Equal Employment Opportunity Commission v. United Ass'n of Journeymen & Apprentices of the Plumbing & Pipefitting Industry, Local 120

9 F. Supp. 2d 836, 1998 U.S. Dist. LEXIS 9799

• Court: District Court, N.D. Ohio
• Decided: June 25, 1998
• Docket: C68-473
• Status: Published

Opinion

OPINION AND ORDER

GWIN, District Judge.

In this employment discrimination case involving claims that a plumbers’ union discriminated against black pipe fitters in job referrals, the Court decides whether the Court should give additional relief to class members. The Court decides whether Plaintiff EEOC and Plaintiff-Interveners show evidence that from 1978 to 1992, Defendant Local Union 120 discriminated in job referrals based on race.

On January 29, 1972, this Court entered a consent decree in this case. In that decree, the United States Department of Justice and the United Association of Journeymen and Apprentices of the Plumbing & Pipefitting Industry of the United States and Canada, Local No. 120 (“Local Union 120”) agreed to union job referral procedures. The case is again before the Court to assess damages resulting from Local Union 120’s claimed noncompliance with the consent decree.

In deciding Plaintiffs’ motion for contempt, the Court first reviews the procedural posture of this case. Second, the Court examines whether this Court should exercise discretion to reexamine earlier findings when strong evidence suggests that earlier Court findings were mistaken. Third, the Court reviews the parties’ burdens of coming forward with evidence proving or disproving discrimination in Local Union 120’s referral system. Finally, the Court discusses what evidence of damage exists in the record before this Court.

For the reasons that follow, the Court denies Plaintiff-Interveners’ request additional relief.

I. Procedural background

In this case, Plaintiff-Interveners say Defendant Local Union 120 pursued racially discriminatory policies and failed or refused to refer black union members for employment on the same basis as the Union referred white union members. Plaintiff-In-terveners say Local Union 120 did this by passing over black members on an out-of-work list.

Local Union 120 is a craft union that represents pipe fitters in the construction industry. It is a signatory to collective bargaining agreements with the Mechanical Contractors Association. Under these agreements, Local 120 is the exclusive referral source for pipe fitters to members of the Contractors Association in the Cleveland area. These employers variously provide HVAC, sprinkler, welding and atomic power construction services.

Because construction work is intermittent, employers frequently supplement their work force with referrals from Local Union 120. In this 1968 litigation, the Plaintiff EEOC claimed Defendant Local Union 120 discriminated based upon race in making these referrals.

As indicated, Defendant Local Union 120 and the Plaintiff EEOC reached agreement over a consent decree. Under this January 20, 1972, consent decree, Local Union 120 was required to maintain out-of-work lists and to generally refer members for work in the order in which they registered for work on the list. As major exceptions to this rule, Local Union 120 was required to allow employers to recall former workers within certain time periods. As further exceptions' to the first registered rule, Local Union 120 was required to respond to requests- for union members with specialized skills.

On March 29, 1990, the Plaintiff EEOC sought an order requiring Defendant Local Union 120 to show cause why it should not be held in contempt for violating the consent decree. In its motion, the EEOC said the union had failed to maintain master registration cards, had failed to maintain registration lists and had failed to make referrals as

required by the January 20, 1972, consent decree.

On September 1, 1992, this Court issued various Orders in this case. 1 In those, the Court found that Defendant Local Union 120 did not maintain a referral register and in 1979 stopped maintaining master registration cards. This Court found Local Union 120 did so without court approval. 2 While finding that Local Union 120 failed to comply with record keeping requirements of the consent decree, this Court found that no adverse inference' may be drawn from its failure to maintain or produce records for these proceedings. The Court refused certain Local Union 120 defenses and denied dismissal of the Mechanical Contractors.Association.

The Court found Local Union 120 in civil contempt. First, this Court found that Local 120 Business Agents could generally predict how long an upcoming job was likely to last. Judge Thomas further found a pattern or practice of discrimination by the union in the operation of its work referral system. 3 As a result, Judge Thomas found that black union members received fewer overall work hours than white members. 4

As a remedy, Judge Thomas directed that Local Union 120 improve record keeping. 5 To remedy individual discrimination, he also ordered that any class member who contends s/he was intentionally “passed over” shall first file a written grievance with the joint hiring committee. If the committee denies the grievance, the complaining class member could file request for review with the Magistrate Judge of this Court. If the black pipe fitter filed such a request for review, the Magistrate Judge would make a de novo review.

Since the time Judge Thomas ordered this grievance procedure be set up in September 1992, no class members filed a grievances with Local 120 claiming that he/she was “passed over” by a work referral given to a white pipe fitter.

Judge Thomas further ordered a coercive fine of $200,000 against Local 120. Of this, $100,000 would be reserved to pay damages ordered under the provision for individual claims described above. Judge Thomas fashioned an algorithm to distribute the other $100,000 among the 106 black class members.

Finally, Judge Thomas ordered the apprenticeship program indenture one black pipe fitter for each white pipe fitter indentured.

In his contempt order, Judge Thomas found that the totality of the evidence presented in the contempt proceedings:

... establishes a pattern or practice of disparate treatment as to Black pipefitters on the part of Local 120 in the operation of its work referral system ... nor does the entire record rebut, Plaintiff-Interveners’ prima face [sic] case of disparate treatment ...

******

But the nature and content of the evidence in the record would not enable any particular black pipefitter to show that one or more business representatives, in making referrals of shorter likely duration than particular pipefitters received. Hence specific loss of pay could not be proved by the particular pipefitter and he or she therefore could not be awarded specific compensatory damages.

The Court also found “loss of pay cannot be awarded to individual members of the class.”

Plaintiff-Interveners moved this Court to amend its December 22,1992, final order and judgment. On February 2, 1993, this Court granted Plaintiff-Interveners’ motion to amend the final judgment. In granting Plaintiff-Interveners’ motion to amend the December 22, 1992, order, this Court ruled:

This Court determines that an additional (second stage) hearing will be held to determine what, if any, “individual relief’ should be awarded.... Hence, this Court questions whether the present state of the record would warrant the application herein of the Teamsters holding.

In his February 2, 1993, memorandum order, Judge Thomas stated that his earlier findings as to a pattern and practice of discrimination were subject to revision after this Court’s Stage II hearing:

[Defendant Local 120 has the burden of proof at the Stage II hearing. However, whether Local 120 has carried its burden in the damages hearing will not be affected by this Court’s finding [the earlier finding that the Local Union 120 had not rebutted Plaintiffs case of prima faca [sic] discrimination.]

* * * * * *

Depending on the state of the record as augmented by the forthcoming hearing, this Court will finally.determine, pursuant to Rule 59(e), (1) whether any of the above referenced rulings of the court in Memorandum and Order II would require modification; and, (2) whether or not the evidence and the .Court’s authority under Rule 59(e) would warrant affirming its conclusion in Memorandum and Order II, page 90, that “specific loss of pay could not be proved by black pipefitter, and he or she could not be awarded specific compensatory damages.”

Thus, under the February 2, 1993, memorandum order, the Court made certain findings. First, Judge Thomas questioned whether the present state of the record would warrant the application of International Bhd. of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977), which would shift the burden to Defendant Local Union 120. 6 Second, Judge Thomas ruled that his earlier finding of a pattern and practice of discrimination was subject to review “depending on the state of the record as augmented by the forthcoming hearing.”

The parties tried this case to the Court for eight trial days between April 29, 1998, to May 7, 1998. After reviewing the evidence and observing the demeanor of the witnesses, the Court finds that Plaintiff-Interveners fail to show entitlement to further relief.

II. Standard required to show “pattern and practice” of discrimination.

Plaintiff EEOC and Plaintiff-Inter-veners suggest this Court is bound by the law-of-the-ease to find that Local Union 120 employed a pattern and practice of discrimination. Determination of whether Local Union 120 engaged in a pattern and practice of discrimination is important because such a finding shifts the burden of the production of evidence.

In “pattern or practice cases,” the plaintiff must first establish “that racial discrimination was the company’s standard operating procedure.” Teamsters, 431 U.S. at 336, 97 S.Ct. 1843; Wooldridge v. Marlene Indus. Corp., 875 F.2d 540, 545 (6th Cir.1989). If a party shows a pattern and practice of discrimination, “the burden of proof shifts to the defendant to show that in individual cases the hiring decision was not due to the discriminatory policy but to some other, legitimate consideration.” Wooldridge, 875 F.2d at 545 (citing Teamsters, 431 U.S. at 359, 97 S.Ct. 1843).

In Teamsters, the Supreme Court described the showing required to make out a pattern and practice of discrimination:

[B]ecause it alleged a systemwide pattern or practice of resistance to the full enjoyment of Title VII rights, the Government ultimately had to prove more than the mere occurrence of isolated or “accidental” or sporadic discriminatory acts. It had to establish by a preponderance of the evidence that racial discrimination was the company’s standard operating procedure — the regular rather than the unusual practice.

Teamsters, 431 U.S. at 336, 97 S.Ct. 1843 (emphasis added). 7

In Stotts v. Memphis Fire Dep’t, 858 F.2d 289 (6th Cir.1988), the Sixth Circuit described the showing required to make out a pattern and practice of discrimination. The Court noted “it is the plaintiffs burden to prove that an employer intentionally ‘treats some people less favorably than others because of their race, color, religion, sex, or national origin.’” Id. at 294 (citing Teamsters, 431 U.S. at 335 n. 15, 97 S.Ct. 1843). In Stotts, the Court noted that:

“Proof of illicit motive is essential, but, especially in cases alleging class-wide discrimination, illicit motive may be inferred from a sufficient showing of disparity between members of the plaintiff class and comparable qualified members of the majority group.” Id. at 295 (quoting Segar v. Smith, 738 F.2d at 1265-66 (citation omitted)). The plaintiffs case must, however, eliminate “the most common nondiscriminatory reasons for the plaintiffs rejection.” [Texas Dept. of Community Affairs v.] Burdine, 450 U.S. [248] at 253—54[ 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981)].

H* ‡ 'l*

The plaintiff who alleges a discriminatory pattern or practice, be it through disciplinary measures or hiring procedures, must ultimately prove “more than a mere occurrence of isolated-or ‘accidental’ or sporadic discriminatory acts. [He must] establish by a preponderance of the evidence that racial discrimination was the Company’s standard operating procedure — the regular rather than the unusual practice.” Teamsters, 431 U.S. at 336[97 S.Ct. 1843]. As Title VII’s legislative history points out: “[A] pattern or practice would be present only where the denial of rights consists of something more than an isolated, sporadic incident, but is repeated, routine, or of a generalized nature.... The point is that single, insignificant, isolated acts of discrimination by a single business would not justify a finding of a pattern or practice .... ” (citations omitted).

******

We find the district court’s finding of a pattern and practice to be clearly erroneous because it is based upon comparisons of a series of factually unique and qualitatively dissimilar incidents and a corresponding series of appropriately individualized disciplinary actions which have been imposed in complete consistency with racially neutral and nonarbitrary guidelines. In sum, plaintiff Jones has failed to eliminate the most common nondiseriminatory reasons for the differences in the fight-related discipline handed down by the Department: the nature of the offenses and the nature of the punishments imposed, particularly in relation to the employees’ personnel records. Accordingly, we conclude that Jones’ evidence of defendant’s fight-related discipline has failed to establish a prima facie case of racial discrimination.

Id. at 295-98 (footnotes omitted).

Judge Thomas’s suggestion of a pattern and practice of discrimination has significant evidentiary consequences. Plaintiff-Inter-veners suggest that this Court be bound by the earlier finding. The Court now examines whether the “law of the ease” requires this finding.

III. Discussion of the “law of the case.”

The “law-of-the-case” doctrine ordinarily directs a court to adhere to prior findings of fact and of law. It is a “rule of judicial comity” designed to promote efficiency,-not one of jurisdictional limitation. See Moses v. Bus. Card Express, Inc., 929 F.2d 1131, 1137 (6th Cir.), cert. denied, 502 U.S. 821, 112 S.Ct. 81, 116 L.Ed.2d 54 (1991). That is, the doctrine does not limit a court’s power to revisit a previously settled issue of law or of fact, it simply “directs a court’s discretion.” Id. (quoting Arizona v. California, 460 U.S. 605, 613, 103 S.Ct. 1382, 75 L.Ed.2d 318 (1983)).

The Sixth Circuit has declined to impose a rigid set of conditions under which a transferee court may review a prior ruling of a coordinate court. See United States v. Todd, 920 F.2d 399, 403 (6th Cir.1990). Instead, “[i]t is within the sole discretion of a court to determine if a prior ruling should be reconsidered.” Id. In the Sixth Circuit, a district court’s reconsideration of a previously decided issue “will be deemed erroneous only if it is shown that the transferee court abused its discretion.” Id. District courts are generally afforded even broader discretion to revisit prior evidentiary issues because “this type of decision turns upon the evidence as developed during the course of a' trial.” Id. Thus in Todd, the Sixth Circuit upheld a district court’s introduction of evidence that established the defendant’s involvement in money laundering when that evidence had been excluded from a first trial. See id. at 404.

Under the law-of-the-case doctrine “it is not improper for a court to depart from a prior holding if convinced that it is clearly erroneous and would work a manifest injustice.” Arizona, 460 U.S. at 618 n. 8, 103 S.Ct. 1382. A finding of fact is clearly erroneous when, although there may be some evidence to support the finding, “the reviewing court is left with the definite and firm conviction that a mistake has been committed.” Kline v. Tennessee Valley Autk, 128 F.3d 337, 341 (6th Cir.1997) (quoting Anderson v. City of Bessemer City, 470 U.S. 564, 573-75, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985)). The “clearly erroneous” standard does not allow an appellate court to modify a'trial court’s findings simply because, had, it been sitting as the trier of fact, it would have found differently. See Anderson, 470 U.S. at 573-75, 105 S.Ct. 1504. However, a reviewing coordinate court has the power to review prior findings under any circumstances. See Todd, 920 F.2d at 403.

This Court finds scant reason to defer to the earlier finding of a pattern and practice of discrimination. First, the September 1, 1992, order was not a final judgment. Plaintiff-Interveners themselves, admitted as much when they moved for amendment of the earlier order, even if on other grounds. Second, in his February 2, 1993, Order, Judge Thomas explicitly stated that his earlier finding would be subject to review on the evidence presented at the stage II hearing. Third, the evidence presented at the stage II hearing shows that adherence to this Court’s earlier order would be clearly erroneous and would work an injustice.

The law-of-the-case doctrine, therefore, permits this court to revisit and modify Judge Thomas’s finding of a pattern of discrimination.

IV. Discussion of statistical evidence

Plaintiff and Plaintiff-Interveners attempt to prove that Local Union 120 discriminated against black members by introducing statistical evidence prepared by Dr. Rebecca Klemm. The Court finds such statistical evidence legally insufficient because it fails to account for recall and special skills, two nondiscriminatory explanations for the alleged disparities. Plaintiffs evidence is also unreliable because it does not accurately recreate workers’ registration dates.

Even if the evidence’s methodology were reliable, it would not show that Local Union 120 discriminated against blacks, because black members routinely received referrals at a rate higher than their percentage in the workforce. Finally, Plaintiffs evidence uses a faulty methodology for determining the hours allegedly lost due to discrimination. For these reasons, the Court concludes that Plaintiffs statistical evidence is insufficient to establish a pattern and practice of discrimination against black union members.

In Stage I of these proceedings, Judge Thomas found that the evidence produced at that time established a pattern or practice of disparate treatment in the operation of Local 120’s referral system. 8 The Court based this conclusion primarily on “three separate findings of significant statistical differences in the total hours of white and black pipe fitters.” Id. The Court made this finding upon statistical recreation of the sign up sheets. 9

Plaintiff EEOC and Plaintiff-Interveners say that Defendant Local Union 120 discriminated against black local union members in making job referrals. Plaintiff-Interveners say this occurred from at least 1978 to 1992.

To determine what, if any, damages resulted from operation of the referral system, the Court ordered that a representative class be chosen. The Court assembled this class randomly, choosing the third person on the class list and every fifth name after that. The resulting sample had twenty-four black class members. After these twenty-four class members were chosen, the Court examined periods of unemployment suffered by these class members from 1978 to 1992.

Local Union 120 did not keep copies of Registration for Employment forms until 1988. 10 Because Local Union 120 had not kept these records, it is difficult to accurately determine whether Local Union 120 referred a white pipe fitter before an earlier registered black class member.

To meet their burden of showing discrimination, Plaintiff EEOC and Plaintiff-Inter-veners relied upon a computer program formula to attempt to recreate registration dates for out-of-work pipe fitters. Using health and welfare payment reports from employers, the program attempted to create “windows” when class members may have been out of work and may have been passed over for referral by a white pipe fitter with a later out of work registration.

Statistical evidence may be used to prove the existence of racial discrimination. See Teamsters, 431 U.S. at 339, 97 S.Ct. 1843. 11 To be legally sufficient, however, the statistical evidence must establish a significant racial disparity and must also eliminate the most common nondiseriminato-ry explanations for this disparity. See Barnes v. GenCorp, Inc., 896 F.2d 1457,1466 (6th Cir.), cert. denied, 498 U.S. 878, 111 S.Ct. 211, 112 L.Ed.2d 171 (1990). Thus, when special qualifications are required to fill particular jobs, “comparisons to the general population (rather than to the smaller group of individuals who possess the necéssary qualifications) may have little probative value.” Ste. Marie v. Eastern R.R. Ass’n, 650 F.2d 395, 400 (2nd Cir.1981) (citation omitted).

The Plaintiff and Plaintiff-Interveners rely upon Dr. Klemm’s computer analysis. But as discussed, that program fails to account for several important non-race related variables. Klemm’s analysis does not account for recalls. Under the collective bargaining agreements, employers have the right to call back any former employee within 100 days (or two calendar months during May 1, 1979 through April 30, 1982) after the employee ceases working for that contractor. Employers use recalls to fill many of their openings. Twenty-two percent of white pipe fitters take positions as the result of recall instead of referral. 12 Because of this contract rule, no employees receive referrals when an employer elects to recall a former employee.

Klemm’s program fails to account for this and creates resulting error.

Klemm’s program also fails to take account of special skills. The evidence credibly shows that many referrals request pipe fitters with special skills or certification (e.g., welders, stainless steel welders, HVAC servicemen, sprinkler qualified pipe fitters). Class members may not have been eligible to 'receive referrals given to the potential white “line cutters” because they were not qualified for the job in question. 13

The Plaintiff and Plaintiff-Interveners’ evidence also fails to account for referrals of apprentices. Local Union 120 does not refer apprentices to positions. The new referral system, effective March 30, 1972, states: “[a]ll Apprentice referrals shall be made by the Apprenticeship Committee-It is rec-

ognized that these apprentices need varied on-the-job training as well as classroom instruction to acquire this expertise, and should be referred separately from journeymen.” By failing to recognize that apprentice referrals are not made by Local Union 120, the Plaintiff and Plaintiff-Interveners create error in their calculations.

Plaintiffs evidence is legally inadequate because it does not eliminate race-neutral explanations for the alleged disparities. But more important, Plaintiff and Plaintiff-Inter-veners fail to show that the computer formula they use to create registration dates is otherwise methodologically accurate. Plaintiffs suggest out of work registration when an individual has worked more than the re registration threshold 14 and has changed employers or a gap of one or more months of employment occurs. This method does not appear to accurately position Local Union 120 members on out of work lists.

A comparison of out of work positions generated by Plaintiffs and Plaintiff-Interveners with actual out of work registration during periods where such actual registrations are available, show great error in the program. A valid program must accurately approximate the rank order of registration for a job referral.

Defendant Local Union 120 shows evidence of major errors in assigning rank ordering. For example, Plaintiffs compare white pipe fitter Cunningham with black pipe fitter Sanchez. Using Plaintiffs’ formula, Sanchez was ranked 1575 places ahead of Cunningham on the out-of-work registration list. However, the actual registration lists exist for periods after 1988. Those actual lists show that Cunningham registered before Sanchez. The Sanchez example shows Plaintiffs’ program does not accurately assign out of work registration dates.

To like effect is window 21, involving class member Coleman. 15 Because actual registration records are available, we know that class member Coleman actually received his referral before white pipe fitter Carter. Yet Dr. Klemm’s program wrongly shows that Coleman received referral after Carter and wrongly says Coleman lost close to $40,-000.00 from Carter’s referral.

Dr. Gerald Barrett gave credible testimony showing significant flaws in the program Plaintiff and Plaintiff-Interveners relied upon. With Defendant’s Exhibit LL, Barrett used Dr. Klemm’s formula but did not suppress windows that showed the impossible widows with opening dates the same or later than closing dates. 16 This obvious- error appeared in 26% of black pipe fitter windows and in 28% of white pipe fitter windows. Plaintiff and Plaintiff-Interveners gave no evidence explaining or rebutting this evidence showing serious error in the program used to approximate out-of-work registration dates.

Defendant’s Exhibit DD further illustrates that Plaintiffs’ program unreliably determines registration dates. In the exhibit, Defendant Local Union 120 used Plaintiffs computer program to decide if it accurately assigned white pipe fitters registration dates. Examining windows 17-22, Defendant Union shows that Plaintiffs computer formula reveals that many white pipe fitters cut line on other white pipe fitters. For example, Plaintiffs’ computer formula for assigning registration dates says - that twenty-nine white pipe fitters cut in line ahead of white pipe fitters L. Gaj during a February 1, 1989, to March 1,1989, window. 17

This evidence leads to either of two conclusions. More likely, it shows that Plaintiffs computer’ creation of registration dates is badly flawed. Less likely, it shows that Local Union 120 made out-of-order referrals that damaged individual white pipe fitters as much as black pipe fitters. If so, Plaintiffs fail to show that Local Union 120 discriminated based upon race.

As described, Plaintiff and Plaintiff-Inter-veners’ computer program does not reliably assign out of work registration dates. However, even if, it did, it does not show a pattern and practice of discrimination.

Several examples highlight the .failure to show a pattern and practice of discrimination by Defendant Local Union 120, even using the unreliable. computer formula to assign workers to slots on out-of-work listings. For example, from January 1979, to December 1992, the Defendant Union referred black pipe fitters for open positions at a rate higher than their percentage of the workforce in 68% (114/168) of the months. During the same time, the Defendant Union referred white pipe fitters at a rate higher than their percentage of the workforce in only 32% (54/168) of the months.

A review of individual windows shows the unreliability of the Plaintiff-Interveners’ computer formula and fails to show discrimination. Plaintiff-Interveners claim that Willie Coleman lost work from June 1, 1988, to July 1989, as the result of racial discrimination. Yet during, those thirteen months, the percentage of black referrals exceeded the percent of black pipe fitters in the union in eight of the thirteen months. Over the whole period, black pipe fitters received 8.77% of work referrals at a time when they made up approximately 7.5% of the workforce.

Review of other windows shows like results. In window 2, Plaintiff-Interveners claim that Local Union 120 discriminated against Ivory Baker in referrals during the period November 1, 1981, to February 10, 1982. Plaintiff-Interveners claim twenty-four white pipe fitters wrongly received re-ferráis ahead of Baker during this period. Yet using Plaintiffs formula, Local Union 120 shows that four black pipe fitters also received referral ahead of Baker. Thus, during this window, black pipe fitters received 14% of the referrals at a time during which black pipe fitters only made up 6.5% of the Local Union 120 workforce. Even if Plaintiffs’ program was accurate, it can make no showing that race motivated any lack of referrals during window 2 when Local Union 120 was referring black pipe fitters at a high rate. Other windows show similar results. 18

Apart from the evidentiary problems described above, Plaintiffs and Interveners use a questionable method of determining lost hours for claimed discrimination. Dr. Klemm multiplied a constant for each window day. The constant purports to reflect the difference of white-black mean work hours per day.

However, factors beyond Local Union 120’s control affect this. White pipe fitters have a higher recall percentage, a factor controlled by employers but not Local Union 120. 19 Because of this higher recall percentage, white pipe fitters have a higher mean of hours worked per day. Plaintiff and Inter-veners use this mean calculation to set lost hours though Defendant Local Union 120 has no control over recalls.

White pipe fitters also have a higher percentage of specialized workers. 20 Again, em ployers and specialization committees control special certification but not Local Union 120. Black pipe fitters with specialization have more hours than similarly specialized white pipe fitters. Because of this higher specialization percentage, white pipe fitters have a higher mean of hours worked per day. Plaintiff and Interveners use this mean calculation to set lost hours though Defendant Local Union 120 has no control over specialization.

The Court finds a more accurate lost hour determination would result from comparison with the actual hours obtained by a similarly situated white pipe fitter whom Local Union 120 referred during the class member’s window. Defendant Local Union 120 made such a comparison of individual hours for all twenty-four sample plaintiffs. This comparison did not show loss of work hours. To the contrary, this comparison showed that black pipe fitters had more hours per referrals than the nearest similarly placed white pipe fitter.

To summarize, Judge Thomas stated that his finding of a pattern and practice of discrimination by Local Union 120 might require modification depending upon the state of the record after the stage II hearings. Even if Judge Thomas had not conditioned his finding of a pattern and practice of discrimination, there is good reason in this record to reexamine that finding.

First, the Court does not find Dr. Klemm’s method of creating registration dates to be credible. Her program results in great error and is not reliable.

Second, even if Dr. Klemm’s unreliable methodology is used, it shows that Local Union 120 did not have a pattern and practice of referring white pipe fitters ahead of black pipe fitters who had earlier registered for referral. Black pipe fitters received referrals at rates higher than their numbers in the Local Union 120 membership in 68% of the months from 1979 to December 1993, more than twice as frequently as white pipe fitters. Such hardly reflects a pattern and practice of discrimination.

Third, Dr. Klemm found black pipe fitters had fewer hours of work resulting from referrals. The Klemm program suggesting this is not reliable. Moreover, the Court credits Dr. Barrett’s testimony that recalls and special skills, factors beyond Local Union 120’s control, account for hours disparity.

For these reasons, the Court finds no pattern and practice of discrimination in Local Union 120’s job referral system during the 1978 to 1992 period.

V. Review of individual claims

As related, the Court therefore finds that Plaintiffs and Plaintiff-Interveners’ evidence is insufficient to establish a pattern or practice of discrimination. Even if the Court did not revisit this prior determination, and shifted the burden of production to Defendant according to Teamsters, 431 U.S. at 359, 97 S.Ct. 1843, Plaintiffs would still not be entitled to relief. Defendant Local 120 can rebut the presumed existence of discrimination with respect to each individual window.

The Court reviews evidence offered as to individual windows. Plaintiff-Interveners presented no direct evidence of discrimination. In reviewing these individual windows, the Court examines whether' circumstantial evidence suggests that Local Union 120 discriminated in job referrals.

In reviewing circumstantial evidence of actions taken up to twenty years ago, the Court first looks to Local Union 120’s referrals. At the time of the window, did Local Union 120 make referrals to white pipe fitters in much higher percentages than to black pipe fitters? Of somewhat less import, the Court examines whether the percentage of black pipe fitters working was greater or lesser than the percentage of white pipe fitters who were then working.

Because actual registration evidence is not available, the Court needs to decide if the subject class member was available for work. Because Local Union 120 use travelers only when local union members were not available for referrals, the Court examines the use of travelers during the subject window as evidence of whether the class member was available for referral. The Court also examines testimony and other evidence to decide whether class members were available for referral.

To begin the review of individual windows, Plaintiff-Interveners have withdrawn all claims for 39 of the original 118 windows. Under these withdrawals, no longer before the Court are windows 7, 13, 14, 15, 16, 17, 21, 25, 27, 29, 30, 31, 33, 34, 35, 39, 40, 41, 43, 46, 47, 50, 59, 60, 64, 68, 79, 81, 84, 85, 89, 94, 98, 108, 111, 112, 113, 114, and 116. 21 Because Plaintiff-Interveners have withdrawn their claims as to these windows, Defendant Local Union 120 should receive judgment as to them.

As related, Plaintiff-Interveners argue that Local Union 120 discriminated against black pipe fitters in job referrals. In support of their claims, Interveners give no direct evidence showing discrimination in job referrals. Without direct evidence, the Court finds most probative the comparison of white pipe fitter referrals to black pipe fitter referrals. Making this comparison, the Court finds that Local Union 120 referred black pipe fitters for employment at a higher rate than white pipe fitters in many months. 22

By showing that black pipe fitters were referred at numbers higher relative to their numbers than white pipe fitters, Local Union 120 rebuts any inference drawn from a pattern and practice of discrimination finding. Obviously, the same showing stops Plaintiff-Intervenors’ claim for damages absent the pattern and practice of discrimination finding.

The Court finds that Local Union 120 can rebut a pattern and practice of discrimination during those periods when the percentage of total referrals received by blacks was equal to or greater than their percentage within the workforce. In other words, the evidence shows there is no discrimination in referrals during those periods where black pipe fitters received higher percentage of referrals when compared with their respective numbers in the workforce.

The Court gives examples that show this. With window 1, Plaintiff-Interveners seek damages for black pipe fitter Ivory Baker for the period August 1, 1981, to September 28, 1981. In support of this request for damages, Interveners give no direct evidence of discrimination. Finding no direct evidence, the Court reviews circumstantial evidence. During window 1, circumstantial evidence shows that black pipe fitters made up 6.35% of the Local Union 120 workforce. During window 1, black pipe fitters received 5 of 45 or 11% of the referrals. Having no other evidence to support an inference of discrimination, Plaintiff-Interveners cannot plausibly make claim for damages for window 1.

A random survey of other windows shows the same result. For example, in window 10, black pipe fitters make up 7.39% of the workforce but received 12.05% of the job referrals. In window 20, black pipe fitters make up 7.39% of the workforce but received 15.09% of the job referrals. In window 30, black pipe fitters make up 7.69% of the workforce but received 9.43% of the job referrals. In window 40, black pipe fitters make up approximately 7.4% of the workforce but received 10.8% of the job referrals. 23 In window 50, black pipe fitters make up 8% of the workforce but received 11.22% of the job referrals.

As described above, the Court has found that Local Union 120 employed no pattern or practice of discrimination in its referral system during the period here involved. However, even if Local Union 120 had been found to have engaged in such a pattern of discrimination, the Court finds that the evidence of high black pipe fitters referrals rebuts the suggestion of discrimination in those windows in which black pipe fitters referrals exceeds their respective percentage of the referral population.

Having no direct evidence of discrimination and having no other credible circumstantial evidence suggesting discrimination, the Court finds that Plaintiff-Interveners should not recover during those windows in which black pipe fitters’ referrals exceed their representation in the workforce.

The Court therefore excludes those windows where black pipe fitters received a greater proportion of referrals than their percentage in the workforce. 24 Under this criterion, the Court finds Plaintiff-Interven-ers can recover no damages for ninety-five windows. The Plaintiff-Interveners have withdrawn claim for some of these windows. 25 Besides the windows withdrawn, this criterion requires exclusion of claims for windows 1, 2, 3, 4, 8, 9, 10, 11, 12, 18, 19, 20, 24, 26, 28, 36, 37, 38, 42, 45, 48, 51, 52, 53, 54, 55, 56, 57, 60, 61, 66, 67, 68, 69, 70, 72, 73, 74, 75, 76, 77, 78, 80, 82, 83, 86, 87, 88, 90, 91, 92, 93, 95, 96, 99, 101, 102, 103, 104, 105, 106, 107, 109, and 115. 26

As described, the Court finds the rate of referral in relation to percentage within the Local Union 120 population most relevant to deciding whether Local Union 120 discriminated in referrals. Although somewhat less probative of the Union’s referral practices, comparison of the percentage of blacks who worked in a period with the percentage of white pipe fitters suggests whether Local Union 120 discriminated in job referrals. Using these criteria, black pipe fitters had a higher percentage working during at least one month during at least 36 windows. 27 Among these 36 windows, most are either withdrawn or no damages can be given because black pipe fitters received high rates of referrals during the window. 28

The Court therefore finds no right to recover for windows in which the percentage of blacks working was equal to or greater than the percentage of whites who worked. That a higher percentage of blacks than whites worked in these windows rebuts the presumption that the Union systematically discriminated against black pipe fitters. Indeed, it suggests that Local 120’s referral policy was more beneficial to blacks than to whites.

Besides the above considerations, the Court finds that it should eliminate windows with a net increase in the number of travelers. If the Union is adding travelers in a given month, the additions suggest not enough local workers had registered as willing to work. As a result, the Union had to go elsewhere to find workers. 29 This in turn shows that Local Union 120 “passed over” blacks because they I had failed to register for work, not because of Union discrimination. In 85 windows travelers were added. Among these windows, a number had also been withdrawn by Plaintiff-Interveners. 30 Also, the Court has found that black referrals exceeded their numbers in the workforce in additional windows. 31 Besides the window excluded for the above reasons, this criterion requires that the additional windows No. 5, 6, 22, 23, 44, 49, 58, 62, 65, 71, 97, 100, and 110 should be excluded. 32

Having found that Plaintiff-Interveners have withdrawn certain windows, having found that no inference of discrimination can be found during months in which black pipe fitters referrals exceed their percentage of the workforce, having found that black pipe fitter class members were not available for work during months during which travelers were added, and having found that any difference in hours worked reflects recalls and different specializations, the Court finds that Plaintiff-Interveners show no right to damages. 33

VI. Conclusion

The Court denies Plaintiff-Interveners’ request for additional relief. This Court’s earlier finding that Local Union 120 practiced a pattern and practice of discrimination in job referrals is subject to review. In his earlier order, Judge Thomas explicitly stated that this finding would be subject to modification based upon evidence produced at this Stage II hearing. Additionally, the evidence produced to this Court causes the Court to find that the earlier finding of a pattern and practice of discrimination is clearly erroneous and would work a manifest injustice.

This Court’s earlier finding of a pattern and practice of discrimination is clearly erroneous. First, the methodology used to approximate out-of-work periods is extremely inaccurate. Because this inexactness has major effect upon placement on out-of-work registration lists, it discredits Plaintiff-Inter-veners’ claims. Second, the methodology completely fails to account for recalls and request for specialized skills. Because recalls and requests for specialized skills make up a large portion of job placements and because Local Union 120 has no control over them, this failure to account for recalls and special skills invalidates Plaintiff-Interven-ers’ claim for damages. Third, even if Plaintiff-Interveners’ methodology were sufficiently accurate, which it is not, Local Union 120 has rebutted Plaintiff-Interveners’ claims for damages.

For these reasons, the Court denies Plaintiff-Interveners’ request for additional relief.

IT IS SO ORDERED.

1 . At the time these orders were issued, then Senior Judge Thomas presided over the case.

2 . Paragraph 4(c) of the consent decree required Defendant Local Union 120 to make note on each employee’s master registration card. The union was required to identify the date of referral, the name of the employer, and the classification of referral.

3 . Judge Thomas however found that black pipe fitters, sometimes received referrals ahead of whites because contractors, operating under government contracts, requested black pipe fitters.

4 . Judge Thomas found that from 1972 to 1978, black union members received substantially equivalent work hours as whites. He found that from 1979 to 1984, 1985 to 1987, and 1988 to 1990, black union members received substantially less work hours than whites.

5 . Judge Thomas ordered revised computerized record keeping. He ordered that written documentation follow employer requests for specific employees. He approved a hiring hall rule saying that employees needed 320 hours, not 80 hours of employment, before they were required to go to the end of the referral list.

6 . In Teamsters, the Supreme Court found that "[t]he proof of the pattern or practice supports an inference that any particular employment de-cisión, during the period in which the discriminatory policy was in force, was made in pursuit of that policy.” 431 U.S. at 362, 97 S.Ct. 1843.

7 . The Teamsters Court, quoting legislative leaders, defined a pattern and practice of discrimination as follows:

Senator Humphrey explained: "[A] pattern or practice would be present only where the denial of rights consists of something more than an isolated, sporadic incident, but is repeated, routine, or of a generalized nature. There would be a pattern or practice if, for example, a number of companies or persons in the same industry or line of business discriminated, if a chain of motels or restaurants practiced racial discrimination throughout all or a significant part of its system, or if a company repeatedly and regularly engaged in acts prohibited by the statute."

"The point is that single, insignificant, isolated acts of discrimination by a single business would not justify a finding of a pattern or practice...110 Cong.Rec. 14270 (1964).

This interpretation of “pattern or practice” appears throughout the legislative history of § 707(a), and is consistent with the understanding of the identical words as used in similar federal legislation. See 110 Cong.Rec. 12946 (1964) (remarks of Sen. Magnuson) (referring to § 206(a) of the Civil Rights Act of 1964, 42 U.S.C. § 2000a-5); 110 Cong.Rec. 13081 (1964) (remarks of Sen. Case); id., at 14239 (remarks of Sen. Humphrey); id., at 15895 (remarks of Rep. Celler).

Teamsters, 431 U.S. at 336 n. 16, 97 S.Ct. 1843.

8 . See September 1, 1992, Memorandum and Order II at 89.

9 . This analysis was flawed, because Dr. Rebecca Klemm’s analysis does not credibly recreate the sign up sheets. See discussion infra pp. 844-845.

10 .As related, Judge Thomas found that no adverse inference may be drawn from Local 120's failure to retain these records.

11 .In Teamsters, the Court instructed:

In any event, our cases make it unmistakably clear that "Statistical analyses have served and will continue to serve an important role" in cases in which the existence of discrimination is a disputed issue. We have repeatedly approved the use of statistical proof, where it reached proportions comparable to those in this case, to establish a prima facie case of racial discrimination in jury selection cases. Statistics are equally competent in proving employment discrimination. We caution only that statistics are not irrefutable; they come in infinite variety and, like any other kind of evidence, they may be rebutted. In short, their usefulness depends on all of the surrounding facts and circumstances.

Teamsters, 431 U.S. at 339, 97 S.Ct. 1843 (citations omitted).

12 . Black class members have a lower recall rate of 13%.

13 . As only one example, Klemm’s program says that on December 2, 1991, white pipefitter Carter cut in front of nine class members with earlier registration under Klemm’s program. But actual records show this referral was for a service man position. Carter was certified for this position. None of the nine class members identified by Klemm has having suffered discrimination were certified as a serviceman. See Defendant's Exhibit NN.

14 . Local Union 120 members fell to the-bottom of the registration list after they had worked 80 or 320 hours, depending upon the year.

15 . Plaintiff-Interveners withdrew claims as to window 21. The Court reviews window 21 because it reflects errors in the program used for all windows.

16 . Dr. Klemm had inserted a command in her software program that suppressed windows with opening dates that were the same or later than the closing date of the window.

17 .Other examples show the same. Plaintiffs’ computer formula for assigning registration dates says that two hundred-ten white pipe fitters cut in line ahead of white pipe fitters L. Clark during a January 1, 1989, to May 1, 1989, window. Plaintiffs’ computer formula for assigning registration dates says that seventy-five white pipe fitters cut in line ahead of white pipe fitters J. Boyle during a May 4, 1992, to August 1, 1992, window.

18 .A random sample of the every twentieth window shows this. Window 20 involves class member Coleman during the period April 17, 1990, to October 2, 1990. Using Plaintiffs data, black pipe fitters received 8% of the referrals at a time during which black pipe fitters only made up 6.5% of the Local Union 120 workforce.

Window 40 involves class member Huff during the period May 1, 1984', to August 1, 1984. Using Plaintiff’s data, black pipe fitters received 8% of the referrals at a time during which black pipe fitters only made up 6.5% of the Local Union 120 workforce.

Window 60 involves class member Kee-Bey during the period August 22, 1989, to September 1, 1989. Using Plaintiff’s data, black pipe fitters receive d 24.49% of the referrals at a time during which black pipe fitters made up only 7.59% of the Local Union 120 workforce.

(Windows 40 and 60 were both withdrawn. Nevertheless, they are still useful to show methodological problems applicable to the remaining windows.)

Window 80 involves class member McCall during the period February 1, 1990, to April 10, 1990. Using Plaintiffs data, black pipe fitters receive d 9% of the referrals at a time during which black pipe fitters only made up 6.5% of the Local Union 120 workforce.

Window 100 involves class member Stewart during the period February 1, 1986, to December 2, 1986. Using Plaintiff’s data, black pipe fitters received 6% of the referrals at a time during which black pipe fitters only made up 6.5% of the Local Union 120 workforce.

19 . Recall is most frequently associated with long work experience. White pipe fitters tend to have greater numbers with long work experience. Thus, only four black pipe fitters had more than thirty years experience while about four hundred white pipe fitters had thirty years experience.

20 . Defendant Local Union 120 gives evidence that only three black pipe fitters (3% of Black pipe fitters) were HVAC servicemen while 129 whites (5.6% of white pipe fitters) were HVAC servicemen. Only 1.4% of three black pipe fitters claimed specialization in air conditioning service work while 8.4% of white pipe fitters claimed such specialization. Only 2.3% of three black pipe fitters claimed specialization in refrigeration work while 11.7% of white pipe fitters claimed such specialization.

21 . Thus Plaintiff-Interveners withdraw the following windows:

[[Image here]]

22 . The Court compares the percent of referrals going to white pipe fitters and black pipe fitters in relation to their respective percentage of the workforce.

23 . Plaintiff-Intervenors withdrew window 40 yet it remains illustrative of whether discrimination occurred as a regular course.

24 . The actual percentage of blacks in the data file on a window-to-window basis is not before the Court. The Court therefore computed for the entire period the average monthly percentage of blacks in the data file. Each month’s percentage of blacks in the data file was added (total 1179.72) and divided by the total number of months (168) to get an average monthly percentage of 7.02%. This percentage was then compared with the percentage of total referrals that went to blacks in each individual window.

25 .As described, Plaintiff-Interveners have withdrawn claim for certain windows. Among the windows withdrawn, the Court would find that Plaintiff-Interveners can not recover damages for windows in which black pipe fitters received higher rates of referrals than their percentage of the Local Union 120 population. Using this criteria, Plaintiff-Interveners cannot recover dam ages for windows 7, 13, 14, 16, 17, 21, 25, 27, 30, 33, 34, 35, 39, 40, 41, 43, 47, 50, 60, 68, 79, 81, 84, 85, 94, 98, 108, 112, 113, 114, 116, 117, and 118

26 . Thus, the Court finds that Plaintiff-Interveners cannot recover for the following windows:

[[Image here]]

27 . Windows Nos. 1, 2, 3, 6, 7, 11, 20, 27, 35, 36, 39, 40, 41, 42, 45, 46, 47, 55, 68, 75, 81, 83, 85, 86, 87, 88, 89, 94, 95, 98, 103, 111, 112, 114, 115, and 116.

28 . Among the 36 windows in which black pipe fitters worked at a higher percentage than white pipe fitters, all were withdrawn or excluded for the above reason except window No. 6.

29 . The Defendant proposes that any window where the number of travelers was above zero can be excluded. However, the pertinent criterion is the months where the number of travelers was increasing. The latter criterion isolates the addition of new travelers, while the Defendant's could reflect carryover from a prior month.

30 . Among the windows where travelers were added, Plaintiff-Interveners had withdrawn claim for windows 7, 15, 27, 29, 34, 35, 39, 40, 41, 43, 46, 68, 81, 84, 85, 94, 98, 111, 112, 114, 116, and 118.

31 . Among the windows where travelers were added, the Court has found Local Union 120 referred black pipe fitters at a higher rate than their numbers in the workforce, in windows 1, 2, 3, 4, 7, 9, 11, 12, 18, 19, 20, 24, 26, 27, 28, 34, 35, 36, 38, 39, 40, 41, 43, 45, 48, 51, 53, 55, 56, 57, 66, 67, 68, 69, 72, 73, 74, 75, 76, 77, 78, 80, 81, 83, 84, 85, 86, 87, 88, 90, 91, 93, 94, 95, 98, 101, 102, 103, 104, 107, 109, 111, 112, 114, 115, 116, and 118.

32 .The Court reached this conclusion by identifying the months with a net increase in travelers over the previous month. The months where the net number of travelers increased were: 3-79, 4-79, 5-79, 6-79, 7-79, 8-79, 10-79, 11-79, 12-79, 1-80, 2-80, 3-80, 6-80, 7-80, 8-80, 12-80, 1-81, 2-81, 3-81, 5-81, 6-81, 7-81, 9-81, 1-82, 6-82, 8-82, 9-82, 10-82, 11-82, 12-82, 2-83, 4-83, 6-83, 7-83, 8-83, 9-83, 10-83, 12-83, 1-84, 3-84, 4-84, 8-85, 9-85, 10-85, 5-86, 6-86, 9-86, 12-86, 3-87, 5-87, 6-87, 7-87, 11-87, 1-88, 4-88, 6-88, 7-88, 9-88, 10-88, 12-88, 1-89, 12-89, 2-90, 6-90, 8-90, 9-90, 10-90, 11-90, 8-91, 2-92, 3-92, 5-92, 9-92, and 12-92. The Court then compared these months with the opening and closing daies of the widows. If the window contained at least one month with a net increase in travelers, it was eliminated.

33 . As described, the Court excludes windows that have been withdrawn, windows where Black pipe fitters received referrals at rates higher than their percentage of the Union population, and windows where there was net increases in travelers. When excluded, only window 63, involving black pipe fitter Nathan Kee-Bey remains. Plaintiff-Interveners make claim on behalf of Kee-Bey for the window from July 6, 1992, to September 2, 1992. But during this time, travelers remained the same in August and increased slightly in September. The Court finds that window 63 should be excluded as well.