Reynolds v. Sheet Metal Workers Local 102

MEMORANDUM AND ORDER BRYANT, Chief Judge. Introduction Plaintiff’s complaint, as amended, alleges widespread discrimination against blacks in the union sheet metal trade in the Metropolitan Washington, D.C. area, in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e, et seq., as amended, and the Civil Rights Act of 1866, 42 U.S.C. § 1981. In particular, plaintiffs attack the selection criteria of the apprenticeship program designed to admit workers into the local sheet metal union, the administration of the pre-apprenticeship program aimed at training applicants not eligible for the apprenticeship program, and the treatment afforded black journeymen who have attained union status or sought direct admission into the union. Defendants in the suit include the international union of sheet metal workers (“SMWIA”); the Metropolitan Washington affiliate of sheet metal workers (“Local 102”); the national association of employers engaged in sheet metal contracting work (“SMACNA”); the Metropolitan Washington affiliate of employers engaged in sheet metal contracting work (“SMACNADC”); the national association of employers and union members promulgating national apprenticeship training standards (“NJATC”); and the local association of employers and union members supervising and administering the local apprenticeship program leading to membership in Local 102 (“JAC”). 1 *955 Since January 17, 1977, plaintiffs have been certified as representative of a class comprised of all black persons in the Metropolitan Washington area who have been or may be “excluded from membership in Local 102,” “denied employment in the sheet metal industry,” or “adversely affected in their terms and conditions of employment [in the sheet metal trade],” because of their race. 2 In November 1979, plaintiffs learned that defendants had begun soliciting applications for a new class of approximately fifty apprentices in the apprenticeship training program. Applications ' were to be accepted during the first three weeks of January 1980, with interviews and selections shortly thereafter. On January 2, 1980, plaintiffs moved for a temporary restraining order and preliminary injunction enjoining the formation of the new apprenticeship class as long as the JAC employed certain selection criteria and procedures. Plaintiffs attacked the high school diploma requirement 3 the interview of applicants by the JAC committee, and the inquiry concerning arrests on the application form. 4 Although applications for the new apprenticeship class were accepted in early January 1980, the defendants voluntarily postponed any interviews or final selection of applicants until after the date of this order. This delay eliminated the need for a hearing on the temporary restraining order; ón January 17,18, 21, and 22,1980, hearings were held on the motion for a preliminary injunction. 5 Plaintiffs’ claim for relief grows out of Griggs v. Duke Power Co., 401 U.S. 424, 429-31, 91 S.Ct. 849, 852-53, 28 L.Ed.2d 158 (1971) and its progeny. Racially neutral selection criteria having a disparate impact on minority applicants give rise to a prima facie case of employment discrimination in violation of Title VII. 6 The defendants must then establish that the selection criteria “bear a demonstrable relationship to successful performance of the jobs for which it was used.” Id. at 431, 91 S.Ct. at 853 (construing 42 U.S.C. § 2000e-2(h) (1976)). Even when such a showing is made, plaintiffs may still prove that “other tests or selection devices, without a similarly undesirable racial effect, would also serve the employer’s legitimate interest in ‘efficient and trustworthy workmanship.’ ” Albemarle Paper Co. v. Moody, 422 U.S. 405, 425, 95 S.Ct. 2362, 2375, 45 L.Ed.2d 280 (1975). This two-step process protects minority applicants qualified for the job or position in question. Plaintiffs have argued that use of the three contested criteria — high school diploma, interview, and arrest record inquiries— has a disparate impact on black applicants and is not job-related. Defendants have attempted to show the opposite. After consideration of the testimony, exhibits, briefs, *956 and memoranda submitted by the parties, this court, pursuant to Fed.R.Civ.P. 52, makes the following findings of fact and conclusions of law. FINDINGS OF FACT Parties 1. Plaintiffs Earl Reynolds, Eugene F. Walker, David Cook, Lumas Reeder, Albert L. Busby and Ernest E. Baylor are each black citizens of the United States. (See NJATC’s Answer to Second Amended Complaint, ¶2, March 22, 1979.) 2. Local 102 is an unincorporated sheet metal workers union with geographic jurisdiction encompassing the Washington, D.C. Standard Metropolitan Statistical Area (“SMSA”), plus certain outlying counties in Virginia and Maryland. Local 102’s business address is 2703 Bladensburg Road, N.E., Washington, D.C. (Local 102’s Answer to Amended Complaint, ¶4, April 6,1976.) 3. Local 102 is the recognized bargaining agent for journeymen and apprentice sheet metal workers hired by SMACNA-DC sheet metal contractors within its geographical jurisdiction. (JAC Ex. 27.) 4. Local 102 has been governed by its own Constitution and By-Laws, and by the Constitution and Ritual of the Sheet Metal Workers’ International Association. (Testimony of Richard Drake; JAC Ex.’s 27, 28.) 5. As of 1975, only 36 of 950 members, or approximately 3.8% of Local 102’s total membership in construction was minority. (PL Ex. 127 and Ex. 4 thereto.) According to Census figures, only 6% of all employees in the Washington D.C. area sheet metal trade (union and non-union) were black as of 1970. (Pl.Ex. 149.) 6. Local 102 has now, and has had since before 1971, a collective bargaining agreement with defendant SMACNA-DC. (Answer of SMACNA-DC to Amended Complaint, ¶6, March 31, 1976; JAC Ex. 27.) 7. SMACNA-DC is an incorporated association of building contractors in the Washington D.C. area who are engaged in sheet metal construction work. (Testimony of Harry Muller; JAC Ex. 27.) SMACNADC is the bargaining representative for its members. (Id.) 8. Defendant JAC is a joint labor-management committee composed of three representatives from Local 102, and three representatives from member companies of SMACNA-DC. The JAC administers Local 102’s and SMACNA-DC’s apprentice selection and training. (Testimony of Clarence Neese; JAC’s Answer to Amended Complaint, ¶8, April 6, 1976.) 9. The JAC operates a four-year apprenticeship program designed to train individuals to become competent journeyman sheet metal workers in the commercial construction trade. (Testimony of Clarence Neese.) Historically, most individuals have been admitted to membership in Local 102 through the four-year apprentice program. 10. The apprenticeship program consists of forty hours per week on-the-job training with a sheet metal contractor, and classroom instruction of three and one-half hours a night for two nights a week, over a nine-month school year. The classroom instruction continues for four consecutive school years. (Testimony of Clarence Neese.) 11. The JAC’s formal apprentice school curriculum includes classroom and shop courses in safety and use of tools, basic and advanced mathematics, sheet metal drafting, basic rectangular layout, parallel line development, blueprint reading, radial line development, roofing, welding, field measure, special fittings, triangulation, air balancing, rigging and related course work. (JAC Ex.’s 10, 11, 12.) 12. The apprentice manpower needs of SMACNA-DC’s members and selection standards are subjects of collective bargaining between Local 102 and SMACNA-DC. Within the trade jurisdiction of Local 102, contractor members of SMACNA-DC employ sheet metal workers who are members or apprentices of Local 102. (JAC Ex. 27.) 13. The JAC anticipates selecting a new apprenticeship class to begin training in *957 early 1980. The JAC advertised for applicants in November and December 1979, and accepted applications for apprenticeship from January 2 through January 17, 1980. (Testimony of Clarence Neese; Testimony of Robert Gawne.) Applicants meeting the requisite standards, see ¶¶(15) — (17) infra, are scheduled for interviews beginning in early March 1980. Apprentice Selection System 14. The procedures and criteria for selection of apprentices used by the JAC were developed jointly and contained in a document entitled “Apprentice Standards” executed by SMACNA-DC and Local 102. (Pl.Ex. 128; see also Pl.Ex. 123 and JAC Ex. 36.) 15. From time to time, according to the needs of SMACNA-DC members, the JAC advertises the availability of apprenticeship positions. In doing so, the JAC currently indicates that applicants must meet certain threshold qualifications, including the following: 7 (a) high school diploma or G.E.D.; (b) age (at least 17 yrs. old); (c) residence within the jurisdiction of Local 102; and (d) physical ability to perform work in the sheet metal trade. (Testimony of Clarence Neese and Donald Alter; Pl.Ex. 1.) 16. Interested persons must meet the qualifications described in ¶15 above, and must appear in person at the JAC office to complete a written application. The written application, which is four pages long, includes a number of questions about the applicant’s background, education and work experience, and also includes an inquiry about arrests. (Pl.Ex. 1.) 17. In addition to completing the written application, each applicant must provide the JAC with copies of the following documents: (a) high school diploma, G.E.D. or equivalency; (b) birth or baptismal certificate; (c) high school transcript; and (d) resume of work experience. (Id.) 18. Only apprenticeship applicants who have fulfilled all of the requirements set forth in ¶¶|15-17 above are scheduled for an in-person employment interview. No apprentice applicant is hired without such an interview. The interview is conducted by a panel of the six members of the JAC and the JAC administrator, and lasts approximately 5-10 minutes per applicant. (Testimony of Clarence Neese.) 19. Applicants are ranked according to a two-part scoring system devised by the JAC. On Part A, the educational portion, a total of 33 points can be awarded. On Part B, the interview portion, a total of 67 points can be awarded, for a total of 100 possible points. (Id.\ Pl.Ex. 1.) 20. Part A is completed by the JAC administrator, based upon the information contained in the written applications. (Testimony of Clarence Neese.) Thirty-three points may be awarded on Part A as follows: - mathematics background 0-3 - drafting background 0-6 - general scholastic performance 0-5 - shop courses 0-8 - schooling beyond high school 0-3 - participation in minority equal opportunity program 0-8 TOTAL 0-33 (Pl.Ex. 1.) Three specific mathematics courses are allocated only one point each. Since mathematics counts only three points, an applicant could score zero on the mathematics portion, but might well be selected. (Testimony of Clarence Neese.) 21. Part B is completed by each of the six JAC interviewers, at the interview. (Pl.Ex. 155 at 63-69.) The average of these six scores is then calculated, and added to *958 the score on Part A, to obtain a total score for each applicant. (Testimony of Clarence Neese.) Points may be awarded on Part B as follows: - physical evaluation 0-10 - work experience 0-30 - conduct record 0- 5 - interest in trade 0-10 - financial situation 0- 4 - attitude 0- 8 TOTAL 0-67 22. Once the total scores for all applicants are obtained, the JAC administrator ranks the applicants and provides the applicants’ names and scores to the JAC. The committee then determines how many applicants to select, based upon the needs of Local 102 and SMACNA-DC members. (Testimony of Clarence Neese.) 23. Applicants are notified by mail whether they have been accepted or rejected. (Testimony of Robert Poole; Rodney Jackson; JAC Ex.’s 8 and 9.) 24. Accepted applicants are normally placed on the job with contractors at approximately the same time they commence apprenticeship training. The JAC determines which newly selected apprentices will work for which particular contractor members of SMACNA-DC. (Testimony of Clarence Neese.) 25. Once an applicant is placed on the job with contractor members of SMACNADC the contractors keep the JAC administrator informed of the apprentice’s performance on the job. (Id.) This practice has been followed consistently over a period of years. (See, e. g., Pl.Ex. 122 at 39-61.) 26. The procedures described in ¶¶15 — 25 above have been in effect throughout the 1970’s, and are the same procedures which the JAC proposes for its 1980 apprentice selection. (Testimony of Clarence Neese, John C. Denton and Donald Alter; Pl.Ex. 1.) Proposed 1980 Apprentice Selection 27. In November 1979, the JAC announced its intention to admit a new class of apprentices in early 1980. It indicated that applications would be accepted on certain specified dates in January 1980. (Pl.Ex. 120.) Under the Jac’s proposal, eligible applicants then would be notified of their right to attend a JAC employment interview; interviews are to be conducted in early March; and selection decisions are to be made shortly thereafter. (Statement of counsel for JAC.) The JAC intends to select between 30 and 50 apprentices, depending on conditions. (Testimony of Robert Keyser.) 28. In its announcement of the 1980 application process, the JAC advertised the high school diploma or G.E.D. requirement, as well as the need to provide copies of a high school diploma and transcript, birth or baptismal certificate, and work resume. (Pl.Ex. 120.) 29. The application form to be used in 1980, as in the past, includes inquiries regarding arrest records, occupation of parents, source of referral (including whether referral came from a member of Local 102), and inquiry as to graduation from high school. (Testimony of Clarence Neese; Pl.Ex. 1.) History of Apprentice Selection 30. The JAC selected its first minority apprentice, a Spanish-American, in 1965, and its first black apprentice in 1966. (Pl.Ex. 122 at 18.) 31. The high school diploma or G.E.D. requirement was instituted by the JAC in the mid-1960’s. A number of Local 102’s members are not high school graduates. (Testimony of Robert Gawne, Donald Alter and Clarence Neese.) 32. During the period 1972-1974, the JAC used a “preapprentice” selection system, which was exclusively for minorities. (Testimony of Donald Alter and Eugene Walker; see also Pl.Ex. 123; Pl.Ex. 156 at 54-55; Pl.Ex. 157 at 21-22.) The program was initiated as a means of increasing the number of minorities counted as sheet metal workers in order to meet government contract requirements. (Testimony of Eugene Walker; Pl.Ex. 156 at 90.) *959 33. Preapprentices received a lower wage rate than apprentices, and received no formal instruction. Preapprentices performed menial tasks, such as hauling trash, sweeping up, loading and unloading trucks, and fetching coffee and sandwiches for more senior workers. (Testimony of Eugene Walker.) Many blacks quit or were terminated from the preapprentice program without being considered by the JAC for the apprentice program. (Id.; Pl.Ex.’s 126-131.) 34. A number of blacks were accepted as preapprentices if the administrator of the program believed they could acquire all of the stated qualifications for apprenticeship, e. g., high school diploma or G.E.D., within a reasonable period of time. Such preapprentices could later apply for apprenticeship only if during their preapprenticeship they satisfied each of the qualifications for apprenticeship, e. g., by obtaining a high school diploma. (Testimony of Donald Alter and Eugene Walker; Pl.Ex. 122.) 35. Apprentice applicants were given points under the “work experience” category of Part B of the apprenticeship application form for their experience as preapprentices. (Testimony of Donald Alter.) Indeed, many blacks who applied for apprenticeship during 1972-1974, and who met all of the initial qualifications, e. g., age, residency, high school diploma or G.E.D., physical fitness, were nevertheless encouraged to spend time first as a preapprentice, obtain “work experience” points, and therefore be more competitive in the JAC interview. (Id.) 36. The preapprenticeship program was terminated in late 1974. 37. There have been few apprentices selected in recent years. In 1977, none was selected (see testimony of Clarence Neese and JAC Ex.’s 10, 11 and 12); and in 1978, just 8 were selected. In the autumn of 1979, 55 apprentices were selected. (PI. Ex.’s 135, 136.) The class contemplated for 1980 is estimated at 30 to 50 positions. (Testimony of Robert Keyser.) The small number of selections in 1977 and 1978 was due to the depressed economic condition of the sheet metal industry at that time. (Testimony of Richard Drake.) 38. Until 1979, the JAC kept no records of persons who inquired about sheet metal apprenticeship, but who did not complete an application form or similar document. (Pl.Ex. 122 at 19-20.) For the 1979 apprentice application process, JAC did keep a log of persons who telephoned or came into the office, but who did not complete any documents. (Pl.Ex. 119.) Impact of Apprentice Selection Procedures 39. Census figures from 1970 indicate that the percentage of blacks in the general population in the Washington SMS A is 25.6%. The percentage of black men, aged 18-34, in the civilian labor force in the same area, adjusted for the national population undercount, is 26.67%. (Pl.Ex. 146.) 40. Plaintiffs base their claim of disparate impact on statistical data for the apprenticeship selection periods from 1971-79, general characteristics of the selection process, and experiences of black applicants. The statistical data consists of applicant flow figures, individual selection qualifications and their projected exclusionary impact, and projections of black representation in the relevant labor market. (a) Adverse Impact Based Upon Actual Applicant Flow Statistics 41. Plaintiffs’ applicant data includes all persons, white or black, recorded as having taken one step in the application process for apprenticeship. This applies to individuals who made preliminary inquiries about the apprentice program, e. g., telephone or walk-in inquiries, or filled out apprentice information sheets. Plaintiffs’ applicant data also includes individuals who participated in, or showed any interest in participating in, the preapprentice program. Preapprentices (actual or prospective) are included under the theory that they were dissuaded from participating in the apprentice program because of the standards under attack in this suit. *960 42. The 1979 application figures prepared by the plaintiff indicate that 44 applicants were black and 80 nonblack, with a black acceptance rate of 38.8% and nonblack acceptance rate of 51.3%. (Pl.Ex. 135.) 8 43. Actual applicant data derived from JAC files for 1971-79 indicates that the nonblack acceptance rate was 31.2%, while the black acceptance rate was 23.4%. (PI. Ex.’s 137, 138). These figures do not include applicants who were prescreened by the AIC and not referred to the JAC or those for whom no JAC record exists. Id. (b) Adverse Impact of Exclusionary Selection Qualifications 44. Defendants have used and will continue to publicize and use the high school diploma or G.E.D. requirement for apprentice selection. (Pl.Ex. 120.) According to the 1970 Census in the Washington SMSA, only 13.4% of nonblacks,'aged 20-34, failed to complete four years of high school, whereas 37.8% of blacks failed to do so. (Pl.Ex. 143.) According to 1976 census data, those figures were 8.7% for whites, and 25.9% for blacks. (Pl.Ex. 144.) 45. Defendants have inquired in the past and will continue to inquire about arrest records in their apprentice selection process. (Testimony of Robert Poole; Rodney Jackson; Pl.Ex. 1; Pl.Ex. 155 at 66-67.) The arrest record inquiry also adversely affects blacks. According to the FBI’s 1978 Uniform Crime Report, 26.4% of persons arrested nationwide were black, while blacks comprised only 12.5% of the population. According to the 1978 Annual Report of the Metropolitan Police of Washington, D. C., 85.5% of persons arrested in Washington D. C. were nonwhite, although nonwhites comprised only 72.4% of Washington, D. C.’s population. (c) Adverse Impact in Terms of Relevant Labor Market 46. The proportion of blacks in the total population of the Washington, D. C. SMSA includes many persons either incapable of or uninterested in training for sheet metal work, such as housewives, students, retired persons, and disabled or infirm persons. (Testimony of Harold Goldstein.) 47. Similarly, gross SMSA population statistics include many members of the labor force, both black and white, who have a college education and are employed in, or seek employment in, professional, technical, managerial, or other white-collar jobs. Most of them would be unlikely to apply for apprenticeship in this trade. (Id.) 48. Approximately 75% of all sheet metal workers in the United States have from 1-4 years of high school; the majority of the remaining 25% have less education (Testimony of Harold Goldstein). Even among the most recent entrants to the occupation, those aged 25-29 in 1970, only 12% have more than a high school education. (Id.) 49. The great majority of apprentice applicants fall within the 18-34 year group. (PI. Ex. 146.) In addition, very few women enter the sheet metal trade; records of Local 102 indicate that no women applied for apprenticeship until after 1977 (Pl.Ex. 152.) 50. Analysis of civilian labor data, adjusted for the proper population under-count, indicates that in 1970, blacks amounted to 39.5% of the male workers in the 18-34 year group with 1-4 years of high school education. (Testimony of Harold Goldstein; PI. Ex. 146.) The 39.5% figure does not account for unemployed blacks likely to constitute a greater proportion of applicants for entry-level jobs such as sheet metal apprenticeship (Testimony of Harold Goldstein; Pl.Ex. 147.) 51. An analysis of the work background listed by recorded applicants for 1967-76 indicates that blacks comprise approximately 45% of those occupations likely to produce apprenticeship applicants (Pl.Ex.’s 133, 148). This figure does not distinguish be *961 tween blacks in the 18-34 year age group in such occupations and older fellow black workers. (d) Additional Eyidence 52. The JAC has had a history of nepotism in its apprentice selection. From 1971-76, an applicant’s relationship to a member of Local 102 was a critical factor, weighing heavily in his favor in the selection process (Pl.Ex. 153). The current procedures still inquire into the occupation of parents and whether the source of referral is a Local 102 member. (Pl.Ex. 1.) 53. Furthermore, the Washington, D. C. area building trade unions have had the reputation in the black community of discriminating against blacks in employment opportunities. (Pl.Ex. 122 at 16.) The sheet metal union and trade has had such a reputation since at least as early as 1972, until the present. (Testimony of Eugene Walker; Robert Poole; Rodney Jackson.) 54. In the interview scoring, defendants continue to place considerable weight on purely subjective factors, such as “attitude” and “conduct.” Further, defendants continue to give considerable weight in interview scoring to sheet metal or construction craft experience. This is done despite the fact that the sheet metal apprenticeship is an entry-level position; during the four-year apprenticeship program, these apprentices receive extended on-the-job training and formal classroom instruction. (Testimony of Clarence Neese.) The median score in the interview evaluations for the 1979 class was much lower for black applicants than for whites. (See Memorandum of Points and Authorities in Support of Plaintiffs’ Motion for a Temporary Restraining Order and a Preliminary Injunction, at 21-22) (calculating interview scores from Pl.Ex.’s 2-118) [hereinafter cited as Plaintiffs’ Memorandum]. 55. In 1979 seven applicants received the full 3 points on Form A (completed by the JAC administrator) for education beyond high school (see ¶ 20, above), although none of the seven had indicated any such advanced education on the application. (PI. Ex.’s 10, 12, 17, 24, 52, 79, 104.) Of those seven, six were white and only one was black. (Id.) Conversely, three 1979 applicants received zero points on Form A for education beyond high school, although each had indicated such advanced education on his application. (Pl.Ex.’s 26, 57, 61.) Of the three, two were black and only one was white. (Id.) 56. Black apprentice applicants have testified to the following: (a) Robert Poole, a 22-year-old black citizen of the United States, applied for apprenticeship in the spring of 1979. (Pl.Ex. 91.) Mr. Poole is a high school graduate, and has had two years of junior college education, including courses in business and mathematics. (Id.; Testimony of Robert Poole.) (b) Along with his application, Poole submitted all of the other necessary materials: high school diploma and transcript, birth certificate, and work resume. (Testimony of Robert Poole.) (c) Mr. Poole was interviewed by the all white JAC panel in May 1979. There were no black or minority interviewers. During the interview Poole was asked not once, but twice, about whether he had been arrested. The answer was “no” each time. Also during the interview, one interviewer indicated Mr. Poole was over-qualified for apprenticeship, and that he should apply instead for a job at a local fast-food restaurant. (Id.) (d) Mr. Poole’s application was rejected by the JAC in the summer of 1979. (Id.) (e) Rodney Jackson, a 24-year-old black citizen of the United States, also applied for apprenticeship in the spring of 1979. (PI. Ex. 61.) Mr. Jackson is a high school graduate, and has had some college courses. (Id.; Testimony of Rodney Jackson.) Jackson, like Poole, submitted all of the necessary materials along with his application. (Id.) Despite college courses, he secured no points for schooling beyond high school. (f) Mr. Jackson was interviewed by the JAC panel in May 1979. Again, there were no black or minority interviewers. During the interview he was asked whether he had *962 ever been arrested, and the answer was “no.” Jackson was told that it was customary on job sites to address apprentices as “boy.” He was asked whether he would object to this practice. Mr. Jackson said he did object to the use of “boy,” but would put up with it if necessary to keep the job. (Id.) (g) Mr. Jackson’s application was rejected by the JAC in the summer of 1979. Validity of JAC’s Selection Criteria 57. Defendants presented expert testimony in support of the high school requirement. Dr. John Denton, an industrial psychologist, testified that he performed a “construct” validity study, which suggested the appropriateness of a high school diploma or G.E.D. requirement for sheet metal apprenticeship. The “construct” used by Dr. Denton was “general learning ability.” 58. Dr. Denton’s “validation” began with the compilation of a “Position Description Questionnaire” (“PDQ”) for the job of sheet metal worker. The PDQ results were then categorized into 21 categories, or “job factors” on which the job of sheet metal worker was compared with all other jobs for which Dr. Denton had compiled PDQ results. The “job profile” that emerged from this comparison was then used to predict, on the basis of data from other jobs, the median salary of sheet metal workers, the median score on the SRA verbal test, and the median score on the Wonderlic test. Denton concluded that the high school requirement was “validated” because the median score on the Wonderlic test predicted by this process was higher than the median score on a national basis for high school graduates. (Denton testimony.) 59. Dr. Denton also applied a “readability analysis” to materials used in the apprenticeship training program and concluded such materials were at a high school level or above. Dr. Denton did not investigate whether the material analyzed was actually taught in the apprentice program. 60. Dr. Denton’s testimony indicated he had not considered less onerous alternatives to the high school diploma requirement. 61. No defendant adduced any evidence to show the validity of the JAC’s arrest record inquiry, or to show that arrest record inquiries had a business justification or were job-related. At the same time, individuals with prior arrests have been selected for apprenticeship (Testimony of Clarence Neesé), and counsel for JAC has suggested that arrest records are not used to disqualify applicants. 62. Defendants did attempt to validate the JAC interview, but their own expert witness found no significant correlation between the rating received in the interview, and performance in apprentice school. (Testimony of John C. Denton; Pl.Ex. 154.) CONCLUSIONS OF LAW 1. Defendant Sheet Metal Workers Local 102 (“Local 102”) is a union and labor organization within the meaning of § 701(d), Title VII of the 1964 Civil Rights Act, 42 U.S.C. § 2000e(d)(1976), and is engaged in an industry affecting commerce. 2. Defendant Joint Apprenticeship Committee (“JAC”) is a joint labor-management apprenticeship committee within the meaning of § 701(d) of Title VII, 42 U.S.C. § 2000e(d)(1976), and is engaged in an industry affecting commerce. 3. Defendant Sheet Metal and Air-Conditioning Contractors National Ass’n, Washington, D. C. Chapter (SMACNA-DC) is an incorporated association of employers engaged in sheet metal construction work in the Washington D. C. area, bargains with Local 102, and is engaged in an industry affecting commerce. 4. This court has jurisdiction over the subject matter of this action and the parties thereto by virtue of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq., as amended. Venue is proper pursuant to 28 U.S.C. § 1391(b) and 42 U.S.C. § 2000e-5(f)(3). 5. Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq., as amended forbids as unlawful employment practices the exclusion of persons from apprentice *963 training programs and union membership because of race, color or national origin. I. Prima Facie Case 6. Under Title VII, plaintiffs must establish a prima facie case of employment discrimination. Griggs v. Duke Power Co., 401 U.S. 424, 429-31, 91 S.Ct. 849, 852-53, 28 L.Ed.2d 158 (1971). In a disparate impact case, such a showing rests on evidence that facially neutral criteria “[select] applicants for hire or promotion in a racial pattern significantly different from that of the pool of applicants.” Albermarle Paper Co. v. Moody, 422 U.S. 405, 425, 95 S.Ct. 2362, 2375, 45 L.Ed.2d 280 (1975). Three general statistical techniques have been developed to identify the appropriate “pool of applicants” and determine if reasonable grounds exist for inferring discrimination. Green v. Missouri Pacific Ry., 523 F.2d 1290, 1293-94 (8th Cir. 1975). The first approach views the general population of minority members as potential applicants. When particular selection criteria, e. g., height or weight requirements, high school diploma, operate to disproportionately exclude minority members as a group, or minority representation in the defendant’s work force differs markedly from general population figures, a prima facie case of discrimination may be established. See, e. g., Dothard v. Rawlinson, 433 U.S. 321, 329, 97 S.Ct. 2720, 2726, 53 L.Ed.2d 786 (1977). The second technique narrows the applicant pool, by developing minority representation in the relevant labor market and comparing it to the racial composition of the defendant’s work force. Hazelwood School District v. United States, 433 U.S. 299, 308, 97 S.Ct. 2736, 2741, 53 L.Ed.2d 768 (1977). The third technique identifies the racial composition of those applicants actually applying for the position in question, and determines if selection rates vary according to race. Id. at 308 & n.13, 97 S.Ct. at 2714. The plaintiffs apply all three approaches to the selection process at issue in this case. A. Adverse Impact of Exclusionary Selection Standards 7. In order to understand the relevance of evidence bearing on the impact of individual selection standards, it is necessary to show how the elements of the application procedure for the apprenticeship program interact. The high school diploma can be described as a “threshold requirement”; applicants who did not complete high school are eliminated immediately from further consideration. See D. Baldus & J. Cole, Statistical Proof of Discrimination § 6.1 (1980) [hereinafter cited as “Statistical Proof”]. The interview, with its allegedly subjective determinations, is the culmination of the process, and results in the applicant’s final rating. The arrest inquiry appears initially in the application form, and according to plaintiffs’ testimony, reemerges during the interview. 8. According to the defendants, the discriminatory effect of individual criteria are irrelevant; all that is important is the number of blacks chosen for the program. When these “bottom line” results are compared to defendants’ version of the proper applicant pool (general population figures), the discriminatory impact is minimal. The defendants’ argument is flawed. First, this is not a ease where applicants without a high school diploma can compensate by performing well on other selection criteria. See Brown v. New Haven Civil Service Board, 474 F.Supp. 1256, 1258-63 (D.Conn. 1979) (applicant for police force eligible for physical agility test regardless of score on written exam; improper to consider disparate impact of written exam when percentage of blacks hired substantially proportionate to percentage of eligible blacks in relevant area). The diploma requirement, and to a lesser extent arrest record inquiries at the time applications are disseminated, are threshold or “knock-out” criteria eliminating actual and prospective black applicants at the onset of the process. Similarly, although the interview is part of the overall application process, it constitutes a distinct and potentially insurmountable barrier to black applicants who have cleared the two preliminary hurdles. 9. It is clear that these three selection criteria have the same role and importance *964 as the “knock-out” criteria at issue in Griggs and Dothard. And their adverse impact is quite apparent. 9 In light of this similarity, plaintiffs ask this court to apply automatically the rule of Dothard to the high school diploma and arrest inquiry: a showing of disproportionate impact based on “general population demographic data” gives rise to a prima facie claim of discrimination. 433 U.S. at 330, 97 S.Ct. at 2727. Such an approach ignores the differences between certain aspects of this case and Dothard. In Dothard, the projected disproportionate impact of threshold criteria (height and weight figures), based on general population figures, was clearly borne out by the actual composition of the defendant’s work force. 10 In this case, all parties admit that the number of blacks in the apprenticeship program approximates the percentage of blacks in the general population. 10. This congruence between work force and general population figures can be explained in two ways. Blacks in the labor force lacking the threshold criteria may have no legitimate interest in the program, with the remaining “qualified” black applicants constituting 26% of the relevant work force. 11 See Statistical Proof, supra, at § 4.2. Under this theory, general population figures indicating the disproportionate effect of the threshold requirement are irrelevant. Or, as plaintiffs suggest, the percentage of blacks in the pool of interested applicants exceeds the percentage of blacks in the general population, and includes individuals with arrest records and without a high school education. Regardless of which argument is convincing, the facts of this case require an additional step in plaintiffs’ argument, a step not present in Dothard. It would be inappropriate, at this point, to find a prima facie case regarding the high school diploma or arrest inquiries, without examining this additional element of plaintiffs’ claim. 12 11. The interview process presents another story. Plaintiffs’ figures are based on actual applicants who reached the interview stage from 1972-76. Plaintiffs’ Memorandum, Table E. Only 19.9% of blacks interviewed were accepted, compared to 41.1% for whites. This is more akin to the situation confronting the Court in Dothard, with an alleged exclusionary procedure producing the expected disparity between individuals interviewed and accepted. The only difference in this case is the use of actual, rather than projected, exclusionary data. Actual applicant data is widely accepted in Title VII cases, infra at 27, and the disparity in selection rates resulting from interviews (21%) is clearly significant, infra at 966-967. With the selection rate for white interviewees more than twice that for blacks, the plaintiffs’ have established a prima facie claim of discrimination at the interview stage of the apprenticeship selection procedure. *965 B. Overall Adverse Impact of Defendants’ Selection Standards Based on Applicant Flow and Relevant Labor Market 12. Plaintiffs rely on actual applicant flow data and relevant labor force figures to show that black representation in the proper pool of applicants was between 35.5-45.5%, and therefore in excess of the percentage of blacks chosen for the apprentice program. See B. Schlei & P. Grossman, Employment Discrimination Law 1165-78 (1976 & Supp.1979) (extensive discussions of both methods of statistical proof) [hereinafter cited as “Employment Discrimination”]. The defendants claim that in this case the applicant pool is either 26.67% or 25.6% black, based on the census figures for the percentage of black males, aged 18-34, in the civilian labor force or general population in the Washington SMSA. When the percentage of blacks selected for the apprenticeship program from 1971-79 (29.0%) or for 1979 (25%) is compared to these SMSA figures, the discriminatory impact on blacks appears minimal. 1. Actual Applicant Figures and Selection Rates a. 1979 Applicant Figures 13. Plaintiffs’ Exhibit (“Pl.Ex.”) 135 analyzes recorded black applicant flow for the September 1979 apprentice class. Forty blacks filed applications; another four were recorded as making telephone or “walk-in” inquiries. Out of this total of forty-four, fourteen blacks were accepted. This results in a black selection rate of 38.8%. The selection rate for nonblacks was 51.3%. This 12.5% disparity in selection rates violates the 80% standard adopted by the EEOC, and a standard chi-square test indicates that the probability such a disparity is due to chance alone is less than 6 in 100 (.058). b. 1971-79 Applicant Figures 14. According to Pl.Ex. 137 and 138, four hundred and eighty-three (483) blacks applied for the apprenticeship program. One hundred and thirteen (113) were accepted. This results in a black selection rate of 23.4%. The selection rate for non-blacks was 31.2%. Again, this 7.8% disparity in selection rates violates the 80% standard adopted by EEOC, and a standard chi-square test indicates that the probability such a disparity is due to chance alone is 3 in 1,000 (.003). c. Relevancy and Significance of These Figures 15. Courts and commentators have expressed a preference for actual applicant flow data in Title VII cases. Hazelwood School District v. United States, 433 U.S. 299, 308 n.13, 313 n.21, 97 S.Ct. 2736, 2742 n.13, 2744 n.21, 53 L.Ed.2d 768 (1977) (remand to District Court in “pattern or practice” suit to determine whether sufficiently reliable applicant flow data available, if so, such data would “be very relevant” to prima facie case); see New York Transit Authority v. Beazer, 440 U.S. 568, 585, 99 S.Ct. 1355, 1365, 59 L.Ed.2d 587 (1979) (no prima facie claim when, inter alia, plaintiffs failed to indicate how many methadone users applied for work with Transit Authority); Statistical Proof, supra, at § 4.1 et seq. (general preference for actual applicant data when representative of work force). 16. Although recognizing this general preference, the defendants present two challenges to the use of applicant flow data in the present case. On a more mundane level, defendants claim that the plaintiffs’ figures for the 1979 class are faulty and that the overall data for 1971-79 includes multiple applications from the same individuals. If defendants’ figures for the 1979 class are acceptedj however, the difference in selection rates for black and white applicants increases to approximately 22%. And while the plaintiffs’ data might involve some double counting because of multiple applications, the defendants have not offered evidence of double counting so pervasive or significant as to undermine the validity of the applicant data. See Robinson v. Union Carbide Corp., 538 F.2d 652, 657-68 (5th Cir. 1976), cert. denied, 434 U.S. 822, 98 S.Ct. 65, 54 L.Ed.2d 78 (1977) (defendants’ witnesses claimed up to six duplicative applications received from each applicant per year). *966 17. Defendants also object to the characterization of preapprentices, and individuals interested in the preapprentice program, as “applicants” in the data prepared by plaintiffs. Defendants assert that this technique inflates the applicant pool and penalizes affirmative efforts to recruit less qualified applicants or individuals not seriously interested in the sheet metal trade. Data prepared by the defendants, based on a more restrictive definition of “applicant,” indicates that between 1971 — 76 the selection rate for blacks was actually 9% greater than whites. (JAC Ex. 42). 18. There does appear to be support for concluding that in certain cases, affirmative action may result in a higher disparity in black-white selection rates than would exist under conditions of normal labor supply. Hill v. Western Electric Co., 12 FEP 1175, 1179 & n.4, 1180 (E.D.Va.1976) (dictum), rev’d on other grounds, 596 F.2d 99 (4th Cir. 1979), cert. denied, 444 U.S. 929, 100 S.Ct. 271, 62 L.Ed.2d 186 (1979); Statistical Proof, supra, at § 4.11. This is not one of those cases. The apprenticeship program is designed to impart training and skills to untrained individuals. While preapprentices, or individuals inquiring about the preapprentice program, might not have met the apprenticeship standards mandated by the JAC, e. g. high school diploma, it would be inappropriate for this court, in determining plaintiffs’ prima facie claim, to equate such standards with successful performance in the apprenticeship program. Spurlock v. United Airlines, Inc., 475 F.2d 216, 218 (10th Cir. 1973). In addition, any applicant who was willing to pursue the rather rudimentary, and at times disagreeable, program afforded preapprentices, supra at 958-959, was clearly interested in the apprenticeship program and sheet metal trade. There is nothing misleading in classifying those individuals routed into preapprenticeship as applicants. To a large measure, the selection criteria under attack in this suit, not a lack of interest or qualifications, prevented these individuals from competing for positions in the apprenticeship program. 13 19. Rejection of the defendants’ criticism does not by itself establish the significance of the plaintiffs’ applicant figures. The plaintiffs bear the burden of establishing their prima facie claim: the statistical proof must support an inference that the disparity in selection rates results from discriminatory activity, rather than chance. The plaintiffs stress that the selection rates for 1979 and 1971-79 violate the “four-fifths rule” adopted by the EEOC. 29 C.F.R. § 1607.4(D) (1979). Although EEOC guidelines are entitled to considerable deference, Albemarle Paper Co. v. Moody, 422 U.S. 405, 431, 95 S.Ct. 2362, 2378, 45 L.Ed.2d 280 (1975), the “four-fifths rule” has recently been criticized as a rather arbitrary standard, failing to account for differences in sampling size and tests results in the applicant population taken as a whole. Comment, Differential Pass-Fail Rates In Employment Testing; Statistical Proof Under Title VII, 91 Harv.L.Rev. 793, 797, 805 (1978). 20. Selection ratios have also been evaluated through the - use of the 5% significance level adopted by statisticians. See Harless v. Duck, 14 FEP 1616, 1623 (N.D. Ohio 1977). The 1979 applicant data lies on the fringes of such statistical significance; even plaintiffs admitted that a chi-square test established that the 12.5% disparity *967 was only “almost statistically significant.” The 1971-79 applicant figures appear to have more statistical significance (0.3%), but plaintiffs have not attributed this result to the selection criteria under attack in this suit, rather than other criteria, such as the GATB or maximum age requirement, also in effect during that period. 21. This is not to say that the actual applicant flow data is of no value. Plaintiffs’ selection rates show that the success rate for white applicants was 1.3 times that of blacks. Other courts have found prima facie claims with similar ratios. See, e. g., Chance v. Board of Examiners, 458 F.2d 1167 (2nd Cir. 1972) (equal protection) (success rate of white applicants 1.5 times that of blacks). And while a significance level of 5% is the conventional standard, certain circumstances may permit a court to infer that differences in selection rates result from discriminatory activity — even if the significance level approaches 6% or more. Cf. United States v. Georgia Power Co., 474 F.2d 906, 915 & n.11 (5th Cir. 1973) (5% level is “desirable goal”, not a prerequisite). 22. In the present case, there is good reason to believe that applicant flow data may not convey the full discriminatory impact of the selection criteria at issue. The high school requirement and arrest record inquiries allegedly dissuaded or “chilled” potential applicants. Consequently, plaintiffs supplement selection ratios by comparing.the percentage of blacks in the relevant labor pool for apprenticeships with the percentage of blacks selected as apprentices. It is important to recognize that a convincing disparity in such a work force comparison would add new meaning to selection ratios already on the verge of statistical significance. Boston Chapter, N. A. A. C. P., Inc. v. Beecher, 504 F.2d 1017, 1020-21 (1st Cir. 1974), cert. denied sub nom., Director of Civil Services v. Boston Chapter, N. A. A. C. P, Inc., 421 U.S. 910, 95 S.Ct. 1561, 43 L.Ed.2d 775 (1975) (“significant disproportionality in minority employment, coupled with even minimal proof of a higher minority failure rate [on selection exams]” establishes prima facie claim). The significance of plaintiffs’ relevant labor pool figures is discussed below. 2. Relevant Labor Pool Comparisons a. Plaintiffs’ Statistics 23. Plaintiffs offer two versions of the relevant labor pool from which individuals are likely to apply for apprenticeships. The first consists of men, aged 18-34, in the civilian labor force with one-four years of high school. (Pl.Ex. 146). Blacks constitute 39.54% of this group. Id. The second version identifies the previous occupation of applicants for apprenticeships from 1967-1976, and then estimates black representation in these previous occupations. (Pl.Ex. 148). Under this approach, blacks constitute 45.49% of the work force serving as a source of apprenticeship applicants. Id. If defendants’ final selection figures for 1979 are examined in light of plaintiffs’ claim that blacks constitute 40 — 15% of the relevant pool, the difference between the actual number of blacks selected (14) and the expected number (22-25) ranges from 8-11 (2.2-3.0 standard deviations). If the final selection figures for 1971-79 are examined in the same manner, the difference ranges from 43-62 (4.4-6.3 standard deviations). b. Relevancy and Significance of These Figures 24. From the onset of this case, defendants have contended that general population figures are the only proper measure of the relevant applicant pool. As this court indicated in its denial of defendants’ motion for summary judgment, recent case law clearly recognizes the use of relevant labor pool statistics in Title VII litigation. Hazelwood School District v. United States, 433 U.S. 299, 307, 308 & n.13, 310-12, 97 S.Ct. 2736, 2741, 2742 & n.13, 2743-44, 53 L.Ed.2d- 768 (1977) (race discrimination); Davis v. Califano, 21 FEP 272, 275-78 (D.D. C.1979) (sex discrimination). Plaintiffs do, however, present a rather novel use of relevant labor pool statistics. Labor pool statistics have been used when the job in question requires skills or expertise not present in the general population. Hazelwood School District v. United States, (1977) (“[w]hen special qualifications are required *968 to fill particular jobs, comparisons to the general population (rather than to the smaller group of individuals who possess the necessary qualifications) may have little probative value”). And such statistics usually have been used by defendants, as a means of rebutting disparities between minority representation in their work force and the general population. See e. g., E. E. O. C. v. Sheet Metal Workers, Local 122, 463 F.Supp. 388, 397-400 (D.Md.1978). 25. This court sees no reason why labor pool statistics cannot be used in certain circumstances to represent interested applicants for “entry-level” positions requiring only general work skills. General population figures tend to be overinclusive; they include individuals who are aged, infirm, or involved in professional careers and are not interested in an apprenticeship program in the sheet metal trade. Lerner, Quantity, Quality and Equality in Employment Testing, 1976 Sup.Ct.Rev. 263, 276 & n.46, 277. Moreover, the statistics in this case support the conclusion that the “figures for the general population might not accurately reflect the pool of qualified job applicants.” International Brotherhood of Teamsters v. United States, 431 U.S. 324, 339, 97 S.Ct. 1843, 1857, 52 L.Ed.2d 396 n.20 (1977); see Brown v. New Haven Civil Service Board, 474 F.Supp. 1256, 1260 n.4 (D.Conn.1979). Since 1971, blacks have constituted, on the average, 29% of all individuals selected for the apprenticeship program. This is in the face of threshold criteria with an acknowledged exclusionary effect on black applicants; blacks are approximately five times more likely than whites not to have the high school requirement or have an arrest record. Supra at 960. With these exclusionary criteria in effect, it is reasonable to conclude that blacks could constitute 29% of all those selected only if the percentage of interested blacks in the applicant pool was considerably greater. 14 Under these circumstances, the defendants’ claim of meeting the percentage of blacks in the general population does not preclude a finding of discriminatory impact. 26. Other courts have recognized that the congruence of general population percentages and racial representation in the defendant’s work force should not bar consideration of more relevant evidence of discrimination. See United States v. International Longshoremen's Ass’n, 334 F.Supp. 976, 978-79 (S.D.Tex.1971), rev’d on other grounds sub nom., E. E. O. C. v. International Longshoremen’s Ass’n, 511 F.2d 273 (5th Cir.), cert. denied, 423 U.S. 994, 96 S.Ct. 421, 46 L.Ed.2d 368 (1975) (job opportunities should be based on “number of Blacks and Whites that makes themselves available for longshore work,” not racial composition of community). This refusal to be tied to general population percentages is consistent with the mandate of Title VII. As the quote from Albemarle cited earlier indicates, supra at 21, Title VII protects qualified individuals from discrimination in employment. It does not require employers to hire minority groups in proportion to their representation in the population. 15 Consequently, general population statistics lose their significance when they are no longer reasonable proxies for the applicant pool. 27. Aside from relying on general population figures, the defendants attack certain premises underlying plaintiffs’ relevant labor pool figures. Defendants claim that PI.Ex. 146 inflates the number of potential black applicants (40%) by restricting the civilian labor force to men with 1-4 years of high school in the Washington SMSA. If the labor force was expanded to include *969 women, individuals with some college education, and the entire jurisdictional area, the percentage of blacks in the relevant labor pool would decrease. Defendants’ contention is simple: plaintiffs’ statistics present an unduly restrictive image of potential applicants; a proper, more expansive, picture would indicate a smaller percentage of blacks. 28. Recent case law goes against the defendants. The proper geographic area for determining the relevant labor pool is where most of the applications originate. United States v. Ironworkers, Local 86, 443 F.2d 544, 551 n.19 (9th Cir. 1971), cert. denied, 404 U.S. 984, 92 S.Ct. 447, 30 L.Ed.2d 367 (1971). In this case it is the Washington SMS A. See E. E. O. C. v. E. I. duPont de Nemours & Co., 445 F.Supp. 223 (D.Del.1978) (choosing Wilmington SMSA based upon pattern of applications to defendant’s plants). The sexual composition and educational background of the current work force are also acceptable grounds upon which to predict the type of individual likely to apply for an apprenticeship position. See E. E. O. C. v. Local 14, International Union of Operating Engineers, 415 F.Supp. 1155, 1170 (S.D.N.Y.1976), modified on other grounds, 553 F.2d 251 (2nd Cir. 1977) (educational background of union members makes it reasonable to assume applicants have high school education or less). Local 102 is all male, and many of its members have had only 1-4 years of high school. The defendants have failed to show that Pl.Ex. 146 offers an unreasonable portrait of the relevant labor pool and the percentage of blacks within that pool. 29. The defendants also attack Pl.Ex. 148 on the grounds that it too establishes a labor pool with an inflated percentage of blacks. Pl.Ex. 148 categorizes the previous occupation of applicants, and the percentage of the total applications each occupation represents, determines the percentage of blacks in each occupational category, and then extrapolates to determine the percentage of blacks in the range of occupations from which applicants were drawn. The defendants contend that many of the occupations that produce applicants, e. g., “laborers,” include a high proportion of older blacks unlikely to apply for apprenticeships. Local 102 no longer maintains a maximum age limit for applicants. Theoretically, older minority members are in the running for apprenticeship positions. There is, however, some truth to the defendants’ contentions. Those individuals most likely to apply to the apprenticeship program would be on the threshold of their careers. See Hazelwood School District v. United States, 433 U.S. 299, 310-13, 97 S.Ct. 2736, 2743-44, 53 L.Ed.2d 768 (1977) (the interest of potential applicants in pursuing job opportunities must be considered in determining relevant labor market). Unfortunately neither the plaintiffs nor defendants has offered an alternative figure based on the “previous occupation” data in Pl.Ex. 148 reflecting the percentage of blacks in a more youthful labor market. 30. This court is unwilling to reject totally the labor market figures derived from Pl.Ex. 148. Because the exhibit rests on the labor patterns of actual applicants, it most likely reflects a more accurate labor market. Perhaps the best approach is to follow the course suggested by the Court in Hazelwood and derive an intermediate figure from Pl.Ex. 146 and 148. -433 U.S. 311-12, 97 S.Ct. at 2743-44. If blacks are assumed to constitute 42% of the relevant labor market, then the difference between the actual number of blacks selected and the expected number is 2.6 standard deviations for 1979 and 5.3 standard deviations for 1971-79. As in the case with actual applicant flow data, the 1979 figure is on the threshold of statistical significance, Hazelwood School District v. United States, 433 U.S. 299, 308 n.14, 311 & n.17, 97 S.Ct. 2736, 2742 n.14, 2743 & n.17, 53 L.Ed.2d 768 (1977); Castaneda v. Partida, 430 U.S. 482, 496-97 & n.17, 97 S.Ct. 1272, 1281-82, 51 L.Ed.2d 498 (1977) (if difference between expected value and observed number is greater than two or three standard deviations it is likely that the outcome is due to factor other than chance). The 1971-79 figure is statistically significant, but, again, additional selection criteria were in use during that period of time. Supra at 957 n.7. *970 3. Combined Effect of Plaintiffs’ Applicant Flow and Relevant Labor Market Statistics 31. The applicant flow and relevant labor market statistics offered by plaintiffs, when viewed together, establish a prima facie claim that the overall apprenticeship selection process discriminates against black applicants. The Court has warned that “precise calculations of statistical significance are [not] necessary in employing statistical proof [in Title VII eases] ..” Hazelwood School District v. United States, 433 U.S. 299, 311 n.17, 97 S.Ct. 2736, 2743-2744, 53 L.Ed.2d 768 (1977). Plaintiffs’ evidence derived from the 1979 apprentice class — actual applicant data and work force comparisons — brings this court to the verge of rejecting the hypothesis that the observed disparity was caused by chance. When this 1979 data is augmented by the 1971-79 data and collateral information about the selection process, plaintiffs’ prima facie claim for all three selection criteria prevails. 32. Although the 1971-79 data may, to some extent, reflect selection criteria no longer in effect, there are reasonable grounds for assuming that a good portion of the additional discriminatory impact arising out of those years is due to all three criteria at issue in this case. The actual disparate impact of the interview has been established; the projected disparate impact of the high school requirement and arrest inquiry is clear. Plaintiffs’ actual applicant data and relevant labor market figures now provide a proper foundation for attributing practical significance to these projected exclusionary rates. Interested black applicants exceeded the percentage of blacks in the general population and percentage of blacks chosen as apprentices. The disparity in the pre-1979 classes between interested black applicants and black apprentices cannot be attributed solely to selection factors no longer in existence, or the interview operating alone. See Pl.Ex. 121 (First Joint Stipulation on Underlying Statistics) (isolating disparate impact of GATB and age requirements); Vulcan Society v. Civil Service Commission, 490 F.2d 387, 392 (2nd Cir. 1973). 33. Moreover, plaintiffs have supplemented the projected exclusionary effect of the high school requirement with actual applicant figures. From 1968-76, blacks were rejected for failing to produce a high school diploma at three times the rate of whites. Plaintiffs’ Memorandum, Table D. Although similar figures were not submitted for the arrest inquiry, it is reasonable to assume that some portion of the projected exclusionary impact was experienced in practice. The defendants have offered no evidence rebutting the conclusion that the present criteria contributed in part to the disparate impact observed in 1971-79 and that the high school requirement and arrest inquiry, in addition to the interview, were responsible for this result. In short, the plaintiffs have established the overall discriminatory impact of the selection procedure and the three components challenged in this action. See 41 C.F.R. § 60-3.4(e) (1979) (if total selection process has adverse impact, individual components should be evaluated). 34. It is also important not to lose sight of collateral evidence adding flesh to plaintiffs’ statistical skeleton. Statistical Proof, supra, at § 9.0. Local 102 has less than 5% minority membership. Supra at 956. Since the mid-1960’s applicants have been subject to screening criteria not applied to older members of the union. Id. 957-958. In recent years, the qualifications of black applicants have not been registered properly by individuals administering the program, id. at 961; more disturbingly, black applicants meeting the threshold qualifications have been rejected from the program in a cavalier manner; id. at 961-962. This actual conduct is another factor underlying the finding of a prima facie claim. II. The “Business Necessity” Defense 35. The defendants must now justify the use of their selection criteria. While Title VII prohibits employment discrimination, it permits an employer to give and act *971 upon “the results of any professionally developed ability test.” 42 U.S.C. § 2000e-2(h) (1976). In Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971), the Court set the scope of this section and the burden an employer must meet to use selection criteria having a disparate impact on minority applicants. 36. This section was construed to apply to tests and other criteria for employment, e. g., high school diploma, that acted as “fixed measures of capability.” 401 U.S. at 433, 91 S.Ct. at 854. This approach has been adopted by the E.E.O.C. which defines “test” to include “specific educational or work history requirements,” “scored interviews,” and “biographical information blanks.” 29 C.F.R. § 1607.2 (1978). The high school requirement, arrest inquiry, and scored interview at issue qualify as the type of selection criteria identified in Griggs and by the E.E.O.C. Second, the Griggs Court required the employer to show that the selection criteria bear a “demonstrable relationship to successful performance of the jobs for which it was used.” 401 U.S. at 431, 91 S.Ct. at 853. In Albemarle Paper Co. v. Moody, 422 U.S. 405, 95 S.Ct. 2362, 45 L.Ed.2d 280 (1975), the Court made it clear that the relationship between selection criteria and job performance was not to be taken at face value; “job relatedness” was to be established only after a thorough examination of the employer’s operations, the history of the selection criterion under attack, and the validation guidelines promulgated by the E.E.O.C. Id. at 427, 430-36, 95 S.Ct. at 2376, 2377-80. Even if selection criteria were found to be “job related,” the Court in Albemarle suggested that less restrictive selection devices might be required. Id. at 425, 95 S.Ct. at 2375. 37. The defendants’ have concentrated on the high school requirement, with the interview and arrest inquiry receiving rather cursory justifications. a. “Job Relatedness” of the High School Diploma Requirement 38. Neither the operations of Local 102 nor the history behind the high school requirement support the job relatedness of the criterion. The requirement was adopted in the mid-1960’s without any validation study, id. at 428, 95 S.Ct. at 2376, and many members of the union without high school diplomas appear to have performed their jobs adequately, Griggs v. Duke Power Co., 401 U.S. 424, 431, 91 S.Ct. 849, 853, 28 L.Ed.2d 158 (1971). Indeed, plaintiffs’ testimony established that the majority of sheet metal workers throughout the country lack high school diplomas. 39. The defendants contend that this educational portrait of the sheet metal industry is misleading. According to the defendants, Local 102 is engaged in the most demanding aspect of the sheet metal trade— commercial sheet metal fabrication and installation in the building and construction industry. Union members are expected to display superior expertise and ability in the performance of their duties, and their salaries reflect the premium attributed to these talents. While some union members without high school diplomas have acquired these skills over the years, only apprenticeship applicants with high school diplomas can be expected to master promptly the classroom exercises, and apply their new knowledge to the demands of the job site. 40. Under Griggs and Albemarle, these general contentions must rest on professional validation studies indicating a significant correlation between the high school diploma and performance in the apprenticeship program. See Albemarle Paper Co. v. Moody, 422 U.S. 405, 430-31, 95 S.Ct. 2362, 2377-78, 45 L.Ed.2d 280 (1975); 29 C.F.R. § 1607.3 (1978) (criteria must be validated); cf. 41 C.F.R. § 60-3.3 (1979) (same) (standards regarding selection criteria used by federal contractors). 41. The defendants have attempted to validate the high school requirement through “construct validation.” This validation process centers on identifying general mental and psychological traits determined to be important in job performance (“constructs”), and then showing that the selection criteria is validly related to the *972 construct. Douglas v. Hampton, 512 F.2d 976, 984 (D.C. Cir. 1975); Vulcan Society v. Civil Service Commission, 490 F.2d 387, 395 (2nd Cir. 1973). As this short description indicates, construct validity involves a two-step process. The construct must be suitably related to the job in question and the selection criteria must reflect the construct. Douglas v. Hampton, 512 F.2d 976, 985-86 (D.C. Cir. 1975). 42. There are two problems with the use of construct validation. First, construct validation is more complex than other validation techniques, and not well recognized as a means of satisfying the Griggs standard of job relatedness. Id. at 985 & n.73; 41 C.F.R. 60-3.14(D)(1) (1979). In Douglas v. Hampton, the Court of Appeals for the District of Columbia held that construct validity may be considered only after a showing that it is infeasible to use criterion validation techniques. The defendants failed to show that criterion validation could not be applied to the high school requirement. 16 43. Even if criterion validation was infeasible, the defendants’ construct validation cannot withstand careful scrutiny. The key to construct validation is insuring that the job is properly represented by the construct, and the construct by the selection criteria. Supra at 971. Defendants used two constructs: “general learning ability” and “readability.” Neither construct properly reflected the job in question, i. e. a position in the apprenticeship program. 29 C.F.R. § 1607.5(a) (1978). “General learning ability” was based on a job profile prepared by sheet metal workers and reflecting tasks performed on the job site. The proper job analyses would have centered on the demands and characteristics of the apprenticeship program. The “construct” reflected job qualifications required of experienced workers, not raw recruits subject to a four-year educational program. Albemarle Paper Co. v. Moody, 422 U.S. 405, 433-34, 95 S.Ct. 2362, 2379, 45 L.Ed.2d 280 (1975); 29 C.F.R. § 1607.4(c)(1) (1978) (candidates should be evaluated for a job at or near entry level when employees’ potential may be expected to change in significant ways). The “readability” construct suffers from the same flaw. The defendants’ expert assumed that all the materials analyzed were taught during the course or used during sheet metal work; yet this assumption was never proven adequately during the hearing. 17 44. Finally, the high school requirement has not been found to be job related in cases involving entry level positions or apprenticeship programs of the sheet metal workers and other trade unions. See *973 Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971) (overall hiring requirement); Donnell v. General Motors Corp., 576 F.2d 1292 (8th Cir. 1978) (apprenticeship program sponsored by UAW); Watkins v. Scott Paper Co., 530 F.2d 1159 (5th Cir. 1976) (high school requirement for intra-company transfer); Pettway v. American Cast Iron Pipe Co., 494 F.2d 211 (5th Cir. 1974), cert. denied, 439 U.S. 1115, 99 S.Ct. 1020, 59 L.Ed.2d 74 (1979) (employer’s apprenticeship program for craft jobs); United States v. Georgia Power Co., 474 F.2d 906 (5th Cir. 1973) (high school requirement for intra-company transfer); Patterson v. Youngstown Sheet and Tube Co., 475 F.Supp. 344 (N.D.Ind. 1979) (employer’s apprenticeship program); Rule v. International Association of Bridge Workers, Local No. 396, 471 F.Supp. 1335 (E.D.Mo.1979) (apprenticeship program of ironworkers); United States v. Sheet Metal Workers, Local 10, 6 FEP 1036 (D.N.J.1973) (apprenticeship program); E.E.O.C. v. Sheet Metal Workers, Local 638, 401 F.Supp. 467 (S.D.N.Y.1975), modified on other grounds, 532 F.2d 821 (2nd Cir. 1976) (apprenticeship program). 18 45. None of the foregoing should be interpreted as prohibiting the use of selection criteria incorporating any high school education. The defendants do not have unlimited resources to fund a training center accepting applicants on a “first come” basis. Defendants obviously desire quick learners who can maintain the union’s reputation for workmanship. Testimony at the hearing established that basic algebra and trigonometry, and general reading skills, are necessary for successful performance in the apprentice program. Some courts have approved of apprenticeship programs incorporating a requirement of a minimum of one year of algebra or geometry, Donnell v. General Motors, 576 F.2d 1292 (8th Cir. 1978), or two years of high school, United States v. Sheet Metal Workers, Local 10, 6 FEP 1036 (D.N.J.1973). In restructuring their educational requirements, the defendants should examine the required course content in area high schools to determine when students are exposed to basic algebra, geometry or trigonometry. Upon further consideration by this court, the year in which such subjects are taught may prove an acceptable educational criteria. b. “Job Relatedness’’ of the Arrest Inquiry and Interview 1. Arrest Inquiry 46. The defendants claim that arrest records were not considered during the selection process. Yet the arrest record inquiry appears on the application form and applicants were questioned repeatedly about arrest records during interviews. Supra at 961-962. This court must conclude that arrest records played a role in the selection process, as a factor in chilling potential applicants and as an element affecting an applicant’s evaluation. The defendants have made no attempt to validate the arrest inquiry as job related; consequently, it must be eliminated. Johnson v. Goodyear Tire & Rubber Co., 491 F.2d 1364, 1371-72 (5th Cir. 1974) (failure to validate compels conclusion requirement was invalid). See Gregory v. Litton Systems, Inc., 316 F.Supp. 401 (C.D.Cal.1970), modified on other grounds, 472 F.2d 631 (9th Cir. 1972) (“no hire” policy for sheet metal mechanic arrested but not convicted not justified by business necessity). 19 *974 2. Interview 47. The interview is the most important part of the application process. It is bifurcated into educational and interview portions, with the latter consisting of a number of broad and undefined criteria. Supra at 957-958. Subjective evaluations have been condemned, whether as a means of admitting applicants into unions or determining those employees qualified for promotions. Watkins v. Scott Paper Co., 530 F.2d 1159, 1191-92 (5th Cir. 1976) (promotion); United States v. Sheet Metal Workers, Local 36, 416 F.2d 123, 135-36 (8th Cir. 1969) (admission into sheet metal union). The interview procedure at issue in this case has many of the undesirable characteristics associated with subjective evaluations: categories such as “attitude,” “interest in trade,” “work experience” are likely to reflect the subjective judgments of the interviewers; no guidelines exist elaborating upon or narrowing the criteria, e. g. what work experience might be important in predicting an applicant’s success in the program; the interviewers’ judgments are unreviewable. Compare with E.E.O.C. v. E. I. duPont de Nemours & Co., 445 F.Supp. 223, 253-54 (D.Del.1978) (disparate impact inherent in subjective criteria reduced by the institution of procedures and standards to control and review subjective judgment). There is also evidence that these criteria were used to favor the sons or relatives of union members, by giving such applicants high ratings in “attitude,” “conduct,” or “work experience.” Supra at 960-961. 20 48. The defendants have not attempted to validate any aspect of the interview; the entire process must be rejected for lack of job relatedness. While this interview procedure is unsatisfactory, more well tailored interviews may be permissible. 29 C.F.R. § 30.5(b)(l)(iv) (1979) (oral interviews permissible if limited to objective questions related to fitness of applicants and not qualifications previously determined). In devising new selection criteria, the defendants should identify those characteristics associated with successful applicants, e. g., drive, ambition, reliability, and the numerous academic, social, or work experiences applicants might have that are indicative of these qualities. Guidelines should then be developed, for use by interviewers and applicants, identifying the desired characteristics and the various ways an applicant may have satisfied them. 21 While it may be impossible and undesirable to eliminate all subjective aspects of the interview, the use of enunciated standards, tailored to the demands of an entry-level program and uniformly applied by all interviewers, will guard against abuses of judgment. III. Issuance of A Preliminary Injunction 49. The plaintiffs have established a prima facie case that certain selection procedures utilized by the defendants violate Title VII. Defendants have failed to rebut that showing and appear unlikely to do so at trial. The merits of plaintiffs’ claim certainly are strong enough to support the issuance of a preliminary injunction. Washington Metropolitan Area Transit Comm’n v. Holiday Tours, Inc., 559 F.2d 841, 844 (D.C. Cir. 1977); Virginia Petroleum Jobbers Ass’n v. FPC, 259 F.2d 921, 925 (D.C. Cir. 1958). 50. Plaintiffs also have met the other standards associated with the issuance of an *975 injunction. Without relief, potential black applicants will be deterred because of defendants’ advertisement of the high school requirement and use of the arrest inquiry. Blacks who apply will suffer the disparate impact of the interview process. There is no guarantee when the next apprenticeship class will be chosen nor any indication of the size of subsequent classes. Failure to act now may preclude proper black representation in the program for an indeterminate period. Continued use of the contested selection criteria will cause irreparable injury to current and potential black applicants. McKenzie v. McCormick, 15 FEP 1310, 1312 (D.D.C.1972). It clearly is in the public interest to protect black applicants from the effects of employment discrimination. Cannistra v. FAA, 20 Empl.Prac.Dec. ¶ 30, 134 (D.D.C.1979). 51. The defendants contend that elimination of the current selection standards will produce severe financial hardship. They envision injunctive relief as resulting in an intolerable delay in the formation of an apprenticeship class and the eventual selection of unqualified applicants. Neither fear is warranted; defendants misconceive the type of injunctive relief necessary at this stage in the proceedings. This memorandum and order gives the defendants considerable leeway in devising new selection criteria. Supra at 973-974. The quality of future applicants, and the prompt administration of the apprenticeship program, rest with the defendants. 52. This approach depends on the good faith efforts of the defendants. This court will not look favorably on a newly tailored application procedure resulting in the same familiar fit. The discretion afforded by an interview is to be used to choose those applicants most likely to succeed in the apprenticeship program. It is not to be used as a subterfuge for reinstituting the high school requirement, favoring the children, relatives, or friends of members of the union, or choosing minimally qualified black applicants unlikely to become journeymen. If the defendants fail to act responsibly, this court will be forced to impose a “goal” system insuring adequate black representation in the apprenticeship program. See Rios v. Enterprise Association Steamfitters Local 638, 501 F.2d 622 (2nd Cir. 1974), on remand, 400 F.Supp. 983 (S.D.N.Y.1975) (42 U.S.C. § 2000e-5(g) (1976) permits imposition of goals to eliminate effects of defendant’s past discrimination and bar like discrimination in the future). 53. For the reasons presented in this memorandum, it is hereby ORDERED that plaintiffs’ motion for preliminary injunctive relief against Local 102, SMACNA-D.C., and JAC is granted, and it is furthermore ORDERED that: (1) Defendants not act on any pending applications until the provisions of paragraphs (2)-(7), listed below, are complied with; (2) Defendants eliminate the requirement that applicants possess a high school diploma or equivalent and the requirement that applicants present a high school transcript; (3) Defendants do not consider subjective personal interview scores in selecting applicants; (4) Defendants do not rely upon arrest record inquiries in selecting apprentice applicants; (5) Defendants take affirmative action to notify the black community of the elimination of these discriminatory hiring criteria, inform the court of the specific actions taken, and postpone consideration of any pending applications until potential black applicants have had a reasonable opportunity to respond to such affirmative action; (6) Defendants afford plaintiffs an adequate opportunity to (a) review the criteria and procedure used to select any apprenticeship class before such criteria or procedure are put in effect, and (b) review the racial composition of that class and bring to this court’s attention any failure to adhere to approved procedures or criteria; and (7) Any new selection criteria or procedures which defendants may choose to employ must be approved by this court before use by the defendants. Footnotes [1] The Amended Complaint also added as defendants the twenty-five employers belonging to SMACNA DC, the local association of sheet metal contractors. [2] The class certification adds that such discrimination must have occurred since July 9, 1971, for Title VII claims, and May 14, 1972, for § 1981 claims. [3] Applicants were required to have a high school diploma and transcript or a certificate of equivalency (G.E.D.). For convenience sake, these provisions are referred to as the high school requirement. [4] Plaintiffs also requested that (a) the defendants take affirmative action to notify the black community of the elimination of these selection criteria, and (b) any new selection criteria be reviewed by the plaintiffs and this court. Plaintiffs did not contest JAC’s minimum age (17), residency (within the jurisdictional area of Local 102), or physical ability (capable of performing the work of the trade) requirements. Although plaintiffs alleged that JAC planned to use a maximum age requirement for the new class, this does not appear to be the case. [5] On January 25, 1980, this court denied plaintiffs’ motion for a preliminary injunction against SMWIA, SMACNA and the local contractor-defendants. Local 102, SMACNA DC, and the JAC remained as parties. [6] Although the underlying complaint alleges a violation of 42 U.S.C. § 1981 (1976), see Macklin v. Spector Freight Systems, Inc., 478 F.2d 979, 993 -94 (D.C. Cir. 1973), the motion for a preliminary injunction is grounded on Title VII. The disposition of the motion on these grounds makes it unnecessary to determine whether the defendants violated § 1981. [7] Before 1975, the JAC required an applicant to pass a General Aptitude Test (“GATB”) administered at the Apprentice Information Center (“A1C”) of the D.C. Department of Labor. Before 1980, JAC also required an applicant to be under 24 years of age, with a four-year military allowance. (Testimony of Donald Alter; Testimony of Eugene Walker and Clarence Neese.) [8] If two nonblack applicants are eliminated from the sample because they were accepted into the apprentice program as a result of the unionization of their shop, rather than direct application to JAC, and one applicant is properly redesignated as black, the nonblack acceptance rate increases to 53.2% while the black acceptance rate drops to 31%. [9] Supra at 960. [10] In Dothard, the Court estimated that the criteria excluded 41.13% of the female population in the United States. Although women in Alabama were 52.75% of the population, they held only 12.9% of the state correctional counselor positions. Thus the projected impact was reflected in the actual figures. 433 U.S. at 329 30, 97 S.Ct. at 2726-27. [11] For the sake of simplicity, this explanation ignores for the moment the separate discriminatory impact of the interview. A more precise approach would set the percentage of “interested” blacks in the applicant pool above 26%, allowing for the discriminatory impact of the interview to reduce black applicants to the actual representation in the defendants' work force. [12] This is not to suggest that evidence of the overall exclusionary impact of the high school requirement or arrest inquiry is irrelevant. Such evidence supports plaintiffs’ claims that potential black applicants were “chilled” from applying and that black representation in the relevant labor market far exceeds general population figures. Infra at 967 -968. It also permits this court to identify further individual discriminatory components of the selection process, in addition to the interview, after a finding of overall discriminatory impact. And when figures indicating an overall discriminatory impact include years when other threshold criteria were in effect, e. g., GATB, the general impact of the high school diploma requirement or arrest inquiry can lead to an inference about their particular effect in the selection process. [13] There is another reason why the use of plaintiffs’ applicant flow data does not penalize the defendants for affirmative minority recruitment. Employers are most likely to suffer such penalization when a plaintiff can establish a high disparity in selection rates and is unable to offer additional proof of an untapped pool of interested and potentially qualified applicants. This was the scenario in Hill v. Western Electric Co., 12 FEP 1175 (E.D.Va.1976), rev’d on other grounds, 596 F.2d 99 (4th Cir. 1979), cert. denied, 444 U.S. 929, 100 S.Ct. 271, 62 L.Ed.2d 186 (1979). In this case, however, plaintiffs have supplemented the applicant flow data with statistics indicating that the percentage of blacks in the relevant labor pool is much higher than the percentage of black apprentices. Defendants’ initial claims of affirmative action can be reexamined in the light of such a showing. If the plaintiffs make a convincing case, the selection policies at issue actually operate to restrict applications from qualified, as opposed to disinterested or unqualified, black applicants. [14] In this case, 31% of blacks, aged 20-24, in the Washington area lack four years of high school. The equivalent figure for whites is 5.58%. Assume that 100 applicants are selected for an apprenticeship program requiring a high school diploma. Thirty of the individuals selected are black, 70 are white. Because of the larger black failure rate, there must be 42 black applicants to insure that 30% of the individuals selected are black. There must be 73 white applicants to insure that 70% of the individuals selected are white. Although blacks constitute 30% of those individuals chosen, they constitute 36% of the applicant pool. [15] 42 U.S.C. § 2000e-2(j) (1976) (expressly stating that general population statistics should not be grounds for inferring employer liability). [16] Criterion validation examines the job performance of candidates who lack the qualification (high school diploma) and determines whether any difference in job performance is significant. In this case, a number of applicants who had not completed high school could have been admitted into the apprenticeship program, or the performance of current members of Local 102 (or similarly situated sheet metal workers) without a high school diploma could have been examined. (PI.Ex. 124.) The defendants’ expert witness, Dr. Denton, acknowledged that the accepted way of validating the high school requirement would have been to conduct a criterion-related validation study and that such a study could have been conducted if even 10% of the sample lacked a high school diploma or G.E.D. [17] The second element necessary for construct validation is a proper fit between the construct and the selection criteria. The defendants’ study also is flawed in this regard. After the defendants’ expert devised a job profile, he used data from other jobs (presumably with similar profiles) to predict the salary and various test scores of sheet metal workers. Supra at 962. He then determined the median score and salary of high school graduates. Because the test scores of sheet metal workers exceeded the same median figure for high school graduates, the high school diploma was said to have been validated. The court is puzzled by this chain of reasoning. The approach rests on identifying predicted characteristics of sheet metal workers (test performance, salary) and comparing them to similar characteristics for high school graduates. Yet no attempt was made to explain why these characteristics were indicators of success in the apprenticeship program, and, more importantly, why a high school diploma had any significant correlation with such characteristics. The case law requires a more convincing showing. Douglas v. Hampton, 512 F.2d 976, 985 -86 (D.C. Cir. 1975); Vulcan Society v. Civil Service Commission, 490 F.2d 387, 395 96 (2nd Cir. 1973). [18] The message of these cases is conveyed by the following quotation from Pettway: [W]e understand the company to mean that a certain reading level and familiarity with study techniques is necessary to participate in the course work of the apprentice program. This cannot be equated with a requirement for a high school education or its equivalent. . The high proficiency level established by this standard not only precludes qualified employees but also is not refined sufficiently to measure the ability sought by the company. 494 F.2d at 237-38. [Emphasis in original.] [19] Arrest inquiries are to be distinguished from inquiries concerning convictions. Use of the latter reflects recognizable employer concerns not now before this court. [20] Previous experience in the sheet metal or construction trades was accorded substantial weight in formulating interview scores. This emphasis perpetuates the effect of past discrimination in the construction trades, United States v. Sheet Metal Workers, Local 10, 6 FEP 1037, 1042 (D.N.J. 1973), and does not appear to be a criteria of overriding importance for admission in an entry-level position, Watkins v. Scott Paper Co., 530 F.2d 1159, 1192 (5th Cir. 1976). Past work experience may be important in highlighting an applicant’s reliability or motivation, but the type of work should not be an overriding factor in predicting an applicant’s success. [21] It is important for interviewers and applicants to realize that experiences other than those in the sheet metal or construction trades can provide a clear indication of characteristics necessary for success in an entry-level program. Supra at 974 & n. 20.