakersfield 185 73 18,346 2,893 15.8 39.5% .72% 2.52%

Monterey Peninsula 90 36 14,003 2,627 18.8 40.0% .47% 1.37%

In the 20 districts accounting for 80 percent of the enrollment of black children, black students comprised about-27.5 percent of the student population and 62 percent of the E.M.R. population.

Statistical analyses of the statewide and distriet-by-district figures indicate the obvious. These apparent overenrollments could not be the result of chance. For example, there is less than a one in a million chance that the overenrollment of black children and the underenrollment of non-black children in the E.M.R. classes in 1976-77 would have resulted under a color-blind system of placement. Hillabrandt 2523. The data collected for 1976-77 on the 20 districts in California that enroll more than 80 percent of the population of black children are again particularly revealing. According to the State Department of Education data, black children were overrepresented in all districts, and in all but three — Inglewood, Long Beach, and Ravenswood — the overenrollment was statistically significant at the .05 level. Berk 9749-50. Comparable data obtained independently by the U.S. Department of Health, Education, and Welfare, confirmed the state’s data, showing a statistically significant overenrollment in 17 of the 20 districts (with data for one not available). Hillabrandt 0992-94.

Data for the earlier years are completely consistent. For 1973, for example, the federally-compiled information revealed statistically significant overenrollments in 15 out of the 18 school districts accounting for 80 percent of the population of black school children. Hillabrandt 2534-36. However, the data are compiled and analyzed, and whether we compare aggregate figures or rates of enrollment, there is no question that a “color-related” factor contributed to the overenrollment. Without belaboring this conclusion, we must consider the reasons for the overenrollment. This opinion will undertake such an examination in a subsequent section, but it is important to emphasize that any explanation must account for the magnitude of the problem, and the defendants’ suggestion that there is a higher incidence of mild retardation among the black population fails to do this. Even if it is assumed that black children have a 15 percent higher incidence of mild mental retardation than white children, there is still less than a one in a million chance that a color-blind system would have produced this disproportionate enrollment. If it is assumed that black children have a 50 percent greater incidence of this type of mental retardation, there is still less than a one in 100,000 chance that the enrollment could be so skewed towards black children. Hillabrandt 2536-38. Perhaps the most telling weakness in defendants’ assertion is that the disproportionate E.M.R. enrollment of black children is not duplicated in the classes for the so-called “trainable mentally retarded” children. See, e. g., Hillabrandt 2561; Watkins 4063-65. There is some thing distinctive about the placement mechanism for E.M.R. classes that has drawn in great numbers of black children since at least the late 1960’s.

E. The Use of I.Q. Tests in E.M.R. Placement

There is no factual dispute about the matters just discussed. E.M.R. classes are “dead-end” classes for children thought to be incapable of learning the material taught in regular classes, and black children can be found in the classes in numbers grossly out of proportion to their representation in the student population. We turn now to the process by which the children are assigned to the E.M.R. classes, and particularly to the role of the I.Q. tests in that process.

1. The state requirement of an I.Q. test. Prior to 1969, California’s education code required that an “individualized psychological examination” be given prior to E.M.R. placement, and the Administrative Code mandated that a “verbal or non-verbal intelligence test” be given. No list of tests was set out, however, and there was at least a possibility that particular school districts could avoid relying on tests that had a discriminatory impact on minorities. This changed in 1969, and for the first time the state adopted a mandatory list. The circumstances surrounding this change reveal much about defendants’ attitude toward minorities in E.M.R. classes.

The year 1969 was an eventful one in the history of I.Q. testing. Concern had been mounting about the tests and the placement of minorities in E.M.R. classes, and that concern had already been expressed in a State Department of Education conference in 1968. In early 1969, the state legislature enacted House Resolution 444, calling for study of the problem of overenrollment of minorities in E.M.R. classes, and later that year there were more hearings around the state. Furthermore, the Diana litigation commenced in 1969 and raised many of these issues on behalf of children from Spanish-speaking backgrounds. And it was also in 1969 that Professor Arthur Jensen added considerable fuel to the I.Q. controversy with his long and detailed brief for the hereditability of “intelligence” as measured by I.Q. scores.

Key officials in the State Department of Education — the central actors in this story — were well aware of these developments; they knew that the E.M.R. classes contained a disproportionate enrollment of minority children, and they were equally well aware of the charge that a cultural bias in the I.Q. tests caused or contributed to that disproportion. E. g., Hanson 3746-52. Wilson Riles, then serving as head of California compensatory education programs, even stated in mid-1969 to the President’s Commission on Mental Retardation that, because of the disproportionate number of minorities in the E.M.R. programs,

Riles 2914. Yet just prior to Riles’ statement about a reassessment of the placement process, the State Board of Education chose for the first time to mandate the use of specific standardized individual intelligence tests for E.M.R. placement. Indeed, as it turned out, the state decision revealed a studied effort to avoid the “second look” suggested by Dr. Riles.

Unfortunately, no state officials called to testify by either plaintiffs or defendants could provide a cogent explanation for the original impetus toward mandatory I.Q. testing. Fred Hanson, the consultant within the department who prepared the list of suitable tests, testified as follows:

Hanson 3783. See also, e. g., Simmons 3349-50. What is certain is that the impetus for change came from within the State Department of Education in the middle of 1969. In July 1969, the Department proposed, and the State Board of Education adopted, an addition to the California Administrative Code requiring that approved I.Q. tests be used as part of the E.M.R. placement process. 5 Cal.Ad.Code § 3401. The I.Q. tests were to be selected from a list prescribed by the Board. The Department then moved very quickly to compile that list.

On July 22, 1969, the Acting Chief of Special Education, Charles Watson, sent a letter to approximately, 100 school psychologists and county offices of education. The letter stated the Board’s intention to develop the list of appropriate tests before the beginning of the next school year in September. The letter “urgently requested” suggestions for appropriate group and individual I.Q. tests. Mr. Hanson had drafted that letter and personally undertook the investigation of I.Q. tests. In addition to examining the some 28 responses to the letter, Mr. Hanson talked to the publishers of the Stanford-Binet, the WISC, and other major standardized tests, and he consulted a standard source book to learn about other supplementary tests. It is admitted by defendants that Mr. Hanson was not himself an expert in I.Q. testing.

In August, Mr. Hanson compiled a general list of tests and used the responses to the letter to tabulate the frequency with which tests were then being utilized by the reporting districts and county superintendent’s offices. It appears from Mr. Hanson’s testimony that the primary criterion for selection of tests for the final list, prepared at the end of the month, was the frequency of use in the field. Hanson 3719-23. As a result, the final list required-that either one of the Wechsler tests, the Stanford-Binet, or the Leiter test be given, with certain other tests permitted on a supplementary basis. The list was not given to the Board for approval until February 1970, but the decision was essentially made in July-August 1969.

The extremely quick and unsystematic method of selecting these tests raises serious questions about the intentions of officials in the SDE. The method resulted in a failure to consider many of the critical issues stemming from I.Q. testing. Mr. Hanson and others in the State Department of Education, for example, chose to ignore letters from field personnel, including a powerful one from Mr. Hocker, co-chairman of the Testing Committee of the California Association of School Psychologists and Psychometrists, that questioned the pro priety of adopting a list of tests with such haste. The Department also failed to consult any outside experts about the preparation of the list except persons employed by the publishers of the standardized I.Q. tests. And despite their awareness of the controversy surrounding the disproportionate enrollment of minorities and the cultural bias of I.Q. tests, the Department personnel did not expressly consider or investigate these problems in making their determination. Hanson 3719-24, 3737; Simmons 3350-52. Indeed, by relying on the most commonly used tests, they opted to perpetuate any discriminatory effects of those tests. Evidently, despite the legislative policy suggested by H.R. 444, minority overenrollment was not considered a problem worthy of serious attention. Mr. Hanson’s testimony is instructive:

Hanson 3798-99.

Not surprisingly, the Department presented the list of tests to the Board with no indication of any problems in the process of selection or in the tests themselves. According to the undisputed testimony of a member of the Board, it was assumed that the I.Q. test selection was done by experts and took into account the question of cultural bias. Kratz 2917-29. See also Sierra 3435-36. Neither assumption was correct, and the Board — which had no staff of its own to rely on — ratified the Department’s list.

Later in 1970, building on the Board’s prior determination of the need for I.Q. testing for E.M.R. placement, the state legislature began to address some of the problems associated with that decision. The general legislative changes have already been discussed, but the most important one affecting I.Q. testing should be highlighted. The legislature in 1970 required that no child be placed in an E.M.R. class unless he or she scored two standard deviations or more below the norm for the test. Cal. Educ.Code § 56506. This enactment lowered the maximum I.Q. score from a range of about' 75-85, depending on the school district, to about 70. When coupled with the provision that all students in E.M.R. classes be retested, this change contributed greatly to the decrease in enrollments that took place after 1969. Of course, as noted before, the reductions in enrollment of black children did not change substantially the percentage of black children in the E.M.R. classes.

The State Department of Education’s policy underlying the adoption of the tests continued well into the 1970’s. It was a policy characterized by a serious neglect of the admitted problems of standardized I.Q. tests and the E.M.R. placement process in general. Subsequent manifestations of this policy, and how this policy should be treated as a matter of law, will be dealt with in the context of the analysis of plaintiffs’ case under the Equal Protection clause of the Fourteenth Amendment.

2. The I.Q. score and the other psychological data for E.M.R. placement. The role of the I.Q. test in the placement process must be determined next. Orders of this court have stopped the administration of the I.Q. tests for placement of black children since late 1974, but the relevant statutes still contain the testing requirement, and defendants wish to resume I.Q. testing for E.M.R. placement. Riles 2786-89.

' The statutory requirement is that an approved “verbal or non-verbal individual intelligence test” be administered (§ 56505), and placement normally is not permitted if the pupil “scores higher than two standard deviations below the norm, considering the standard error of measurement.” Cal.Educ. Code § 56506. In addition to the I.Q. scores, state law provides as follows:

Cal.Educ.Code § 56506. Another provision requires parental permission for placement into the E.M.R. program, Cal.Educ.Code § 56506, and there also are a number of “due process” guarantees set out in the statute. Cal.Educ.Code § 56507. Place ment thus in theory can be permitted only if certain procedures are followed and a low score on the I.Q. tests is “substantiated” by other evidence.

Defendants argue that, despite the critical importance of the I.Q. tests, in practice the test scores are not unduly relied upon for placement into the special classes. According to Mrs. Alice Bryant, Specialist, Psychological Services, of the Los Angeles Unified School District,

Bryant 8875-K. Similarly, Mr. Jack Schuetz, Director of Special Education for Kern Union High School District, testified as follows:

D-93, at 4. Mr. Hanson for the State Department of Education testified further that the I.Q. test was certainly one of the prime determinants, and he could think of none more important, but he did not believe it was the prime determinant. Hanson 3548-49.

It is thus not disputed that the I.Q. scores played at least a substantial role in E.M.R. placement. Other testimony, however, suggests strongly that this assessment underestimates the real significance of the tests in many if not most instances. Robert White-neck, the Director of Special Education of the Sonoma County Office of Education and the Director of Special Education from 1965 to 1975 of the Berkeley schools, testified as follows:

Whiteneck 4241-42. Professor Alice Watkins, a professor at California State University in Los Angeles with considerable expertise with E.M.R. classes, also testified that prior to the moratorium the I.Q. test score was the chief determinant for placement. Watkins 4076-77.

The only systematic empirical research into the placement process further documents the heavy emphasis in practice on I.Q. scores. In 1969, Professor Jane Mercer of the University of California at Riverside investigated the disproportionate enrollments of minorities in Riverside. She found that the overenrollment of minorities was directly attributable to the disparities in I.Q. scores between white and minority children. Mercer 1431-32

More recently, the State Department of Education’s only investigation of the E.M.R. placement process, undertaken from July 1975 through June 1976, made important findings with respect to how the children in E.M.R. classes had been placed there. The study, reported as “Systematic Program Development for Mentally Retarded Pupils: A State-Initiated Project 1975-76,” examined 26 school districts that had more than 15 . percent black overenrollment in E.M.R. programs. The moratorium on I.Q. testing was already in effect, but the information on I.Q. scores was available in the files of many children, most of whom had been placed when the I.Q. tests were permissible.

The report stated that, even using a “liberal . . . interpretation of what constituted an estimate of adaptive behavior,” about one-third of the pupil records contained no estimates of pupil adaptive behavior. Furthermore, 16 of 25 districts admitted that they made little or no use of adaptive behavior information in the placement process. The report also revealed that more than one-quarter of the files surveyed revealed no “developmental history.” In San Francisco, for example, the study team found that a request for adaptive behavior information was found in 23 of the 47 files sampled, developmental histories were located in 35 of the 47, and records indicated the I.Q. scores in 46 of the 47. rpjjg record 0f the I.Q. score was clearly the most scrupulously kept record, and it appears to have been the most important one. The lure of the single, seemingly precise and objective measure, is no doubt strong.

The available data suggest very strongly that, even if in some districts the I.Q. scores were not always determinative, they were pervasive in the placement process. This is hardly surprising. As pointed out by Judge Wright in his detailed discussion' of this problem, “Although test publishers and school administrators may exhort against taking test scores at face value, the magic of numbers is strong.” Hobson v. Hanson, 269 F.Supp. 401, 489 (D.D.C. 1907). Moreover, the entire placement process revolves around the I.Q. determination. Retardation is defined in terms of the I.Q. tests, and a low score in effect establishes a prima facie case of retardation. The statutory scheme provides that the score must only be “substantiated.” As a result, it is the I.Q. score that is religiously compiled and recorded. If the I.Q. tests are discriminatory, they inevitably must bias the entire process.

3. The moratorium on I.Q. testing. Since January of 1975, I.Q. examinations no longer have been part of the E.M.R. placement process in California. The state testing moratorium has forced school districts to rely primarily on other factors for placement decisions. This reliance on alternatives sets up a kind of test of the I.Q.’s role in the overrepresentation of black children in E.M.R. classes. Unfortunately, the data do not clearly prove or disprove the hypothesis. It appears that the percentage of black children in E.M.R. classes has not changed substantially, and defendants cite this fact to depreciate the importance of the I.Q. test scores in creating those percentages. Plaintiffs suggest, however, that in view of the admitted tendency of schools to keep children in E.M.R. classes on a more or less permanent basis, the relevant data to examine are those concerning new E.M.R. placements. The only data on this point can be found in an exhibit provided by plaintiffs and covering the 20 school districts with 80 percent of the black student population. Plaintiffs’ chart is reproduced below as Table 3.

TABLE 3

TWENTY DISTRICTS ACCOUNTING FOR 80% OF CALIFORNIA’S BLACK POPULATION

_ District % Black In District 1976-1977 % Black of New EMR Placemts. 1976-1977 % Black of New EMR Placemts. 1973-1974 % Black In District 1973-1974

Bakersfield Elem. 15.76 15.62 41.4 (36/87) 16.1

Berkeley Unified 43 0 100 (4/4) 43.9

Vallejo City 28.06 25.0 57.1 (60/105)

Fresno Unified 10.34 25.0 27.3 (23/84) 9.7

Compton Unified 84.87 88.52 93.4 (71/76) 88.2

Pomona Unified 29.69 28.57 56.7 (136/240) 26.7

San Diego Unified 15.54 33.62 20.9 (35/167) 13.3

Pasadena Unified 42.78 33.33 42.9 (3/7) 40.0

Monterey Peninsula 18.76 22.22 22.7 (5/22) 18.3

Richmond Unified 35.22 42.85 55.0 (11/20) 31.8

San Bernardino Unified 15.24 29.54 40.2 (39/97) 15.9

Los Angeles Unified 24.08 50.1 50.7 (604/1192) 25.4

Oakland Unified 67.11 83.15 84.8 (112/132) 64.4

Inglewood Unified 79.8 76.92 -- No Data -- 59.7

Lynwood Unified 48.25 53.33 51.0 (25/49) 27.1

Long Beach Unified 15.04 18.98 24.6 (17/69) 11.9

Stockton 16.43 35.48 18.0 (11/61) 15.3

Ravenswood City 84.60 100 100 (4/4) 84.5

S.F.U.S.D. 31.05 43.47 79.4 (23/34) 30.5

Sacramento 20.30 --No Data -- 31.5 (12/38) 17.7

The data are somewhat fragmentary for a number of districts, and the results are not uniform, but uncontradicted expert testimony based on the totals indicates that the four percent drop in the placement of black children into E.M.R. classes — 50 percent to 46 percent — is statistically significant and therefore not likely to have oc curred by chance. Berk 9748-49. Thus, while firm conclusions cannot be drawn, the evidence is consistent with the hypothesis that I.Q. tests have contributed to the disproportionate enrollment of black children. This is not to say that such evidence by itself establishes the contribution of the I.Q. examinations, but it does bolster the conclusion already suggested by the analysis of the role of the I.Q. in the placement process.

IV. INTELLIGENCE TESTS

A. Introduction: The Impossibility of Measuring Intelligence

A large proportion of the 10,000 pages of transcript in this case concerns the I.Q. tests themselves, how they operate, and what they can and cannot .measure and predict. While many think of the I.Q. as an objective measure of innate, fixed intelligence, the testimony of the experts overwhelmingly demonstrated that this conception of I.Q. is erroneous. Defendants’ expert witnesses, even those closely affiliated with the companies that devise and distribute the standardized intelligence tests, agreed, with one exception, that we cannot truly define, much less measure, intelligence. We can measure certain skills but not native intelligence. Professor Robert Thorndike of Columbia, defendants’ first expert witness, confirmed that the modern consensus represents a change from that held in the early years of the testing movement: “Everybody would acknowledge we would have no conceivable way of directly measuring native ability.” Thorndike 4869. Dr. Leo Munday, Vice President and General Manager, Test Department, Houghton Mifflin Co., publishers of the Stanford-Bi-net, similarly concluded that, “It is safe to say that ... no one in aptitude testing today believes that intelligence tests measure innate capacity.” Munday 5085.

I.Q. tests, like other ability tests, essentially measure achievement in the skills covered by the examinations. As stated by Professor Leon Kamin, Professor of Psychology from Princeton University,

Kamin 880. This point ought to be an obvious one, but it sometimes gets overlooked because of the mystique associated with the concept of I.Q. The tests measure the skills tested, and each of the tests subject to this litigation assesses very similar skills.

In addition, the tests have been constructed in such a way that the numerical scores will fit along a normal probability curve of the type utilized by statisticians. Designers of the tests have selected questions that cumulatively are of a degree of difficulty that, when given to a sample population, will result in scores that correspond to a bell-shaped curve. According to that curve, as applied to standardized intelligence tests, the mean or average score will be 100, and the other scores will cluster equally on either side of the mean. As explained in the Manual for the Wechsler Intelligence Scale for Children — Revised,

Thus the tests are designed “[i]n consonance with the theory that the least ambiguous way to specify intelligence levels is to define them statistically.” Tests are developed according to the plausible but unproven assumption that intelligence is distributed in the population in accordance with a normal statistical curve.

The I.Q. tests must be recognized as artificial tools to rank individuals according to certain skills, not to diagnose a medical condition. The ranking is done in a manner useful for statistical analysis. The recognition of what the tests can and cannot do does not vitiate the use of I.Q. tests, but it is hecessary to clear up some of the confusion surrounding the tests. It cautions us to look very carefully at what the tests do measure and exactly how they were validated for determining mental retardation. Before probing deeper into these test issues, however, it is necessary to discuss the test scores of black and white children.

B. Scores of Black and White Children on I.Q. Tests

It has generally been known since at least the 1920’s that black persons do worse than white persons on standardized intelligence tests of the kind developed earlier in the century. Kamin 910-13; Thorndike 4966. The undisputed testimony of experts from both sides of the case demonstrated this disparity graphically. For example, the figure below shows that black children aged 6V2 to IOV2 have a mean score of 88 on the WISC-R, while white children have a mean score of 103. Doppelt 5428-33; Thorndike 5037-38 (the score for Hispanic children is also low).

FIGURE 1

Black children score, on the average, one standard deviation below white children. Translating this finding into the premoratorium criteria used in California for E.M.R. placement, approximately two percent of the total population fall below the two-standard deviation cut-off, while about 15 percent of black children fall below that level. Hilliard 298-300. The use of the standardized tests thus necessarily leads to the placement of more black than white children in E.M.R. classes.

C. Reasons for the Disparity in I.Q. Scores

Disparities in I.Q. test scores between black and white children can be caused by various factors. One explanation could be that the test itself is inherently discriminatory. An unbiased test that measures ability or potential should yield the same pattern of scores when administered to different groups of people. If variations in test scores arise, the test may be revised to remove the unwanted bias. An earlier version of at least the Stanford-Binet I.Q. test was modified in this way because the test yielded different scores for boys and girls and the testing experts assumed such differences were unacceptable. E. g., Kamin 875-77; Hilliard 290. No such modification on racial grounds, however, has ever been tried by the testing companies according to the testimony at this trial. Rather, the experts have from the beginning been willing to tolerate or even encourage tests that portray minorities, especially blacks, as intellectually inferior.

Equalizing scores of minorities on intelligence tests may not be an easy task. It raises difficult questions regarding which skills are appropriate for our schools and society, and which abilities, or lack thereof, should dictate E.M.R. placement. Such modifications in the test may not be necessary, however, if the disparate scores are attributed not to an unjustified bias in the testing instruments but to actual, relevant, differences between the groups tested, in this case, between white and black children. Several arguments may be advanced to explain the test score differences, including the notorious genetic argument that evidently has found, new respectability since the late 1960’s.

1. The genetic argument. The simplest explanation for the I.Q. score disparity is the genetic argument expounded most notably by Professor Jensen and given support at trial by Professor Gordon. A number of key state officials, in addition, testified that they were familiar with Professor Jensen’s writings, and they would not rule out the genetic explanation for disparities between black and white children in I.Q. scores and E.M.R. enrollment. E.g., Brinegar 4691-4703; Hanson 3800-02 ; Simmons 3486-88. The genetic argument, stated bluntly, is that natural selection has resulted in black persons having a “gene pool” that dooms them as a group to less intelligence. Of course exceptions are recognized, but the intelligence curve is lower for blacks as a group. This curve coincides with the bell-shaped curve of I.Q. scores, which as noted before tends to fall one standard deviation lower than that of whites.

The first problem with the genetic argument is that defendants were unwilling to admit any reliance on it for policy-making purposes. Superintendent Wilson Riles, the person with authority to take a position on this issue, was asked whether he thought that genetic inferiority could possibly explain the high proportion of black children in E.M.R. classes. He testified, “The whole concept is reprehensible to me. No, I do not agree.” Riles 2751. Thus, even if it may appear that certain decisions were made on the basis of the genetic argument or at least in harmony with that argument, the official policy rejected the genetic explanation for mild mental retardation.

The rather weak evidence on the record in support of the genetic explanation tends to rest on the disparities in I.Q. scores, which obviously overlooks any possible bias of the tests themselves. Indeed, it appears that both assumptions underlying the genetic argument are highly suspect. First, it is not at all clear whether “intelligence,” whatever that may be, is inherited rather than the product of the environment, or at least whether the environment can substantially overcome any inherited weaknesses. Second, even if intelligence is inherited, it is not to be assumed that black persons are less intelligent as a group than white persons as a group. That conclusion depends upon one’s view of the objectiveness of the I.Q. tests, and, as noted before, at the very least the testing personnel do not claim to have examined systematically whether the tests are as valid for black children as for white children in measuring innate capacity to learn. We cannot, accordingly, rely on the genetic argument to explain the I.Q. test score disparities between black and white children.

2. The socio-economic argument. Dr. Riles preferred to admit that the I.Q. tests were biased against the poor — black and white alike — not because of inferior genes or divergent cultures but because of their inferior home and neighborhood environments. Again, according to Dr. Riles,

Riles 2720. Poverty is clearly a handicap for children taking I.Q. tests. Powell 1211-12. But the question before the court is whether such disadvantage leads not only to poorer test and school performance, but also to the condition of mental retardation thought to justify placement in E.M.R. classes.

Dr. Herbert Grossman of U.C.L.A. testified that poverty puts individuals at a “greater risk” for all kinds of diseases, including the “disease” of mild mental retardation. Grossman 5669. He was unable, however, to specify how this risk operated, nor was he able to detail the extent of its asserted effect on intelligence as measured by I.Q. scores. Grossman 5695. In addition, this kind of analysis fails to explain adequately why more severe mental retardation, such as that consistent with placement into classes for the “trainable mentally retarded” children, also does not occur in greater proportions among the poorer sectors of the population, particularly among blacks. Grossman 5669-70, 5597-98; Hillabrandt 2555-61.

Further, plaintiffs’ expert from U.C.L.A., Dr. Gloria Powell, explained convincingly that recent studies, including the Collaborative Research Study of the National Institute of Neurological Diseases and Blindness, have demonstrated the insufficiency of this “at risk” explanation of low I.Q. scores and increased mental “disease.” The studies she cited demonstrated that unless there is “severe” malnutrition of a kind rare or non-existent in this country, less than ideal nutrition during pregnancy and the first six months of life “does not affect later performance on I.Q. scores or in cognitive processes.” Powell 1249. The relatively low scores of black children do not result from mental disease attributable to the physical risks of poverty even if school performance, as opposed to mental retardation, does vary somewhat according to socio-economic status. Powell 1180-82, 1211-12.

Moreover, it is clear that socio-economic status by itself cannot explain fully the undisputed disparities in I.Q. test scores and . in E.M.R. placements. As one of defendants’ experts, Professor Lloyd Humphreys of the University of Illinois, testified, “there are a very great many studies in which . . . the objective socio-economic status indices have been held constant. The almost uniform finding is that blacks still score lower than the whites.” Humphreys 6342-43. The government’s statistical experts confirmed this observation in some detail. Hillabrandt 10026. The I.Q. disparities of some 15 points (one standard deviation from the mean) and the overrepresentation of black children in E.M.R. classes cannot be explained away solely on the grounds of the generally lower socio-economic status of black children and their parents. It is necessary, therefore, to inquire further into the cultural biases of I.Q. tests.

3. The cultural bias of standardized I.Q. tests. The insufficiency of the above explanation leads us to question the cultural bias of I.Q. tests. The first important inferential evidence is that the tests were never designed to eliminate cultural biases against black children; it was assumed in effect that black children were less “intelligent” than whites. Those who did not make that assumption recognized the problem of cultural bias, even if they did not try to investigate it systematically. Dr. David Wechsler in 1944 stated this problem quite strongly in his introduction to the WISC, which he developed:

Wechsler’s observations apply with equal force to the other standardized tests. The tests were standardized and developed on an all-white population, and naturally their scientific validity is questionable for culturally different groups. Black childrens’ in-1 telligence may be manifested in ways that the tests do not show, so that the existing! tests developed on the white population obviously would be inadequate.

Testing organizations did become sensitive to the absence of minority children in the standardizing samples. The StanfordBinet was restandardized in 1972 to include black children in the same proportions that they were represented in the population at large. The 1970’s version of the revised Wechsler Intelligence Scale for Children (WISC-R) was similarly restandardized. The problem with these efforts is that they did not address the main point of Wechsler’s statement. Mixing the populations without more does not eliminate any preexisting bias. If the tests were inappropriate before, and, as the testimony indicated, the test items were not changed significantly after the mixing, the tests would remain inappropriate.

Another type of indirect or inferential evidence of cultural bias is the manner in which I.Q. scores of black children may be improved. For example, there was testimony that black children raised in a cultural environment closer to the white middle-class mainstream tend to perform better on the tests, that black children raised by white parents tend to do better, and that early intervention programs can dramatically improve testing results. Mercer 1637-46, 1664-66; Thorndike 4991-92; Albee 545-46.

Beyond these kinds of evidence of bias against black children, the specific qualities of the tests must be scrutinized. Unfortunately, we are not able to base such an inquiry on as much information as we would like, because the testing companies, in the words of Dr. Munday of Houghton • Mifflin, have not done a “systematic analysis of how blacks and whites perform on each item of the tests.” Munday 5248; see also Doppelt 5368-84, 5483-91. The data that are available however, are consistent with a finding of bias against black children.

The intelligence tests with which we are concerned typically comprise a number of different sub-tests, usually organized into a “verbal scale”-and a “performance scale.” The verbal scale tests items of general information that the test designers have found to be known by a certain percentage of the standardizing sample. The verbal scale also tests knowledge of similarities and differences, general comprehension, and vocabulary. The performance component of the examinations includes mazes, picture completion, recollection of numbers in sequence and block design tests.

At the outset, it is undeniable that to the extent black children speak other than standard English, they will be handicapped in at least the verbal component of the tests. We must recognize, as Dr. Asa Hilliard testified, that “Vocabulary is not standard, even when people use the same word. So even before you are able to use your mentality to figure out the answer to a problem, you first have to ensure that the examiner and the examinee understand the word in the same way. . . .” Hilliard 165. Dr. Hilliard and other witnesses pointed out that black children are more likely to be exposed to nonstandard English, and that exposure will be reflected in I.Q. scores. E. g., Albee 448-49; Hilliard 794-95; Meyer 8663-64; Goldstein 7756-57. This conclusion appears to be corroborated by the only major study brought to the court’s attention on how different races score on different sub-parts of I.Q. tests. The Collaborative Perinatal Research Project, utilizing a sample of 2,000 black and 2,000 white children, showed the largest disparity in scores in vocabulary, where the difference was approximately 12 I.Q. points. In contrast, the difference in the coding of numbers subtest was only two points, although the black children still scored worse.

Cultural differences can also be found in specific test items. Some of these items have in fact become rather notorious, such as the “fight item” on WISC tests. This question asked children what they would do i if struck by a smaller child of the same sex. ‘ The “correct” answer is that it is wrong to strike the child back. Doppelt 5484-85. Young black children aged six and seven “missed” this item more than twice as often as their white counterparts. Doppelt 5488. The difference can only be attributed to a cultural variation at that age. Similarly, it may be that such questions as who wrote Romeo and Juliet, who discovered America, and who invented the light bulb, are culturally biased.

At a more subtle level, such skills as “picture arrangement” may be tested in a biased fashion if the pictures, which generally are of Caucasian persons, relate to situations more typical of white, middle class, life than the experiences of many black children. In short, a number of indicators point to the existence of a cultural bias against black children. Dr. Riles again admitted this phenomenon, even if he seemed to emphasize the importance of socio-economic status. He testified that “The instruments are biased and . . . they tend to reflect cultural bias.” Riles 2710-11. Only one of defendants’ experts, Professor Gordon, felt that there were no cultural differences between blacks and whites. Otherwise, there was general agreement by all sides on the inevitable effect of cultural differences on I.Q. scores. Put succinctly by Professor Hilliard, black people have “a cultural heritage that represents an experience pool which is never used” or tested by the standardized I.Q. tests. Hilliard 370.

D. The Relevance of Cultural Bias and the Disparity in I.Q. Scores

The discovery of the cultural bias of I.Q. tests, which is hardly disputed in this litigation, can lead to a variety of contradictory conclusions. The conclusion reached depends upon what one assumes the tests measure, and for what purposes they are designed and used. We have seen quite clearly that the tests cannot give a simple number corresponding to an innate trait called “intelligence.” The tests instead give a statistically correlated score indicating the current performance of children on the items chosen and utilized by test designers. On this finding alone one might want to question some of the fundamental assumptions of the E.M.R. classes built on I.Q. testing. In particular, we could question the assumption that there is a definite and more or less fixed group of children who are not mentally retarded in the sense that they have a definable disease, but rather are normal except for the fact that they are incapable of mastering the academic skills taught to normal children. Judge Wright in Hobson v. Hansen emphasized the importance of that assumption and its relationship to the I.Q. tests in the Washington, D.C. schools:

269 F.Supp. at 475 . While the existence of these E.M.R. classes — analogous to the track system in Washington, D.C. — is not directly in question in this litigation, we cannot avoid noting its relationship to somewhat tenuous assumptions about the ability of I.Q. tests to measure a student’s potential to master normal academic skills.

Assuming, however, that the I.Q. tests are valid for measuring appropriate skills and potential for white middle-class children, one might also conclude that the tests are inappropriate for black children, especially poor black children. To the extent that a “black culture” — admittedly a vague term — exists and translates the phenomenon of intelligence into skills and knowledge untested by the standardized intelligence tests, those tests cannot measure the capabilities of black children. This important conclusion will be considered very carefully in relation to the legal analyses developed below.

Another possible conclusion, which should be set out here in a preliminary fashion, is based on a slightly more complicated analysis. It may be most illuminating to phrase this approach to the I.Q. problem in terms consistent with defendants’ major argument. Let us assume that the tests are culturally biased, and that black children accordingly score poorly. Assume also that intelligence is not an innate and fixed trait. Still, the argument goes, we must recognize that the culture of the United States (and perhaps that of other developed countries) is the culture reflected in the I.Q. tests. Success in terms of the modern culture, whether in school or later in work, depends on the acquisition of skills that the I.Q. tests can measure. The I.Q. tests, moreover, not only measure those skills, but also predict how those taking the tests will succeed in the future in learning those skills. Those who will not succeed belong in special classes where they can at least learn some very basic skills that will enable them to “get along” better than if they were left in normal classes. E.M.R. classes arguably serve this purpose, and if black children get assigned more often to those classes, that is perhaps lamentable, but it merely indicates that black people have not yet been brought sufficiently into the modern culture necessary for them to advance at a more rapid rate in schools and in careers.

We also shall be scrutinizing in some detail this last justification for I.Q. tests, and it will be seen that its evident logic does not fit the facts. To give up so easily on black children is simply not justified. Again, however, it would not be appropriate to draw factual conclusions in the abstract. This level of defense and additional factual material must be examined in light of the various legal theories and burdens of proof that derive from them. With the factual issues thus framed, we turn to the law.

V. LEGAL ANALYSIS

A. Introduction

Plaintiffs and the U.S. Department of Justice, as amicus curiae, have mounted a massive “attack on E.M.R. classes built on I.Q. testing. While plaintiffs’ focus is on the I.Q. tests themselves, it can be seen that they challenge a number of related practices, including the disproportionate enrollments in E.M.R. classes, the E.M.R. placement process, the central role of I.Q. tests in that process, and the state decisions to mandate and continue I.Q. testing for E.M.R. placement. These concerns revolve around the question of what the Constitution compels in educational equality and what Congress, with its own interpretation of that concept, has meant to proscribe.

There are a number of sources for plaintiffs’ contentions. Plaintiffs and the Department of Justice invoke the Equal Protection Clauses of the United States and California Constitutions; Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d et seq. (hereinafter “Title VI”) and the Emergency School Aid Act of 1972, 20 U.S.C. § 3191 et seq., protecting racial minorities against discrimination by recipients of federal assistance; and the analogous protections for the handicapped — including those treated as handicapped — under section 504 of the Rehabilitation Act of 1973, 29 U.S.C. § 794 (“section 504”), and the. Education for All Handicapped Children Act Amendments of 1975, 20 U.S.C. § 1401 et seq. (“the E.H.A.”). Each of these legal theories dictates a particular analysis and approach and a particular assignment of evidentiary burdens. There is a significant overlap, however, which will be apparent' from the following organizational scheme. We shall begin with the federal laws that place most emphasis on the objective results. — the effects — of I.Q. testing and E.M.R. placement.

B. Federal Statutory Claims

1. Private right of action and exhaustion of administrative remedies. Defendants rely very heavily on several procedural arguments to defeat plaintiffs’ federal statutory claims. They insist first that plaintiffs lack a private right of action to enforce their federal claims, and second, that any private rights cannot be enforced because of plaintiffs’ failure to exhaust administrative remedies. Neither of these arguments requires detailed analysis.

There is no longer any question about the existence of a private right of action to enforce plaintiffs’ statutory claims. The issue has for all practical purposes been resolved by the Supreme Court’s decision in Cannon v. University of Chicago, 441 U.S. 677 , 99 S.Ct. 1946 , 60 L.Ed.2d 560 (1979). In Cannon , the Court reversed the Seventh Circuit and held that, despite the absence of state action and of any statutory language expressly creating a private right of action, Title IX of the Education Act Amendments of 1972, 20 U.S.C. § 1681 (“Title IX”), creates an implied private right of action. The Supreme Court’s lengthy opinion leaves no doubt that such implied rights of action also are available to enforce the provisions of Title VI of the Civil Rights Act of 1964 and section 504 of the Rehabilitation Act of 1973. As the Court emphasized, Title VI and Title IX are very closely related:

441 U.S. at 694-96 , 99 S.Ct. at 1956-57 , 60 L.Ed.2d at 574-75 (footnote omitted). The Court has therefore approved the analysis favored by the appellate courts, e. g., Bossier Parish School Board v. Lemon, 370 F.2d 847, 851 (5th Cir. 1967), cert. denied, 388 U.S. 911 , 87 S.Ct. 2116 , 18 L.Ed.2d 1350 , and consistent with the Court’s own implicit holdings based on Title VI in Lau v. Nichols, 414 U.S. 563 , 94 S.Ct. 786 , 39 L.Ed.2d 1 (1974), and Regents of University of California v. Bakke, 438 U.S. 265 , 98 S.Ct. 2733 , 57 L.Ed.2d 750 (1978). A private cause of action must be considered established under Title VI.

The decision and reasoning in Cannon cogently applies also to the private right of action question under section 504 of the Rehabilitation Act of 1973, even though that provision was not expressly mentioned in the Cannon opinion or interpreted for this purpose in the Supreme Court’s more recent decision of Southeastern Community College v. Davis, 442 U.S. 397 , 99 S.Ct. 2361 , 60 L.Ed.2d 980 (1979). This conclusion follows necessarily from the close relationship of section 504 to Title VI and Title IX.

Section 504, like Title IX, tracks the language of Title VI, except that instead of substituting “sex,” for “race, color, or national origin,” Congress inserted language applicable to “otherwise qualified handicapped individual[s].” Moreover, this provision, like Title IX, originated in an abortive attempt to amend Title VI when it was originally going through Congress. See Lloyd v. Regional Transportation Authority, 548 F.2d 1277 , 1280 n.9 (7th Cir. 1977); Cannon v. University of Chicago, 441 U.S. at 694 n.16, 99 S.Ct. at 1956 n.16, 60 L.Ed.2d at 574 n.16. Given this similarity in derivation, the analysis undertaken in Cannon and derived from Cort v. Ash, 422 U.S. 66 , 95 S.Ct. 2080 , 45 L.Ed.2d 26 (1975), compels the implication of a private right under section 504.

Further, one key factor under the Cort test, the legislative history of section 504, is especially persuasive. Amendments to the Rehabilitation Act in 1974 made it clear that Congress expected and intended that the words similar to Title VI would be interpreted as they impliedly had been in Lau v. Nichols, 414 U.S. 563 , 94 S.Ct. 786 , 39 L.Ed.2d 1 (1974), decided earlier that year. In the words of the Committee Report, those words would “permit a judicial remedy through a private action.” See analysis in Lloyd v. Regional Transportation Authority, 548 F.2d at 1285 . Not surprisingly, all circuits who have faced this issue have ruled in favor of the private right of action. Lloyd v. Regional Transportation Authority, supra; Leary v. Crapsey, 566 F.2d 863,865 (2d Cir. 1977); Davis v. Southeastern Community College, 574 F.2d 1158, 1159 (4th Cir. 1978), rev’d on other grounds, 442 U.S. 397 , 99 S.Ct. 2361 , 60 L.Ed.2d 980 (1979); United Handicapped Federation v. Andre, 558 F.2d 413 (8th Cir. 1977). On the basis of this authority and on the strong signal of the Supreme Court in Cannon , private rights of action to enforce Title VI and section 504 must be recognized.

Similarly, Congress expressly recognized a private right of action to enforce the requirements of the Education for All Handicapped Children Act of 1975. See 20 U.S.C. § 1415 (e)(2), 45 C.F.R. § 121a.511. The Act, however, evidently first requires individuals to exhaust the administrative machinery set up to review the merits of the “individualized education plan” determined pursuant to the law. Accordingly, at least one court has declined to entertain an individual child’s action under the Education for All Handicapped Act because the administrative machinery had not been invoked. Stubbs v. Kline, 463 F.Supp. 110 (W.D.Pa.1978). But see New York State Association v. Carey, 466 F.Supp. 479 (E.D. N.Y.1978) (time of essence so no deference to administrative machinery). Whatever the merits of the Education for All Handicapped Act’s normal exhaustion requirement, however, we find exhaustion inapplicable here. Despite the failure of any class members to exhaust administrative remedies, it is abundantly clear that such exhaustion would have been meaningless and superfluous.

First, it must be remembered that this case began strictly as a constitutional claim, and the equal protection claim has remained central. Indeed, a preliminary injunction based entirely on that claim is still in effect. It is important that we get to the merits of that claim, and it would be wasteful to avoid the statutory claims that have accrued in the meantime. Judge Weinstein’s recent decision in Lora, based on a number of claims similar to this case, appropriately made short shrift of the exhaustion arguments: “This is essentially a constitutional and not a statutory case. Defendant’s strong reliance on exhaustion cases is, therefore, inappropriate.” Lora v. Board of Education, 456 F.Supp. 1211 (E.D.N.Y.1978).

Second, the Supreme Court’s decision in Cannon suggested that exhaustion will not be a serious barrier to suits under Title VI and the Rehabilitation Act. Both provide for the same administrative scheme before the Court in Cannon . They provide by regulation that any person may file a written complaint alleging violation of the statute within 180 days of the occurrence of the discrimination. After investigation, HEW will seek compliance, formally or informally, or inform the complainant that further agency action is unwarranted. 45 C.F.R. § 80.7 (bXc) (Title VI); § 86.71 (Title IX); § 84.61 (section 504). The Supreme Court refused to require exhaustion of this procedure for private Title IX claims, and the reasoning is very strong. Noting the absence of private participation in those proceedings, the Court emphasized that “it has never withheld a private remedy where the statute explicitly confers a benefit on a class of persons and where it does not assure those persons the ability to activate and participate in the administrative process contemplated by the statute.” Cannon v. University of Chicago, 441 U.S. at 707 n. 41, 99 S.Ct. at 1963 n. 41, 60 L.Ed.2d at 580 n. 41. Once again the same results are dictated for section 504 as for Title VI.

Given the close relationship in this litigation of Title VI, section 504, and the E.H.A. and our finding that the first two do not require exhaustion, it would be absurd as a practical matter to deny redress under the E.H.A., which took effect on October 1, 1977. It is true that the drafters of the E.H.A. regulations took no position on the propriety of a direct appeal to the courts if, inter alia, a class action was involved, but in a case based on multiple statutes and charging systemwide testing and placement biases, we frankly do not see the purpose of individual exhaustion. As one commentator observed:

Note, Enforcing the Right for an “Appropriate” Education: The Education for All Handicapped Children Act of 1975, 92 Harv. L.Rev. 1103, 1116 (1979).

There is a final reason why exhaustion would be inappropriate. The Department of Justice, which participated very actively in this litigation since it was granted status as amicus curiae, never urged any deference to administrative machinery. More significantly, John Wodatch of the Office of Civil Rights of the Department of Health, Education, and Welfare told the court prior to trial that the office would await the outcome of this litigation before taking any administrative action against California. In this situation, exhaustion would obviously be just a time-consuming exercise in futility. This lawsuit must be resolved on the merits.

2. Title VI of the Civil Rights Act of 1964: One application of an “effects test.” Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d, provides a very strong statutory basis for plaintiffs’ challenge. The focus is precisely on the effects of I.Q. testing and E.M.R. placement decisions, and the application of this law is still rather simple, despite uncertain hints from the Supreme Court suggesting that some “intent” requirement may eventually be applicable here. Regents of University of California v. Bakke, 438 U.S. 265, 353 , 98 S.Ct. 2733, 2780 , 57 L.Ed.2d 750, 810 (1978); see Guadalupe Organization v. Tempe Elementary School, 587 F.2d 1022 , 1026 n. 2 (9th Cir. 1978). Cf. Board of Education v. Califano, 584 F.2d 576 , 588 nn. 38-39 (2d Cir. 1978), cert. granted, 440 U.S. 905 , 99 S.Ct. 1211 , 59 L.Ed.2d 453 (1979).

Title VI provides that

There is no question that defendants receive substantial federal assistance. E. g., Riles 2707-10; Brinegar 4678-83. Regulations issued under the authority of section 602 of the Act state that recipients of federal aid

45 C.F.R. § 80.3 (b)(2).

The Supreme Court’s signal Title VI case is Lau v. Nichols, 414 U.S. 563 , 94 S.Ct. 786 , 39 L.Ed.2d 1 (1974), holding specifically that “[discrimination is barred which has that effect even though no purposeful design is present." (Emphasis added.) 414 U.S. at 568 , 94 S.Ct. at 789 . In Lau , the Court found that Title VI and implementing regulations were violated by San Francisco’s failure to provide English language courses for students of Chinese ancestry. That failure, according to the Court, violated Title VI, since it had an adverse impact on Chinese-Amerieans and “students who do not understand English are effectively foreclosed from any meaningful education.” 414 U.S. at 566 , 94 S.Ct. at 788 . See also Serna v. Portales Municipal Schools, 499 F.2d 1147 (10th Cir. 1974) (following Lau for Spanishsurnamed children); cf. Guadalupe Organization v. Tempe Elementary School, 587 F.2d at 1029 and n.7; De La Cruz v. Tormey, 582 F.2d 45 (9th Cir. 1978), cert. denied, 441 U.S. 965 , 99 S.Ct. 2416 , 60 L.Ed.2d 1072 (1979) (effects test for Title IX suggested, citing Lau).

The same kind of analysis is applicable to students placed in E.M.R. classes on the basis of criteria that result in a grossly disproportionate overenrollment of black children. Indeed, plaintiffs in this case rely on the very regulations that were found valid in Lau , 45 C.F.R. § 80.3 (b)(2). Placement mechanisms for E.M.R. classes operate with a discriminatory effect that, to paraphrase Lau , effectively forecloses a disproportionate number of black children from any meaningful education. On the basis of their different cultural background, which results particularly in lower scores on I.Q. tests, black children are subjected to discrimination analogous to that borne by many San Francisco Chinese, who, because of their cultural background, could not communicate effectively in English. Certainly many Chinese Americans would succeed in those schools even without remedial English. Nevertheless, the failure to provide English language teaching foreclosed substantial numbers of students from any meaningful educational opportunity. This same result occurs from the use of I.Q. tests and a biased placement process.

Lau does not dictate affirmative remedial education for all divergences from the majority culture, but at a minimum it prohibits confinement of blacks to E.M.R. classes because of their culture. The overwhelmingly disproportionate placement of black children into E.M.R. classes penalizes black children and leads to the same kind of drastic culture-specific deprivation addressed in Lau . The Ninth Circuit Court of Appeals, citing Lau , recently emphasized the relevance of this additional fact. In refusing to require “bilingual-bicultural” education for Mexican-American and Yaqui Indian children under Title VI when remedial English was available, the court carefully noted the absence of any hint that the remedial program “operates ‘as an educational deadend or permanent track.’ ” Guadalupe Organization v. Tempe Elementary School, 587 F.2d at 1030 (citing 414 U.S. at 568 , 94 S.Ct. 786 ). Such a track is involved here, making Lau all the more persuasive. On this authority it is clear that plaintiffs have established their prima facie case.

The findings can be corroborated further by reference to “interpretive guidelines” analogous to the guidelines utilized by the Supreme Court in Lau . An HEW memorandum dated August 1975, and entitled “Identification of Discrimination in the Assignment of Children to Special Education Programs,” was sent to all Chief State School Offices and local school district superintendents. It specifically interpreted Title VI (and the Education for All Handicapped Children Act) as it applies to placement criteria for special classes. According to the memorandum, any one or more of the following practices may violate Title VI “where there is an adverse impact on children of one or more racial or national origin groups, . . . ”

Memorandum for Chief State School Officers and Local School District Superintendents, August 1975, at 3. Again, this HEW memorandum, which particularly addresses testing and evaluation materials, places exclusive reliance on discriminatory impact. The specific provisions logically follow Lau and the earlier regulations and guidelines and therefore merit strict enforcement.

The next issue is whether that impact alone dictates a decision for plaintiffs. Lau and the various guidelines followed here do not discuss how, if at all, defendants can rebut the prima facie case caused by the disproportionate deprivation of educational opportunity. No such rebuttal was made in Lau and the Court suggested that none could be made. Nevertheless, the Lau situation was to some extent different, and some defenses would have been tenable here. If defendants could have shown here that disproportionate enrollment in E.M.R. classes actually reflected and tapped a greater incidence of mild mental retardation in the black children population, and not “cultural bias,” then defendants’ E.M.R. placement process could have survived. Cf. Lora v. Board of Education of City of New York, 456 F.Supp. 1211, 1277 (E.D.N.Y.1978). Further, if defendants could somehow have demonstrated that the intelligence tests had been “validated” for the purpose of E.M.R. placement of black children, those tests could have been utilized despite their disproportionate impact. As discussed elsewhere in this opinion, however, defendants did not make these showings. If the “effects test” of Title VI, as followed by the Supreme Court in Lau v. Nichols, is still the law, and we believe it is, then defendants’ pre-moratorium procedures for assignment to E.M.R. classes must be invalidated on the basis of Title VI alone.

3. The Rehabilitation Act of 1973 and the Education for All Handicapped Children Act of 1975 — another “effects test.” These two federal statutory schemes closely overlap and embrace an educational philosophy contrary to that which has at least until very recently characterized California’s treatment of E.M.R.’s. “Mainstreaming” has become the theme of federal law, while California has opted to isolate E.M.R.’s in special day classes.

The first of the two laws, section 504 of the Rehabilitation Act of 1973, 29 U.S.C. § 794 , as amended in 1974, prevents discrimination against the handicapped in programs that receive federal funding. It covers persons who are actually handicapped or “who are regarded” as handicapped, such as the plaintiffs in this case. Section 504 extends to these persons the same broad protections against discrimination already provided against discrimination on the basis of race by Title VI of the Civil Rights Act of 1964.

It is clear from the legislative history of the law, as well as the interpretative regulations which have been in effect since June 3, 1977, 45 C.F.R. Pt. 84, that one abuse sought to be corrected was the erroneous denial of admission into regular classes.

The Education for All Handicapped Children Act of 1975, 20 U.S.C. § 1401 et seq., with its emphasis on procedural protection in programs funded under its terms, also addresses classification into special classes. In addition, the act provides that in order to qualify for federal assistance, which California is now receiving and which has led to much of the statutory reform embodied in the Master Plan, a state must, inter alia, establish

20 U.S.C. § 1412 (2)(D)(5). The emphasized language, it should be noted initially, restates the concern with racial impact set out in Lau , Title VI, and the August 1975 memorandum. The analysis undertaken in the previous section suggests the same conclusion. Beyond this overlap, however, the E.H.A. and the Rehabilitation Act provide a more specific concern with the validity of evaluation materials used for diagnosis of educational needs and placement into special classes.

The basic aims of the E.H.A. and Rehabilitation Act, as relevant to this litigation, are to ensure the proper classification of children and promote, to the maximum extent possible, the “mainstreaming” of children into regular classes. Accordingly, virtually identical regulations under the two statutes describe allowable practices for decision on proper classroom placement and curriculum design. Both sets of regulations provide, with respect to testing, that state and local educational agencies

45 C.F.R. § 84.35 (b); 45 C.F.R.

§ 121a.532(a). These regulations therefore require that the tests and other evaluation materials be “validated for the specific purpose for which they are used.” We must examine in some detail the meaning of validation in the school placement setting.

4. The question of validation despite discriminatory effects. “Validation” is the determination of whether the placement tests or other evaluation materials are suited for the purposes for which they are used. It is a concept familiar in the testing literature and has frequently been used to evaluate tests used for employment qualifications. To date, however, there are no cases applying validation criteria to tests used for E.M.R. placement. The problem of lack of authority is compounded by some confusion about what exactly the I.Q. tests covered by this litigation are supposed to do. As has already been seen, one can no longer rely on I.Q. scores as objective measures of innate capacity for learning. Prudence dictates, therefore, that we address the question of the validity of the tests from several different perspectives.

The important threshold question is whether “validation” of I.Q. tests and E.M.R. placement mechanisms requires the same showing that is required of employment tests in employment discrimination cases. Defendants argue that tests such as the “Test 21” used by the District of Columbia to screen police candidates are basically general intelligence tests. Defendants thus rely on the Supreme Court’s finding in Washington v. Davis, 426 U.S. 229 , 96 S.Ct. 2040 , 48 L.Ed.2d 597 (1975), that, despite the disproportionate impact of Test 21 on black applicants, the test was validated for the purpose of predicting success in the police training program.

The analysis in Washington v. Davis on this issue was taken from other employment discrimination cases and focused on the question of whether the test scores predict job performance or at least performance in the job training program. As stated in the leading case of Griggs v. Duke Power Co., 401 U.S. 424, 432 , 91 S.Ct. 849, 854 , 28 L.Ed.2d 158 (1971), tests shown to have a discriminatory impact cannot be utilized unless the employer meets “the burden of showing that any given requirement has a manifest relationship to the employment in question.” If the employer meets the burden, the plaintiffs may then show that alternative selection devices exist that would serve the employer’s legitimate interests without discriminatory effects. This three-part test frequently has been cited with approval in recent cases under Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e. See e. g., Dothard v. Rawlinson, 433 U.S. 321, 329 , 97 S.Ct. 2720 , 53 L.Ed.2d 786 (1977); Albemarle Paper Co. v. Moody, 422 U.S. 405, 425 , 95 S.Ct. 2362 , 45 L.Ed.2d 280 (1975); Blake v. City of Los Angeles, 595 F.2d 1367 (9th Cir. 1979); 29 C.F.R. § 1607.4 .

We recognize that this allocation of the burden of proof for “validation” should be followed in the present case, given the disproportionate impact of the tests on black children, but the notion of predicting “job performance” cannot be effectively translated into the educational context. Tests used to determine employee qualifications can, if validated, “serve the employer’s legitimate interest in ‘efficient and trustworthy workmanship.’ ” Albemarle Paper Co. v. Moody, 422 U.S. at 425, 95 S.Ct. at 2375 . Employers are entitled to hire workers who, on the basis of the hiring criteria, can be expected to perform well on the job. Our economy could not survive for long if employers were forced to ignore this crucial factor in economic efficiency.

With schools, however, we obviously have a very different situation. Compulsory attendance of educational institutions is required by the state, and the schools are supposed to take children from different backgrounds and teach them the skills necessary for adaptation and success in our society. This points out a fundamental difference between the use of tests in employment and education, at least in the early years of schooling. If tests can predict that a person is going to be a poor employee, the employer can legitimately deny that person a job, but if tests suggest that a young child is probably going to be a poor student, the school cannot on that basis alone deny that child the opportunity to improve and develop the academic skills necessary to success in our society. Assignment to E.M.R. classes denies that opportunity through relegation to a markedly inferior, essentially dead-end, track.

Furthermore, in the employment sphere, predictions of performance must be made before entrusting a person with a particular job. In the educational setting as it relates to E.M.R. classes, the child is already in school and already doing poorly, so it can be predicted without tests that the poor performance will continue in the absence of remedial attention. Indeed, past performance in school is a better predictor than tests of future school performance. See, e. g., Kamin 885-86; Humphreys 6333-34; Mercer 1650-55; Thorndike 4955.

Finally, it is only to be expected that children placed in E.M.R. classes will fall increasingly behind in their performance in academic skills, which the E.M.R. curriculum deemphasizes. Placing children with lower I.Q. scores in these classes thus necessarily is a self-fulfilling prophecy. The results of such placements cannot be the basis for test validation in an educational setting.

Nevertheless, a more useful analysis of validation in the educational setting can begin with the burden of proof allocation development in the employment discrimination area. We must inquire whether the tests have been validated for the black children who are members of the plaintiff class, given the disproportionately low test scores of those children and the critical effect of those scores in keeping black children from full participation in regular classes. We are not concerned now with predictions of performance, but rather whether the tests are validated with respect to the characteristics consistent with E.M.R. status and placement in E.M.R. classes. E.M.R. classes exist, it will be recalled, “for people whose mental capabilities make it impossible for them to profit from the regular educational program.” “Mental retardation” is the touchstone, and retardation must make it “impossible” to profit from the regular classes, even with remedial instruction. Defendants have the burden of showing validation of intelligence tests with respect to these characteristics.

The experts testifying in this case agree that I.Q. tests cannot measure innate intelligence, but they disagreed on whether the tests can measure the “mental ability” of children. Compare Albee 513 and Mercer 1550-52 with Doppelt 5315-17 and Munday 5129. There is some question whether this distinction between “intelligence” and “mental ability” makes conceptual sense. For example, several experts, most notably Professor Mercer, persuasively showed how the statistical measure of I.Q. repeatedly has been confused with the “medical model” of retardation as a mental disease. We need not, however, enter into that controversy. We will assume that the tests can accurately measure the mental ability of white children placed in E.M.R. classes, and ask if the tests as utilized for E.M.R. placement have been validated for black children.

The answer, as should be clear from the earlier discussion of the history and biases of I.Q. tests, is that validation has been assumed, not established, for blacks. The tests were developed and standardized in the United States on white, essentially middle-class groups, and if they were then validated with independent indicia of mental retardation justifying placement in isolated, dead-end classes, it was done on the basis of that population. Although several leading figures in the early testing movement saw no problem with this situation, given their own racial prejudices or at least strong ethnocentrism, others, including Dr. Wechsler, the developer of the Wechsler battery of tests central to this litigation, frankly admitted the problem. Wechsler stated in 1944 that the tests were not valid for “the colored population of the United States.” His tests, like the others involved here, had not been made suitable for minority racial and ethnic groups. The tests had been adjusted, for example, to eliminate differences in the average scores between the sexes, but a comparable effort was not made and has never been made for black and white children.

Furthermore, although particular items of the tests reveal a racial and cultural bias, the test items have not undergone the kind of scrutiny consistent with validation for determining retardation for members of the plaintiff class. Defendants have never made such a study, and the testing companies upon whom they relied also admitted this problem. Despite the one standard deviation difference in test scores between the races, no effort has been made to investigate why this disparity occurs or whether it can be corrected by testing redesign. It is instructive, for example, that when the bias of the famous “fight item” was brought to the attention of Dr. Wechsler, he conceded that the item was biased. Yet the test designers have not searched for bias in other test items.

Under these circumstances, defendants have failed to take the steps necessary to assure the tests’ validity. They have committed a serious error that Title VII regulations warn against in the employment situation: “Under no circumstances will the general reputation of a test, its author, or its publisher, or casual reports of tests utility be accepted in lieu of evidence of validity.” 29 C.F.R. § 1607.8 . Whether or not the tests in fact do what they are supposed to do, the law is that defendants must come forward and show that they have been validated for each minority group with which they are used. Cf. 29 C.F.R. §§ 1607.4 (b), 1607.5(b)(5). This minimal burden has not been met for diagnosing the kind of mental retardation justifying E.M.R. placement. See also, e. g., Doppelt 5455-56; Humphreys 6372-73; Munday 5273-75; Sattler 7039-41; Thorndike 5041-45 (defense testimony showing an unwillingness or inability to reach conclusions about validation for E.M.R. placement).

The failure to meet that burden is not just a technical error. The evidence suggested very strongly that the result has been the misplacement of many black children into these special E.M.R. classes. Dr. Watkins, for example, reported from her extensive experience with E.M.R. classes, that

Watkins 4087-88. Similarly, Dr. Powell noted, on the basis of her comprehensive clinical diagnoses, that

Powell 1186-87. See also, e. g., Albee 437-38; Kamin 888; Watkins 4087-88 (quoting Arthur Jensen). The tests have not been validated for black children in E.M.R. classes, and many black children have had to pay the price of that omission.

We now consider defendants’ evidence that the predictive validity of I.Q. tests has been established. Although we are not persuaded that a finding on this evidence is essential to the legal outcome of this case, it may be helpful to develop these findings in view of the large amount of evidence presented on this point. Defendants’ defense of I.Q. testing relied considerably on the assertion that the predictive validity of the tests could be established. We must again lament the paucity of probative studies in this area. Even Professor Sattler’s exhaustive review of the testing literature failed to turn up more than a few studies precisely on point. D-46. The question for predictive validity is not whether the standardized intelligence tests predict school performance generally, but whether they predict such performance for black elementary school children, as opposed to white elementary school children, who score between two and three standard deviations below the mean on intelligence tests. This question has not been the subject of much study. For example, the well-known article by Professor Humphreys and his colleagues published in 1975 as a direct response to the critics of I.Q. biases examined systematically only data relevant to college grade predictions based on high school grades and scores on college entrance examinations. Other studies of I.Q. predictive validity concern only white or mixed populations, and very few indeed compare black and white populations.

Another substantial proportion of these validity studies correlate I.Q. scores with scores on achievement tests, not grades. These was substantial disagreement about the importance of correlations with achievement tests. Plaintiffs and their experts argued vehemently that measurements of achievement “autocorrelate” with I.Q. tests, since they are based essentially on the same pool of test items, and that they do not accurately measure the classroom performance of black children. E. g., Mercer 1527; Jones 4513. Several defense experts, on the other hand, insisted that grades were too “subjective” to assess performance accurately. E. g., Thorndike 4887-88. Despite the admitted subjectivity of grades, however, adequate correlations with grades must be required. In addition to the problem of autocorrelation with the more “objective” achievement tests, we must recognize that grades in our schools are the accepted measure of school performance.

It turns out that very few studies have correlated scores on the standardized intelligence tests mandated in California until the testing moratorium with subsequent grades for black and white children. Sattler 6990-7010. At least one of the major studies casts some doubt on the ability of the tests to predict grades of black as opposed to white children. This Goldman and Hartog study, based on a sample of 201 blacks and 430 whites, investigated correlations between the WISC and grade point average, ■ and found the correlation to be .14 for blacks and .25 for whites. It concluded that the WISC “appeared to be differentially valid for white vs. minority children,” and that, “The WISC has little or no validity for predicting the scholastic performance of black or brown children.” Sattler 6998-7004. In addition, Professor Mercer reexamined the extensive data that she had collected in the late 1960’s in Riverside, California, and she also found that the correlation was both low and racially different for classroom grades in 1968 compared to WISC scores in 1967. The correlation was .20 for black children and .46 for white children. Mercer 1527-32. See also Thorn-dike 4927-32, 5002-04.

The key finding of these studies of grade predictions is not the extent of the correlations between I.Q. test scores and school performance (although the correlations are not great), but rather that the I.Q. tests are differentially valid for black and white children. As noted before and consistent with the standards in the employment testing area,-tests cannot be validated on black persons if they are differentially valid. Cf. 29 C.F.R. § 1607.5 (b)(5). Differential validity means that more errors will be made for black children than whites, and that is unacceptable.

If we consider how I.Q. scores correlate with achievement tests or simply take the informed opinions of the experts, it is still apparent that the tests are differentially valid. Professor Thorndike, for example, admitted this proposition while testifying for the defense:

Thorndike 4927-32. See also, e. g., Doppelt 5507-09; Humphreys 5302-03, 6334-35; Lambert 6693-98 (all testifying for the defense); Albee 437-38 (for plaintiff). Whether correlations are sufficient generally to satisfy the standards of predictive validity, the telling point is that they are substantially worse for black than for white children. This conclusion, consistent with what one would expect given the historical development of the tests, compels a finding for plaintiffs on the issue of predictive validity. Even if predictive validity could form the basis of validation for E.M.R. diagnosis and placement, defendants would still be unable to prevail.

5. Alternatives to I.Q. testing for E.M.R. placement. Even if defendants had discharged their burden of showing a sufficient relationship of the test-centered placement process to E.M.R. status, plaintiffs would still be entitled to prevail if they could show that alternative devices for placement exist and would serve defendants’ legitimate interests without the same discriminatory effect. See, e. g., Blake v. City of Los Angeles, 595 F.2d 1367 (9th Cir. 1979). Plaintiffs have met this burden.

Since the state moratorium stopped I.Q. testing in California for E.M.R. placement in early 1975, we can initially focus on the alternatives to the tests that have been utilized to place E.M.R. students since that date. The testimony of those close to the schools and the E.M.R. placement process provides a clear answer to this inquiry. The alternatives have worked reasonably well. For example, Mr. Dean, the Director of Special Education for the San Francisco Unified School District, testified that, although he would prefer to reinstate I.Q. testing, he did “not believe that students have been misplaced as a result of not having had the individual tests of intelligence.” Dean 4361. See also Boyce 4390-92. Mrs. Bryant, the Specialist for Psychological Services with the Los Angeles Unified School District, similarly felt that “adequate assessments” were being made without the tests. Bryant 8894. And Mr. Schuetz, testifying about the Bakersfield schools, stated that “we have not noticed any qualitative difference in the kind of kids we are getting in the program.” Schuetz 9526-27.

Moreover, the evidence showed that many schools are taking more time and care with their assessments for E.M.R. classification, relying more on observational data gathered both in and out of the classroom. See, e. g., Bryant 8894; Schuetz 9526-27; Joyce 7999-8001. School psychologists, teachers, and others involved in the process are now making decisions based on a wide number of factors, and the evidence suggests that the results are less discriminatory than they were under the I.Q.-centered standard. Evaluations can and are taking place through, inter alia, more thorough assessments of the child’s personal history and development, adaptive behavior both inside and outside of the school environment, and classroom performance and academic achievement. Certain tests other than I.Q. tests may be useful in that proc ess. For example, Professor Watkins testified that, from a teacher’s perspective, “the instruments that would be most useful would be an instrument that yielded very specific information relative to what a child could and could not do in specific skill areas.” Watkins 4174. See also, e. g., Powell 1183-1204. Therefore, even in the absence of validated, culturally neutral intelligence tests, which the work of Professor Mercer and others (c/. Hilliard 352-56; Lambert 6576-77) suggests may be developed, California can proceed to make adequate assessments for placement into E.M.R. programs. Indeed, California agreed voluntarily in the Diana litigation to undertake such a program for Hispanic children, and several other states already have developed new approaches.

The present alternatives to I.Q. tests, however, are not automatically valid. We can infer from the available data and our knowledge of the impact of I.Q. tests that there has been improvement. Less discriminatory alternatives have been developed, but they have apparently not been validated, and they are still producing intolerable disproportions in E.M.R. classes. If E.M.R. classes are to continue in their present form, the state will have to make a better effort. In particular, Dr. Watkins’ suggestion will need to be taken seriously. Consistent with the recent federal emphasis on individualized education plans and mainstreaming, the state will increasingly need to utilize tests and other materials that do not simply provide a label, but enable the development of curricula that respond to specific educational needs. E.g., Albee 459-60; Hilliard 347-355; Kamin 913; Lambert 6609-10; Mercer 1598-1608; Munday 5177-78; Sattler 6974; Thorndike 4949-54; Watkins 4170-74. Cf. Jones 4609-10.

C. The Equal Protection Clause of the Fourteenth Amendment

This decision could rest on the federal statutory claims, but in this case it would be inappropriate to conclude our analysis there. The Education for All Handicapped Children Act and the Rehabilitation Act are relatively new, with a corresponding paucity of judicial interpretation fleshing out the contours of the legislation, and the provisions of Title VI may be interpreted differently in the future. Four Supreme Court justices in Regents of the State of California v. Bakke, 438 U.S. 265 , 98 S.Ct. 2733 , 57 L.Ed.2d 750 (1978), already have indicated that they felt Title VI should be construed to require some showing of intentional discrimination rather than simply discriminatory impact, and the position of one or more other justices may now be the same. At this point we simply cannot be sure.

We are therefore brought to the claim of invidious discrimination under the Equal Protection Clause of the Fourteenth Amendment, which is where this litigation commenced. The earlier injunctions rested on this constitutional ground, and on that basis I.Q. testing for E.M.R. placement of black children has been enjoined in San Francisco since 1973 and throughout the state since late 1974. The law of equal protection has evolved considerably in recent years, and therefore the analysis undertaken here is slightly different than that pursued in 1973 and 1974. The difficult question of intent has moved to center stage. It will be the focus of this part of the opinion.

The quest for intent, however, is evidently still not the only path for judicial scrutiny that exceeds the demands of the rational relationship test. The Ninth Circuit recently has affirmed an alternative approach, relying on the so-called “intermediate standard” of review, and a second focus of this discussion of equal protection will concern how that approach fits the facts of Larry P.

1. Intentional discrimination.

a. Determining the standard. It is now established that disproportionate impact, the central fact of the I.Q. tests and the E.M.R. placement process generally,

Washington v. Davis, 426 U.S. 229, 242 , 96 S.Ct. 2040, 2049 , 48 L.Ed.2d 597 (1976). Since the late 1960’s, beginning especially with the Supreme Court’s first case involving “de facto” segregation in a northern school district, Keyes v. School District No. 1, 413 U.S. 189, 205 , 93 S.Ct. 2686 , 37 L.Ed.2d 548 (1973), the Court has increasingly emphasized that the strict scrutiny applied to racial classifications is triggered only by actions that can “ultimately be traced to a racially discriminatory purpose.” Washington v. Davis, 426 U.S. at 240 , 96 S.Ct. at 2048 .

Ascertaining that intent, however, has not been a simple matter, and the Supreme Court has already had to address this question at length on several occasions. See, e. g., Village of Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252 , 97 S.Ct. 555 , 50 L.Ed.2d 450 (1976); Personnel Administrator of Massachusetts v. Feeney, 442 U.S. 256 , 99 S.Ct. 2282 , 60 L.Ed.2d 870 (1979); Columbus Board of Education v. Penick, 443 U.S. 449 , 99 S.Ct. 2941 , 61 L.Ed.2d 666 (1979); Dayton Board of Education v. Brinkman, 443 U.S. 526 , 99 S.Ct. 2971 , 61 L.Ed.2d 720 (1979). It has been particularly difficult, as these cases show, to agree on a formula relating foreseeable discriminatory effects to discriminatory intent. In view of the importance of that relationship in the present litigation, it may be useful to explore the problem of effects vs. intent in evolving equal protection doctrine.

The simplest method for ascertaining administrative or legislative intent is to move in the direction of an “objective” test based on the foreseeáble effects of a challenged decision. Justice Stevens’ concurring opinion in Washington v. Davis, 426 U.S. at 253 , 96 S.Ct. at 2054 , lends some support to this approach:

A completely “objective” standard of intent, however, based on foreseeable effects, would clearly be inconsistent with recent Supreme Court holdings. It would collapse the Supreme Court’s carefully developed distinction between de facto and de jure segregation and place a duty on school boards to eliminate racial disproportions in neighborhood schools even if those schools reflected only local housing patterns and no actual discriminatory or segregative intent. See, e. g., Keyes v. School District No. 1, supra. Compare Crawford v. Board of Edu cation, 17 Cal.3d 280 , 130 Cal.Rptr. 724 , 551 P.2d 28 (1976).

While rejecting this purely objective “effects” standard, however, a majority of federal appellate courts have thus far moved only one step towards subjective determinations of intent. They have created a presumption that school boards and others charged with discriminatory intent intend the natural and foreseeable consequences of their actions. The Second Circuit Court of Appeals recently explained this test as follows:

Arthur v. Nyquist, 573 F.2d 134, 142-43 (2d Cir. 1978), cert. denied, 439 U.S. 860 , 99 S.Ct. 179 , 58 L.Ed.2d 169 .

In Nyquist , the Court of Appeals rejected the Buffalo Board of Education’s contention that racial imbalance resulted from

573 F.2d at 144 . The board’s burden of proof was not met, and it was accordingly charged with a subjective intent to discriminate. This type of analysis, which in practice often gives substantial weight to the presumption, has been adopted with slight variations in the majority of circuits, and they have reaffirmed their approaches after Washington v. Davis and Village of Arlington Heights. See, e. g., Austin Independent School District v. United States, 532 F.2d 380 (5th Cir. 1976), vacated and remanded sub nom. United States v. Texas Educational Agency, 429 U.S. 990 , 97 S.Ct. 517 , 50 L.Ed.2d 603 (1977), on remand 564 F.2d 162 (5th Cir. 1977), petition for rehearing denied, 579 F.2d 910 (5th Cir. 1978) (en banc), cert. denied, 443 U.S. 915 , 99 S.Ct. 3106 , 61 L.Ed.2d 879 (1979); NAACP v. Lansing Board of Education, 559 F.2d 1042 (6th Cir. 1977), cert. denied, 434 U.S. 997 , 98 S.Ct. 635 , 54 L.Ed.2d 491 (1977); Penick v. Columbus Board of Education, 583 F.2d 787 (6th Cir. 1978), affirmed 443 U.S. 449 , 99 S.Ct. 2941 , 61 L.Ed.2d 666 (1979); United States v. Board of School Commissioners, 573 F.2d 400 (7th Cir.), cert. denied, 439 U.S. 824 , 99 S.Ct. 93 , 58 L.Ed.2d 116 (1978); United States v. School District of Omaha, 521 F.2d 530 (8th Cir. 1975), vacated and remanded 433 U.S. 667 , 97 S.Ct. 2905 , 53 L.Ed.2d 1039 (1977), prior judgment reinstated per curiam, 565 F.2d 127 (8th Cir.) (en banc), cert. denied 434 U.S. 1064 , 98 S.Ct. 1240 , 55 L.Ed.2d 765 (1977). Cf. Richardson v. Pennsylvania Department of Health, 561 F.2d 489 (3d Cir. 1977).

The Ninth Circuit Court of Appeals, however, has evidently adopted a more stringent test for intent, requiring a “determination that the school authorities have intentionally discriminated against minority students by practicing a deliberate policy of racial segregation.” Soria v. Oxnard School District, 488 F.2d 579, 585 (9th Cir. 1973); Johnson v. San Francisco Unified School District, 500 F.2d 349 (9th Cir. 1974); Berkelman v. San Francisco Unified School District, 501 F.2d 1264 (9th Cir. 1974). Since this circuit has not spoken recently, however, we cannot state with confidence whether evidentiary “presumptions” such as those used in other circuits could be entertained to prove intent. A very recent case outside of the school desegregation area suggests that there may be some relaxation of the standard.

In De La Cruz v. Tormey, 582 F.2d 45 (9th Cir. 1978), cert. denied, 441 U.S. 965 , 99 S.Ct. 2416 , 60 L.Ed.2d 1072 (1979), the Court of Appeals reversed the district court’s dismissal of a lawsuit challenging the lack of child care facilities in the San Mateo Community College District. The court found that “plaintiffs have alleged a course of conduct by defendants susceptible of an inference of intentional discrimination.” It then cited the following pleadings:

While the De La Cruz opinion made clear that it was a close case, and the only issue was whether to dismiss the complaint, it is instructive that the allegations which permitted “an inference of intentional discrimination” amounted only to evidence that defendants knowingly failed to revise their policies and spend money on child care when the existing policy had the “direct effect” of excluding or hindering substantial numbers of women from educational opportunities.

If the Ninth Circuit is willing to follow the courts that have developed a presumption of intent, and the Supreme Court ultimately acquiesces, then the case against defendants is a very simple one. In Lora v. Board of Education, 456 F.Supp. 1211 (E.D. N.Y.1978), Judge Weinstein utilized this test to invalidate New York City’s system of special public and private schools for the “educationally handicapped.”

Lora v. Board of Education, 456 F.Supp. at 1285 . The court noted further that “[a] screening process that passes Whites through and leaves minorities behind is prima facie unconstitutional.” 456 F.Supp. at 1284 . The state failed to meet its burden since it could not “show that less segregative referral procedures are unavailable,” or that it had taken “ ‘affirmative steps . to insure the availability of all its programs to blacks who may choose to participate.’ ” 456 F.Supp. at 1284 , quoting Brumfield v. Dodd, 425 F.Supp. 528, 532 (E.D.La.1976).

The analytical approach taken in Lora would compel the same legal result here. Indeed, Larry P. appears to be an easier case to resolve on these grounds. As explained before, the screening process for E.M.R. classes to a great extent passes white children through while leaving minorities trapped. This result was natural, probable, and foreseeable, and the actions of the state, especially the adoption of the particular I.Q. testing requirement, predict ably reinforced the result. Plaintiffs, finally, have affirmatively shown that less segregative referral procedures are available.

Uncertainty in the law, however, compels us to develop the analysis further. The Ninth Circuit has not approved of this presumption approach, and perhaps more significantly, while the other circuit decisions are still in effect, three very recent Supreme Court cases suggest that further refinement in any presumption approach may be necessary.

In Personnel Administrator of Massachusetts v. Feeney, supra, a 7-2 majority of the Supreme Court reversed a three-judge court’s holding, 451 F.Supp. 143 (D.Mass. 1978), that the Massachusetts veterans’ preference law for civil service employment was unconstitutional. The veterans’ preference resulted in a very substantial overrepresentation of men in the Massachusetts civil service. The Supreme Court, however, held that overrepresentation was not discrimination in this instance. In so holding, the Court explicitly rejected the objective, tort standard of intent.

442 U.S. at 279, 99 S.Ct. at 2296 (footnotes omitted). This statement underscores the importance of a finding of subjective intent, but it should not be construed to invalidate completely the presumption analysis.

Feeney counsels us that “inevitability or foreseeability of consequences” can permit “a strong inference that the adverse effects were desired,” but “an inference is a working tool, not a synonym for proof.” 442 U.S. at 279, 99 S.Ct. at 2296 n. 25.

442 U.S. at 279 n. 25, 99 S.Ct. at 2296 n. 25. As the quotation suggests, the rejection of the inference or presumption in Feeney rested on two central findings: first, as the trial court held, all the evidence showed only a desire to favor veterans, not a desire to “accomplish the collateral goal of keeping women in a stereotypic and predefined place,” 442 U.S. at 279, 99 S.Ct. at 2296; and second, the adverse impact upon women was an unavoidable consequence of the desire to favor veterans. The presumption analysis is not therefore necessarily precluded, but it cannot be utilized mechanistically to reject probative evidence and find a subjective intent grounded solely on discriminatory impact.

This conclusion is reinforced by the two equal protection school desegregation decisions handed down against school boards in the summer of 1979. Columbus Board of Education v. Penick, supra; Dayton Board of Education v. Brinkman, supra. Both cases affirmed appellate court decisions that school boards had operated dual school systems at the time of Brown v. Board of Education, 347 U.S. 483 , 74 S.Ct. 686 , 98 L.Ed. 873 (1954), and that the discriminatory effects of those dual systems had not been eliminated or had even been exacerbated since that time by subsequent board decisions. The lower court findings, and the Supreme Court rulings, depended on inferences drawn from the existence and maintenance of segregated schools, but the Supreme Court majority expressly refused to endorse an effects presumption such as that presently employed in the majority of the circuits. Noting the lower court application of the test articulated in Oliver v. Michigan State Board of Education, 508 F.2d 172, 182 (6th Cir. 1974), cert. denied, 421 U.S. 963 , 95 S.Ct. 1950 , 44 L.Ed.2d 449 (1975), the Court advised:

443 U.S. at 536 n. 9, 99 S.Ct. at 2978 n. 9. See also, Columbus Board of Education v. Penick, 443 U.S. at 465, 99 S.Ct. at 2950 . There is thus no “general” presumption that can “routinely” be applied to affect the burden of proof or burden of producing evidence. Nevertheless, the use of presumptions — better termed inferences in this context — may be required in particular factual settings, and effects of challenged conduct may be particularly telling when evincing a failure to eradicate the consequences of prior purposeful discrimination.

Under the present state of the law, therefore, a factual showing of discriminatory impact provides no simple shortcut to the ultimate fact — discriminatory intent. As Justice Powell indicated in his majority opinion in Village of Arlington Heights, the situation “demands a sensitiye inquiry into such circumstantial and direct evidence of intent as may be available.” 429 U.S. at 266 , 97 S.Ct. at 564. As that opinion also indicates, there are a number of evidentiary sources available for ascertaining intent, and those will necessarily guide our inquiry below. Before proceeding to that discussion, however, it will be useful to clarify what is meant by discriminatory intent in this case.

b. Applying the law. Discriminatory intent does not necessarily mean an intent to harm black children. See, e. g., United States v. Texas Education Agency, 579 F.2d 710 (5th Cir. 1978); Stark v. Shell Oil Co., 450 F.2d 994 (5th Cir. 1971). Since Brown v. Board of Education, 347 U.S. 483 , 74 S.Ct. 686 , 98 L.Ed. 873 (1954), an intent to segregate minority children in separate schools has sufficed to prove a prima facie violation of the Fourteenth Amendment. An intent to segregate into special, all day classes is functionally indistinguishable. Minority children — black children under the present posture of this case — are isolated for virtually the entire school day from mainstream children, and E.M.R. classes provide an inferior education for children assigned there on the basis of biased methods of placement. Moreover, even more than segregated schools, disproportionate enrollment of minorities in E.M.R. classes stigmatizes those in the classes and serves inevitably to perpetuate invidious stereotypes based on the superiority or inferiority of “racial stocks.” Even a well meaning intention to adopt and retain procedures that place disproportionate numbers of minorities in these special, isolated classes will therefore serve to prove that “a discriminatory purpose has been a motivating factor in the decision.” Village of Arlington Heights, 429 U.S. at 265-66 , 97 S.Ct. at 563 . Once such a purpose is shown, even if it was not the sole or predominant purpose behind the actions in question, then defendants must show that the actions would have resulted anyway in the absence of the prohibited motive. If such a showing is not made, then “judicial deference is no longer justified.” Id.

That deference cannot be justified in this case. The statement of facts at the beginning of this opinion provides a picture of what happened, and the legal conclusion from these facts follows even more clearly if we turn specifically to the factors suggested by Justice Powell for ascertaining intent. They include, in addition to (1) the impact of the official actions, (2) the historical background of the decisions, (3) the specific events leading up to the challenged decisions, (4) departures from normal procedural sequence, (5) substantive departures that suggest the outcome might have been different, and (6) direct statements by SDE officials. The decision by the State Department of Education in 1969 to compel the use of standardized I.Q. tests clearly fails this test and reveals the impermissible intent to discriminate.

First, it is obvious that the decision to require the administration of one of the approved I.Q. tests prior to E.M.R. placement had profound discriminatory effects. It doomed large numbers of black children to E.M.R. status, racially imbalanced classes, an inferior and “dead-end” education, and the stigma that inevitably comes from the use of the label “retarded.” That impact was not only foreseeable but foreseen, and appropriate inferences must be drawn. Moreover, it is established that less discriminatory alternatives could have been utilized.

The historical background of the I.Q. decision provides a second essential source of evidence. As noted before in some detail, the history of I.Q. testing and special classes for E.M.R. in California is a tainted one. Notions of racial inferiority permeated the classes established after World War I, and I.Q. scores were used generally to justify the isolation of so-called inferior ethnic or racial groups thought to contain large numbers of “feebleminded.” The differences between black and white childrens’ I.Q. scores were known, but that simply reinforced the stereotypes of racial inferiority. These assumptions, explicitly taken for granted at the time that the first E.M.R. classes and the early I.Q. tests were developed, have had a serious impact on California’s black children, and obviously they have never been overcome. Admittedly, this history extends back almost 50 years, but “we reject any suggestion that remoteness in time has any relevance to the issue of intent.” Keyes v. School District No. 1, 413 U.S. at 211 , 93 S.Ct. at 2698 . Racially neutral tests or other neutral criteria have not been developed, and in fact no effort has been made by defendants to develop them. On this basis alone, we could find, under the two-stage test of Dayton Board of Education v. Brinkman, 443 U.S. at 536 n. 9, 99 S.Ct. at 2978 n. 9, that there has been a constitutionally prohibited “failure to fulfill the duty to eradicate the consequences of prior purposefully discriminatory conduct.” And there is still considerably more upon which to base the finding of intent.

The circumstances surrounding the decision by the State Department of Education to insure the administration of a prescribed I.Q. test prior to placement into an E.M.R. class are especially instructive. As de tailed earlier, the process of adopting the requirement in 1969 was riddled with procedural and substantive irregularities. In the midst of substantial public controversy in 1969 about the overenrollment of minorities in E.M.R. classes and the fairness of the standardized intelligence tests — indeed, after a legislative resolution calling for study of the problem — the State Board of Education, dependent on the State Department of Education and with no independent staff of its own, adopted the resolution proposed by the SDE requiring the use of approved I.Q. tests for E.M.R. placement. The state has offered no explanation whatsoever for this decision, or why it suddenly became so important.

The actual compiling of the list of approved I.Q. tests was even more peculiar. The list was put together in two months, July and August 1969. In response to the SDE’s urgent request for the names of suitable tests, a number of letters objecting to the procedure being followed were received. They were ignored, and the only “outside” sources consulted were persons associated with the testing companies. Ultimately the question of bias never was considered, and choices were made solely on the basis of frequency of use at that time. No hearings on the merits of the choices were held, no investigation was undertaken, and no representatives of minorities or independent experts were consulted.- The SDE officials knew perfectly well about the questions raised by the I.Q. tests; their attention had even been called to the problem by a legislative resolution. Yet these vital concerns were overlooked by the SDE and never brought to the attention of the Board of Education. The decision to require tests that would inevitably perpetuate or even exacerbate the very serious overenrollment of minorities in E.M.R. classes was thus made in a very suspicious and inadequate manner. The inference of discriminatory intent — an intent to subject black children disproportionately to the stigma and educational limitations of the E.M.R. classes — -is inescapable, and once again defendants have proffered no explanation to rebut this inference or even permit another inference.

Defendants naturally object to this method of analysis, contending essentially that the selected I.Q. tests were generally known and had long been accepted “all over the world.” Defendants’ Proposed Findings of Fact, No. 128. The tests’ general reputation, however, cannot negate the inference of discriminatory intent. This justification fails to explain the timing of the decision to require the selected tests, and it ignores the fact that the tests had not been validated by the state for black children. Defendants’ complete failure to ascertain or attempt to ascertain the validity of the tests for minority children cannot be ignored. Rather defendants’ actions resulting in the adoption of the I.Q. requirement and the short list of accepted I.Q. tests can only be explained as the product of the impermissible and scientifically dubious assumption that black children as a group are inherently less capable of academic achievement than white children.

Key officials of the State Department of Education, moreover, actually corroborated this explanation. They testified that they believed the overenrollment of black and Chicano children in the E.M.R. classes accurately reflected the incidence of mental retardation among those children. E.g., Hanson 3798-99 . That belief was counter to state legislative policy and to the statements of the person in charge of the SDE, the Superintendent of Public Instruction, but it evidently was the basis for these crucial state decisions about E.M.R. placement. The state itself is accountable for those discriminatory decisions, which legislative declarations and other procedural reforms failed to overcome. As the Supreme Court recently reiterated in one of the Ohio school eases, “Even actions of state agents that may be illegal under state law are attributable to the State.” Columbus Board of Education v. Penick, 443 U.S. at 485 n. 5, 99 S.Ct. at 2946 n. 5.

Also attributable to the state are acts and omissions subsequent to the I.Q. decision, which provide further confirmation of the discriminatory intent crucial to this equal protection analysis. These actions were not discussed before, but they merit attention because they continued the pattern of SDE complacency in effective opposition to legislative concern about biased testing and disproportionate enrollments. This legislative concern, described earlier, during the 1969-71 period resulted in a number of statutory changes. In addition to moving the I.Q. requirement from the Administrative Code to the Education Code, the legislature lowered the I.Q. score cut-off and signaled the existence of the problems of minorities with E.M.R. classes. Disproportionate enrollments were explicitly condemned, problems of bias with intelligence tests were noted, and reporting requirements about minority enrollment and due process guarantees were instituted. While the legislature failed to offer an enforceable program to overcome these problems, it did clearly signal its concern.

The SDE treatment of these reforms has not been inspiring. Throughout the period covered by this lawsuit, the SDE set policy as if the legislative findings about testing and disproportionate enrollments meant nothing, and the Board in turn has compelled no investigation nor indeed taken any position about these problems. The SDE, for example, has generally complied with the clear legal requirement that it submit a report to the Board on school districts with significant variances in racial and ethnic composition in E.M.R. classes, but it has done nothing at all to inquire into the reasons for these variances. E. g., Law 6097; Simmons 3485-86. While several of the reports suggested that, in response to litigation, minority enrollments were declining absolutely, and perhaps even relative to white enrollment, the reports gave no support for this possible trend. Rather, the problem perceived and reported by the SDE was that E.M.R. classes might not have been serving the estimated 2 percent of children who were presumed to be retarded. Thus, the only normative comment about minority disproportions in the 1972-73 report was that “no ethnic group is heavily over-represented with respect to the expected incidence of mental retardation.” P. 3. Again, the SDE was perfectly willing to assume greater minority retardation.

In addition, while the SDE has dutifully collected the reports legally required from school districts about local disproportionate enrollments, and it has required local boards to comply with the law and explain the variances, these reports have been treated in a manner calculated to maximize tolerance of minority disproportions. First, the SDE interpreted the statutory requirement of a report when there is a variance of “15 percent or more from the percentage of such children in the district as a whole” to mean that a district was “entitled” to fifteen free percentage points added to the minority representation in the population. Parity was therefore not even sought, and the literal reading of the statute meant that a district with 2 percent black children in the population could have 17 percent black enrollment in the E.M.R. classes. A district with 85 percent black children would in turn be allowed 100 percent of the E.M.R. enrollment. Beyond these remarkably tolerant interpretations, Berk 9767-75, the SDE never evaluated the adequacy of the local explanations for the disproportions and never sought to utilize the data or explanations to obtain a sounder understanding of the problem, much less to work toward its solution. Superintendent of Public Instruction Wilson Riles testified that he did not know of this practice by his subordinates in the SDE, Riles 2737-40, but there is no question that the legislatively mandated reports were simply filed away. Hanson 3605-09 . The SDE acted merely as a “custodian of records.” Hanson 3618 ; Brinegar 4707-20; Law 6077-6100. Nothing was done to further the goal implicit in the reporting requirement.

Finally, not only was essentially nothing done to monitor and follow up on the districts with disproportionate enrollments, but also the SDE failed to monitor the implementation of the procedural protections imposed by the legislature in 1970 and 1971. The only effort to investigate was a 1975-76 SDE survey of the 26 districts with the highest disproportion of black enrollment in E.M.R. classes. The Uslan Report based on this investigation found that the files of students placed in E.M.R. classes contained I.Q. scores — although the testing moratorium precluded new testing — but little or no evidence of adaptive behavior information despite the clear legal importance of that information. The Uslan Report also revealed inadequacies in developmental information. The considerable data there about E.M.R. placements, however, were again ignored. The SDE did nothing but file this report, even where it revealed severe deficiencies in compliance with statutory norms, and the particular results were not even conveyed to the Board of Education. Brinegar 4784.

Admittedly a number of reasons can explain a lack of attention to a problem such as that of disproportionate enrollment, but the clear pattern cannot be ignored. The facts permit but one inference, and the state has not offered evidence that permits any other inference. Despite the admitted problems with the I.Q. tests, and despite disproportionate enrollments which had even been condemned by the legislature, the SDE’s actions revealed a complacent acceptance of those disproportions, and that complacency was evidently built on easy but unsubstantiated assumptions about the incidence of retardation or at least low intelligence among black children. Coupled with the affirmative decision to adopt the requirement of particular I.Q. tests in 1969, that complacent acceptance must be seen as a desire to perpetuate the segregation of minorities in inferior, dead end, and stigmatizing classes for the retarded.

This factual situation, therefore, is easily distinguishable from that in the two leading Supreme Court decisions allowing the sur vival of testing requirements challenged on equal protection grounds. In Washington v. Davis, 426 U.S. 229 , 96 S.Ct. 2040 , 48 L.Ed.2d 597 (1976), the Court had only the continued use of “Test 21” from which to infer any discriminatory intent. Further, the Court found that even if

426 U.S. at 246 , 96 S.Ct. at 2051 . Even aside from the obvious and important distinction between E.M.R. placement and employment, the other factors found by the District Court in Washington v. Davis are clearly absent here. Racial compositions have not been changing in E.M.R. classes, at least prior to the moratorium on I.Q. testing, the tests have not been validated, and the affirmative effort to end disproportions in the District of Columbia police force has no parallel for E.M.R. classes in California.

In United States v. South Carolina, 434 U.S. 1026 , 98 S.Ct. 756 , 54 L.Ed.2d 775 (1977), the Supreme Court affirmed per curiam the decision of a three-judge court, United States v. South Carolina, 445 F.Supp. 1094 (D.S.C.1977), that the use of the National Teachers’ Examination in South Carolina did not violate equal protection. Here, too, even aside from differences between the employment and educational settings, there are profound dissimilarities between facts negating any inference of intent in that ease and the facts proving discriminatory intent in Larry P. First, the racial composition of the South Carolina teacher force, despite the disproportionate impact of the test used for screening teacher candidates, “closely paralleled the racial composition of the state’s population.” Id. at 1101 . Second, the particular acts alleged to reveal discriminatory intent in South Carolina proved a lack of any purposeful discrimination. A 1945 decision to institute a new certification system was based on a two-year study resulting in a four-volume report. A 1956 decision to adopt an “absolute score system” resulted in the exclusion of very few blacks, and “significantly, this system was maintained for over 11 years.” Id. at 1103 . In 1969, the year of the hasty and unsystematic adoption of an I.Q. requirement in California, the South Carolina State Board raised the minimum test scores, but in doing so it followed the recommendations of a high level, racially integrated, special committee formed to study the certification process. And in 1976, the South Carolina Board set up new minimum scores for the same test, but this action was taken only after authorizing and obtaining the results of an extensive validity study by the Educational Testing Service. The study produced a two-volume, 300-page report, and the Board adopted the principal recommendation that the minimum scores be set by teaching fields. No such validity study has been undertaken in California. And finally, unlike the plaintiffs in the present case, plaintiffs in United States v. South Carolina could come up “with no other reasonably appropriate criteria upon which certification may be properly based.” Id. at 1104.

Neither of these “testing” cases, therefore, provides support for defendants’ position. Plaintiffs have met their burden of proving discriminatory intent. Since there is no evidence that the state would have acted the same “had the impermissible purpose not been considered,” Village of Arlington Heights v. Metropolitan Housing Corp., 429 U.S. 270 n.21, 97 S.Ct. 566 n.21, we must now apply the “strict scrutiny test.”

Under the strict scrutiny or compelling interest test, defendants cannot prevail unless they are able to justify the racial classification by showing that it was necessary to the accomplishment of a compelling state interest. See, e. g., McLaughlin v. Florida, 379 U.S. 184 ,. 192-93, 85 S.Ct. 283, 13 L.Ed.2d 222 (1964). Defendants can establish no compelling state interest in the use of the I.Q. tests nor in the maintenance of E.M.R. classes with overwhelming disproportions of black enrollment. With respect to the tests, defendants assert the need to avoid misclassification, their interest in fiscal control, and the lack of alternatives for E.M.R. placement. These post hoe justifications merit little analysis in light of earlier discussions. Fiscal control merely means that I.Q. scores are predictable for a large group, therefore enabling fiscal planning, but that reason does not explain why biased tests or other criteria should be used. With respect to misclassification, the tests necessarily compound the problem for black children, who on the average score one standard deviation below white children, and less discriminatory alternatives are available both for fiscal control and for classification. There is, furthermore, no compelling justification — indeed, no reasonable justification — for keeping large disproportions of black children in classes that presume a fixed intellectual capability and prescribe a limited, dead-end curriculum. Defendants have shown the lack of a compelling interest by voluntarily agreeing to eliminate such disparities for Hispanic children. Defendants’ placement processes, in short, tend to defeat rather than promote the general interest in a quality education for all, not to mention the legislatively proclaimed interest in avoiding “disproportionate enrollment of any socioeconomic, minority, or ethnic group pupils in classes for the mentally retarded.” The state in this litigation is trying to cling to a practice that “spites its own articulated goals.” Stanley v. Illinois, 405 U.S. 645, 653 , 92 S.Ct. 1208, 1213 , 31 L.Ed.2d 551 (1972). Plaintiffs have proven their case under the Fourteenth Amendment.

2. An “intermediate” standard of review. Plaintiffs also prevail under an intermediate standard of review. While recent Supreme Court case law does not provide much guidance on the applicability of an equal protection standard of review between strict scrutiny and the very permissive “rational basis” test, decisional law in the Ninth Circuit provides guidance for this case. In Berkelman v. San Francisco Unified School District, 501 F.2d 1264 (9th Cir. 1974), the Circuit Court of Appeals upheld the admission standards for Lowell High School against an equal protection attack. The analysis, however, did not stop with a finding of no intent to discriminate or segregate. The court went on to apply the following test:

501 F.2d at 1267 . While Berkelman was decided before Washington v. Davis, which disapproved of a case relied upon as authority for this intermediate test, the Berkelman approach has not lost its validity. Very recently, in fact, its viability was reaffirmed by the Ninth Circuit in De La Cruz v. Tormey, 582 F.2d 45 , 59 n. 10 (9th Cir. 1978). We shall therefore apply that approach to the facts of Larry P.

In Berkelman , the Court of Appeals considered the obvious advantages of an “academic” high school such as Lowell, and it found that the admissions criterion — junior high school grade point average in college preparatory subjects — substantially furthered that purpose of operating an academic high school. The court thus concluded “that the district’s legitimate interest . outweighs any harm imagined or suffered by students whose achievement had not qualified them for admission to that school.” 501 F.2d at 1268 . The court further suggested, however, that it was “troubled by the underrepresentation of some racial and ethnic groups at Lowell,” and it carefully noted that the outcome of its balancing test depended on its findings that students not admitted to Lowell are not relegated to a “tracking” system, are not “isolated” on the basis of past achievement, and are not subject to the “psychic injury” of an inferior track; “students not admitted to Lowell are neither denied a quality education nor relegated to an inadequate school.” 501 F.2d at 1268 . All these factors absent in Berkelman are present in Larry P.

Unlike in Berkelman , there is little evidence here that the I.Q. tests or the perpetuation of “troubling” disproportionate enrollments substantially furthers the purpose of providing the best possible education for California public school students. The evidence already discussed points in the opposite direction. When balanced against the stigma, isolation, and educational inferiority of the E.M.R. classes, there is no question about the result. Even if discriminatory intent had not been shown, plaintiffs under the authority of Berkelman have still proved a violation of the federal equal protection clause.

D. State Claims

Finally, plaintiffs allege several state claims over which this court has pendent jurisdiction, and we agree that plaintiffs have proved a violation of California’s constitutional provision guaranteeing equal protection of the laws. Article I, § 7. Under prevailing California constitutional law, disproportionate enrollments in. E.M.R. classes, and I.Q. tests with a discriminatory impact, need not be scrutinized under a specific intent test. Under the California Constitution, as interpreted by the California Supreme Court, school districts have a constitutional duty to eliminate de facto as well as de jure segregation in the schools. As stated clearly by that court in Crawford v. Board of Education, 17 Cal.3d 280, 292 , 130 Cal.Rptr. 724, 731 , 551 P.2d 28, 35 (1976),

This challenge to school segregation applies with equal force to segregation through disproportionate enrollments in E.M.R. classes, and that conclusion is substantiated by the California Supreme Court’s reasoning in Crawford’s predecessor, San Francisco Unified School District v. Johnson, 3 Cal.3d 937, 950 , 92 Cal.Rptr. 309 , 479 P.2d 669 (1971). The court quoted the following language from a federal study to justify its decision to erase the legal effect in California of the distinction between de facto and de jure segregation:

3 Cal.3d at 949 , 92 Cal.Rptr. at 315-316 , 479 P.2d at 675 -676 (quoting the 1967 Report of the United States Commission on Civil Rights: “Racial Isolation in the Public Schools”). This description of segregated schools is even more accurate when applied to E.M.R. classes heavily loaded with minority children. The law of California requires the state to take “reasonable feasible steps to alleviate school segregation, regardless of its cause.” Since such efforts have not been made, plaintiffs have also prevailed under California’s constitutional equal protection law.

E. Adequacy of Representation

Defendants devoted considerable time in the closing argument to the eases of the named plaintiffs serving as the representatives of the class. Defendants’ contention was that the named plaintiffs were in fact retarded. Even though the named plaintiffs were assigned to E.M.R. classes some time ago, defendants placed them on the witness stand, presumably for the court to make its own determination whether the earlier E.M.R. assignments were appropriate. The evident theory behind defendants’ tactics was that if plaintiffs could not prevail as individuals, they could not represent the class, and there could be no classwide relief. We reject this legal theory, just as it was rejected before. See, e. g., Order of September 21, 1977.

Plaintiffs were adequate representatives of the class of black children who have been or in the future will be wrongly placed and maintained in special classes for the educable mentally retarded. The individual plaintiffs were assigned to E.M.R. classes greatly overrepresented by black children, and their assignments were based on racially and culturally biased criteria, notably the I.Q. tests. Plaintiffs obviously were no longer in such classes at the time of the trial, but they were either in E.M.R. classes or subject to reassignment at the time the class was certified and the preliminary injunction granted. Moreover, their particular situations would not affect the outcome of this class action. The case was vigorously contested, with more than adequate representation of the interests of the unnamed class members. The issues at stake were of vital importance to unnamed class members. No doubt many refrained from seeking legal remedies in reliance on the preliminary injunctions and in the hope and expectation that plaintiffs and their counsel in this well-publicized class action would marshal the tremendous expertise and resources necessary to expose the biases of the tests and the plight of those assigned unfairly to E.M.R. classes. If the issues vital to that group of unnamed class members are not addressed, the class members will be forced to start again, and in the meantime there will be further misclassifications and further irreparable harm.

It is now well-established in the case law that the interests of unnamed class members do not depend on the particular validity or even mootness of the claims of the individual class representatives. See generally 3B Moore’s Federal Practice, ¶¶ 2304(2), 2307; cf. Kuahulu v. Employers Insurance of Wausau, 557 F.2d 1334 (9th Cir. 1977). In Franks v. Bowman Transportation Co., 424 U.S. 747 , 96 S.Ct. 1251 , 47 L.Ed.2d 444 (1976), for example, the single named class representative could not himself claim retroactive seniority because he had been validly discharged by the defendant for cause. Nevertheless, the Supreme Court allowed seniority relief to the class. Similarly, in Zablocki v. Redhail, 434 U.S. 374 , 382 n. 9, 98 S.Ct. 673 , 679 n. 9, 54 L.Ed.2d 618 (1978), the Court made clear that the class of persons needing court permission to marry could prevail even if the sole class representative had married: “regardless of the current status of appellee’s individual claim, the dispute over the statute’s constitutionality remains live with respect to the members of the class appellee represents, and the Illinois marriage took place well after the class was certified.” Consistent with prevailing authority, class relief does not require this court to undertake the task of evaluating in retrospect the cases of named plaintiffs who themselves seek no individual relief. Plaintiffs were adequate representatives of the class, and the class is entitled to prevail.

Finally, while we see no reason to enter into the cases of the named plaintiffs, we can observe that the only relevant evidence on their cases indicated that they were not retarded.

F. Legal Conclusions

On the basis of the preceding discussion, plaintiffs are entitled to prevail on the merits and obtain a declaration to that effect. First, in violation of federal statutory law, including Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d; section 504 of the Rehabilitation Act of 1973, 29 U.S.C. § 794 ; and the Education for All Handicapped Children Act of 1975, 20 U.S.C. § 1401 et seq., defendants have: (a) utilized I.Q. tests that are racially and culturally biased, have a discriminatory impact on black children and have not been validated for the placement of black children into E.M.R. classes; and (b) generally allowed the placement of black children in grossly disproportionate numbers into E.M.R. classes by procedures that have not been validated.

Second, plaintiffs have shown that defendants’ tolerance of the disproportionate enrollment of black children into the E.M.R. classes and their use of placement mechanisms, including the I.Q. tests, which perpetuate that condition, violate state and federal guarantees of a right to equal protection under the laws. These findings do not depend on a finding of intentional discrimination.

Third, defendants’ intentional fostering of the vast overenrollment of black children in E.M.R. classes, evinced throughout the history of E.M.R. classes and I.Q. testing in California and revealed more recently by such actions as the hasty and unexplained decision to prescribe a list of “approved” I.Q. tests in 1969, provides a further ground for proving a violation of plaintiffs’ right to equal protection of the laws. Many black children have been isolated, stigmatized, and provided inadequate educations on the basis of unwarranted and impermissible assumptions.

Finally, the private plaintiffs have been adequate representatives of the class of black children who have been or in the future will be wrongly placed and maintained in special classes for the educable mentally retarded. In view of the obvious and well-documented irreparable harm that comes from misplacement into E.M.R. classes, it is clear that injunctive relief is appropriate for the class represented by plaintiffs.

Before detailing the remedy, however, a few qualifying observations must be made. First, we recognize that California’s educational system is undergoing a transformation as reflected in the provisions of the Master Plan and in the movement toward compliance with federal statutory law. Progress is no doubt being made in some of the areas challenged by this lawsuit. Such progress must be considered in the remedy, but progress in curing illegality does not prevent court action — it “does not oust the court from its jurisdiction to declare the constitutional wrong.” Hobson v. Hanson, 269 F.Supp. 401, 498 (D.D.C.1967).

A second essential caveat is that there should be no illusions about our capacity as a court to require educational systems to transcend societal inequalities and provide black children generally with the kinds of skills necessary for educational and social advancement in our country. That would take a major social commitment, not a court order. We are perfectly aware of the complexity and interrelationship of educational problems, and indeed of the danger that an attack on E.M.R. classes, in the absence of effective remedial education, could even hurt those individuals who, despite the aims and approach of E.M.R. classes in general, actually benefited from what appears in retrospect to have' been misplacement. Nevertheless, that consideration cannot allow us to sanction a labeling process that unjustifiably blames educational failure on the ostensible mental retardation of black children and dooms disproportionate numbers of black children to a program de signed to keep their performance below normal.

Third, our decision in this case should not be construed as a final judgment on the scientific validity of intelligence tests. What has been found in that respect is that the state has acted contrary to the law by requiring the use of tests that are racially and culturally discriminatory and have not been validated for labeling black E.M.R. children and placing them into special classes for the mentally retarded, and by requiring the use of the I.Q. tests in an I.Q.-centered process which was being used intentionally to perpetuate and legitimate a vast overrepresentation of black children in those special classes. Despite a legislative policy opposed to disproportionate enrollments, defendants have deliberately chosen a consistent policy in favor of black overenrollment, and the I.Q. tests have been a key component of that policy. Our constitutions and implementing laws forbid such a racially discriminatory policy unless necessary to accomplish an overriding state interest. The relationship of the tests to any legitimate interest involved here, however, has not been established through the validation of the tests, and in any event their use for E.M.R. placement has not been shown to be necessary. Whatever the general scientific merits of the tests, therefore, defendants have failed to show a valid, legal justification for their use for black E.M.R. placement. Their use will have to be enjoined as part of the remedy discussed below.

VI. REMEDY

A. Introduction

The injunctive relief must address two specific problems — testing abuses, and the disproportionate enrollment of black children in special E.M.R. classes. While the state is evidently moving toward reform of the procedures for E.M.R. placement and strengthening due process protections for children and their parents, specific injunctive relief must be ordered for the class represented by plaintiffs. Regardless of whether the Master Plan is fully implemented and funded, the state has indicated that its intention to maintain special classes for E.M.R. children — even if such children are not classified by precisely that term. They may be included as part of the “learning handicapped” category created by the Master Plan. Unfortunately, at present, data are not available on the minority percentages in the various categories and statuses created under the Master Plan. While that data will have to be collected, a remedy can nevertheless be directed to the violations uncovered by this litigation.

B. Permanent Injunctive Relief

1. Intelligence tests. Defendants are enjoined from utilizing, permitting the use of, or approving the use of any standardized intelligence tests, including those now approved pursuant to Cal.Admin.Code § 3401, for the identification of black E.M.R. children or their placement into E.M.R. classes, without securing prior approval by this court. For the purposes of this order, “E.M.R.” includes the category presently defined by Cal.Educ.Code § 56500, or a substantially equivalent category, and the term, “E.M.R. classes,” includes those special classes presently provided for by Cal.Educ.Code § 56511 or other special classes serving substantially the same functions.

In order to obtain court approval of the use of any standardized intelligence tests, the State Board of Education must take the following steps:

(1) make a written request to the court that the standardized intelligence test or tests be approved. The request must be authorized by a majority of the Board, and it must state whether the Board has determined that the test or tests

(a) are not racially or culturally discriminatory,

(b) will be administered in a manner which is nondiscriminatory in its impact on black children,

(c) have been validated for the determination of E.M.R. status or placement in E.M.R. classes;

(2) accompany their application with supporting evidence produced under oath, which must include:

(a) statistics on the average or mean scores of blacks and whites on the test or tests,

(b) statistics or other data, reported separately for blacks and whites, which form the basis for the Board’s determination that the tests are validated for the purpose(s) for which they are to be used,

(c) a certification that the State Board has held open public hearings on the proposed test or tests before reaching its determination.

2. Disproportionate placement. Defendants are hereby ordered to monitor and eliminate disproportionate placement of black children in California’s E.M.R. classes. Specifically, defendants shall:

(a) obtain annually from each school district a report specifying, separately, (i) the number of children enrolled, and (ii) the number of children newly identified or enrolled in each district on the date selected for the February Child Count, by race/ethnicity as the latter terms are defined on Office of Civil Rights forms 101 and 102, in each of the following categories and subcategories:

(b) prepare a report that aggregates and reports separately for (i) statewide, (ii) all Master Plan school districts, (iii) each Master Plan Responsible Local Agency and (iv) non-Master Plan school districts, the information called for in each of the subsections of paragraph (a) above;

(c) file by June 1 of each year the reports required by paragraphs (a) and (b) of this section with the court, and serve copies on counsel for plaintiffs; on attorneys for the Justice Department; and on the Office of Civil Rights of the Department of Health, Education, and Welfare;

(d) direct each school district that reported for the 1977-78 school year a rate of black E.M.R. pupil enrollment one standard deviation above the district rate of white E.M.R. pupil enrollment, to prepare and adopt a plan to correct the imbalance in rates of enrollment; the plan shall specify the measures that will be adopted for reducing the disparity over a period of three years from the date of this order;

(e) when a disparity of one standard deviation in the rate of E.M.R. class placements of blacks above the district rate for whites continues or occurs after the three-year period, the state defendants must bring the imbalanced district to the attention of this court, and the court after hearing may order such further relief as it deems appropriate.

3. Reevaluation and other class relief. To remedy the harm to black children who have been misidentified as E.M.R. pupils and to prevent these discriminatory practices from recurring in California with respect to a similarly situated class of youngsters in the future, the defendants shall direct each school district to reevaluate every black child currently identified as an E.M.R. pupil, without including in the psychological evaluation a standardized intelligence or ability test that has not been approved by the court in accordance with paragraph VI.B.l, supra. Such reevaluation shall include information derived from:

(1) diagnostic tests designed to reveal specific learning needs and to prescribe specific pedagogical approaches;

(2) adaptive behavior observation;

(3) the child’s developmental and health histories.

Following the reevaluation, an “individual educational plan” that specifies the type of supplemental assistance needed to allow the child to return to a regular classroom shall be prepared for each child reevaluated and found to have been misdiagnosed. Such assistance shall then be provided.

C. The Necessity of Limiting the Scope of the Remedy

In view of the complexity of this area and the need to preserve flexibility wherever possible, we have sought to tailor the remedy as carefully as possible to the specific legal findings that have been made. Thus, while we have followed many of the suggestions in plaintiffs’ proposed remedy memorandum, we have not opted for as broad a remedy as was there advocated. Two examples merit particular attention. First, plaintiffs proposed that the court correct any disparities in “E.M.R. pupil identification,” even if the identification does not result in placement in a special class. The court will obtain data on those identifications, and it may be that at some future time further court action will be warranted, but at present it would not be appropriate to intervene with a relatively rigid requirement. The I.Q. tests that necessarily lead to such disparities have been enjoined, and the remaining problems resulted largely from the particular nature of the E.M.R. classes. On the facts now before the court, it would not be useful to go beyond testing and E.M.R. classes to regulate conditions of “mainstreamed” E.M.R. students about whom we know very little.

Second, plaintiffs sought a court order mandating supplemental assistance to black disadvantaged children who are now in regular classes. While we recognize that California’s educational system is failing to educate adequately a vast number of minority and disadvantaged students, that matter is not before the court. It is not the role of the court to reach out to order what would amount to a massive expenditure of funds for supplemental assistance. As we have noted several times in this opinion, educational reform of that scope will depend on action by other branches of government.

The remedy in this case, therefore, has been confined to the facts and issues found at the trial. The injunctive relief will place some burdens on defendants, but those burdens should not be difficult to meet. Defendants have already survived for several years without the use of I.Q. tests for E.M.R. placement, and the remedy to correct the disproportionate enrollment of black children is very similar to that voluntarily accepted for Hispanic children as a result of the Diana settlement in the early 1970’s. We are thus confident that defendants can comply with the remedy, conform their E.M.R. system to applicable law, and get on with the task of providing a quality education for all of California’s children.

VII. CONCLUSION

This court has been forced to intervene in a complex and controversial area. Despite the prodding of a preliminary injunction, California state defendants have not been willing to remedy the problems of E.M.R. classification and placement addressed by this litigation. Nevertheless, we should recognize that other changes are taking place that may ultimately soften the impact of this court’s order.

It may be that E.M.R. classes and the philosophy underlying them are educational anachronisms. They focus on a label — retardation — derived for the most part from arbitrary cut-off scores on standardized I.Q. tests, and that label is used to justify academic isolation in special dead-end classes.

A contrary educational philosophy has been gaining adherents. The state’s Master Plan, still not fully implemented, embodies many of the concerns of this philosophy, ánd it underlies recent federal statutory reforms concerning the education of the handicapped. This philosophy aims not at labeling a child and confining the “retarded” in special classes. Rather, it aims to diagnose learning problems, to take the child as he or she is, and to encourage remedial action through an individualized educational program. The idea is also to keep children whenever possible in regular classes with nonhandicapped children.

This philosophy may not eliminate special education, but it certainly contemplates a lesser role for classes such as those for the E.M.R. Despite the potential of the Master Plan, those classes still do exist in California, forcing this court to act to protect the rights of black children assigned to them, but such classes are clearly not as important as they once were. Enrollments are now much less than one-half of what they were a decade ago. Change is taking place. The movement is away from labeling and isolation and toward education to meet individual needs. The future of special education may be a brighter one.

Whatever the future, however, it is essential that California’s educators confront the problem of the widespread failure to provide an adequate education to underprivileged minorities such as the black children who brought this lawsuit. Educators have too often been able to rationalize inaction by blaming educational failure on an assumed intellectual inferiority of disproportionate numbers of black children. That assumption without validation is unacceptable, and it is made all the more invidious when “legitimated” by ostensibly neutral, scientific I.Q. scores. We have refused to allow the continuation of E.M.R. policies consistent only with that assumption, and it is hoped this will clear the way for more constructive educational reform.

State Department of Education herein referred to as “SDE.”

SOURCE: SDE EMR Ethnic Survey Printout — 1976-1977

. “Educable mentally retarded” (E.M.R.) children are defined as incapable, “because of retarded intellectual development,” of mastering the skills necessary to advance beyond a minimal educational level. See Cal.Educ.Code § 56500. . Individual intelligence tests are given to one person at a time and involve both oral and written responses to the questions posed by the examiner. Group I.Q. tests are entirely written, can be given to large numbers of persons, and do not involve direct communication between examiner and examinee. Group I.Q. tests are no longer statutorily authorized in California. 1975 Cal.Stats. ch. 460. Both kinds of examinations result in a score purporting to measure intelligence, defined often as “general intellectual functioning,” with an average intelligence rated with a score of about 100. . This case concerns particularly the tests that were on the State Board of Education’s approved list at the time of the moratorium in January 1975. The list was divided into basic tests, one of which had to be given for E.M.R. placement, and supplemental tests. The basic list included the Stanford-Binet, and WISC (Wechsler Intelligence Scale for Children, now superceded by the revised WISC-R), the WAIS (Wechsler Pre-School and Primary Scale of Intelligence), the WPPSI (Wechsler Pre-School and Primary Scale of Intelligence), and the Leiter International Performance Scale. The supplemental list included (1) Author Point Scale of Performance Test, revised form II, (2) Catell Infant Intelligence Scale, (3) Columbia Mental Maturity Scale, revised edition, (4) Draw-a-Person (Goodenough), (5) Full Range Picture Vocabulary Test, (6) Gessell Development Schedule, (7) Goodenough-Harris Drawing Test, (8) Merrill-Palmer Pre-School Performance Test, (9) Peabody Picture Vocabulary Test, (10) Raven Progressive Matrices, (11) Slosson Intelligence Test, and (12) Van Alstyne Picture Vocabulary Test. . A preliminary injunction limited to the San Francisco Unified School District was issued in 1972. Larry P. v. Riles, 343 F.Supp. 1306 (N.D. Cal. 1972), 'aff’d 502 F.2d 963 (9th Cir. 1974). The statewide order was entered on December 13, 1974, and in January 1975 the state imposed a moratorium on all I.Q. testing for E.M.R. placements. The time lag between the two orders reflects the unsuccessful appeal of the earlier preliminary injunction. Larry P. v. Riles, 502 F.2d 963 (9th Cir. 1974). . Another case, Diana v. Board of Education, C-70-37 RFP, was filed earlier and made essentially the same claims as Larry P. on behalf of Hispanic children. The issues of that case were resolved through a stipulated settlement approved by the court on June 18, 1973. . U. S. Department of Health, Education, and Welfare, Background Papers for the Conference on Problems of Education of Children in the Inner City 2-3 (1969). a. Mercer testimony,- Reporter’s Transcript 1753-59 (testimony is cited hereafter in the form: Mercer 1753-59). . At least two other courts, however, have addressed similar concerns. In 1967, Judge Wright held in a powerful opinion that the ■ “tracking system” in use in the Washington, D. C., schools violated the equal protection of the laws. Hobson v. Hansen, 269 F.Supp. 401 (D.D.C.1967), affirmed, sub nom. Smuck v. Hobson, 132 U.S.App.D.C. 372 , 408 F.2d 175 (D.C.Cir. 1969). More recently, Judge Weinstein found that New York City’s special day schools for emotionally handicapped children infringed important federally guaranteed rights. Lora v. Board of Education of City of New York, 456 F.Supp. 1211 (E.D.N.Y.1978). These pathbreaking decisions built on the recognition that guarantees of equal rights in education can be rendered meaningless through the assignment of minority students to special classes. . Detailed information on the history of the I.Q. test and of special education in California can be obtained from the following two works: L. Kamin, The Science and Politics of IQ (1974); J. Simmons, A Historical Perspective of Special Education in California, Vol. 1 (unpublished doctoral dissertation, 1973). Both Professor Kamin and Dr. Simmons testified at the trial. . L. Terman, The Measurement of Intelligence 6-7 (1916). . Id. at 91-92. . See generally L. Kamin, supra, note 8, at 19-51. Professor Kamin testified that, “The early leaders in the I.Q. testing movement all seemed to agree that the tests were measuring some fixed innate capacity that was determined by the genes.” Kamin 849. One of the original translators of the Binet into English, Henry Goddard, tested a number of immigrants in 1912, and he concluded from his testing that “83 percent of Jews, 80 percent of Hungarians, 79 percent of Italians and 87 percent of Russians were feeble-minded,” which he believed to be an inherited trait. Kamin 850-51. Probably the best example of the assumptions and approach of the testing movement was the program of widespread testing of army draftees in World War I on the basis of a suggestion'by Robert Yerkes, then President of the American Psychological Association. Terman and Goddard also participated in this ambitious project, and the results were published in detail in a book by Carl Brigham entitled A Study of American Intelligence (1923). Brigham used the data to make conclusions such as the following: The Nordics are . . rulers, organizers and aristocrats individualistic, self-reliant, and jealous of their personal freedom . . as a result they are usually Protestant . . The Alpine race is always and everywhere a race of peasants . The Alpine is the perfect slave, the ideal serf the unstable temperament and the lack of coordinating and reasoning power so often found among the Irish we have no separate intelligence distributions for the Jews our army sample of immigrants from Russia is at least one half Jewish Our figures, then, would rather tend to disprove the popular belief that the Jew is intelligent ... he has the head form, stature, and colour of his Slavic neighbors. He is an Alpine Slav, quoted in L. Kamin, supra note 8, at 38. Brigham warned of the problems that liberal immigration policies could cause, and he linked his concern to his test-based opinion about the intelligence of black Americans. We must face a possibility of racial admixture here that is infinitely worse than that faced by any European country today, for we are incorporating the negro into our racial stock, while all of Europe is comparatively free from this taint . . . The decline of American intelligence will be more rapid than the decline of the intelligence of European national groups, owing to the presence here of the negro. Interestingly, Brigham later became secretary of the College Entrance Examination Board, where he “designed and developed the Scholastic Aptitude Test.” Id. at 39. Fortunately he recanted some of his earlier views. . L. Terman, supra note 9, at 91. . According to Dr. Simmons, “In keeping with the thrust of the testing movement, Los Angeles City Schools established a Division of Psychology in 1917, and San Francisco began a Department of Psychological Services in 1925.” J. Simmons, supra note 8, at 40. . See J. Simmons, supra note 8, at 49. . Id at 61. . Id . Id. at 136-46. . At present, Cal.Educ.Code § 56502(a) reads as follows: (a) Admission of a pupil to a special educational program for the mentally retarded established under the provisions of Sections 56500 to 56534, inclusive, and in Sections 1880 to 1887, inclusive, shall be made only on the basis of an individual evaluation according to standards established by the State Board of Education and upon individual recommendation of a local admission committee which shall include an administrator in charge of special education programs in the district or county superintendent of schools, an experienced special education teacher, a school nurse, and a school psychologist who has examined the child under consideration for, eligibility and placement. . See J. Simmons, supra note 8, at 201-02. No action resulted from conference discussions or proposals. . See section III.E.l., infra. . State Department of Education, Placement of Underachieving Minority Group Children in Special Classes for the Educable Mentally Retarded (1970). . State Department of Education, Placement of Pupils in Classes for the Mentally Retarded (1970). . 1970 Cal.Stats. ch. 1543. Retesting resulted in the reduction of E.M.R. enrollments by some 11,000 to 18,000 students. See Watson 4089; Meyers 8649. . The requirement is specifically the following: If the percentage of children from any minority ethnic group in such classes varies by 15 percent or more from the percentage of such children in the district as a whole, an explanation for such variation shall be attached to the report to the Department of Education. . The Master Plan was established as a pilot program for no more than ten local educational agencies. 1974 Cal.Stats. ch. 1953. In 1977, the Master Plan was revised to provide for the eventual phase in of the plan throughout the state. 1977 Cal.Stats. ch. 1247. Sufficient funds are necessary, however and if not appropriated by the legislature, school districts “may continue to operate special education programs” such as those now in effect outside of the Master Plan. Cal.Educ.Code § 56315(d). Under the Master Plan, governing boards of local school districts or other “responsible local agencies” are to develop comprehensive plans for the education of children with special needs. Cal.Educ.Code §§ 56315 et seq.. The provisions of the Master Plan do not preclude the operation of classes to meet these special needs, which are functionally equivalent to present E.M.R. classes. Cal.Educ.Code § 56332(a)(1). (“Special classes and centers which enroll pupils with similar and more intensive educational needs” must be part of the plan.) The definition of E.M.R., however, appears to be superceded by the more inclusive one of “learning handicapped.” Cal.Educ.Code § 56302(g). The Master Plan also follows the dictates of new federal legislation, especially the Education for All Handicapped Children Act of 1975, 20 U.S.C. §§ 1401 et seq., that emphasize procedural guarantees for parents and children and placement in the “least restrictive environment.” With respect to I.Q. testing, the Master Plan provides as follows: (i) Procedures and materials for assessment and placement of individuals with exceptional needs shall be selected and administered so as not to be racially, culturally, or sexually discriminatory. No single assessment instrument shall be the sole criterion for determining placement of a pupil. Such procedures and materials for assessment shall be in the individual’s mode of communication. Procedures and materials with [sic] use with non-English-speaking and limited-English-speaking pupils as defined in subdivisions (d) and (e) of Section 52163, shall be in the individual’s primary language. As will be seen, however, this language does not indicate any practical change in the state’s attitude to standardized intelligence tests. See infra note 107 and accompanying text. . For the school year 1976-77, for example, of $10 million spent on E.M.R. education, only about $20,000 was for educating such students in regular classes as opposed to special segregated classes. . State Department of Education, Programs for the Educable Mentally Retarded in California Public Schools (1974). . Id. at 26 . . Id. at 27 . According to Clifton Shryock, Consultant with the State Department of Education in the area of mental retardation, The State Board said that — and the legislature — that it was very important that this curriculum be different. . . . It is also one of the stipulations . that it not be a watered down academic curriculum, should not follow the regular curriculum in any way as far as requirements are concerned. Shryock 9143. See also Hanson 3581-85 . . Hanson 3585 . See also, e. g., Watkins 4166-67 (“the longer the child remains in the segregated program, the more out of touch he becomes with the curriculum demands of a regular program, and the lower the probability of his performing effectively when returned”), 4037-42. . D. Uslan, R. Russell, K. Anselm, Final Report: Systematic Program Development for Mentally Retarded Pupils. A State Initiated Project 1975-76, at 9 (Report to Mr. Leslie Brinegar, Director, Office of Special Education, State Department of Education) (August 1977). See also Watkins 4082. . Divergences between the two sets of data evidently resulted primarily because of some underreporting of E.M.R. enrollment in Los Angeles in response to the Office of Civil Rights survey. . These figures in practice mean that E.M.R. classes are often almost entirely composed of minority students, chiefly black and Hispanic children. From the 1972-73 academic year until at least 1974-75, a majority of San Francisco’s E.M.R. classes had 80 percent or more minority students. Dean 4347-50. In Los Angeles, black children during the period 1973-74 to 1976-77 made up about 24-25 percent of the school population and 56-59 percent of the E.M.R. population. Hispanic children represent 21-23 percent of the E.M.R. students and 25-32 percent of the school population. . The rate of enrollment for black children in the 20 districts accounting for 80 percent of California’s black population is generally two to four times higher than for non-blacks. E. g., Hillabrandt 2517. . Prior to the 1969 change, 5 Cal.Admin.Code § 182 read as follows: Eligibility of Pupils. The eligibility of a minor for admission to any such school or class shall be determined by a psychologist or psychometrist after such pupil has been given an accepted verbal or non-verbal intelligence test. 12 Cal.Admin.Reg. No. 5 (April 27, 1948). Neither side introduced any evidence regarding the adoption of this earlier and more flexible requirement. The cut-off score for E.M.R. placement generally was about 75, depending on local policies, and it reached 85 in some instances. Simmons 3497-98. . See note 5 supra. . Jensen, “How Much Can We Boost I.Q. and Scholastic Achievement?” 39 Harv.Educ.Rev. 1-123 (1969). . The new code section specified as follows: 3401. Eligibility of Pupils. The eligibility of a minor for placement in a special training school or class for mentally retarded minors shall be determined as provided in Education Code Sections 6908 and 6909 and after the minor has been given verbal or nonverbal individual intelligence tests and after other pertinent information has been collected and considered. Group intelligence tests may be used as screening devices. All individual and all group intelligence tests administered pursuant to this section shall be selected from a list approved by the State Department of Education. . The list is in note 3 supra. . Mr. Hocker’s letter to Dr. Joseph Rice, Chief of the Bureau for Educationally Handicapped and Mentally Exceptional Children, stated, inter alia, I would like to respond as co-chairman of the California Association of School Psychologists and Psychometrists’ Testing Committee and also as a psychologist and educator whose role regularly involves a review of special education class placement procedures. I am particularly concerned with the urgency expressed in Mr. Watson’s letter in which he mentioned both the intention to have an “approved list” ready by August 1969 and a “tentative list” before the opening of school in September. It would seem to me that a matter as important as this could deserve a bit more time, particularly since most of the persons involved in implementing the law will not know of its existence until after the list has been prepared. The time is certainly ripe for assuring that children placed in special education programs for the mentally retarded have been properly identified and the merits of the individual cases carefully considered. You and I are personally familiar with the concerns expressed by John Chandler and John Plakos of the State Department of Education, and by Steve Moreno and his group within the Association of Mexican-American Educators, that the labeling process at present is insufficient to accurately identify retardates among certain ethnic and racial minorities. The research of Dr. Jane Mercer of the University of California at Riverside has indicated that in the case of minority groups, individual IQ’s play a less significant role in determining of eligibility than do certain other indices. As I am sure you are also aware, some pressure has been exerted upon school districts to depart from traditional clinical practice in making the essential determination that a child is mentally retarded. I think it would be safe to say that most psychologists within the State would approve of the requirement that only individual intelligence tests of the stature of the WISC, or recent revision of the Stanford Binet, be employed for the purpose of obtaining clinical data which will contribute to the labeling process. It is also obvious however, that many psychologists, special educators, and persons of varying levels of sophistication who have a particular interest in the education of the culturally disadvantaged child will disagree as to what kinds of alternative or additional psychometric information ought to be included. It is perhaps unfortunate that we find ourselves in the process of adding to the status of classical IQ instruments at a time when psychology is re-evaluating the appropriateness of IQ as a simple basis for classification. Letter from Robert W. Hocker to Dr. Joseph Rice, July 30, 1969. . Other letters also raised questions. For example, a letter of July 30, 1969 from Dr. Steve Moreno, Evaluation Specialist, San Diego City Schools, urged that “only tests which distribute all groups equally should be used by any psychologist.” Less than 30 letters were received. . See note 23 supra. . See section III.D. supra. . § 56507. Rights of parents and pupils; identification, placement, and program (a) Both a parent and a pupil are guaranteed and may initiate procedural due process protections in any decision regarding, and resulting from, the pupil’s identification as an individual with exceptional needs; the pupil’s assessment and the implementation of the individualized education program; and the denial, placement, transfer, or termination of the pupil in a special education or related services program. Such protections shall include, but are not limited to: (1) Written notice to the parent of his or her rights, in language understandable to the general public and in the primary language or other mode of communication of the parent; (2) The right of the parent to initiate an assessment of his or her child within a specified period of time; (3) The requirement that written parental consent must be obtained before any assessment of the pupil is conducted; (4) The right of the parent to participate in the development of the individualized education program and to be informed of the availability under state law of free appropriate special education programs, both public and private; (5) The requirement that written parental consent must be obtained before the pupil is placed in any special education program; and, (6) Procedural due process by a fair and impartial administrative hearing before a fair hearing panel. . See generally, J. Mercer, Labeling the Mentally Retarded 91-123 (1973). . See D. Uslan et al., supra note 31. . Id. at 14. . Id. at 32. . Id. at 23. . This information is from the summary report prepared by Uslan, et al. on the San Francisco Unified School District’s E.M.R. administrative and placement procedures. Plaintiff’s Exhibit P — 46(c). . This is of course true even if, as several knowledgeable witnesses testified, there are disproportionate initial referrals of black children for screening for E.M.R. placement. Watkins 4067; Bryant 8884. Defendants suggest that I.Q. tests save those who are not retarded from E.M.R. placement, since the tests are administered only to “confirm” the suspicion of retardation raised by school failure and teacher observation, but clearly a disproportionate number of white children will be saved because of their I.Q. scores. Defendants also contend that the requirement of parental consent for E.M.R. placement overcomes any deficiencies caused by bias in the placement process. The evidence shows, however, that consent is rarely withheld, particularly by minorities, since the mystique of teacher authority and I.Q. scores tends to overwhelm parents. See Watkins 4081-82; Dean 4312-13; Brinegar 4761-62. Several witnesses also suggested that I.Q. tests should be and can be interpreted such that they are more fair to black children. It appears, however, that there is no systematic method for doing this, and all that these witnesses — including the representatives from the testing companies — could suggest was that, in interpreting the scores, examiners should in some general manner “take into account” different racial and cultural backgrounds. E. g., Doppelt 5352-55; Lambert 6659-67; Munday 5160-61. There is no suggestion that this after-the-fact taking into account can negate the impact of scores for black children that are, on the average, about 15 points lower than those of white children. Taken from SDE Ethnic Survey Report Taken from P-49, SDE Report on Racial Distribution of Pupils & Staff in California, Public Schools, Pall 1973 . See section V.B.5., infra. . So-called achievement tests are not very different in approach from intelligence tests. According to Professor Mercer, [A]cademic achievement tests are . practically the same thing as the I.Q. test. The items in these two types of tests are drawn from the same item pools. They are essentially measuring the same thing. Mercer 1527-28. Similarly, Professor Humphreys, who testified for defendants, wrote, “all ability tests (whether called intelligence, aptitude, or achievement tests) measure current performance. There are no measures of innate capacity.” Cleary, Humphreys, Kendrick, and Wesman, “Educational Uses of Tests with Disadvantaged Students,” 1975 American Psychologist 15. See also, e. g., Thorndike 4806. The major difference between the intelligence and achievement tests lies in their specificity and the functions they serve. Achievement tests are not used to provide a single, overall score upon which a finding of more or less irremediable retardation may be based. Achievement tests are supposed to measure learning in particular subjects. The customary uses of achievement tests are not questioned by plaintiff, even though black children also tend not to do well on these tests. . The correlations are especially strong among the most popular of the tests, which are those on California’s required list — the Wechsler tests, the Stanford-Binet, and the Leiter test. Hilliard 265-73. See also Kamin 883; Sattler 6825. . Originally an I.Q. of 100 was supposed to mean that the child had a “mental age” 100 percent equivalent to the child’s chronological age. Thus a child of five who had a mental age of an average four-year-old would be considered to possess an I.Q. score of 80 — four-fifths of 100. I.Q. scores are no longer designed to represent mental ages. . D. Wechsler, Manual for the Wechsler Intelligence Scale for Children — Revised 24 (1974). . Id. at 25. . The arbitrary character of a statistical definition is evident from the fact that in the early 1970’s the American Association of Mental Deficiency changed the definition of retardation. The 1961 AAMD Manual on Terminology and Classification in Mental Retardation defined mental retardation as “subaverage general intellectual functioning which originated during the developmental period and is associated with impairment in adaptive behavior.” “Subaverage intellectual functioning” meant an I.Q. score lower than one standard deviation below the mean. In the 1973 revision of the AAMD Manual, prepared by Dr. Grossman, the definition was changed to “significantly subaverage general intellectual functioning,” now meaning scores lower than two standard deviations below the mean. This influential decision essentially reduced the number of mentally retarded Americans from 15 percent to 2 percent. See generally, Filler, et al., “Mental Retardation,” in Issues in the Classification of Children 194 (N. Hobbs ed. 1975). The confusion of the statistical definition of retardation with the medical or pathological mo del is an especially problematic result of “diagnoses” based on test scores. It has been estimated that 15 percent of the total of those below the statistical cut-off suffer from some organic, identifiable medical condition, such as Down’s syndrome. Albee 420-23; Grossman 5679-80. The rest of the “retarded” are labeled primarily on the basis of substantially below average school and test performance. Despite the lack of any medically known pathology, however, the medical model tends to be applied to the E.M.R. label. According to Professor Mercer, When evaluations using a statistical model are conducted in conjunction with evaluations using the pathological model — as is true in the case of mental retardation — there is a tendency to think in terms of one model while operating with the other. Behavioral patterns are translated into pathological signs. The implicit logic that underlies this transformation is as follows: Low IQ = “bad” in American society; a social evaluation. “Bad” = pathology in the pathological model. Therefore, low IQ = pathology. Thus, IQ, which is not a biological manifestation but is a behavioral score based on responses to a series of questions, becomes conceptually transposed into a pathological sign carrying all the implications of the pathological model. Statistical abnormality is equated with biological pathology without any evidence based on functional analysis that this statistical sign is related to the biology of the organism or that it has any functional relationship to system maintenance. J. Mercer, supra note 45, at 5-6. See also Mercer 1461-64; Meyers 8704. . Interestingly, the authors of one of the classic studies in this area found this variation in test scores of black and white children on the Stanford-Binet, and they concluded on that basis that “the results . . . should give serious concern to systems using the Binet IQ as a cut-off score for special classes, legal definitions of responsibility, and mandatory sterilization.” Kennedy, Van de Riet, and White, “A Normative Sample of Intelligence and Achievement of Negro Elementary School Children in the Southeastern United States,” 28 Monographs of the Society for Research in Child Development, No. 6, at 69 (1963). . See notes 64-69 infra and accompanying test. . As Professor Kamin observed, If one looks at what happened, it is perfectly clear the testmakers have a preconception that males and females do not differ in intelligence. It seems quite clear from the fact they continue to include in the tests items which do discriminate between races or even social classes . . . [that] at the least, they do not have a preconception that races and social classes are equal in intelligence. Kamin 877. . See L. Kamin, supra note 8; Kamin 909-10. . D. Wechsler, The Measurement of Intelligence 107 (1944); Doppelt 5467. . The 1916 Stanford-Binet was standardized exclusively on white persons in California. The sample comprised 1,000 school children and 400 adults, the latter of which included 30 businessmen and 150 “hobos.” Kamin 872-73. The test was modified somewhat and restandardized in 1937 on a sample of 3,000 school children, all of whom were white. In the report on the 1937 restandardizing, Terman and Merrill stated that they eliminated some test items for which there was a gross discrepancy in the ability of boys and girls to answer correctly.. They left in the test some items which tended to favor one sex, but they balanced such items so that boys and girls would both have an average score of 100. Kamin 874-75. The next revision of the test was in 1960, but no new standardization was undertaken. Finally, in 1972, the Stanford-Binet was again modified slightly and restandardized, this time on children in seven districts — Placer County, California; Denver, Colorado; Shelton, Connecticut; Valparaiso, Indiana; Fort Worth, Texas; Houston, Texas; and Salt Lake City, Utah. Munday 5106. It is presumed that black children were represented in the sample in proportion of their representation in the population of those districts, but the test administrators did not identify the test-takers by race. Thus, the restandardization process did not yield any data useful for a comparative analysis of the performance of black and white children on particular items. Munday 5247^18. The Wechsler Intelligence Scale for Children (WISC) was standardized in the late 1940’s on a sample of 2,200 white children. The WISC was revised in the early 1970’s (WISC-R) and restandardized on a population that was 15 percent non-white. The revisions did result in the elimination of a few specific items in response to objections that they were racially or culturally inappropriate, such as the Romeo and Juliet question, Doppelt 5326-27, but this was not a systematic attempt to validate test items for minorities. See also, e. g., with respect to the problem of boys and girls, Doppelt 5367-68; Humphreys 6182-83; Hilliard 290. . See note 64 infra. Standardizing on a mixed population, with such different scores may even exacerbate the discriminatory effect of the tests, since it shifts the curve up for white children in the sample by two or three points. Mercer 1565-66; Kamin 921. . The study was completed in the late 1960’s. It also found that black children scored 11 points behind white children in picture arrangement, showing that black children may also do substantially poorer in the “performance” component of the exams. Kamin 890-93. . Dr. Wechsler stated in an interview that one ought to “give the child credit if he comes from Harlem” and thinks he should fight back, Doppelt 5497, but the manual has not been changed. Dr. Hilliard noted the significance of this apparently off-hand concession by Dr. Wechsler: [T]he importance of Dr. Wechsler’s answer there was that when confronted with an actual concrete illustration of a cultural difference in terms of either information or values, then he yielded the point.- ... If you have to take it into account there, then, of course, you have to take it into account throughout the entire test. Hilliard 336. . ■ See generally Hilliard 174-97. Even certain apparently neutral skills may reflect cultural biases. Kamin 892-93. Dr. Powell testified about her research in Uganda, which shows how bias may creep in: We tried using the WISC-R in Uganda, because we were trying — kind of struggling with how we could made assessments of children mentally retarded. And what we found was that children who were doing very very well in Ugandan schools, which is a British system, and they have the same kind of criteria for performance, were not doing well on-the WISC-R because some of the tasks that they were asked to do, which were some matrices, you know, find your self — wind yourself through what they called mazes, and some of the block building tasks were very foreign culturally to them. However, we had given them the task of sorting rice in different kinds of kernels of rice, and different kinds of sizes of kernels of com, they performed very, very well and very, very superior than our children. We have done now cross-culturally, looking at several groups of children in Africa, in the continental U.S. and in the Caribbean, and have found that many cognitive tests are very culturally prescribed and depends upon what the child’s cultural experience has been in terms of judging his competency. . See, e. g., the following testimony by defendants’ experts: Humphreys 6301-03; Munday 5137-39, 5176, 5197-98; Thorndike 4948-49, 4964-65, The fact of cultural bias, indeed, appears to be increasingly recognized as self-evident. Justice Rehnquist, for example, recently suggested a willingness to take judicial notice of this conclusion: It is well documented that minorities do not perform as well as Anglos on standardized exams — principally because of cultural and socioeconomic differences. Columbus Board of Education v. Penick, 443 U.S. 511 , 449 n. 17, 99 S.Ct. 2941 , 2964 n. 17, 61 L.Ed.2d 666 (1979). The real differences between plaintiffs’ and defendants’ positions turn on the interpretation and legal effect of that cultural bias. . Admittedly “black culture” is not easily defined and is not homogeneous in this country. Nevertheless, it is clear that black Americans are distinguished in our society because of their common history of African heritage, slavery, and pervasive racial discrimination, coupled with widespread poverty, inadequate educational systems, and isolation in effectively segregated pockets of our urban centers. This has produced a unique “experience pool.” See also Mercer 1581-86. . According to Professor Humphreys, for example, The purpose of the schools is to impart the culture of the society to the young. . [T]he intelligence test does predict, does measure certain kinds of skills . , . that we value in our society. And I’m sure that a measure of inner-city coping skills would have its own pattern of validity, but these would . . differ from those of the intelligence test. Humphreys 6201. See also Thorndike 4948-49; Munday 5137-39. Dr. Powell, who testified for plaintiffs, made a similar observation: “By and large most I.Q. tests test those kind of skills that we are asking kids to have in order to perform in class or in schools, in Western schools, and to do problem-solving in and around those kinds of tasks.” Powell 1243. Of course, different witnesses had different ideas as to how much our schools should seek to impart only a “majority” culture, and whether “minority” cultures need be inconsistent with high quality performance in the skills deemed valuable to our society as a whole. That issue obviously cannot be resolved here. . While plaintiffs did not argue this theory, it appears also that a private right of action exists under 42 U.S.C. § 1983 to enforce the . federal statutory rights, since state action is involved. See, e. g., De La Cruz v. Tormey, 582 F.2d 45 (9th Cir. 1978) cert. denied, 441 U.S. 965, 99 S.Ct. 2416 , 60 L.Ed.2d 1072 (1979); cf. Chapman v. Houston Welfare Rights Organization, 441 U.S. 600, 99 S.Ct. 1905 , 60 L.Ed.2d 508 (1979). . Also significant are the changes made by Congress in 1978. A new provision, 29 U.S.C. § 794a, was added. It states, inter alia, that, (a) (2) The remedies, procedures, and rights set forth in title VI of the Civil Rights Act of 1964 shall be available to any person aggrieved by any act or failure to act by any recipient of Federal assistance or Federal provider of such assistance under section 794 of this title. (b) In any action or proceeding to enforce or charge a violation of a provision of this sub-chapter, the court, in its discretion, may allow the prevailing party, other than the United States, a reasonable attorney’s fee as part of the costs. These amendments further show the Congressional intent to have section 504 construed in the same manner as Title VI, and they also reflect Congressional recognition of a private right of action — encouraged through an award of attorney’s fees to private parties. . Pretrial conference of September 23, 1977. Reporter’s Transcript, p. 39-41. The Department of Health, Education, and Welfare has undertaken compliance activity in California only against the Bakersfield Elementary School District. . Plaintiffs have also sought to bring an “effects” challenge under the Emergency School Aid Act, 20 U.S.C. § 3191 et seq. This law specifies the conditions under which funds will be made available to school districts to enable them to come into compliance with desegregation standards, and it closely related to Title VI. This act contains an express statement that [N]o educational agency is eligible for assistance under this subchapter if it has, after June 23, 1972 . (C) in conjunction with desegregation or the conduct of an activity described in this section, had in effect any procedure for the assignment of children to or within classes which results in the separation of minority group from nonminority group children for a substantial portion of the school day, except that this clause does not prohibit the use of bona fide ability • grouping by a local educational agency as a standard pedagogical practice. 20 U.S.C. § 3196 (c)(1). See also 45 C.F.R. § 185.43 . Plaintiffs, however, have not demonstrated the applicability of these provisions to California schools, nor the propriety of a private action. In any event, since we do not think that this law adds significantly to the Title VI claim, we make no specific findings on this provision. On its relationship to Title VI, see Board of Education of City of New York v. Califano, 584 F.2d 576 , 587-90 (2d Cir. 1978), cert. granted, 440 U.S. 905 , 99 S.Ct. 1211 , 59 L.Ed.2d 453 (1979). . The Court took the time to reject language in the appellate court decision suggesting that courts should not intervene when schools themselves are not responsible for inequalities manifested in the schools. The disapproved language stated: “ ‘Every student brings to the starting line of his educational career different advantages and disadvantages caused in part by social, economic and cultural background, created and continued completely apart from any contribution by the school system.’ ” Lau v. Nichols, 414 U.S. at 565 , 94 S.Ct. at 788 , quoting 483 F.2d at 797. Accord, De La Cruz v. Tormey, 582 F.2d 45 , 57 n. 8 (9th Cir. 1978). The existence of societal inequalities, therefore, while inevitably reflected in schools, cannot be used as a glib justification for a failure to attempt to provide equal rights in education. . In Lora , the court held A prima facie violation of Title VI would be established if it were shown that referrals and assignments to the special day schools are the result of policies, practices and the use of vague and subjective criteria which have a racially discriminatory effect as evidenced by the overrepresentation of minority students in these schools. This showing would shift the burden to defendants to rebut the inference that their actions are a substantial cause of the racial disparity and deficiency in educational opportunity for minority' students in the special day schools. Lora v. Board of Education, 456 F.Supp. at 1277 . . See section IV.C.2., supra and section V.B.4., infra. . See 29 U.S.C. § 706 (6); 45 C.F.R. § 84.3 (j). Thus, the fact that plaintiffs claim not to be retarded — handicapped—does not preclude them from asserting the protections of this law. . According to the Senate Report by the Labor and Public Welfare Committee, the Act was designed to protect “physically or mentally handicapped children who may be denied admission to Federally-supported school systems on the basis of their handicap.” S.Rep.No. 93-1297, 93d Cong., 2nd Sess.; 1974 U.S.Code Cong. & Admin.News, pp. 6373, 6389. The , Report also highlighted Congressional concern with the classification or labeling process, stating that “racial and ethnic factors may contribute to misclassification as mentally retarded.” Id. . The Senate Report by the Labor and Public Welfare Committee expressly noted the relationship of this law to the regulations then being issued under section 504 of the Rehabilitation Act of 1973, and the Report included the following significant remarks about classification: The Committee is deeply concerned about practices and procedures which result in classifying children as having handicapping conditions when, in fact, they do not have such conditions. The Committee is alarmed about the abuses which occur in the testing and evaluation of children, and is concerned that expertise in the proper use of testing and evaluation procedures falls far short of the prolific use and development of testing and evaluation tools. The usefulness and mechanistic ease of testing should not become so paramount in the educational process that the negative effects of such testing are overlooked. The Committee bill provides that the Commissioner shall, in carrying out the provisions of this Act, issue, amend, and revoke such rules and regulations as may be necessary and that no other less formal method of implementing the provisions is authorized. Specifically in this regard, the Committee directs the Commissioner to issue regulations which assure that State and local educational agencies shall establish procedures to ensure that certain conditions are met with regard to testing and evaluation. These regulations should assure that: (1) tests and other materials used for placement have been properly and professionally evaluated for the specific purpose for which they are being used; (2) no single test or type of test or procedure is used as the sole criterion for placement and that all relevant information with regard to the functional abilities of the child is utilized in the placement determination; and, (3) tests and other evaluation procedures include assessment of specific areas of educational need so that provision of special education and related services can be limited to areas directly related to the child’s need, so that broad and unspecific classifications of handicapping conditions do not occur. S.Rep.No. 94-168, 94th Cong., 1st Sess., at 26, 29; 1975 U.S.Code Cong. & Admin.News, pp. 1425, 1450, 1453. . According to the explanation of the regulations under the Education for All Handicapped Children Act of 1975, the two laws relate generally as follows: Basically, both require that handicapped persons be provided a free appropriate public education; that handicapped students be educated with nonhandicapped students to the extent appropriate; that educational agencies identify and locate all unserved handicapped children; that evaluation procedures be adopted to insure appropriate classification and educational services; and that procedural safeguards be established. 45 C.F.R. Pt. 121a, App. A at 500. The basic approach revolves around I.E.P.’s or Individualized Education Programs, which are to be developed for each student found to be handicapped under the acts. I.E.P.’s are to foster mainstreaming — the education of the handicapped to the extent possible in regular classes. I.E.P.’s are therefore supposed to ensure: (1) That to the maximum extent appropriate, handicapped children, including children in public or private institutions or other care facilities, are educated with children who are not handicapped, and (2) That special classes, separate schooling or other removal of handicapped children from the regular educational environment occurs only when the nature or severity of the handicap is such that education in regular classes with the use of supplementary aids and services cannot be achieved satisfactorily. 45 C.F.R. § 121a.550(b). See also 20 U.S.C. §§ 1412 (5)(B), 1414(a)(l)(C)(iv); 45 C.F.R. § 84.34 . . See text accompanying note 28, supra. . While the terminology is not particularly helpful in this case, it may be noted that the validation inquiry generally is based on assessments of “content validity,” “construct validity,” or “predictive validity.” See generally American Psychological Association, Standards for Educational and Psychological Tests (1974); 29 C.F.R. Pt. 1607; Hilliard 213-20; Sattler 6799-6803; Thorndike 4881-82. The only predictive validity asserted by defendants is that borrowed from the employment testing area, and we have already denied the relevance of predictions of academic performance to validation for E.M.R. placement. Content validity, according to the expert testimony, determines whether all aspects of a given subject are measured. The practical utility of content validity assessments is generally limited to achievement tests. Thorndike 4881. One of the essential aspects of E.M.R. status is that it presupposes mental retardation, not merely a limited academic achievement. Defendants nevertheless argue in their closing memoranda that the content validity of the I.Q. tests has been established. That argument, however, is made in the context of employment testing and is on the basis of predictive validity materials. We are thus apparently left with construct validity — whether the tests measure the domain of skills that comprise mental ability. Thorn-dike 4882; Sattler 6799-6800. The definitions are sometimes confused, however, and the important concern is the practical one of establishing the relationship between I.Q. test scores of black children and the characteristics justifying their placement into special E.M.R. classes. . See note 58, supra. . See note 63, supra, and accompanying text. . See note 64, supra. . See note 67, supra. . See Cleary et al., supra note 53. . Cf. note 53, supra. . Cf. Cleary et al., supra note 53. . See J. Mercer, System of Multicultural Pluralistic Assessment: SOMPA. Conceptual and Technical Manual (unpublished manuscript, February 1977). . Diana v. Board of Education, C-70-37 RFP. The consent decree of June 18, 1973 provides that the SDE must send letters and ultimately investigate school districts maintaining a “significant variance” — later defined as one standard deviation — in percentage of Chicano children enrolled in E.M.R. classes. Districts must develop plans to eliminate the disproportion, and the court can order sanctions if the disproportion remains. See Order of May 27, 1974 (defining significant variance). Clifton Shryock, Special Education Consultant with the SDE, indicated by affidavit filed with the court in the Diana case on February 22, 1977, that at that time only 17 districts out of the total of 1,054 in California had significant variances, and these variances were minimal considering the actual numbers of children involved. See also Riles 2748-49; Brinegar 4649-61; Bryant 8897-8917. . Evidently Massachusetts no longer uses the tests, and Martin Gerry testified about the work of his consulting firm in helping New York and Louisiana develop fair placement criteria. Gerry 9582, 9594. . This alternative approach was taken in Guadalupe Organization v. Tempe Elementary School, 587 F.2d 1022 (9th Cir. 1978). . Law review commentary is extensive on this subject. For a summary of the various viewpoints, see 15 San Diego L.Rev. 925-1183 (1978). . In Lora , the special classes were not dead-end classes, but rather were intended to allow students to follow the regular course of study and return to regular classes. The educationally handicapped classes and day schools also did not have the stigma of classes for the “retarded.” A third distinction is that in Lora it was more difficult to ascertain the existence of culturally biased placement mechanisms, such as the I.Q. tests challenged in Larry P. . See section III.D., supra. . See section V.B.5., supra. . The Court made a subtle distinction to reject the finding of the lower court and the dissent that a “range of less discriminatory alternatives” were available. 442 U.S. at 285, 99 S.Ct. 2282 . Those alternatives amounted to schemes where veterans were given less of a preference. The Court reasoned, however, that the question of intent turned on whether a desire to favor veterans was a permissible intent. If it was, as the Court held, then it presumably did not matter how much of a preference was extended. 442 Ü.S. at 277 & n. 23, 99 S.Ct. 2295 & n. 23. . We note that there is a line of cases questioning or prohibiting the use of “ability groupings” based on standardized tests. See, e. g., United States v. Gadsen County School Dist., 572 F.2d 1049 (5th Cir. 1978); McNeal v. Tate County School Dist., 508 F.2d 1017 (5th Cir. 1975); cf. Spangler v. Pasadena City Board of Education, 311 F.Supp. 501, 519-20 (C.D.Cal. 1970). These cases are instructive in that they recognize how seemingly “neutral” groupings can perpetuate discrimination, but the circumstances of the present case make judicial intervention even more compelling. First, unlike in the above cases, the bases of the “ability group” have been impeached, and second, we do not have a simple ability group, but rather special classes that necessarily doom the children in them to an inadequate and inferior education, not merely a less advanced curriculum. . See section III.E.l, supra, for more details. . See section III.A., supra. . As noted by Professor Brest, the timing is an important factor to consider: In considering the question, “Why did the decisionmaker adopt this practice?” the court may properly consider the subquestion, “Why did the decisionmaker adopt this practice at this time?” The sequence of events may thus support the inference that the decisionmaker’s objective was to do covertly that which he was forbidden to do overtly. The strength of the inference will also be affected by the tenacity of the decisionmaker’s past commitment to the forbidden rule, the extent to which the innovation marks a departure from traditionally established practices, and the existence of other decisions that seem designed to serve the same illicit objective. [Footnotes omitted.] Brest, “Palmer v. Thompson: An Approach to the Problem of Unconstitutional Legislative Motive,” 1971 Supreme Court Review 95, 123. . One longstanding member of the State Board, which has never taken a position one way or the other on I.Q. tests and disproportionate enrollments (except to support this case), stated that the Board was merely a “facade whose meetings were concerned with trivia.” Sierra 3442-44. See also Schmidt 2846-47. For example, the Board was never told by the SDE — its only “staff” — of the negative reactions of school psychologists to the hasty adoption of the testing requirement in 1969. Krotz 2920-25; Sierra 3434-35. The Board also was never supplied with a copy of the specific results of the important Uslan Report, discussed earlier. Brinegar 4782-84. The picture that emerges, therefore, at least in this area of concern, is of a Board that has had little to say itself, essentially delegating its responsibilities to the SDE. . None was filed, however, for the 1971-72 academic year, and there was a ten-month delay between the completion of the 1972-73 ethnic survey and its presentation to the State Board in April 1974. . One further and more recent example should be noted. Under the 1974 Master Plan the Board was required to get involved in the search for nondiscriminatory placement procedures. 1974 Cal.Stats. ch. 1532 (former Cal. Educ.Code § 7021(e)): (e) Any diagnostic procedure used for placement of individuals with exceptional needs shall be appropriate to the individual’s ethnic, cultural, and linguistic background. All such procedures shall be approved by the board. No procedure shall be approved by the board which discriminates against any minority or ethnic group or fails to account for the home experience and environment of minority and ethnic groups. The SDE did no work in this area, and the Board took no action. See, e. g., Brinegar 4630-32. The 1977 Master Plan enactments for no apparent reason deleted this provision. Cal.Educ.Code § 56337. . In the words of Professor Brest, defendants have committed “the constitutional accounting error of treating the infringement of a constitutional right as a benefit rather than as a cost.” Brest, “Reflections on Motive Review,” 15 San Diego Law Review 1141, 1144 (1978). . See sections V.A.4. and V.A.5., supra. . We do not believe that the evidence shows violations of California statutory law, as plaintiffs contend. The laws relied upon by plaintiffs — CaI.Educ.Code §§ 56302(c) and 56504-are a definition and an expression of legislative intent. They are useful in that they show that the policies of the SDE and the State Board contradict the interests proclaimed by the legislature in culturally fair tests and in E.M.R. classes that are not overrepresented with minority students, but we do not think they are enforceable in this forum. . Defendants argue on the basis of a footnote in Johnson v. San Francisco Unified School District, 500 F.2d 349 , 352 n.2 (9th Cir. 1971), that this court should abstain from deciding the California constitutional claim. Since plaintiffs have consistently pressed this legal theory, and since the case law is now fairly well-settled in California on the question of de facto versus de jure segregation, we see no reason to abstain. . See notes 5 and 93, supra.