EQUAL EMPLOYMENT OPPORTUNITY COMMISSION, Plaintiff-Appellant, v. ATLAS PAPER BOX COMPANY, Defendant-Appellee

3308, 745, 103 S.Ct. 77 L.Ed.2d 987 the U.S. attorney to draw to appointed an (1983). prevailing Certainty of attention. court’s Naturally, this required. appeal on is not squarely within the Lombard falls neces — pos- defendant every not mean that Ohio, does sary protections of Penson v. The Anders appeal. a meritorious U.S. -, 346, sesses 102 L.Ed.2d 300 109 S.Ct. indeed, very level a low California, v. requires, (1988) Anders threshold and 1396, (1967). 738, 18 L.Ed.2d 493 arguability. 87 S.Ct. Thus, any flirting even with the discussion turn, case, in The Anders error in our prejudice language of Strickland’s to opportunity the state court denied analysis unnecessary. harmless error ap- Anders brief adequate an review Penson, I 109 S.Ct. at 354.4 therefore con arguable counsel if it found point new sepa joyfully cur the result write Therefore, preju- any discussion claims. rately express my to that this court belief inap- dice, error is possibilities, or harmless decisively decide on side of should which Lombard situation. plicable Lombard’s and Jones v. Barnes the Anders/Penson v. of Penson purview falls within I find fence Lombard’s case falls. Because presumed prejudice. Ohio’s standard constructively denied that Lombard was appellate the assistance of counsel alto presence appel- of an physical formal presumed prejudice I gether, believe the appellate counsel. attorney is not late standard of Penson applies thereby obviat which attorney filed a document Lombard’s ing any prejudice or harmless the need the mer- contained no arguments going to Clarke, v. error discussion. Accord Evans case. Under Anders its of Lombard’s v. (8th Cir.1989); Sanders 868 F.2d 267 Penson, criminal defendant is entitled Clarke, on F.2d 488 attorney unless an advocate’s brief — remand U.S. from action, seeks to from the which withdraw (1989). -, 102 L.Ed.2d 964 attorney to file an An- requires itself ders brief. appellate document attorney “Appellant’s filed as

Lombard’s essence, An-

Brief”, was, inadequate

ders brief without even the attendant mo- It is inade-

tion to withdraw as counsel. “arguable” quate it set forth no because EQUAL EMPLOYMENT OPPORTUNI claims, allegation conclusory COMMISSION, TY Penson, 109 S.Ct. at 350-51. merit. Plaintiff-Appellant, withdraw, addition, a motion to without placed the undesirable sit- Lombard attorney arguing govern-

uation of his COMPANY, PAPER BOX ATLAS filing attorney ment’s case without Defendant-Appellee. motion to the motion withdraw. Without No. 87-5421. withdraw, court the state Appeals, Court of United States appoint prompted to consider whether to Sixth Circuit. appellate counsel. This dif- Lombard new Lombard and the Penson ference between Argued March 1988. and Anders cases makes Lombard’s predic- Decided 1989. Feb. than Penson egregious even more ament Rehearing April Denied 1989. cry and Anders. Lombard’s case is a far the “case in counsel fails to from argument appeal.”

press particular Barnes, citing,

Penson Jones — U.S. -, Texas, (1967), and Satterwhite inappropriate apply therefore either It is (1988). prejudice requirement of Strickland or the 108 S.Ct. 100 L.Ed.2d 284 — Ohio, -, Chapman Penson v. analysis U.S. 109 S.Ct. harmless error v. Califor nia, 353-54, 102 L.Ed.2d 300 Kamenetzky,

Lawrence J. Calvin Tenn., Williams, E.E.O.C., Memphis, Jef- EEOC, frey (argued), Washing- C. Bannon ton, D.C., plaintiff-appellant. for (argued), Humphreys, Frank P. Pinchak Chattanooga, Moseley, Hutcheson and Tenn., defendant-appellee. for NORRIS, Before and WELLFORD COOK, Judges, Circuit District Judge.* WELLFORD, Judge. Circuit Plaintiff EEOC’s claim this Title VII case1 based on treat- both disparate impact, the ment and former re- quiring an intentional racial discrimination finding. The district court found no basis concluding intentionally that Atlas had potential against discriminated black office employee applicants, and clerical 1184. The focus this case is upon the use of the Personnel Wonderlic screening job appli- Test in such cants, emphasis appeal and the is wheth- er the district court committed error in rejecting regarding EEOC’s contentions disparate impact theory liability light of the evidence. significant

There are several factors that upon plaintiff’s bear contention the chal- lenge employment practices the Atlas question. The first is that the district court found “some evidence that the test was administered in a biased fashion.”2 The second factor is the district court’s finding that “Atlas has never had a black person working Third, in its office.” 25, contrary selected a cut-off score Cook, Jr., *The Honorable United did not Julian Abele make Wonderlic’s recommended bias; adjustment respect possible Judge cultural States District District of for the Eastern adjustments it did make certain for older white Michigan, sitting by designation. blacks; applicants, but not for and a number of more than one seq. 1. See 42 U.S.C. 2000e et § pass chance to take or the test. in 1980 scores recommendations, posi- Two whites were hired for the Wonderlic’s though a score of 21 court even less than district noted tions in clerks, 18 for file typists, 19 for “it is unfortunate that Atlas did not suggested. As a operators telephone upon opportunity that same seize [in choosing high initial consequence of highly intelligent year] to hire this of white point, almost three-fourths cut-off Wilson, applicant,” Beverly applied percent applicants and well over 25, equal during July and scored to another to meet this criteri- applicants failed black who was hired. Two other white white stat- expert testified for Atlas on. An who applicants scored than 25 were who less methodology test favored ed that this respectively. hired in 1981 and in 1982 to one by at least a three white candidate rejected in Three of four black ratio. yearly All 1983 were not even tested. found a fact information was as in Finally, experts who testified all of court, evidence, by finding summarized if district who agreed that statistical the case 1969, hired scored less when the Wonderlic that whites were who evaluated from utilized, 1984, Wonderlic, the date first until than 25 on the four with scores test was significant 20, dis- hearing, would reflect a less than while seven blacks scored parate impact. Atlas contended employees better than these four white but analysis proper period for was were not hired. 1984, produced expert witness-

through findings factual We believe that these period for this opined es who that the data proof presented and other were sufficient significance.3 lacked statistical prima out to show that EEOC made a court concluded that whether The district case of as indicated significant factors were taken or not these court, particularly in by the district view account, produce plaintiff had failed to into single person the fact that not a black was prac- adequate pattern that a or considering in office. In hired to work of racial discrimination was shown. It tice the case further whether rebutted clear, however, that at least one black by plaintiff, made the district court con- not applicant “pass” did the test was in a cluded clerical workforce Furthermore, the district court not- hired. Atlas, “any attempt the one at small as 20, 1971, July applicant, ed that black would have validate the Wonderlic test ... Kendrick, scored a and was hired meaningless.” statistically Hutson, employee, “scored while a white disagree conclusion reached We with the 16, August only 21 1971 test and] [on the relevant sta- by the district court that actually temporary A ... was hired.” information, related tistical data and other in clerk who also scored was hired white itself, statistically noted the court was 1973, employee In white Dicker- October. test meaningless whether or not the after she was son took the test “two weeks question attempted to validated was be hired.” were hired Two whites on a study.” a “formal It is evident that 21, had scores of 17 and and a “black who many years Atlas number of occasions over 1975, In three applicant was not tested.” did not test a of black even number than 25 of the six whites hired scored less and, moreover, many ap- cants hired white (two less). 1976, In the same scored plicants not tested at the time of who were eight thing happened; two of whites hired requisite hiring or scored less than the who test. In scored 20 or less on the Wonderlic required. allegedly Under minimum score 1977, applicant scored a 23 but black Davis circumstances, a defi- we are left with Later, applicant not hired. a white firm that the conclusion (In 1978, nite and conviction who scored 23 was hired. upon weighing of reached the relevant applicant scored 20 was not hired and hired). applicant who scored 19 was factors was error. See Anderson employees Cunningham experts’ and Smith did ed the Atlas inclusion of these two em- 3. Black work, purposes ployees employees not in the as clerical some clerical but office. accept- proper analysis. statistical district court did not indicate whether it 564, 572, City, 470 U.S. job related in the case of clerical hires. Bessemer City of test, 1511, (1985); 1504, 84 L.Ed.2d 518 If a defendant uses such a it must 105 S.Ct. Industries, Inc., Raymark “procedure Yung v. show that used measures Cir.1986). (6th skills, abilities, important knowledge F.2d necessary per are for the successful regarding firm rule is no fixed and There Black Law Enforce job.” formance related to analyzing studies criterion Akron, ment Ass’n v. Officers upon by relied cognitive ability test Cir.1987). short, (6th generally They must be evaluated Atlas. employer testing must show that “important elements of by examination job performance.” is “related to comprise or are rele- work behavior Griggs v. Duke Power 401 U.S. Paper job.” Albemarle Co. v. vant to the 91 S.Ct. Moody, Griggs, (1971). use of the Wonderlic It is sufficient L.Ed.2d 280 general plant and maintenance say study was made at this here to question, a different situation from by any of the ex- particular Atlas office employees. application of the test clerical presented. there was perts Atlas plant employ testing of Similar Wonderlic court’s conclu- for the district little basis generally, employees, ees not office was screening purported test de- sion that the the Court in Albemarle by also considered “more reliable than vice was Paper Moody, Co. success.” In the single predictor (1975). that the statisti- instant we conclude period 1969-84 cal data for the entire summary, we are convinced that the material, that this and and statisti- relevant following compel circumstances of this case comparison and between cal information (1) single a reversal: not a black office regard to and hiring practices with white span worker was hired over a of fifteen properly applications cannot black years office white workers were “meaningless” due to the size deemed hired; (2) 29 of the 121 whites hired did not find, of the Atlas office workforce. We purported required meet the minimum moreover, the evidence based on the entire test; (3) score of 25 on the Wonderlic six-year fifteen-year period rather than the in some workers instances received favor- Atlas to period by used be relevant and adjustments on their test scores able and meaningful regard. in this not; (4) did recommendations as to blacks acceptable typing, fil- minimum scores for Looking statistical data the at the and ing, telephone answering and skills were adopted by Atlas, testing practices as well not used. discussed, previously the other factors including the failure to test and the dis- respect to a remand in We note relatively high regard scores expert by case that witness offered applicants, by indicates to achieved defendant, Hunter, John failed to and visit of bias in the administra- us some evidence inspect and the Atlas office never studied tests. must therefore re- tion of the We nature and content the Atlas clerical and remand the case to the district verse jobs validity and office involved. The was, then, There oth- court. statistical and generalization theory by utilized proof er of racial in the respect expert testimony to this under practices employment utilized Atlas. appropriate. these circumstances is not There indicate was evidence which would Linkage similarity jobs dispute or testing purported of a and use this case must be shown such on site testing may cut-off score of 25 have been a investigation justify application of such pretext mere to exclude blacks from em- theory. ployment in the office. may A test used in other than judgment

We no professional- make this record as those for which it has been whether, in theory, proper ly signifi- use of the validated if there are “no Wonderlic test not be demonstrated to cant differences” between the studied I jobs. CFR unstudied and 1607.4(c)(2). study in this case § January re- On trial court of, analysis of the attributes involved no jected Equal Employ- a contention in, the particular skills needed or the (EEOC) Opportunity ment Commission accordingly groups. There is studied (Atlas) Paper Company the Atlas Box signifi- concluding that “no basis hiring guilty of racial discrimination among lines exist cant differences” of office and clerical workers between 1969 among distinct progression, 1984, in of Title violation VII of the pro- lines of groupings within the studied Rights Act of 42 U.S.C. 2000e Civil § Indeed, gression. study’s checkered seq. et trial court and EEOC fo- compel opposite appear results cused their attention on Atlas’ use of the conclusion. Test, cognitive Wonderlic Personnel abili- Moody, Paper Co. v. Albemarle test, ty screening job applicants. in the 2378-79, 45 L.Ed.2d rejected dispar- The trial court the EEOC’s impact disparate theories ate treatment district court Upon we direct the remand liability. plaintiff has made to consider whether the Chattanooga, Atlas is a Tennessee com- prima case of race discrimina- out a boxes, pany paper, that makes school com- or dis- tion on a basis envelopes, mercial related At- items. parate treatment under the circumstances. plant generally las’ workers are unionized. must then determine The district court contrast, its office are non-un- workers defendant Atlas has articulated whether ion. The from clerical staff numbers twen- reason for sound and reasonable business ty-five thirty to The EEOC’s alle- workers. its actions under the standards of Texas gations solely the clerical relate to staff at Department Community Affairs Burdine, expected Atlas members to whose were variety assignments. perform L.Ed.2d 207 If the defendant has a of work worker, example, justifiable not demonstrated a business ba- As an a who had been practices, particularly light expected in type, sis for its hired to was also to file or disparities pointed shipping necessary. herein out Tr. when assist testing inap- procedures employed judge proce- and the 139. The trial noted that the validity generalization the- plicability hiring employees new were not dures ory, judgment then should be entered for same, in always the that: proper remedy pro- plaintiff EEOC and a positions rarely Atlas advertises clerical vided. rule, agencies. As a employment or uses needed, help brings it for fur- extra REVERSED AND REMANDED when perma- proceedings conformity temporary and offers them help ther herewith. positions they nent if out well. work COOK, ABELE District JULIAN permanent position applicant Each for a Judge, concurring in part, and questionnaire fill expected out a and dissenting part. provide typical ed- information about ucation, Ap- past experience, etc. work my distinguished I concur with col- given typing key- plicants are also a leagues, who have concluded that test, test, punch the Wonderlic However, cause should be it is reversed. However, the record from the interview. my inappropriate that a remand is belief that, many years issue from indicates under the circumstances this case. time, quick judgment is made Moreover, time to I believe that the trial record reject either to hire or and the tests or much crit- mandates a broader review and There is no hiring prac- interviews be omitted. analysis Appellee’s ical complete hiring proce- testing procedures tices and than that provided prevailing dure has truncated from time to which has been opinion. applicants. time for black and white both support hired first kinds of evidence this statement. people have even been

Some First, decides it When Atlas did not follow the Wonderlic and tested later. somebody, usually selects to hire it testing needs manual recommendation that a very applied in the from those who have adjustment score be made on the basis of does not review the earli- past recent possible in order to cultural race eliminate kept on file. applications er testing. Second, although the Com- bias pany scoring adjustments to cer- awarded Paper Box v. Atlas (E.D.Tenn.1987). tain older white on the basis of age, gave adjust- their it no such favorable judge also found that Atlas The trial Third, applicants, ments to black id. Test in began its use of the Wonderlic appli- that ten trial evidence indicated 1969, and added: typing to take the test cants were allowed unsystematic ex- company did some on more than one occasion. No black test, trying it out perimentation with the opportunity. cants were second family members and management, Fourth, there was evidence that twelve rel- employees in the clerical staff whose applications from candidates bore already performance known. ative Fifth, marks of “B” or “Black” on them.1 attempt No was ever made formal significant number of white candidates verify validity of the Wonderlic job performance. positions despite hired for predictor as a clerical having achieved a score of 25 on the test.2 Id. at 1186. pointed He out that judge expressed also The trial his view of the “[t]he management try that it decided would hiring practices net results of Atlas’ for all people hire who scored 25 or better procedures “[ajlthough he stated that positions.” Significantly, its Id. many employed in there are blacks although the trial observed plant, person Atlas has never had a black hiring suggests Wonderlic manual score working (empha- in its office.” Id. at 1185 help, of 25 for it recom- secretarial added). Moreover, sis a former President typists, 19 for file mends scores of Atlas, Robinson, Jr., Harry T. also con- clerks, telephone operators. At- and 18 for Company ceded at trial that the had never relatively high las contended that it set a employee hired a black in the of- work cut off score in order to secure a staff of According fice. Tr. 143.3 to Atlas’ statis- qualified persons extremely who would be tics, persons 148 of 699 took the who productive. consequence, As passed requisite Wonderlic with the Test judge found no evidence that “this initial (27.4%) score of 25 whereas 3 of 52 racially decision motivated.” Id. (5.8%). applicants passed black Atlas’ ex- Hunter, pert, acknowledged John Interestingly, very in the next sentence decision, generally “[tjhere of his he that Wonderlic Test favored the white stated margin. some adminis- candidate at least a 3 to 1 evidence that the test was Tr. tered in a fashion.” Id. Several biased 112.4 hand, important 1. The district did not address this issue. it contends that the issue However, any personnel it should be noted that Atlas denies is whether black worked as non- any markings. responsibility managerial, non-plant employees. for these Atlas identi- Virginia Cunningham fied and Ben Smith as falling category.

2. The number of hires from 1969 to in this Atlas described them as who test, employees black who function more as clericals did not score a 25 or take the was 48% plant though exceptions —not workers—even their work persons hired. All of the were stations were not on the first and third floors of white. Tr. 176-177. building. the administration Atlas asserts any composition that statistics on the racial argues regarding 3. Atlas that the EEOC statistics Cunning- clerical workers should have included employees misleading. Although the office are ham and Smith. acknowledging that were no black em- there ployees with offices on the first and third floor building, of the 4.The issue of whether Atlas’statistics administration Atlas discounts should having any significance. accepted this factor as On the will be discussed infra. the determination of whether a consisted held that testimony at trial Much of the discriminatory intent respective in defendant acted with experts offered their who clearly such deference. The erro evidence. deserved terpretations of the statistical on standard has been extended to dis agreed the results neous All them Test, parate impact e.g. cases. See Atonio v. if evaluated the basis Wonderlic Co., Inc., Packing period, reflected a Core 1969 to 1984 Wards the entire test, impact. Under this significant disparate statistically reviewing out that court must affirm the trial However, experts pointed Atlas’ court unless it is “left with the definitive passing in the rates would any difference a mistake has significance period if the and firm conviction that been lack statistical analyzed. 1984 was committed.” United States United 1978 and between 364, 395, Co., conclusion, they Gypsum Vir 333 U.S. reaching included States this among Ben 92 L.Ed. 746 How ginia Cunningham and Smith S.Ct. ever, legal views of the trial court are staff.5 the clerical appeal subject to a de novo review on even testimony expert introduced Atlas also under Pullman. a certain purportedly demonstrated which Test and between the Wonderlic correlation Supreme Griggs Court in Duke These jobs clerical at other businesses. Power opined sample the small

experts (1971) Paper L.Ed.2d 158 and Albemarle precluded compila- at Atlas work force Moody, Co. v. S.Ct. meaningful job relatedness any (1974) tion L.Ed.2d 280 set forth the burden However, experts study. proof standard in a Title VII assessment, asserting that critical of Griggs case. and Albemarle are compiled by some- had been these studies highly of these instructive because both Company one had neither visited employment cases the exact involved analyzed the tasks involved the plant nor issue here—the Wonderlic Test. More- performed there.6 evidence, over, has the nature of to validate the proffered gist judge’s reasons for trial tests, similarities here. have distinct rejecting the EEOC claim is summarized his assertion that there was Supreme unani Griggs, Court “quali- applicant would have prohibits the mously ruled that Title VII any if employment at Atlas even fied” dis employment tests which have a use any of the test or biased administration employer criminatory effect unless the removed. 680 other area had been showing meets “the burden at 1189. given requirement ... a manifest re [has] lationship employment question.” to the II 432, 91 at 854. In 401 U.S. at Albe marle, explained that: decision of a district court as to The arises, course, only after a defendant has violated Title VII This burden whether complaining party or class has made generally subject appellate to an review the discrimination, prima a facie case of clearly under the erroneous standard. out Swint, i.e., question in has shown that the tests Pullman-Standard applicants promotion for hire or in select Pullman, significantly In the different pattern a racial judge, making only judge explicitly trial never resolved their 6. The trial in the briefest 5. although expert testimony, status as members of the clerical staff the stated comments about correctly concluded that there were no black that, he employees among any extremely "It is difficult to make sense building administrative out of the evidence on the issue of adverse However, office staff. this issue need not be impact.” Id. at 1186. assume, I resolved at this time because will arguendo, that Atlas is correct in its assertion Cunningham and Smith should be counted as clericals. case, applicants.... In the pool from that of the instant the trial court made a ambiguous regarding somewhat statement employer If does then meet the bur- prima the EEOC had met its whether “job proving den of its tests are facie demonstrating disparate impact: burden related,” open complain- to the it remains As the Court understands the statistical ing party to show that other tests or proof case, offered in this has devices, similarly un- selection without prima made a case that the use of effect, also desirable racial would serve screening the Wonderlic test for Atlas’ employer’s legitimate interest in “ef- statistically sig- clerical had a trustworthy workmanship ficient and impact nificant adverse on black showing be evidence ...” Such a would 1978, although ap- cants at least before using employer its tests parently years. recent merely “pretext” as a for discrimination. added). (emphasis at 1188 Be- however, case, we are present In the judge’s “apparently” cause trial use of wheth- concerned with suggests impact that an adverse had not er has shown its tests to be Albemarle during period, shown the later job related. upon point argue seizes this (citation U.S. at 95 S.Ct. at judge impact trial found adverse after omitted).7 addition, Company argues 1978. plain- major difference between a improper it that would have been for the disparate impact proof tiff’s burden of upon any trial court to have relied case and a treatment case is that prior to 1978. proof discriminatory intent need not be Assuming, arguendo, that Atlas can val- regard shown in the former context. With idly point though assert this even it did not issue, Judge to the Pierce any appeal make formal on cross this is- Freeman, Lively, Lynch 817 F.2d sue,9 this decision court trial must wrote: finding be construed as a im- adverse such a the trial court con pact. opinion, In my any ruling consequences cerned with “the of em would be error. ployment simply moti practices, not There why are several reasons such a (ci Griggs vation.” v. Duke Power Co. justifiable. First, conclusion is the trial omitted) original). (emphasis tations judge explicitly determined never Disparate impact typically cases are con EEOC had failed to establish adverse facially practices cerned neutral though after 1978.10 Even he held place standards that fact work to “apparent” post pe- doubts about the disproportionate discrete burden riod, analyzed pre- still group employees protected are text issue. the most reasonable in- under Title VII. terpretation that, of his decision is al- Accord, though Assn. County finding impact prior Kent adverse Sheriffs County, retaining Kent 826 F.2d at 1492-1493.8 1978 and some doubts about the predecessor objective practices 7. Several of our Sixth Circuit Court volved the well known Appeals panels adopted tripartite Griggs issue in and Albemarle. analysis. County See Kent Assn. v. Sheriff’s (6th Cir.1987); County, Kent 826 F.2d prevailing party appeal A can those issues Black Assn. v. against Law specific which were decided it in certain Enforcement Officers Akron, particular the decision on instances if issue OSHA, R.T. was erroneous. Vanderbilt Co. v. — Trust, 728 8. In F.2d 815 Watson v. Fort Worth Bank and -, 2777, U.S. 108 S.Ct. 101 L.Ed.2d 827 (1988), vibrancy the post-1978 period, expert, Court showed the 10. Even for the Atlas' disparate impact approach by extending Geraghty, of the it passing Thomas found that the rate of beyond objective type subjective employ- only tests to was 65% white practices. ment The Court also discussed vari- cants. Tr. 52. This satisfies the 80% adverse rule, However, peculiarities approach impact opined ous used n. 12. he infra regarding subjective practices. significance. This the result lacked statistical case in- that

1495 willing 402, post-1978 period, he to assume 106 S.Ct. 92 L.Ed.2d 315 (1985)(Brennan, proof impact concurring); for the entire J. that adverse Black Law period established on City 1969to 1984 had been Assn. v. Ak Enforcement Officers ron, 483; had 824 assumption that the defendant sat- F.2d at v. Walker Jefferson requirement. Home, 1554, (11th County isfied the relatedness 726 F.2d Cir.1984) (“In case, Second, frag if meant clearly may court pri- consider suggests, proofs ment the EEOC’s as Atlas discriminatory or acts if such evidence is ruling totally such with would independently relevant to show actionable An out foundation and erroneous. exami occurring statutory peri conduct within the for arguments segregating nation of Atlas’ od”); City Guardians Assn New York period that the statistical will demonstrate Commission, 232, v. Civil Service 633 F.2d unjustified each them one of amount to (2d Cir.1980), aff’d, 582, 463 U.S. See, fragmentations Capaci of the data. v. 3221, (1983); S.Ct. 77 L.Ed.2d 866 Harless 647, (5th Besthoff, 711 F.2d 654-56 Katz & Duck, (6th Cir.), cert. Cir.1983), denied, 927, 104 cert. 466 U.S. (1984). Hence, (1980); L.Ed.2d 92 Police Officers for discussing there is no for the adverse basis Equal Rights Columbus, any period for other than 1969 (S.D.Ohio 1985). through toto. In the employ- instant the same Atlas asserts that evidence subse- ment test was from 1969 to 1978 as quent probative to 1978 or 1979 is to this period for the after 1978. The results of original controversy charging par- since the certainly part “ongoing this test are of an ty filed her In claim 1980 policy” period. the entire support position, argues of this relevant, pre-1978 properly results are con- backpay awards Title VII cases are limit- siderable, fully admissible. 2000e-5(g) covering ed U.S.C. § year period immediately prior the two argument, posits another filing charge. Chattanooga Even if system longer Atlas has school preserved appeal, argu- experienced the issue for this segregation effects of af- 1978, and, therefore, ment must fail any pre-1978 because the limitation of a ter dis- plaintiff’s damages periods parate impact to certain does evidence was due to the infe- schooling not mean that evidence of previously discrimination rior which had been prior periods provided cannot be admitted and to black students. evaluated. argument This must also fail it because lacking totally supporting Held Oil 684 F.2d 427 is documenta- Gulf plain- ruled regarding that a tion. No statistical studies tiff Chattanooga system maintain a Title VII action school based were intro- prior incidents which occurred to the 180 duced into record or examined day filing charge Moreover, limit for an EEOC if trial court. no statistics were “subsequent produced acts of identifiable discrimina- which demonstrated the absence tion period any lingering occur within the critical time Finally, effect. the Court provided and are to the related time barred inci- studies which added). (emphasis dent.” Id. at 430 addressed the correlation between test segregated schooling. scores and This doctrine was reaffirmed in the con age Significantly, running text of an discrimination counter to this case Jani argument, following revealing cross ex- Corp., kowski v. Bendix (6th Cir.1987) Geraghty, expert amination of Thomas an “ongoing where discrimi Atlas, place: witness for took natory policy” was the same before and during filing period. QUESTION: Evidence of the you saying Are policy period beyond effect that blacks had achieved a level day filing the 180 limit equivalent was relevant. their to those of white coun- also, Friday, terparts Bazemore v. testing and therefore would not There are sev- not be a therefrom are reliable. race would

not be—their tests, argument. problems serious with this test scores? eral factor in their First, support record does not this time, prob- that by I ANSWER: said full dec- conclusory assertion that almost a ef- the adverse ably my judgment, experiments concededly unscientific ade its had worked prior fect of education Robinson, place. Harry T. Jr. testi- took way through system. 1970’s, early part “in the fied that any studies QUESTION: Has there been evaluating may the test and we we made of this? anybody kind of a have hired to see what may been or There have ANSWER: got type of performance we from that not have been. Thus, testimony his score.” Tr. 160. even QUESTION: you any stud- Are aware testing “early part” limits the to the ies? suggests experimen- 1970’s. This No. ANSWER: only until 1972 or 1973. tation continued any publi- QUESTION: you Do know Second, applicants’ per- black that because of cation that has shown period during formance that earlier is rele- integration and the result of inte- gives evidence that Atlas contin- vant of blacks on gration the scores rely upon ued to utilize and the Wonderlic improved and scored tests Third, despite Test results. parity their white coun- reached nothing in there is the record which would terparts? any cut off other than 25 was indicate seen—I cannot cite I have ANSWER: Finally, if by used Atlas. Atlas is assert- specific I have seen refer- citations. ing that the Wonderlic Test had no real cut improve- newspaper ences use, during years nine of its the arbi- off students. ment of scores trary nature of its exceeds administration parity or they have reached Whether findings the factual of the trial court. testify. I I haven’t— cannot my interest. that’s not reasons, I for all the above believe field of arguments against measuring added). that Atlas’ (emphasis Tr. 53-54 impact of the Wonderlic Test for the Notwithstanding reasons which have period 1969 to 1984 must fail. Since ex- Atlas, there is evidence been advanced perts controversy from both sides of this picked in the as a cut record that 1978 was agree that the Wonderlic Test had statisti- expert off date at the direction of one significant disparate impact cally for the Moreover, attorney. Atlas’ Tr. 62. period, e.g. to 1984 Tr. see inclusion selection this date ensured the (Atlas’ expert impact admits adverse Cunningham sample. in the Tr. 57. An- period), clearly entire the EEOC estab- expert, Mickey Petty, stated prima lished its case.11 that his selection of cut off dates between facie arbitrary.” admittedly 1969 and 1984 “is Complete Transit, In Chrisner v. Auto Tr. 212. Inc., 645 F.2d 1251 conflicting rulings addressed some Finally, argues only it had regard to a experimented prior Test to with defendant’s burden once a the Wonderlic and, hence, prima case has statistics which result been demonstrated. facie (1981); prima under L.Ed.2d The EEOC met its burden Chris rule, Transit, Inc., C.F.R. Complete the so-called 80% Vs ner v. Auto 1607.4(D), sig level of statistical § nificance, and the .05 signifi n. The .05 Grossman, Employ B. and P. Schlei probability cance means that the (1983). The ment at 1372 Discrimination Law disparity occurring by chance is 5%. Based on Employee Government’s Guidelines on Uniform statistics, passing Atlas’ own rate for black Selection Procedures state that adverse 27.4% that of their white presumed pass protect will be rate for the if the counterparts. signifi Tr. 170. The statistical group ed is less than 80% for probability cance of the results was .012 or the group. most favored Numerous courts have that this result occurred chance was 12/iooo. approvingly used or referred this test. See Tr. 171. Teal, e.g. Connecticut v.

1497 case, adopted sliding In that a sort of we Guardians Assn New York City v. Civ of Service, 79, (2d that Cir.1980), scale was based “considerations il 93 jobs 940, which are a function of the demands.” cert. 452 U.S. (1981). Clearly,

Id. at 1262. L.Ed.2d the clerical positions at issue in this require case particular, In this Court held that it stringent scrutiny job of Atlas’ relatedness would be somewhat more lenient in scruti- justifications.12 nizing employment practices jobs for In Black implicate public (e.g., Law safety airplane pilots) Enforcement Officers Akron, Assn v. City spoke this Court positions. Regarding than for other those obligation: a defendant’s jobs require guarantees that did not such safety, quoted we from a Tenth Circuit defendant meet this burden Appeals Court of case and held that: establishing procedure that used important skills, measures abilities, joba requires When a small amount knowledge that are training necessary for consequences skill and and the performance successful job. hiring unqualified See applicant are (1986). 29 C.F.R. 1607.5 insignificant § the courts should examine closely any pre-employment standard or 824 F.2d at 482. This case makes it clear against criteria mi- which discriminates that there must abe measurable correla- case, norities. In such a the employer people tion perform between how on a test should a heavy burden to demon- performance and their job posi- level at the strate to the court’s satisfaction that his tion. employment job criteria are related. In the instant correctly Id., citing, Spurlock Airlines, v. United argues that the trial applied never F.2d Accord legal the correct standards when he con- Corp., EEOC Ball F.2d 541 & cluded that proven job Atlas had related- (6th Cir.1981) n. 20 (Regarding jobs which key ness. section of Opinion, his he do public safety, not affect defendant in wrote: rebuttal must show “that use of the em- While no formal study validation ployment condition or selection device is performed itself, at Atlas it is obvious compelled by necessity”). business any attempt to validate the Wonderl- approach premised

This argu ic test on a clerical workforce as small as ment justification there is less for the one at Atlas would have been statisti- employment difficult cally meaningless. standards where the Much evidence was particularly complex involved is neither to show that cogni- the use offered likely public nor to involve the welfare. ability tive applicant test as an screen- Thus, standards in these more basic ing device is more reliable than require scrutiny. even stricter Other cir single predictor job success. cuits have also drawn this kind of distinc The Court is satisfied that Atlas effec- e.g. Dallas, tion. Davis City tively prima rebutted the case and 210-11 nothing cert. de suggest there was nied, 106 S.Ct. the use of the Wonderlic test was a mere (1986)(discussing L.Ed.2d 656 cases pretext where for racial discrimination. A close diploma requirements were found hiring not to be review of the decisions made over factory related for workers but were many years at issue shows a concert- police officers); related for attempt Walker v. ed qualified to hire the most Home, County 726 F.2d at employees. While there were a few Jefferson (court required); examines the risk applicants and skill met Atlas’ strin- tests, 12. The result would be no different if the less if the which are stringent approach less, advancement, had been used. Neverthe- hire or have been inserted as le- important I gitimate believe that it is employment show the bases for or barriers to skepticism Bartholet, with which the case Application law views em- minorities. See Title ployment Places, relatively jobs. High tests for unskilled VII to Jobs in 95 Harv.L.Rev. 945 always This Court must be zealous to determine Swint, 456 standards, they passed never Pullman-Standard gent qualified U.S. at over less well findings

cants. are infirm Court held that “where law, erroneous because of an view of added). (emphasis proper remand unless the is the course with this rea- problem The most obvious only one the permits record resolution of dis- soning judge is that trial failed to the In the we factual issue.” instant using manner of Atlas cuss whether the already have determined that “cognitive (not the abili- the Wonderlic test misapplied legal regard- judge the standard tests) par- manifestly related to a ty” ing “job However, relatedness.” a remand generally). (not job “job success” ticular validity necessary is not Atlas’ because important a clerical particularly This is generalization theory sufficiently defi- that the job as Given discussed above.13 reasonably cient a trial court could not that wrong for judge applied standards trial the Company’s conclude that the use of the obliged relatedness, was not job the EEOC job tests related even if the correct prove tests less to that with alternative legal by applied.15 standard had been Several impact have been used adverse could appellate courts have held stage of the burden of that: Atlas—the third proof.14 plaintiff pri- Where the has established case, and the fails to ma defendant addition, trial my the it is belief show relatedness of an examina- job the assuming judge also erred discriminatory impact, tion the is- qualified people attempted to hire the most open sues are not for redetermination on Company on the basis of a claim the usually picked person with best test remand. the in this very question issue score. Jacksonville, Nash Consolidated controversy is whether testing procedure, was used which Bullard, also Bunch v. 795 F.2d Atlas, permissible method of is a measur- (5th Cir.1986); Elec v. Western Crawford ing qualifications job under Title VII. (11th Co., Inc., tric 745 F.2d legality legitimacy Since the and the of the Cir.1984). issue, judge Wonderlic Test is at trial Duck, F.2d at 616 n. Harless scores, should not used the which have discussed test validation: test, applicants on were obtained to may employer any use of three logic legitimacy. begs evaluate Such cri- methods validate examination: dramatically. rather This Court validity, validity, terion construct or con- Vukovich, held in Williams v. validity_ validity tent Criterion com- (6th Cir.1983),that, “Ranking can- pares job per- of employees’ evaluations according performance didates to their on a formances with test non-job does scores. Construct related examination not neces- validity degree sarily qualified result in selection of determines to which applicants possess candidates.” traits are con- which disparate judge’s impact 13. The trial not relevant to this assertion these tests are case be- good “more reliable” devices is not the than other cause even a motive cannot excuse a test “manifestly proclaiming same as that the test job with an adverse which is not related. necessity job or a as related" required by virtual attempts The fact Atlas made its best to hire only a case law. A test "qualified” employees is not relevant to this and, relationship minuscule dispute. time, possess relationship same reliable a more than other devices. especially inappropriate 15.Remand is also here complexity of the issues and the trial 14. The trial made a fundamental doctrinal that, judge’s refreshingly candid “It is admission error he ruled that the EEOC had not extremely pretext stage difficult to make sense out of the shown third at the because attempt impact.” had made “a on the issue of concerted hire the most adverse qualified employees." ruling, Id. In so he con- at 1186. The relatedness issues are disparate fused the treatment im- complex. certainly no less pact theories. The nature of Atlas’motivation is analysis of, job. involved no of the important for the Content attributes sidered ability in, validity particular assesses the or the skills needed perform specific cants to tasks job groups. accordingly studied There is performed job. must be concluding signifi- basis that “no among exist cant differences” the lines general intelligence tests Traditionally, progression, among construct validation distinct have been validated such methods instrumentalities groupings pro- because within the studied lines of “necessarily Test measure the Wonderlic gression. study’s Indeed the checkered regardless ability, for an inferred appear compel opposite results Assn. New York context.” Guardians conclusion. Commission, City v. Civil Service 422 U.S. at 95 S.Ct. at 2378-79. referencing (specifically F.2d at 94 n. 13 Supreme specific the Court concluded that test). Gillespie v. State of the Wonderlic findings relating validity to the of one test Wisconsin, 1035, (7th 1040 n. 3 771 F.2d generalized cannot be from that of others. Cir.1985). case, gener validity In this the The Albemarle result is particularly strik studies criterion related alization involved ing jobs appeared the at because issue gener premise validity The studies. very they only similar since involved theory, alization as advocated Atlas’ ex points progression. different Neverthe Hunter, intelligence pert, John is that tests less, any job analyses the absence made always are valid. See also Watkins inadequate. the studies major problem validity The first with a Co., Paper 1159, v. Scott 1188 generalization approach is is radical- that it denied, cert. (5th Cir.), 861, 429 97 Paper ly at odds with Albemarle Co. v. (1976) (where 50 139 S.Ct. L.Ed.2d Moody, supra, Griggs v. Duke Power expert proof produced previously circuit, supra, relevant case law within this jobs studied similar to the one were at Guidelines, and the EEOC all of re- which issue, “unwilling presume the court was quire showing actually that a test is require that all maintenance crafts predictive performance specific job. at a skills”); Dickerson v. United job same validity generalization approach simply The Corp., States Steel F.Supp. 472 1335 dispenses similarity with that or manifest (E.D.Pa.1978), vacated and remanded on relationship requirement. Albemarle and (3d Cir.1980) grounds, 616 F.2d 698 Griggs are particularly important prece- (“Proper job analyses required are if a dents since each of them involved the Won- company jobs seeks to use tests for similar derlic Test. study”).16 in the not included validation Albemarle, Burger Chief Justice validity generalization theory ig- Atlas’ wrote: of Albemarle teachings by imply- nores the jobs A be used other than ing linkage similarity that no or between professional- those for which it has been jobs previously those which had re- ly signifi- if only validated there are “no and those at the searched Hunter Com- cant differences” between the studied pany similarity The jobs. need be shown. was and unstudied 29 CFR 1607.4(c)(2). study simply The this case assumed.17 § Davis, 229, 249-52, Washington distinguished Washington 16. In several cases have on (1976), grounds. e.g., S.Ct. 48 L.Ed.2d these Guardians Assn Newof Service, Supreme accepted Court a criterion related v. Civil F.2d at 92 n. 12. York study being good pre- as that validated a test cases, person training only Atlas cites no circuit court and dictor of how well a for a fire- 17. job training. approves ap man’s would do one district court this accepted although proach, Pegues Employment Court the test it was never v. Miss. State Ser vices, (N.D.Miss.1980), regarding job just aff’d, validated training. the actual — Cir.1983), Yet, (5th Washington readily is distin- 699 F.2d 760 rt. ce guishable from the instant case because it L.Ed.2d in- U.S. Moreover, S.Ct. relating portion Pegues, upon volved an issue to the Fourteenth which relies, require- Amendment —not Title VII case. The is not instructive because the trial stringent plaintiffs proven ments under Title VII are more had not court held that the quality held that overall the work force.” Id. Griggs, background similarity The use of between had not shown their the defendants adopted in Griggs the tests which were job relat- Test to have been the Wonderlic striking. case employ- Griggs plaintiffs ed. The “operating” departments of the ees of the regulations The has also set forth Company. 401 Duke U.S. at Power permit the use studies else- which done de- repeatedly S.Ct. at 851. The Court a different site where validate tests at purported as scribed the Wonderlic Test specific findings of but on distinct “general intelligence” test. Id. regulations similarities the work.18 Court, Moreover, the read: disputing although not the view that (2) similarity. The Job incumbents general intelli- Test measured Wonderlic job and the incumbents in the user’s relat- gence, found it not to have been jobs validity group or on which the ed. study perform conducted substan- behaviors, major tially same work Burger held that: Chief Justice Warren analyses by appropriate job shown both evidence, however, em- shows group which high completed

ployees, who have validity study performed and on school or taken the tests continued procedure job for the selection satisfactorily make perform to be used. progress departments for the which 1607.7B(2). 29 C.F.R. § high are now school and test criteria effect, analyses generally job must be promotion present used. The record of at that done both sites order ensure who not be to meet employees would able are conducted the studies which elsewhere suggests possi- the new criteria thus the out arise of similar situations the site requirements may the not be bility that similarity has been found issue. No such purpose of needed even for the limited here. policy preserving the of advance- avowed company. ment within the regulations provide, part, The EEOC as follows: 432, 91 at 854. Id. at strategies showing validity for New the noted the ab- Supreme Court The procedures be of selection will evaluated general intelligence, of stract measures they psycho- accepted as the become supposedly obtained which had been logical profession. test, job through the did not correlate 1607.5(A). 29 C.F.R. § It the need performance. also contrasted rela- “meaningful study” job determining for a major whether requirements accepted “the tionship expert’s to the fact that theories become have colleagues they is Company’sjudgment his whether were instituted recogni- “significant degree” of improve the received a they generally would complex long they plaintiffs as with impact. had not area so are consistent adverse Since the burden, e.g. Paper obli- applicable met never their the defendant was case law. See Albemarle 2369; gated conse- As a to show relatedness. Moody, v. 422 U.S. at 95 S.Ct. at Co. quence, theory validation was dicta. this of Griggs v. Duke Power Moreover, Appeals of Circuit Court Fifth 855; City Chicago, 549 S.Ct. at United States v. of ever lower decision without affirmed the court’s (7th Cir.), cert. F.2d validity addressing issue. relatedness (1977), Brunet 54 L.Ed.2d simply agreed appellate tribunal Columbus, F.Supp. City v. had not been trial court adverse (S.D.Ohio 1986). specific Finally, requirement of shown. this regu- recently This has held that EEOC Court particularly strong in- in cases relatedness is guidelines.” "operate as Yates v. AVCO lations type Complete volving jobs. clerical Chrisner (6th Cir.1987). Corp., ex- For an Transit, Inc., supra. Auto employ courts should cellent discussion how standards, New see Guardians Assn the EEOC Although regulations do not bind the EEOC Commission, Court, F.2d v. Civil Service sub York shown this numerous courts have technically at 90-91. in this stantial deference to them professional tion or whether literature credentials as fellow APA does not essentially concept. supportive sufficiently counter this evidence of unac- Distler, United States ceptability. my It is belief that he has case, In this 961-62 John upon far too thin a reed approach fundamentally at Hunter’s base departure which to a radical from standards, odds with the APA’s were Albemarle. Albemarle, generally regard- endorsed potentially The kind of Kafkaesque re ing generalization. validity sult, if intelligence which would occur tests APA 1.16 that: Standard states always valid, to be assumed dis adequate When local validation evidence cussed in Young, Van Aken available, is not related criterion evidence 448, (E.D.Mich.1982), 750 F.2d 43 aff'd validity specified for a test use (6th Cir.1984). Hunter, as an testified validity generalization based on from a expert witness for the Plaintiffs in that studies, set prior provided validity generalization advanced his specified test-use situation can be con- theory defending certain exams sidered to have drawn from the department. fire The trial cri then population same of situations on which tiqued testimony: Hunter’s generalization validity was conducted. specific He admitted knew nothing he (Primary) Department about Detroit Fire Acad- Comment: curriculum, emy and its and stated that validity generaliza- Several methods of ability fighter depended be a fire tion and simultaneous estimation have just upon reading ability as much fast proven methods, integ- useful. all agility ability to climb ladders and do rity depends de- inference physical fighters tasks that fire gree similarity between local situa- perform. He explain was unable to how prior tion and the set situations. wheelchair, person get in a who could prior Present and situations can be highest exam, grade on a written judged similar, example, to be for accord- fighter. be could an active fire He fur- ing to factors such as the characteristics high ther testified that there a corre- of people and functions involved. cognitive lation between ability and the (correlations, Relational measures re- good ability picker. be a lemon rates, gressions, etc.) success should be potential exactly These absurdities were carefully appropriate selected to be what in Griggs and the inference to be made.19 sought by requiring Albemarle to avoid a contrast, In theory requires Hunter’s job analysis detailed in validation studies. proof “similarity of between local situa- law, gen- As a of validity matter Hunter’s prior tion and the set of situations.” The theory totally unacceptable eralization is APA represent professional- standards profession- under the relevant case law and ly acceptable standards for statistical stud- al standards. satisfy ies. He did not these standards.20 Atlas, Hunter are jobs why never visited studied the There additional reasons at or Company even read the studies which cannot demonstrate relatedness other formed the basis for validity general- validity generaliza- than deficiencies in the compilation Mickey ization Petty. theory.21 general of His tion proposition, As a a 19. Tr. years 426. would reflect that Atlas went with fifteen hiring a out black clerical worker. Numerous (Donald 20. Tr. See also Schwartz notes that overwhelming have held courts that such statis Atlas). specific job study Hunter did a never at virtually impossible tics are to rebut in defend Falls, See also 531, v. Rivera Wichita 665 F.2d ing disparate impact examples claim. For (5th Cir.1982) (jobs 538 n. be must zero,” referring cases to the “inexorable see In similar). substantially U.S., ternational Brotherhood Teamsters v. Virginia Cunningham If and Ben Smith are 431 U.S. 23, 342 n. 1858 n. at workers, (1977); not as included clerical Capaci the record &Katz in extensive as occurred support a

plaintiff can demonstrating through byor statistics case case.23 in disparities substantial existence of In the instant the test cutoff score make that are used to procedures those points higher than Albe- was seven In Motor hiring EEOC Ford decisions. higher cutoff here clear- marle. score rev’d ly disqualified more blacks. The Wonderlic 219, 102 S.Ct. grounds, 458 U.S. on other of 18 Company recommended scores or (1982), Fourth 73 L.Ed.2d up clerical to 25 for a for some Appeals ap- held that the Circuit Court Yet, Atlas, in positions. few other estab- “held” for six plications of women were lishing every position, score of 25 employment proce- “loose months mouth, figure by asserting that justified the all dures,” hiring by such word of disparities hiring. expected per- to the contributed its clerical workers were which, turn, multiple form functions Corp., 532 Motors v. General Senter Company with a more provide would (6th Cir.), cert. F.2d great productive utility. work force and 182, 50 L.Ed.2d 150 However, (1976), utility factors that readi not we discussed five the benchmark. minority pro ly to fewer contributed necessity. The benchmark No is business vague promotional stan such as motions suggested the trial record has dards, opportunities, notice of lack of for, necessary score 25 was safeguards. no anti-discrimination to, positions. manifestly “related” these also, Greenspan v. Automobile Club of argument support does its (E.D. 1033-37 Michigan, 495 gener- relying only theory on Hunter’s Grossman, Mich.1980); Em B. Schlei & P. intelligence always al tests are valid.24 Law at ployment Discrimination brief, Atlas its asserts that: (1983).22 Finally, there is no that the use Appeals for The Court of the Second compliance Test the Wonderlic York Circuit Guardians Assn New Employ- with the Uniform Guidelines Service, supra, held that: City v. Civil ee Selection Procedures. Even were one produces disparate an exam racial reject completely validity the fact of results, requires adequate a cutoff score generalization, Petty Dr. did review the justification and cannot used at positions unreliability job descriptions its of the clerical point where has such 662; explains only Besthoff, Pegues State of .25 that a test 5% to 6% F.2d Miss. means Yet, Services, generally performance. "Courts ac Employment 768 n. 9. cept reli correlation coefficients above as +.30 Clady County Angeles, able.” v. Los 770 F.2d examples disparities 22. are of such There other 1421, denied, 1431-32 cert. 475 instance, points in this case. For were added 1109, 1516, U.S. 106 S.Ct. 89 L.Ed.2d 915 applicants. scores of some Some of Accord, NAACP, Beecher, Chapter Boston Inc. v. taking were hired without 504 F.2d at 1024n. 13. This Court need not rule Astonishingly, of those hired did test. 48% figure juncture adopt at this on the that it will being of 25 hired. For a receive score before Nonetheless, bare minimum correlation. examples, recitation other see at 1504. infra higher the Court also notes that correlations are great sought impact. often when there is adverse Rogers Paper F.2d See Int'l 630 at 106. also id.; Clady County Angeles, Los Guardians 1340, (8th Company, Cir. F.2d n. 17 510 1347 Service, Assn York v. Civil New 630 F.2d Beecher, NAACP, 1975); Chapter Inc. v. Boston Thus, despite great at 105-06. pact adverse im 1017, (1st Cir.), 421 F.2d 1025 cert. here, significantly the correlations fall be 1561, (1974) accepted. generally low those arbitrary). (cut off score of on a test attempts by to rehabilitate the studies having explain Petty greatly Hunter how under- problems theo- 24. There are with Hunter’s correlations. estimated Hunter’s "corrections” invalidity highlight of the ry which further problems, extensively by as discussed average Petty argument. computed example, Richard Barrett. For Hunter excluded predictive making .25 con- to be when correlation for the studies the lower studies in correc- tions, predictive. enhancing A correlation thus the result. current .15 Atlas, they Assuming, and he concluded that arguendo, the validity sufficiently positions similar to the generalization studies have similarities to examined the other validation studies jobs, Atlas’ it has similarities to about of the Wonderlic test.25 of the remaining because the 50% half regard were hired without to the test re- flawed, argument job This in that the quirement. Such an arbitrary hiring meth- descriptions, by which were Pet reviewed od cannot be found to be similar or validat- ty, way accepted fell norms. below metastudy. ed The adverse employees interviewed at Atlas were sim of the method used cannot be shown to be ply job responsibili asked to describe their job job related when the relationship study Paper ties. In Albemarle Moody, Co. v. only applies to about of the hired 50% U.S. at analy- employees. inadequate ses were found to be when con by supervisors vague ducted because of County Walston v. School Board of ness and bias. The absence of a careful Va., Cty., Nansemond 492 F.2d job analysis, job descrip which evaluates (4th Cir.1974), the Fourth Circuit Court of tions as well as assessments of the relative Appeals offered various criticisms on the importance tasks, of various has been general use of a intelligence emphasized by rejecting several courts in defendant: validation studies. James v. Stockham it cannot be used as a tool of discrimina- Fittings Valves & 559 F.2d tion; it cannot be proper used without (5th Cir.1977); City Chicago, U.S. v. job analyses validation and ... Nor can it (7th Cir.1977); Rogers v. Int’l be capriciously administered in der- 1351; Paper Company, 510 F.2d at Dozier ogation guidelines promulgated of the Chupka, (S.D.Ohio so as facially disparate ETS to have a 1975). Those cases where criterion related impact. accepted studies have invariably been in In the capricious instant adminis- job analysis. clude such a detailed e.g. tration of the test impos- makes validation Corp., Bernard v. Oil Gulf sible.27 (5th Cir.1988) (supervisors impor rated tance of 117 duties and 37 abilities when summation, my it is belief that no performing analysis). criteria necessary remand is in this case because Atlas’ efforts to show relatedness fall Nonetheless, I do not my base conclusion Court, applicable Supreme far below the problems very with these informal Appeals profes- Sixth Circuit Court of job descriptions.26 job descrip- if Even guidelines. sional Consequently, I believe tion is sufficient to similarity show a judge, conclusion Atlas, study jobs to it still faces the over- determined that had not whelming problem nearly a full half established, should reversed with- (48%) employees, who were hired out remand. during period question, did not take

the test or failed to achieve a score of 25 being explains

before hired. Atlas this ac- III. by claiming tion hiring that immediate deci- Yet, case, necessary. sions were often Unlike the this is an astoundingly large and EEOC must discriminatory unsubstantiated show intent in number of sudden prevail decisions. disparate order to on this treatment 25. Atlas brief at 26. Appeals 27.The Fourth Circuit Court of also expressed displeasure givers its that the test as job descriptions supplemented by 26. These were teachers, subject, sumed that all no matter what Pettys positions assessment that Atlas’ are sim- could be evaluated with the same test. Walston metastudy surpris- ilar to those of the ingly and the County Cty., School Board Nansemond Vir expert, similar conclusion of the EEOC ginia, 492 F.2d at 926. Richard Barrett. Tr. 511-512. U.S., already at 431 U.S. 335 the EEOC’s statistics have Teamsters

claim.28 Moreover, rejected. at n. 15. trial the EEOC has also n. clearly subject to the judge’s many specific decision instances of dis- enumerated plaintiff A level of review. parate erroneous supplement treatment to the statis- demonstrating a initial bears the burden tics: Teamsters, In the Su- case. prima facie applicants 1. or white were able that: preme held Court typing a second take test. pattern systemwise a alleged it because 2. older white re- Several enjoy- to the full practice of resistance or age to their ceived additions rights, the Government ment of Title YII scores. prove more than the ultimately had to exceptions require- Many 3. to the test or “acciden- mere isolated occurrence only made ments were discriminatory act. It sporadic tal” or exceptions. from cants benefited by preponderance of establish had to some 4. Racial notations existed on racial discrimination the evidence that applications. black candidates operating company’s standard was the regular rather than the procedure clearly prima —the The EEOC has met its facie practice. unusual burden on the basis of the statistics 1855. particular Id. at S.Ct. these instances. method, has been uti- A common upon response, In Atlas relies the trial satisfy prima his plaintiff lized judge’s hiring conclusion no individual proof, is to offer statistical burden racially (only decision was motivated

facie disparity. The evidence of “qualified” hired) most assert- District v. United Hazelwood School ing that none of the individual incidents 299, 307-308, States, 433 U.S. S.Ct. hiring in the made difference determina- (1977) 2736, 2741-2742, 53 L.Ed.2d 768 held tion. disparities” “gross statistical need be Having carefully examined the trial disparate treatment context shown in the judge’s ruling, clearly I it to believe discriminatory This re- motive. to show erroneous. Atlas never offered a nondis- disparate quirement higher than criminatory explanation individual impact context. Where Although incidents of treatment. circumstances, incidences, discriminatory had Atlas asserts that these incidents history are shown in addition or upon composition effect the racial statistics, grossness disparity force, explanation nothing work does great. Bernard v. Oil need not be Gulf discrimi- to remove the inference that racial at 568. Teamsters v. Corp., 841 F.2d responsible nation was differential U.S., over 431 U.S. at S.Ct. treatment. No other reason has been of- other instances were forty specific Although fered. these incidents alone record. might racially not rise of a to the level prima has met its Once the they discriminatory pattern practice, or cer- burden, must it defendant rebut testing do tainly with the combined attacking by providing or a “non statistics practices, and the statistics. discriminatory explanation appar for the racial Atlas’ assertion that no motivation ently discriminatory Payne result.” “quali- since the most Inc., was involved Laboratories, Travenol premised on the (5th Cir.) fied” were hired is as- cert. performance sumption best L.Ed.2d quali- synonymous being the most case, discredit test is Atlas’ efforts to the instant racially at issue and does not find discrimi- concluded that the In this relying upon natory practice. pattern its burden after EEOC had not met hiring: year by year of Atlas’ his assessment at 1188. The Court has careful consideration years hiring decisions over the Atlas’actual arguendo, assuming, Even that the fied.

highest governing scores criteria qualification applicant’s

and established an work, using departed the test from hiring workers, on of the hired all of 48%

whom were white. Tr. 176-177. When the

statistics test are combined with the

strikingly practices different used for hir-

ing compared black workers as to white

applicants, my it is conclusion trial clearly holding erred in proven prac- had not pattern hiring, discriminatory

tice of and his deter-

mination this claim should be reversed remand,

without as well. America,

UNITED STATES

Plaintiff-Appellee, COOPER, Sylvester

Otis Graham and Hoyer,

Dennis R.

Defendants-Appellants. 87-2118,

Nos. 87-2119 and 87-2183. Appeals,

United States Court of

Sixth Circuit.

Argued 14, 1988. Nov.

Decided March

Case Details

Case Name: EQUAL EMPLOYMENT OPPORTUNITY COMMISSION, Plaintiff-Appellant, v. ATLAS PAPER BOX COMPANY, Defendant-Appellee

Court Name: Court of Appeals for the Sixth Circuit

Date Published: Apr 4, 1989

Citation: 868 F.2d 1487

Docket Number: 87-5421

Court Abbreviation: 6th Cir.