Full opinion text
OUTLINE I. Background of the Cases...................... II. Jurisdiction................................. . III. Historical Background......................... . A. Tattnall County.......................... B. The School System......................... C. The History of Segregation in Tattnall County .. - D. The Tracking System....................... IV. The Test................................... . A. Its Inception............................. . B. Its Implementation ....................... . C. The Racial Impact of the CAT Requirement------ D. The CAT Itself........................... . E. The CAT in Tattnall County................ . P. The Benefits............................. . V. Special Placement............................ . VI. The Legal Conclusions.......................... A. The Racial Discrimination Claims............ . 1. The Equal Protection Claims............. . 2. Statutory Claims....................... B. The Due Process Claims.................... . C. The Claims of Those Students Classified as Mentally Retarded ............................ 1. The Effect of the Diploma Policy on Children Accurately Classified.................... 2. The Effect of the Diploma Policy on Those Misclassified as Mentally Retarded......... 3. The Equal Protection Claims of the Handicapped .............................. . VII. Summary................................... . ORDER B. AVANT EDENFIELD, District Judge. I. Background of the Cases, the Claims and Causes Involved The impetus for this litigation was the institution by the Tattnall County School District of an exit examination. Beginning with the graduating class of 1978, all diploma candidates have been required, in addition to other already existing diploma requirements, to achieve a grade equivalency score of 9.0 on both the mathematics and reading sections of the California Achievement Test (CAT). Claims were initially raised in an administrative proceeding which progressed from the local school board, to the State Board of Education, and ultimately to the Georgia Court of Appeals. The issues raised in that proceeding were: (1) whether the Tattnall County School Board had the authority under state law to adopt an additional graduation requirement; and (2) whether a violation of the equal protection clause had occurred since a greater burden was placed on the students of Tattnall County than on the students in other Georgia counties. The first suit in this Court challenging the exit exam, styled as Wells v. Banks, CV478-138, was then filed asserting due process and equal protection claims as well as the state law claims. An additional claim concerning irregularities in the School Board voting districts was voluntarily dismissed by the plaintiffs before trial. While the case was pending here, the authority of the School Board to require the examination was upheld in the Georgia courts. The Georgia courts also found no equal protection violation in the fact that children in other counties were not subjected to the requirement. There were no black plaintiffs remaining in the case in this Court at time of trial. The case was restyled Anderson v. Banks. In addition, plaintiff’s counsel adopted in his proposed pretrial order the outline of legal issues set forth by plaintiff’s counsel in Johnson v. Sikes, thus further narrowing the scope of issues in this case to the due process claim. Since Anderson v. Banks is a suit for damages, it was agreed among all counsel and the Court that liability only would be tried presently and a separate trial on the issue of damages would be held later if necessary. 2. The second action here consolidated for trial was filed in October, 1979, by Kathy Norris Johnson. In October, 1979, she moved to proceed in forma pauperis and to consolidate her case with Wells v. Banks. The motions were granted. The Court certified the following classes: 1. All black children who have attended, are attending, or will attend public schools in Tattnall County, Georgia, and who have completed, will complete, or are eligible to complete all valid and legal requirements for receipt of a high school diploma established by Defendant Board of the Georgia State Board of Education,' but who did not or will not achieve a particular score on the California Achievement Test, and who, as a result of having failed to achieve a certain score on said test, have been or will be denied a high school diploma by Defendants. 1. a. All black children who have attended, are attending, or will attend public schools in Tattnall County, Georgia; and who have completed, or will complete all requirements for receipt of a high school diploma established by Defendant Board or the Georgia State Board of Education, other than achieving a particular score on the California Achievement Test, and who, solely as a result of having failed to achieve a certain score on said test, have been or will be denied a high school diploma by Defendants. 2. All children who have attended, are attending, or will attend public schools in Tattnall County, Georgia, who have been or will be classified by the Tattnall County School District or Board as Educable Mentally Retarded, and who have been or will be foreclosed from receiving a high school diploma by Defendants because of said classification. The issues in the Johnson case are as follows: (1) whether the exit exam violates the equal protection clause of the Fourteenth Amendment; (2) whether the diploma requirement is violative of due process; (3) whether the policy violates Title VI; (4) whether the policy violates 20 U.S.C. §§ 1703,1706; and (5) whether the policy of excluding educably mentally retarded (EMR) students from the possibility of obtaining a diploma violates Section 504 of the Rehabilitation Act of 1973. The case was brought to trial on August 7, 1980. After a thorough analysis of the issues, the Court sent a copy of its proposed order to the parties for comment. The Court then reworked the order in light of these very helpful comments. The Court has been very cautious in this matter which is of vital importance to the educators of this state. The recent opinion in Debra P. v. Turlington, 644 F.2d 397 (M.D.Fla.1981), has caused the Court to reconsider major portions of its reasoning. II. Jurisdiction This Court has jurisdiction pursuant to 28 U.S.C. § 1343(3) and (4) over plaintiff’s claims asserted under the Fourteenth Amendment and 42 U.S.C. § 1983; Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d, et seq.; the Rehabilitation Act of 1973, 29 U.S.C. § 794; and the Equal Educational Opportunities Act, 20 U.S.C. § 1701, et seq., because each constitutional or statutory provision provides for the protection of equal or civil rights. Cases interpreting the Rehabilitation Act of 1973 have noted the strong similarities in statutory language and legislative history to civil rights legislation. See University of Texas v. Camenisch, 616 F.2d 127 (5th Cir. 1980), review granted, 449 U.S. 950, 101 S.Ct. 352, 66 L.Ed.2d 213 (1980), vacated and remanded on other grounds, —— U.S.-, 101 S.Ct. 1830, 68 L.Ed.2d 175 (1981). Jurisdiction over the Equal Opportunities Act claim is also found under 20 U.S.C. § 1708. Defendants have strenuously argued that the plaintiffs’ case is basically a challenge to an academic program and academic regulations, which, in the absence of a bad motive, are not subject to judicial scrutiny. As support they cite Mahavongsanan v. Hall, 529 F.2d 448 (5th Cir. 1976). Yet in Mahavongsanan the Fifth Circuit ruled that plaintiff had no action on the merits of the case. The Court did review plaintiff’s arguments and decided that no rights of the plaintiff had been violated. In underscoring the reluctance of the courts to review purely academic decisions, the Fifth Circuit did not suggest that the district court should have refused to hear the case at all because it concerned academic matters. A substantive rule of law rather than a jurisdictional rule is announced there. In addition, the diploma of a graduate student rather than a high school student was at stake. Arguably, a high school diploma represents a very different interest, in light of compulsory attendance laws and the realities of our society which would not be applicable to a graduate degree. In addition, a large portion of plaintiffs’ claims here rest on allegations of racial discrimination which, even in an academic setting, are clearly subject to judicial scrutiny. See United States v. Gadsden County School System, 572 F.2d 1049 (5th Cir. 1978). Insofar as the claims allege that particular statutes giving rise to private litigation have been violated the claims are clearly justiciable. Insofar as the due process claims are concerned, in the absence of a protectible liberty or property interest, defendants’ argument may have some validity. In that case, plaintiffs fail on the merits. So while agreeing with defendants that, in general, purely academic decisions are not within the province of the judiciary, as a threshold consideration, discussion of this point adds little. III. Historical Background A. Tattnall County Evidence was submitted concerning the history of race relations in Tattnall County and the sociological makeup of the area. The Court admitted evidence of this nature for the purpose of the determination of whether discriminatory intent played a role in the implementation of the exit exam policy and for the purpose of evaluating the Board’s prior efforts in desegregating the schools. However, the Court is also mindful that for the most part the social and racial history of Tattnall County is a history shared with all of the South and standing alone sheds limited light on these recent events. Tattnall County is located in the coastal area of Georgia. In 1970, its population of 16,557 was 30.8% black. While the population of Tattnall County as a whole is poor with more than 35% having incomes below the poverty line, the black population is poorer according to the 1970 census. More than 64% of the blacks had incomes below the poverty line. Median income for blacks lags behind that of whites. This state of affairs is not peculiar to Tattnall County. B. The School System The Tattnall County Board of Education is an elected body which conducts the business of the Tattnall County School District. The superintendent is Ben F. Sikes who is also elected. Mr. Sikes was superintendent at the time the Tattnall County School District abolished separate schools for whites and blacks in the 1970-1971 school year and has held that post to the present. The Tattnall County Board of Education operates two high schools, Reidsville High School and Glennville High School. These schools are located in the towns whose names they bear. Reidsville is also the location of Tattnall Elementary School which contains classes from kindergarten through sixth grade. Two additional schools are located in Glennville. These are Seckinger Primary School, housing grades kindergarten through fourth grade, and Glennville Middle School which contains grades five through eight. A third primary school is located in Collins, the Collins Junior High School, which, despite its name, houses grades kindergarten through eight. The Tattnall County School District also operates the Ben Sikes School which is a separate school for the trainable mentally retarded students. C. The History of Segregation in Tattnall County Plaintiffs introduced evidence to illustrate that racial prejudice was pervasive in Tattnall County prior to and at the time when the Tattnall County School District abolished the dual system. For example, newspaper clippings did indicate a less than substantial dedication to the equality of the races. However, such items are only marginally helpful in assessing the conduct of the School Board in implementing the exit exam policy fifteen years later. The emergence of Tattnall County into the era of integration was not an easy transition. The schools were totally segregated until 1965 when the School Board agreed to introduce a Freedom of Choice program. Four black students applied for admission to the all-white high schools. These were Gail Strong, Roxie Ann Smiley, Johnny Harrison, and Ronald Johnson. At the School Board meeting which took place on September 7, 1965, directly before the plan was to be implemented, the principals were asked to talk to their students and explain “the situation they are facing, stressing the fact that there will be no preferential treatment for any child.” School Board Minutes, Plaintiffs’ Exhibit 58. One of these students, now Mrs. Gail Strong Small, testified that she believed that she and her friend Roxie had been chosen to be the ones to integrate the all-white high school. Mrs. Small had graduated from Collins Elementary in 1962, a black school that was closed in 1970 when the dual system was abolished because it was substandard. Mrs. Small was attending Tattnall Industrial High, the all-black high school at the time of her decision. However, she was induced to change her mind about changing schools. She and her friend were informed at a meeting at Collins High School that because their scholastic training had been so inferior, they would have to be put back a year, or perhaps two, even though Mrs. Small was an honor student at Tattnall Industrial High School. The students decided to remain in the black school. Mrs. Small testified that during this period she was subjected to harassing phone calls and that strangers shot at her house. Miss Ophelia McIntosh testified to the condition of Collins Elementary School throughout this period. Classes were crowded with 40 to 60 children in a classroom. The books were old and written in. There were never any maps or other teaching aids. Mrs. Small testified to the poor plumbing facilities and the lack of playground facilities. Another black child who attempted to attend a white school under the Freedom of Choice plan was Ronald Johnson. His father testified that during the period when Ronald attended a white school under the Freedom of Choice plan shots were fired into the house from passing cars. Harassment culminated in the burning of the Johnson’s barn which destroyed their livestock. The Federal Bureau of Investigation began investigation and the harassment ceased. Ronald Johnson spent only that school year in the white school. The dual system remained intact in the 1966-1967 school year. On January 20, 1967, the Board voted to forego federal funds rather than integrate the schools. On May 11, 1967, Tattnall County was found to be in violation of Title VI and federal funds were cut off. There were no federal funds received by Tattnall County in the 1968 — 1969 school year. Ben Sikes, the present superintendent of schools in Tattnall County, was first elected to his post in 1968 after having been principal of Glennville Middle School for five years. He was superintendent at the time the dual system was abolished. The final impetus for the abolition of the dual system came from a court order. The Georgia Department of Education was under court order to withhold state school funds to any district which had not submitted its desegregation plan. The letter sent by the Board of Education to the citizenry, dated July 14, 1970, clearly spells out that the reasons for integration were financial. The 1970-71 academic year was the first in which blacks and whites attended the same schools. In the reorganization, Collins Elementary School, the formerly all-black elementary school, was closed as substandard. Mr. Sikes testified that once the decision had been made, he attempted to implement the plan in a positive manner. He called a meeting of all the teachers and informed them that the dual system was at an end and that anyone who felt unable to work in the integrated system “had better go elsewhere.” A university course designed to ease the transition to the unitary system was offered locally but, except for the administrators, few attended. As a measure to aid in the transition to an integrated system, summer school for Title I children was held in the summers of 1970 and 1971. Both blacks and whites attended this program which was financed through Title I funds. Several hundred children attended. However, Superintendent Sikes’ files contained minutes of a meeting with the principals on January 19, 1970, the winter preceding integration. According to the minutes, the principals were informed that attendance laws would not be as important in the biracial schools as they were in the dual system. At that same meeting, according to the minutes, it was suggested that two separate tracks be initiated — one for college-bound students and the other for non-college bound students. Even after the abolition of the dual system, the bus transportation to Reidsville High School from Collins, Georgia continues to be racially segregated. Two buses driven by black drivers transport the black students while two buses driven by white drivers transport the white students. Paul Walden, a black student at Reidsville High School, testified that Mr. Rhodes, the principal of Collins Junior High, had instructed the students that they were to ride a bus containing students of their own race. Buses carrying students to extracurricular activities are not segregated. Except for the programs discussed above, no special effort was made to insure that the establishment of integration in the schools would be successful. The black students who had attended the substandard black schools were incorporated into the schools without any effort at that time to provide remediation. Dr. Spencer, plaintiff’s expert on child psychology and development, testified that it would be very difficult for a child to overcome his early training in a substandard segregated school. In addition, if a child perceives himself as inferior and a low achiever, he tends to divorce himself and his energies from the source of the unpleasant information. Thus, as Dr. Spencer testified, if a child is exposed to those who believe black is inferior or if he is exposed to those who brand him as a low achiever, he is likely to protect his self esteem by ignoring the source. Here, the black children attended inferior schools before they were permitted to attend integrated schools. Both factors worked against the black children and tended to retard their achievement. They had received a poorer education in the segregated schools and they were perceived by themselves and others as inferior. A possible example of this phenomenon is Kathy Johnson, a named plaintiff in the litigation, who had a nonverbal IQ score which fell from 112 to 58 during her school career in Tattnall County. D. The Tracking System While a remedial program for the black students was not instituted, a tracking system was adopted in 1970-1971 school year. Defendants characterized this grouping as by achievement rather than by ability but as plaintiff’s expert, Mrs. Bryan, testified, the distinction between achievement grouping and ability grouping is an illusory one since it is impossible to measure raw ability. Tracking was employed in the elementary and junior high schools with the exception of Collins Junior High School which employed heterogenous grouping. The tracking system often resulted in racially identi-, fiable classrooms. For example, at Seckinger Elementary School in 1971, while the first grade classes were equally proportioned, the second grade includes a 100% white class, a 100% black class, and an identifiably white class. The Fourth grade which consisted of five classes included all white classes and two predominantly black classrooms. In 1972, almost half of twenty-four classes significantly departed from the one-third black, two-thirds white makeup which one might expect. Racially identifiable classes persisted at Seckinger until 1979-1980 when the school system abandoned its tracking system as a result of an investigation by the Office of Civil Rights. Similar patterns consisting of a large number of racially identifiable classrooms occurred in the other Tattnall County schools which employed the tracking system. Mr. Sikes testified that it was not intended that the black children be placed in the lower tracks — it simply turned out to be the case. That a disproportionate number of blacks were enrolled in the slower moving classes is even more surprising in light of the IQ data which came to light during the litigation. In examining the files of the students who have had to meet the CAT requirement in order to graduate, it was discovered that some files contained IQ scores from 1968-1969. There were 137 scores in the files. Ninety-five of the scores were for white children. Defendants strenuously objected to the use of these scores as a basis for any conclusions in the case on the grounds that the test given was a group test and was therefore unreliable and on the basis that there was no way to judge whether the sample composed of the students for whom early IQ scores are available is representative of the population of students who were subjected to the CAT requirement in order to graduate. Ms. Driver, a psychometrist employed by the Tattnall County School District, testified that she did not believe that group scores were reliable and' would make no decisions based on group scores. Dr. Leon Hurley, plaintiff’s expert, testified that while he agreed that the placement of an individual child should never be based on an IQ obtained from group testing, his opinion was that the scores were reliable for the group. He also testified that an individual child was likely to do better in an individual testing situation than in a group testing situation. The Court agrees with Dr. Hurley that the group scores are reliable on a group basis. The problem of the representativeness of the sample for whom IQ scores were available presents a more difficult problem. The number of children for whom scores were available, according to plaintiffs’ statistical expert, Mr. Cole, was numerous enough for meaningful statistical analyses to be performed on the data. But the fact that the data are numerous enough does not necessarily indicate that the data are also representative. However, support for their representativeness comes from the data themselves. When the individual IQ scores were plotted, they were normally distributed with a mean of 102, almost exactly duplicating the distribution in the general population. Before the Court, then, is the situation in which the distribution of scores for the 137 students whose files contained early IQ scores is known. The scores of the 137 students appear to be representative of the scores of the general population. If the 137 scores are not representative, then the early IQ scores of all the students who had been subjected to the CAT graduation requirement are not similar to the scores of the population at large. Instead of believing that the missing scores differ from both those of the entire population and the Tattnall County students for whom scores are available, it is more likely that the scores which are available are representative of those of the group scores which are missing. Analysis of these early IQ scores according to Mr. Cole and Mr. Huberty, defendant’s expert on statistics, demonstrated that race and IQ were not related. The Court therefore concludes that the black and white children who were later subjected to the CAT diploma requirement began their academic careers with generally equal abilities. In view of this fact, the disproportionate number of black children in the lower groups is difficult to justify. Mr. Sikes testified that children were not placed in the lower tracks by IQ scores. Achievement, not ability, was the criterion used for placement according to Mr. Sikes. However, even ignoring the difficulty discussed above of distinguishing meaningfully between achievement and ability, it appears that the black students in Tattnall County have not been placed fairly according to any objective criterion. The Court here relies on the factual findings of the Office for Civil Rights (OCR) based on its investigation of student class assignments conducted December 10-13, 1979, after a complaint had been filed alleging that the grouping practices violated Title VI of the Civil Rights Act of 1964. Defendants’ present tracking policies were implemented in the ,1972-1973 school year. Before integration was implemented, achievement grouping was not practiced. The primary determinant for placement of children entering the first grade is the Scott Foresman Initial Placement Score. Possible scores range from 0 to 57. Teacher judgment also plays a heavy .role in placement. The CAT given to the kindergarten children in the eighth month of their kindergarten year plays a role primarily in the placement of borderline cases. The children were placed in one group for all subjects. OCR found that black students were over-represented in the lower achievement groups and under-represented in the higher groups. This occurred in many cases despite rather than because of the scores achieved by the children. There were no absolute cutoffs for particular groups and the over-lap worked to the disadvantage of the black students. At Seckinger Elementary there were many examples of black children who had been placed in lower groups than white children who had the same scores. Black children were placed as many as three classes below their white counterparts with the same scores. At Tattnall Elementary School, first graders were placed by adding their Scott Foresman scores to their CAT scores. However, the only use which was made of the composite scores was to make an absolute cutoff to determine the A and B groups. Those falling below the cutoff were placed in C, D, and E groups essentially without additional regard to their abilities. These grouping practices at Tattnall Elementary resulted in racially identifiable classrooms on the first grade level and the pattern of racially identifiable classrooms continued in each grade. The tracking system as employed moved the lower groups through the same material which was utilized for the high achievement groups. No particular remediation or enrichment was offered. Students were promoted on the basis of their chronological age rather than their having successfully completed the requirements for a particular grade. Plaintiffs bolstered their attack on the tracking system with statistical evidence. Two types of statistical analysis were used throughout the trial. The first type of analysis compares the actual distribution of data with its expected distribution. A chi square is the term used to describe the analysis of two variables, each of which is divisible into two mutually exclusive categories. A particular combination of these variables is compared to the expected value. A distribution is considered statistically significant if it would occur no more than five times out of one hundred. This .05 is the level of statistical significance. The second type of analysis is the correlation or linear regression. This analysis seeks to determine the existence of a relationship between two continuous variables or between one continuous variable and two mutually exclusive categories. This relationship is expressed by “Pearson r.” Pearson r measures the strength of a relationship. If the value of Pearson r is squared (“r squared”), the resulting figure indicates the extent to which one variable is dependent on or can be predicted by the other variable. Another way of expressing this type of relationship is the probability or “P value.” In order to be statistically significant, the P value must be less than .05; in other words, in order to be considered statistically significant, the relationship can be expected to occur by chance fewer than five out of 100 times. Plaintiffs determined for each child for whom an early IQ score was available, the “race percent average.” This figure was obtained by figuring the percentage of black children in a particular student’s classes each year after the dual system was abolished and then averaging these figures. Dr. Huberty, defendant’s expert on statistics, at one point in his testimony, strongly disagreed that race percent average was a proper variable because it lacked independence. Dr. Huberty argued that since all of the children in a particular class would share the same percentage for that year, the final figures for the race percent average would not be independent — many children would necessarily share the same numbers. Dr. Shapiro, plaintiff’s expert, disagreed that the data were not independent and argued that if Dr. Huberty was correct, the data for race percent average when plotted should line up in columns. When Dr. Shapiro plotted the coordinates, the data were scattered. Thus, the Court concludes that race percent average is an appropriate variable for use in their correlation studies. Plaintiffs’ experts determined the correlation between the early IQ scores and race percent average for blacks and whites. The results support the inference that the black children were not placed according to their abilities while the white children were. For white students there was a significant negative correlation between their early IQ scores and being in a predominantly black class. In other words, for white children, the higher their early IQ scores, the fewer black children were likely to be in the same classroom. The actual data respecting this comparison were likely to occur by chance fewer than once in a thousand instances. However, for black children, the correlation between early IQ scores and being in a predominantly black class was not significant. Thus, even if one posited a generally lower level of intelligence for the black children, an assumption which the evidence in no way supports, one would expect that high IQ black children would have fewer black students in their classes. But the correlations run on the actual figures support the conclusion that, while early IQ scores were not correlated to race, black children were placed disproportionately in the lower levels, placements which were not supported by early IQ scores or the achievement tests administered. Achievement grouping was not attempted at Collins Junior High which used a system of stratified homogenous grouping. Since two of the elementary schools in Tattnall County, Seckinger and Tattnall Elementary, profess to group by achievement levels, a comparison can be made between the performance of the students who were grouped and the performance of those who were not grouped. Two objective bases of comparison exist in the Georgia Criterion Reference Test and the CAT. A criterion referenced test is one which tests specific skills which the students presumably have been taught. Such a test differs from a norm referenced test. A norm referenced test compares the performance of the test taker to the performance of the test-taking population. The score is essentially a relative one. An individual score on a norm referenced test indicates how overall performance compares to that of the group on which the test was normed. The score does not indicate what particular skills are possessed by the student. A criterion referenced test indicates which specific skills have been mastered by the test taker. While a criterion referenced test tells only the skills which have been mastered, obviously these results can be compared between schools or between school districts. The Georgia Criterion Reference Test (CRT) is given to many Georgia school children every year. This test is constructed to evaluate students’ mastery of particular concepts appropriate to the students’ particular level. Plaintiffs submitted CRT results for eighth graders in Tattnall County. Illustrated by means of bar graphs, plaintiffs’ exhibits compared the average results achieved in 1977, 1978, and 1979 by the eighth graders in three schools; Collins Junior High, the non-tracked elementary and junior high school; Glennville Middle School, the tracked middle school which is fed by Seckinger Elementary, also tracked; and Reidsville High School which is fed by Tattnall Elementary School which is tracked. In the mathematics sections of the CRT in the category of sets, in 1977, the three schools scored closely together with Reidsville slightly ahead of Collins which was slightly ahead of Glennville. However, in 1978 and 1979 Collins was dramatically ahead. In operations, Collins was well ahead for all three years. In measurements, Collins was ahead in all three years. In probability and statistics, Reidsville was slightly ahead of Collins in 1977 but Collins was the leader in 1978 and 1979. In relations and functions, Reidsville was slightly ahead in 1977, Collins was ahead in 1978 and 1979. In geometry Collins was dramatically ahead in all three years. See Plaintiffs’ Exhibit 122. The reading areas of the CRT appear to reflect similar results for the eighth graders in the three schools. In word recognition, Reidsville High was slightly ahead of Collins in 1977, but Collins was dramatically ahead in 1978 and 1979. In comprehension, Collins was well ahead in all three years. In language usage, Collins was ahead in all three years. In study, Collins was dramatically ahead for all three years. Defendants did not contradict the above stated evidence but attempted to refute the inference that students learned more efficiently when they were not grouped by achievement' by submitting additional evidence. One such item of evidence was the drop in scores of the eighth graders at Collins in 1980 when a new edition of the CRT was introduced. In addition, students at Collins are using a different curriculum based on different teaching materials. Defendants introduced comparative scores made on the CRT for second and third grade students in 1979 and 1980. The pattern apparent in the scores for both the reading and mathematics sections is that the second graders at Collins do relatively poorly on the CRT while third graders at Collins do very well. The suggested explanation of this discrepancy in the scores of the Collins students offered by plaintiffs was that the different curriculum used by Collins introduced particular concepts at different times than the curriculum used at the other two schools. This explanation makes a great deal of s.ense. Although the eighth grade figures appear to support strongly the premise that the non-tracked children out-perform the tracked children, the sixth grade CRT scores apparently demonstrate the opposite. The Collins students performed poorly. Mr. Akins, principal of Tattnall Elementary School, performed research and came up with some interesting figures. Reidsville High School graduates students from both the non-grouped Collins Junior High line and the grouped Tattnall Elementary line. He determined that more black students who had progressed through tracked programs achieved 9.0 on the CAT. The precise figures for black students meeting the CAT requirement were as follows: 1980 — 80.5% of tracked, 66.7% of non-track e¿l; 1979 — 81.8% of tracked, 63.6% of non-tracked; 1978 — 77.8% of tracked, 66.7% of non-tracked. On the basis of all the evidence recited on the merits and non-merits of achievement grouping, I am unable to reach any firm conclusion on whether such grouping is beneficial. The strong showings of the Collins third graders and eighth graders on the CRT appear to demonstrate the benefits of non-grouping and the sixth grade scores and experience of ex-Collins students in passing the CAT appear to show otherwise. The differences in the curriculum used at Collins further cloud the question. Throughout the comparisons, Seckinger, at which achievement grouping is practiced, generally performed poorly, and more significantly, not similarly to Tattnall Elementary where achievement grouping was also practiced. Tattnall County operates only three elementary schools. Since the non-tracked school also uses a different curriculum, I find it impossible to accurately evaluate the effects of achievement grouping. In addition, since children were not grouped in a fair manner, it is difficult to reach any conclusion on the operation of a fairly administered tracking system. I see no basis, however, for concluding that achievement grouping as practiced in Tattnall County has obvious benefits. IV. The Test A. Its Inception The diploma requirement in Tattnall County that each graduating student must perform at ninth grade level in the mathematics and reading portions of the California Achievement test was first conceived by Mr. Akins, principal of Tattnall Elementary School, and Mr. Russell, principal of Reidsville High School, on an automobile trip. They were concerned that some of the schools in Tattnall County were considered substandard by the State of Georgia. At that time, there was no formal existing procedure to identify those who were not performing at grade level in high school. No remedial programs existed and there was no method of requiring students to enroll in particular courses in high school to improve their basic skills. Mr. Akins and Mr. Russell passed their idea concerning competency testing onto other teachers and the idea was eventually presented to Superintendent Sikes and the- School Board. The School Board adopted the policy in Spring, 1976, but did not impose the diploma sanction until Spring, 1978. Previous to the implementation of the CAT requirement, receipt of a diploma had been conditioned on successful completion of a certain number of credit hours and sufficient attendance. Superintendent Sikes and the School Board did not perform some of the preliminary steps suggested as appropriate by plaintiffs. They did not examine the Technical Bulletin published by CTB McGrawHill, the test publishers. They did not perform a local validation of the CAT. They did not specifically address the effect that the CAT would have on students in the lower tracks or students who had previously attended all-black schools. However, Superintendent Sikes testified that he did look up the test in the Mental Measurement Yearbook, the authoritative source on testing according to plaintiffs’ expert on testing, Mrs. Bryan. That source gave a favorable review to the 1970 edition, including a favorable review by Mrs. Bryan. When questioned about the choice of test, Superintendent Sikes testified that the CAT was chosen because the school district had had good experience with the test. A norm referenced test was chosen rather than, for example, a particular score on the CRT, because the Board wanted to graduate students who compared favorably to national achievement levels. It was felt that in light of the mobility of contemporary society, the school system should prepare its graduates to compete on a national scale. The school system had had success with its Title I programs. By administering the CAT during the high school years, students who needed remedial help could be identified and required to attend remedial classes in reading or math as indicated. By employing the diploma sanction, the Board and administration hoped to spark student motivation in both the sense that consequences attached to failure to achieve and that the diploma, for the first time in years, would meaningfully represent a particular level of achievement. Mrs. Bradley, head of the math department at Reidsville, testified that she and others believed that part of the students’ poor achievement was due to an overwhelming lack of motivation on the part of the students. Although Mr. Sikes testified that he did not expect that every student would pass the CAT and acquire a diploma, there was no evidence that the policy was adopted with racial intent. There was no evidence that defendants actually foresaw that the policy would work to the disadvantage of the black students. The testimony of the School Board members was that the policy was intended to improve student performance and not to discriminate in any way. The Court agrees with plaintiffs in that had the Board and administration directed their attention to the possible racial impact of the policy, it would have been apparent that considering the history of the dual system in Tattnall County, a disproportionate number of blacks would be denied diplomas. Thus, the policy adopting the CAT diploma requirement and establishment of remedial classes was accepted by the School Board in 1976 although implementation was deferred. Evidence was conflicting on when students and parents were officially notified of the new policy. Loretta Wilcox, a named plaintiff who was a member of the class of 1978 who failed to meet the CAT requirements, testified that she found out about the requirement when she was in the eleventh grade. James Farmer who failed the CAT twice testified that he found out about the CAT requirement by overhearing a conversation when he was in the eleventh grade. However, Mr. Akins testified that the students were informed at the end of the 1975-1976 school year. This testimony was corroborated by Mr. Greg Moran, assistant principal at Reidsville High School. The Court credits the testimony presented by defendants that parents and students were specifically notified. It is incredible that the School Board would adopt a policy of such importance and fail to notify the parents and students. It is more likely that a particular family might not recall the letter or misplace it. However, the notification still occurred only two years before the diploma sanction was imposed. B. Its Implementation Testing of students was begun at the beginning of the 1976 — 1977 school year. Remedial classes were established for those who failed to achieve 9.0. The same procedure was employed at the beginning of the 1977- 1978 school year. The 9.0 requirement was applied to the class of 1978 and thirty students, out of a class of 219, received a certificate of attendance instead of a diploma as a result of the imposition of the requirement. Seventeen of those who received a certificate were black. In the graduating class of 1979, twelve students, out of 192, received a certificate of attendance for failing to meet the CAT requirement. Ten of these were black. In 1980, six students, out of 192, received a certificate of attendance due to failure to pass the CAT. All six were black. The graduating class of 1978 was exposed only to the 1970 edition of the CAT. In the 1978- 1979 school year, the school system switched to the 1977 edition of the CAT, Form C & D. This test is currently used in the school system. Students who fail to achieve 9.0 on the CAT by graduation are permitted to retake the CAT and to enroll in further course work if they wish. C. The Racial Impact of the CAT Requirement The evidence was overwhelming and essentially uneontradicted that the CAT requirement had a disproportionate impact on the black students in Tattnall County. Mr. Cole, who was qualified as an expert in statistics, testified about the methods employed by a statistician in order to determine whether or not certain data can occur by chance. The statistician begins by positing that no relationship exists between the two variables. The actual values are then compared to the data which could be expected if there were no relationship between the variables. The chi square is used to measure proportionality for two sets of non-overlapping data. For example, a chi square analysis would be appropriate if one were cheeking to determine whether there was a'relationship between religious belief and political affiliation by constructing a square with four cells, one for Catholic/Democrat, one for Catholic/Republican, one for Protestant/Democrat, and one for Protestant/Republican. These actual values would be compared to what would be expected in the absence of any relationship. It is agreed that a value of 3.841 for chi square is the level of statistical significance. The probability value tells the likelihood of the occurrence in the absence of a relationship. The level of significance is .05, which means that a result is significant if it can be expected to occur by chance five times or less out of every hundred instances. Any value equal to or less than .05 demonstrates statistical significance according to established statistical convention. The standard deviation is a measure of the dispersion of data from the mean. The measure of standard deviation is significant at between two and three standard deviations. Dr. Cole, a statistical expert, testified for plaintiffs. According to his calculations, the statistical analyses of the racial differences in receiving or failure to meet the CAT requirement are as follows: Year Chi Squared P. Value Standard Deviation 1978 12.223 .0005 3.496 1979 24.697 .0000 4.970 1980 15.474 .0001 3.934 1978-80 49.020 .0000 7.001 The significant racial impact of the CAT requirement is obvious. Mr. Cole also performed correlation studies on the data. A correlation study, as opposed to those statistical analyses which detect the existence of a relationship, measures the strength of the relationship between two variables. For the purpose of the analyses contained in plaintiffs’ Exhibit 91, “education” denotes the receipt of a diploma. A strong correlation, as noted above, existed between race percent average and meeting the graduation requirement, the more blacks in a student’s classes, the less likely he was to meet the requirement. In addition, for whites, the early IQ scores were positively related (.3541) to success on the CAT. For blacks, there was no significant relationship (.1039) between early IQ scores and success on the CAT — and what slight relationship appeared was in a negative direction for blacks. Those with higher IQ’s were slightly less likely to perform well on the CAT. A multiple regression analysis was also performed by plaintiffs. A multiple regression analysis makes use of the Pearson or correlation coefficient to analyze the roles played by several variables in an attempt to explain a particular phenomenon. The variables are regressed, that is, considered in order of statistical importance. The r2 is the percentage of the phenomenon accounted for by a particular variable. As the variables are regressed in order of statistical importance, the change in the 2%r represents only that amount of variance attributed solely to the newly regressed variable. Plaintiffs’ expert performed such an analysis correlating High CAT Reading and High CAT Math with race percent average, early IQ scores, special education placement and race. Race percent average, the average of the percent of black children in a student’s class from the third to eighth grades, was the most potent predictor of CAT scores. The r squared for race percent average and High CAT reading was .48636. This means that the racial composition of a student’s classroom alone accounted for 48% of the variance in scores on the CAT reading section between blacks and whites. Race percent average accounted for 33% of the variance between whites and blacks on the math portion of the CAT. The Court finds these figures fairly remarkable in that of all the universe of possible influences race percent average is such a potent predictor. Of course, these figures do not necessarily indicate a causal relationship. They do indicate a high correlation. The effect of the racial composition of the classroom on CAT scores is especially remarkable in light of the early IQ scores of the students. Those scores accounted for only 1.4% of the variance in the reading section and .5% of the variance in the math section for the black students. Nor can the IQ factor be perceived as subsumed in the race percent average figure since IQ was not related to placement for blacks. Since, as discussed above, placement for blacks was not carried out on an acceptable, consistent rationale and blacks were often placed in spite of rather than because of their achievement scores, the fact that race percent average can account for such a large portion of the variance is remarkable. D. The CAT Itself The California Achievement Test is a norm referenced test of achievement in reading and math. National norm-referencing means that by testing a large random sample of students and determining what the normal level of achievement is, a basis for comparison is established so that an individual’s scores can be compared to his peers’ across the nation. CTB McGraw-Hill used a stratified random sampling procedure in test development. First, examples of curriculum were solicited from state boards of education and many school districts across the country. Then experts are asked to develop test items after studying the curricular objectives obtained. The test is tried out in a pretest on a group which is not as large as the group used for norming purposes. After the pretest, item analyses are performed. The item analyses include checking for item validity. The concept of item validity is discussed more fully below. The test is then standardized on a large stratified random sample of students. School districts are identified as urban, small, etc. and each region of the country is represented. Within these established strata, groups of students are selected randomly. This method insures that school systems representative of the types of systems existing in the country are represented in the norming group. Although there were no Georgia students included in the 1977 norming groups, students in the Southeast were well represented. For the 1970 edition, 203,684 students from 36 states participated in the standardization. Approximately 200,000 students were involved in the standardization process for the 1977 edition. After the standardization process, grade equivalency scores were calculated. Grade equivalency means the typical performance of a student in the standardization sample for that grade. Although the CAT was not given to the norming group in every month, the scores were extrapolated to give a grade equivalency score for each month. For example, the test might have been administered to the norming group in November and May. The grade equivalent scores for the third and ninth month would actually correspond to the typical performance of those in the norming sample. Steady progress was assumed, according to Ms. Bryan, in order to provide grade equivalency scores for those months in which the test was not actually administered to the norming group. The methods of test construction for the 1970 and 1977 editions were similar. However, the 1977 edition was constructed with more attention to the elimination of bias in the test. The national norms for both editions included proportional representation of minority groups, but in preparing the 1977 edition, particular attention was paid to reviewing items for racial and ethnic bias. Both editions of the CAT were well-received in the testing community and the 1970 edition received favorable reviews in the Mental Measurements Yearbook. Both tests were constructed in accord with the standards for educational and psychological testing developed by the American Psychological Association. Further discussion of the test is aided by the understanding of some terms used in the testing field. The first of these is the standard error of measurement. The concept of the standard error of measurement reflects the fact that it is impossible to say with certainty that a particular score represents the exact achievement level of a particular student. The score instead is indicative of a range in which the student’s real score falls. One can be, according to Ms. Bryan, 67% sure that the real score falls between one standard error of measurement above and one standard error of measurement below the observed score. One can be 96% sure that the actual score is between two standard errors of measurement below and two standard errors of measurement above the observed score. According to Dr. Findlay, defendant’s expert on testing, the standard error of measurement is less important in the situation where the highest score only is of import and a student takes the test repeatedly, since he will not always fall in the lower part of his range. “Validity” in the testing field indicates whether a test measures what it is supposed to measure. The degree of difficulty for a particular item is one relevant factor in assessing the validity of the test. The degree of difficulty is the percentage of children who answered the item correctly. If almost none or almost all of the children answer correctly the item is of little worth since it does not differentiate between the test takers. Ms. Bryan testified that an acceptable range of difficulty is from .20 to .90 with the mean on a multiple choice test of .60. The point biserial is a measure of the validity of a particular item. It measures the relationship between success on a particular item to success on the test as a whole. The point biserial is used in the testing industry as a whole and by CTB McGraw-Hill in order to determine that validity. Ms. Bryan testified that, in her opinion, an item should be discarded if its point biserial is below .30. CTB McGrawHill uses .20 as its cutoff figure. Bias in the test items affects the validity of the test. “Bias” as' it is used in the testing industry means results do not have the same meaning for different groups. In other words, when an item or test is biased it is measuring different things for different groups. This concept is distinct from the possibility that one group may not achieve as well as another group. Except as already noted, no serious attacks were raised by plaintiffs on the test construction and validation performed by CTB McGraw-Hill. E. The CAT in Tattnall County Plaintiffs have taken the position that, while viewed alone the CAT is a respectable test instrument, the circumstances of its use in Tattnall County, even apart from problems created by historical racial discrimination, cause the test to be unreasonable and unrelated to any legitimate educational goal. One of these objections to the use of the CAT in Tattnall County is that defendants have not taken into account the standard error of measurement when assessing a student’s performance. Ms. Bryan testified that had one standard error of measurement been taken into account in evaluating their scores that at least eight students who were denied diplomas in 1978 and 1979 would have graduated. The Court is reluctant to place much weight on this particular factor since the School Board’s policy might be described as requiring a slightly higher level of performance. The 9.0 points represent one standard error of measurement below the actual required score. In addition, since the CAT is administered to the students on several occasions, the Court credits the testimony of Dr. Findlay that the errors in measurement are reduced over successive administrations. Thus, that the announced policy set the cutoff standard at 9.0 when, taking into account the standard error of measurement and the fact of repeated test administrations the actual cutoff point was slightly higher than 9.0, does not seem to be of much legal significance. A second criticism of the CAT’s use in Tattnall County is that no local validation study was performed. A validation study is not normally necessary before a standardized test is used by a local district since such a requirement would negate the reasons for selecting a norm referenced test in the first place. Dr. Shapiro testified that when a test is used for a purpose other than that for which it was designed, it should, according to APA guidelines, be revalidated locally. Mr. Donald Green and Mr. Robert Long, testing experts employed by CTB McGraw-Hill, both testified by deposition that the CAT had not been designed to be used as an exit examination. To some extent, the analysis depends on the definition of “exit exam.” Mr. Sikes testified that the CAT was chosen in order to compare the achievement of students in Tattnall County to the achievement of students nationwide. Essentially, according to the testimony of Mr. Sikes, the CAT was employed for its usual function — to compare children to the norm group. It was this comparison which was then used to make decisions about the students; i. e., whether or not they would receive diplomas. The local validation performed by plaintiffs in preparation for trial clearly demonstrates that the CAT did not perform as well in Tattnall County as it had on the groups selected by CTB McGraw-Hill. While, according to the testimony of Mr. Green, the CAT had been examined for bias by the publisher and biased items removed, the test was not free from bias when administered in Tattnall County. CTB McGraw-Hill utilized a point biserial analysis to determine the existence of bias. The point biserial relates to the student’s performance on the test as a whole or on a relevant subtest. A high point biserial — a strong relationship between doing well on an item and doing well on the test — indicates that the test is valid. The specific standard utilized by the test publisher in rejecting items appears on page 141 of the Technical Manual for CAT 1977: Point biserials under .20 were considered unsatisfactory for either the “Black” or “other” group, and differences between the groups in excess of .20 were considered evidence of bias. Thus, the item was rejected if either group had a point biserial less than .20 or if the difference between the point biserials was greater than .20. Dr. Huberty, defendant’s expert, had a different interpretation of the standard which the Court rejects in light of the plain language of the Technical Bulletin. Plaintiff’s expert, Dr. Shapiro, performed an item analysis on the performance of Tattnall County students on the CAT 1970 and CAT 1977. He defined bias and utilized the point biserial analysis as defined in the publisher’s Technical Bulletin. He found that for students at Reidsville High School, 52.9% of the questions on the reading portion of the 1970 CAT and 69.4% of the questions on the math portion of the 1970 CAT were biased by CTB McGrawHill’s standards. At Glennville, 62.4% of the questions on the reading portion and 62.2% of the math questions on the 1970 CAT were biased. In 1979, both schools used the same form of the CAT 77 and 60% of the questions on the reading portion and 45.9% of the questions on the math section were biased. Examining the scores of the Tattnall County students only for validity, that is, point biserials equal to or greater than .20 for the items, many items do not appear to be valid. Plaintiffs’ Exhibit 95 lists the point biserial calculated. At Reidsville in 1978, more than 50% of the items had biserials below .20. At Glennville in 1978, more than 44% of the items had unacceptably low point biserials. In 1979, almost 42% of the items had unacceptably low point biserials. The Court is uneasy noting these facts. Plaintiffs’ evidence on the bias and validation points was not successfully challenged at trial. Yet it is difficult to believe that the results obtained in Tattnall County can be so different from those obtained by CTB McGraw-Hill. Ms. Bryan testified that there must be significant differences between the students in Tattnall County and the national sample. She did not elaborate on the possible explanation for such a phenomenon. Plaintiffs also challenged the CAT graduation policy on grounds that the CAT did not test what was being taught in the Tattnall County schools. To some extent, plaintiffs appear to be arguing that an achievement test should not be used to determine who can be awarded a diploma. In arguing that a nearly exact correlation should exist between what the stud