Stender v. Lucky Stores, Inc., 803 F. Supp. 259 (N.D. Cal. 1992)
August 18, 1992
Nancy J. STENDER, Diane Skillsky, Julie Valentine-Dunn, Reba Barber-Money, Irma Hernandez, Anita Martinez and Jon Gold on behalf of themselves and all others similarly situated, Plaintiffs,
LUCKY STORES, INC., Defendant.
United States District Court, N.D. California.
*260 *261 *262 *263 *264 Brad Seligman, Jocelyn D. Larkin, Donna M. Ryu, Saperstein, Seligman & Mayeda, Oakland, Cal., for plaintiffs.
Kirby Wilcox, Kathleen V. Fisher, James E. Boddy, Jr., Portia R. Moore, Morrison & Foerster, San Francisco, Cal., for defendant.
TABLE OF CONTENTS TO FINDINGS OF FACT AND CONCLUSIONS OF LAW FINDINGS OF FACT..............................266 I. LUCKY STORES INC. ............................................................. 266 Corporate Structure of Lucky .................................................. 266 Lucky's Northern California Division .......................................... 266 Districts in the Northern California Division ................................. 267 United Food and Commercial Workers Union Contracts ............................ 267 Retail Food Stores ............................................................ 268 Gemco Stores .................................................................. 269 II. LUCKY'S EMPLOYMENT PRACTICES .................................................. 269 Plaintiffs' Witnesses ......................................................... 269 Defendant's Witnesses ......................................................... 271 Initial Placement ............................................................. 271 A. Hiring .................................................................... 271 B. Placement ................................................................. 272 Promotion ..................................................................... 273 A. Lucky's Job Ladders ....................................................... 273 B. The UFCW Contract Provisions on Promotion ................................. 273 C. Promotion Practices ....................................................... 274 D. The Valley Posting Program ................................................ 280 E. Training .................................................................. 280 F. Movement between Departments .............................................. 282 Movement from Part-Time to Full-Time .......................................... 283 A. The UFCW Contract Provisions on Movement to Full-Time ..................... 283 B. Movement Practices ........................................................ 283 Allocation of Additional Hours/Step-Ups ....................................... 285 A. The UFCW Contract Provisions on Allocation of Hours ....................... 285 B. Allocation of Hours Practices ............................................. 285 C. Step-Up Practices ......................................................... 286 Wage Rates .................................................................... 286 Grievance Procedures .......................................................... 287 Comparison of Lucky's NCD and SCD ............................................. 288 Lucky's Affirmative Action Efforts ............................................ 288 A. Before 1986 ............................................................... 288 B. The 1986 NCD Affirmative Action Plan ...................................... 289
*265 C. Discontinuation of the 1986 Affirmative Action Plan ....................... 291 D. Lucky's 1988 Affirmative Action Efforts ................................... 291 E. Notes from the Specter Meetings ........................................... 292 F. Discontinuation of the 1988 Affirmative Action Efforts .................... 294 G. Attitudes of Senior Management ............................................ 294 III. EXPERT WITNESSES .............................................................. 294 Dr. Drogin .................................................................... 294 A. Initial Placement ......................................................... 296 B. Promotion ................................................................. 296 C. Movement from Part-Time to Full-Time ...................................... 297 D. Movement off the Bid Lists ................................................ 297 E. Allocation of Hours ....................................................... 297 F. Earnings .................................................................. 298 G. Dr. Drogin's Conclusions .................................................. 298 Dr. Pencavel .................................................................. 298 Dr. Bielby .................................................................... 301 Dr. Daum ...................................................................... 303 A. Dr. Daum's Work Interest Survey ........................................... 304 B. Dr. Daum's Work Hours Interest Survey ..................................... 309 Dr. Haworth ................................................................... 310 A. Dr. Haworth's Initial Placement Database .................................. 311 B. Dr. Haworth's Conclusions ................................................. 313 Dr. Hoffman ................................................................... 314 A. Critique of Dr. Daum's Job Interest Survey ................................ 314 B. Cluster Analysis .......................................................... 317 C. Critique of Dr. Daum's Work Hours Interest Survey ......................... 317 CONCLUSIONS OF LAW ................................ 318 I. JURISDICTION AND VENUE ........................................................ 318 II. LIABILITY PERIOD .............................................................. 318 III. THE PLAINTIFF CLASS ........................................................... 318 IV. LEGAL STANDARDS ............................................................... 318 Section 1981 .................................................................. 318 Title VII Generally ......................................................... 318 Title VII Disparate Treatment ............................................... 318 Title VII Disparate Impact .................................................. 320 The Use of Statistical Evidence in Title VII Cases ............................ 322 Punitive Damages under Title VII .............................................. 324 California Fair Employment and Housing Act .................................... 325 V. THE USE OF INTEREST SURVEYS ................................................... 325 VI. FINDINGS OF EXPERT TESTIMONY .................................................. 327 VII. DISPARATE TREATMENT ........................................................... 330 Evidence of Intent ............................................................ 330 A. Evidentiary Rulings ....................................................... 330 B. Subjective Decision Making ................................................ 331 C. Knowledge ................................................................. 331 D. Discriminatory Attitudes .................................................. 332 E. Findings .................................................................. 332 Initial Placement ............................................................. 332 Promotion ..................................................................... 333 Movement from Part-Time to Full-Time .......................................... 334 Allocation of Hours/Step-Ups .................................................. 334 VIII. DISPARATE IMPACT .............................................................. 335 Subjective Decision Making .................................................... 335 Failure to Follow Bid Procedures .............................................. 336 IX. PUNITIVE DAMAGES .............................................................. 336 X. FEHA CLAIM .................................................................... 336 CONCLUSIONS ........................ 336
*266 FINDINGS OF FACT AND CONCLUSIONS OF LAW
PATEL, District Judge.
Plaintiffs have brought this class action against Lucky Stores, Inc. on behalf of Black and female employees working in retail stores within Lucky's Northern California Food Division. Plaintiffs allege discrimination on the basis of race and sex in initial job placement, allocation of work hours, movement of part-time employees to full-time positions, and promotions. Claims are brought pursuant to Title VII of the 1964 Civil Rights Act, 42 U.S.C. § 2000e et seq.; 42 U.S.C. § 1981; and the California Fair Employment and Housing Act, Government Code §§ 12900-12996.
The court hereby enters its findings of fact and conclusions of law as to plaintiffs' claims. To the extent that any findings of fact are included under conclusions of law they shall be deemed findings of fact, and to the extent that any conclusions of law are included under findings of fact they shall be deemed conclusions of law. This order incorporates all of the previous orders of this court which are cited herein.
FINDINGS OF FACT
I. LUCKY STORES INC.
Corporate Structure of Lucky
1. Lucky Stores, Inc. ("Lucky") is a retail food chain headquartered in Dublin California. Joint Statement ¶ 1 (filed May 3, 1992). During the liability period, Lucky operated through a number of divisions. The number of these divisions ranged from a low of two to a high of nine. Joint Statement ¶ 2.
2. During the liability period, see infra ¶¶ 419-421, Lucky conducted its operations in California through two divisions, the Northern California Division ("NCD") and the Southern California Division ("SCD"). Lucky sold or liquidated its other divisions between 1984 and 1989. Joint Statement ¶ 3.
3. During the liability, period Lucky had, and continues to have, a General Management Committee comprised of senior management from Lucky Corporate and the divisions. This Committee dealt with a broad range of issues, including employment issues. Joint Statement ¶ 4.
Lucky's Northern California Division
4. Lucky's NCD is headquartered in San Leandro, California. Joint Statement ¶ 5.
5. Officers and managers within the San Leandro headquarters of the NCD include: a Training and Development Manager who reports to the Vice President, Human Resources; an EEO Manager who reports to the Vice President, Human Resources; a Vice President, Human Resources who reports to the Senior Vice President of Administration; a Manager of Labor Relations who reports to the Vice President, Industrial Relations Manager; a Vice President, Industrial Relations Manager who reports to the Senior Vice President of Administration; a Senior Vice President of Administration who reports to the Division President; and a Division President who reports to the President of Lucky Stores. Each of these positions has existed for different periods of time during the liability period and each position, when in existence, experienced one or more changes in its precise job title and reporting relationships. Joint Statement ¶ 5.
6. During the liability period, relevant NCD supervisory positions were held by:
a. President: Richard Goodspeed (1985 to present); Jack Hoover (until 1985).
b. Senior Vice President, Director of Operations: Louis Gloyne (1986-1989); James Scoggins (until 1985).
c. Senior Vice President of Administration: John Plotts (1989 to present); Tim Martin (until 1989).
d. Vice President, Operations: Robert Grant (1980 to present); Terry Privott (1989 to present).
e. Vice President, Industrial Relations: Walter Herkal (1979 to present).
*267 f. Vice President, Human Resources: Virginia Javier (1990 to present).
g. Training and Development Manager: Bruce Gentile (1988 to present); David Atwood (1985-1988); Jane't Noriega-Ailor (until 1985).
h. Manager of Labor Relations: Robert Gill (1986 to present).
i. Manager of Human Resources: Virginia Javier (1989-1990); Jane't Noriega-Ailor (1985-1988).
j. EEO Manager: Mark Foley (1989 to present) and Virginia Javier (1986-1989). Joint Statement ¶ 6.
Districts in the Northern California Division
7. During the liability period, Lucky assigned each of the retail food stores in the NCD to a district. The number of districts ranged from a low of seven to a high of nine. There are currently eight districts in the NCD. Joint Statement ¶ 7.
8. During the liability period the number of retail food stores in the NCD ranged from a low of approximately 150 to a current high of approximately 185. Joint Statement ¶ 7.
9. District level personnel currently include: a Grocery Merchandiser who reports to the District Manager; a Produce Manager who reports to the District Manager; a Deli/Bakery Trainer who reports to the District Manager; a General Merchandise Merchandiser who reports to the District Manager; and a District Manager who reports to the Vice President of Operations. Joint Statement ¶ 8.
United Food and Commercial Workers Union Contracts
10. With the exception of Store Managers and certain Assistant Store Managers, plaintiff employees are represented for purposes of collective bargaining under the terms of the National Labor Relations Act as part of a multi-employer bargaining unit. The bargaining representative for these employees is the United Food and Commercial Workers Union ("UFCW"). Joint Statement ¶ 24.
11. Within the NCD there are twelve separate UFCW Locals which have collective bargaining agreements with Lucky ("UFCW Contract"). Joint Statement ¶ 25. For the purposes of this action, the UFCW Contracts for the period 1980 through 1992 between Lucky and Retail Clerks Local 775 are representative of the collective bargaining agreements for the same time period between Lucky and the other UFCW Retail Clerk locals representing employees in Lucky's retail food stores. However, for the agreements for the period 1989 though 1992 involving Locals 588 and 1288 with regard to the issues of job posting and movement of Deli/Bakery and General Merchandise Clerks into Journey Food Clerk vacancies, the UFCW Contract between Lucky and UFCW Retail Clerks Local 588 is representative of the UFCW Contract between Lucky and Local 1288. Stipulation Regarding Collective Bargaining Agreements (filed May 29, 1991).
12. Prior to 1986, all Assistant Store Managers were in the UFCW collective bargaining unit. From February 1986 through February 1989, the UFCW collective bargaining unit included Assistant Store Managers who worked in stores of 30,000 square feet or less. However, since February 1989 the collective bargaining unit has excluded all Assistant Store Managers except for those who have elected to remain within the UFCW. Thus, the current UFCW Contract covers all store employees, except Store Managers and most Assistant Store Managers. Joint Statement ¶ 26.
13. Under the terms of the UFCW Contract, seniority is measured by continuous service at Lucky. Under certain circumstances, the seniority of an employee may be broken or temporarily suspended. Joint Statement ¶ 27.
14. The UFCW Contract includes a nondiscrimination clause:
*268 The Employer shall not discriminate against any person in regard to hire, tenure of employment or job status because of race, creed, religion, color or national origin, nor shall age, physical handicap unrelated to the job duties, veteran status or sex under any circumstances be a basis for rejection or termination of an otherwise qualified employee or applicant for employment.
UFCW Contract § 2.4.4. Joint Statement ¶ 28.
15. Wages for all of the positions in the retail stores, with the exception of Store Manager and Assistant Store Manager, are set forth in the UFCW Contract. Joint Statement ¶ 29.
16. The UFCW Contract provides that all employees, except Courtesy Clerks, who work between the hours of 7:00 p.m. and 7:00 a.m. shall receive a "night premium" wage rate for those hours of work. The contract provides that Courtesy Clerks shall receive a night premium for work performed after 9:00 p.m.. Joint Statement ¶ 29.
Retail Food Stores
17. There are approximately 18,000 employees in the NCD, 15,000 of whom work in the retail stores. Reporter's Transcript ("R.T.") at 4-623 (Grant). 124 of the 188 stores in the NCD are open twenty-four hours a day. R.T. at 4-665 (Grant).
18. Each retail store contains the following departments: Grocery, Produce, and General Merchandise (also referred to as Non-Food). Joint Statement ¶ 11. Approximately 125 of the stores in the NCD have Deli/Bakery departments. R.T. at 4-649 (Grant). The Grocery department, the largest department in each of the stores, is responsible for 65% of Lucky's sales. The Meat department is responsible for 12% of the sales, Produce is responsible for 6-10%, General Merchandise is responsible for 8%, and Deli/Bakery is responsible for 3.5%. R.T. at 4-652 (Grant). Work in the Grocery department is available twenty-four hours a day, Night Crew work is from about 11:00 p.m. to 9:00 a.m., Produce department work is from 4:00 a.m. to 11:00 p.m., Deli/Bakery work is from 3:00 a.m. to 10:00 p.m., and General Merchandise work is from 5:00 a.m. to 4:00 p.m. R.T. at 6-997-98 (Hoffman). See figure 1.
19. Depending on the size of the store, each store in the NCD has Department Head or Head Clerk level personnel which include: a General Merchandise Department Head who reports to the Store Manager; a Deli/Bakery Department Head who reports to the Store Manager; a Produce Manager who reports to the Store Manager; a Head Clerk/Front-End Clerk who reports to the Store Manager; a Head Clerk/Fifth Person who reports to the Store Manager; a Head Clerk/Fourth Person who reports to the Store Manager; a Head Clerk/Third Person who reports to the Store Manager; and a Head Clerk/Night Crew Manager who reports to the Store Manager. Each store also has an Assistant Store Manager who reports to the Store Manager and a Store Manager who reports to the District Manager. Joint Statement ¶ 12.
20. The Store Manager, Assistant Store Manager, Third Person, Fourth Person, Fifth Person, Front-End Clerk, and Night Crew Manager positions are all in the Grocery Department. Joint Statement ¶ 13.
21. In addition to the Grocery Department Head positions, other Department Head level positions in the Grocery department may include Receiving Clerks (later replaced by Direct Store Delivery or DSD Clerks.) Joint Statement ¶ 14.
22. Other personnel in the Grocery Department include Courtesy Clerks, Utility Clerks, Apprentice Clerks, and Journey Clerks. Joint Statement ¶ 15. The duties of Courtesy Clerks are specifically described in the UFCW Contract and include bagging of merchandise, cleaning up and price checks. Courtesy Clerks are not permitted to work at the checkstand. They *269 are primarily part-time employees. Joint Statement ¶ 16. Utility Clerks essentially perform janitorial duties. Joint Statement ¶ 17.
23. In each of the departments relevant to this case, there are Apprentice Clerks, Journey Clerks, and a Department Head. However, Courtesy Clerks only work in the Grocery department. Joint Statement ¶ 20.
24. Journey Clerks and Apprentice Clerks in the Grocery Department may be assigned to a variety of tasks including: working at the cash register (also know as checking) and stocking (also called floor work). Joint Statement ¶ 18. Stockers take the merchandise that is delivered to the store and put it on the store shelves. R.T. at 1-74 (Gill). Journey Clerks and Apprentice Clerks in the Grocery Department can be assigned work hours during any time that the store is in operation. Joint Statement ¶ 19.
25. Journey Clerks and Apprentice Clerks in the Produce Department may be assigned to a variety of tasks, including stocking and arranging displays. Joint Statement ¶ 21.
26. Journey Clerks and Apprentice Clerks in the General Merchandise Department may be assigned to a variety of tasks, including stocking. Joint Statement ¶ 22.
27. Journey Clerks and Apprentice Clerks in the Deli/Bakery Department primarily prepare and sell products. Joint Statement ¶ 23.
28. During the liability period Lucky operated a number of Gemco stores, including Gemco stores within the geographic scope of the NCD. Each Gemco store included a retail food department and a retail variety goods department. Those individuals employed in the retail food department of Lucky's Gemco stores were considered NCD employees and were members of the same UFCW Locals as were other employees in the NCD. Joint Statement ¶ 9.
29. Until approximately the last year of their operation, Gemco's retail food departments were assigned to districts in the NCD according to the same geographic boundaries used to assign the stand-alone retail food stores in the division. Joint Statement ¶ 9.
30. In November 1986 Lucky closed all of its Gemco stores, including the Gemco retail food departments. The employees in the retail food departments were given the right to apply for positions in the NCD based on their seniority according to the terms of the UFCW Contracts. While many former Gemco employees were absorbed into NCD stores, some employees were laid off, as were some Lucky retail food store employees. Joint Statement ¶ 10.
II. LUCKY'S EMPLOYMENT PRACTICES
31. Mark Foley has been the Equal Employment Opportunity ("EEO") Manager for Lucky since 1989. In that capacity he responds to internal complaints and administrative complaints from the Department of Fair Employment and Housing ("D.F.E.H.") and the Equal Employment Opportunity Commission ("E.E.O.C."). Foley also administers Lucky's affirmative action programs and works with the Bockman consent decree. R.T. at 8-1308. Foley worked as Lucky's EEO Recruiter from 1988 to 1989 and worked in the Human Resources Department prior to that. He helped Russell Specter with the EEO training that Specter provided to the District and Store Managers. R.T. at X-XXXX-XX. Foley was first hired by Lucky as a Courtesy Clerk in 1977. R.T. at 8-1297.
32. Robert S. Gill, has been the Manager of Labor Relations of Lucky's NCD since February 1986. Gill represents Lucky in grievances, arbitrations, and negotiations of UFCW Contracts. Before 1986 he was the Personnel Manager for the NCD. Gill was designated under Fed. R.Civ.P. 30(B) (6) as the person at Lucky most knowledgeable about the duties and minimum qualifications for job classifications, about the criteria and procedures for obtaining work in different departments and shifts, and about the criteria and procedures for getting promotions. His immediate *270 superior is Walter Herkal, the Vice President, Industrial Relations Manager. R.T. at 1-51-53.
33. Richard Goodspeed has been the President of Lucky's Northern California Food Division since 1986. Until 1989, he reported to Leon Roush, the Corporate Senior Vice President of Operations. He currently reports to Larry Del Santo, the Corporate President of Lucky. He was first hired by Lucky as a part-time Apprentice Clerk in the SCD in 1953. From 1960 to 1966 he worked as a Store Manager in Lucky's SCD. R.T. at XX-XXXX-XX. Goodspeed tries to visit a few stores once or twice a week. He talks to store management, employees and customers. R.T. at XX-XXXX-XX.
34. Robert John Grant, is currently the Vice President, Operations Manager of the Bay Area Lucky stores. He began his career at Lucky in 1952 working as a Courtesy Clerk. R.T. at 2-281-86. Grant is responsible for all of the stores in the Northern Food Division in California and Nevada. He reports to the Director of Operations. R.T. at 2-284. One of his duties is to visit stores to see if store practices are in accordance with management policies. He visits stores two days a week and visits about fifty stores throughout the districts in a month. Ex. A-37; R.T. at 2-287. He meets with and talks to the five District Managers under his supervision regularly. R.T. at 4-639-40. In addition, he has had breakfast meetings on a regular basis with different Lucky employees for the past four or five years. R.T. at 4-633-34.
35. Walter Herkal has been the Vice President, Industrial Relations for the NCD since 1979. He was first hired by Lucky in 1977 as the Industrial Relations Manager for the Northern Region of Lucky Stores. Until 1990, Herkal was the primary person responsible for personnel, including affirmative action and EEO matters. R.T. at XX-XXXX-XX.
36. Alvin M. Hoffman was the District Manager for District 5, the Valley Stores, from 1985 through 1990. R.T. at 6-934. He was first hired by Lucky as a Courtesy Clerk in 1952. He worked as a Store Manager from 1968 to 1979. Since 1990 he has rotated between Districts 4 and 8. R.T. at 5-875-77. He spends thirty to forty hours a week visiting the stores in his Districts. R.T. at 6-967.
37. Virginia Javier has been the Vice President of Human Resources in Lucky's NCD since 1990. She has primary responsibility for administering EEO and affirmative action policies in the NCD. R.T. at 10-1651. Javier was first hired by Lucky in 1986 as an EEO Manager to work on administrative implementation of the Bockman consent decree. In 1989 she became Manager of Human Resources. R.T. at XX-XXXX-XX. Javier initially reported to Jane't Noriega-Ailor. She then reported to Walter Herkal. R.T. at XX-XXXX-XX.
38. Jane't Noriega-Ailor left Lucky in 1988. R.T. at 8-1405. Lucky hired her in 1983 to be the Human Resources and Training Manager. Her assignment was to create and implement a full service Human Resources Department. Noriega-Ailor reported to Walter Herkal, the Vice President, Industrial Relations Manager. R.T. at 8-1406. In 1984 she worked as a Labor Relations Specialist at Lucky. In 1985 she was given the title of Human Resources Manager. Her duties were personnel and payroll administration, training development, equal employment opportunity, communications and non-union payroll. R.T. at 9-1596.
39. Russell Specter was hired by Lucky in 1988 "to design and present a series of training programs regarding non-discriminatory promotion of females into management and sexual harassment." Ex. A-117; R.T. at X-XXXX-XX. He had previously represented Lucky as an attorney in the Bockman litigation, and had represented Lucky before the E.E.O.C.. R.T. at X-XXXX-XX.
40. Rosalind Speaker Thompson has been the Director of Human Resources at Lucky since March 1989. From 1985 to *271 1989, Thompson held the position of Corporate Personnel Manager at Lucky. Her responsibilities were to design a Global Management Development Program to prepare individuals to move into middle and senior level management positions at Lucky. From 1982 to 1984 Thompson was director of Human Resources at the Atherton Division of Lucky. Thompson Depo. at 17-19, 27.
41. D. Bruce Frazier has been the Employment Manager at Lucky since 1989. He is responsible for the hiring of administrative employees, the resolution of discrimination cases, and other employment related matters. From 1987 to 1989 he was a training supervisor in the Human Resources Department, and from 1984 to 1987, he was a Store Manager in the SCD. R.T. at XX-XXXX-XX.
42. Willi Fumero has held the position of Payroll and Benefits Manager since 1980. Her responsibilities are maintaining personnel records, verifying and compiling employee work schedules, creating the seniority list for allocation of additional hours, and compiling the bid list for movement to full-time. R.T. at XX-XXXX-XX.
43. Bruce Gentile has been the Manager of Training and Development at Lucky since 1988. Gentile's responsibilities are to recommend, evaluate and develop training programs. Gentile was first hired by Lucky in 1972 as a Courtesy Clerk. R.T. at 21-3637. He has held the positions of Fourth Person, Third Person, Assistant Store Manager, and Labor Scheduling Coordinator in the NCD. R.T. at 21-3630.
44. Milton Hardy has been a District Manager at Lucky since 1985. Hardy is responsible for the profitable operation of all the stores in District 1. His duties include staffing, merchandising of stores and representing the district at labor grievances. He tours each store in his district two or three times a month. Hardy was promoted to Store Manager in 1973 and held that position until 1977. From 1977 to 1981 Hardy was a Produce Merchandiser and from 1981 to 1985 he was a Grocery Supervisor. From 1965 to 1973, Hardy held the positions of Journey Clerk, Fourth Person, Third Person and Assistant Store Manager. He joined Lucky in 1965 as an Apprentice Clerk. R.T. at XX-XXXX-XX.
45. Peter Garrett Michon has been the Produce Merchandiser in Lucky's NCD since 1991. Between 1990 and 1991 he attended the Food Management Program at the University of Southern California. From 1986 until he went to USC, Michon was a Store Manager in the NCD. His District Manager is Milton Hardy. He has also held the positions of Fourth Person, Third Person and Assistant Store Manager. R.T. at XX-XXXX-XX.
46. Lucky places advertisements for Deli/Bakery Apprentices. Those advertisements are not specifically directed at women. R.T. at 4-696 (Grant); 6-985 (Hoffman). Lucky advertises for help in all departments upon the opening of a new store. R.T. at 6-985 (Hoffman). After opening, Lucky does not usually place advertisements for openings in departments other than Deli/Bakery or Pharmacy. R.T. at 5-851 (Grant). Lucky has advertised for help in the Grocery, Meat, Produce, and General Merchandise departments only occasionally. R.T. at 6-985 (Hoffman). With the exception of Deli/Bakery, the primary method by which people are recruited at Lucky is through walk-ins, not in response to advertisements. R.T. at 10-1693 (Javier).
47. One to three days a week, Lucky stores accept applications for new hires. R.T. at 8-1296 (Foley); 4-697 (Grant); 10-1695 (Javier). Applicants can only get a job application form from the person in charge of the store. R.T. at 5-852 (Grant); 10-1694 (Javier). Most, if not all, hiring is done through the application process. R.T. at 4-698 (Grant). 90% of the people who apply for positions in the Deli/Bakery department are women. R.T. at 6-991, X-XXXX-XX (Hoffman). Grant testified that he "was told" that more women than men respond to advertisements for Deli/Bakery positions. R.T. at 4-718 (Grant).
*272 48. There is one standard job application form. Generally, the completed applications are retained at the store of application for about a year. Each store has its own policy for keeping completed applications. R.T. at 4-701-02 (Grant). After hiring decisions are made, the completed job applications are sent to the Human Resources Department and kept as part of the employee's personnel file. R.T. at 21-3602 (Fumero); 5-854 (Grant).
49. A Store Manager or an Assistant Store Manager verbally gives applicants for jobs information about the positions that are available. R.T. at 1-144 (Gill). An applicant responding to an advertisement might or might not be informed about other openings in the store. R.T. at 5-855 (Grant).
50. Neither Store Managers nor Assistant Store Managers are given instructions about how to interpret job applications. R.T. at 5-853 (Grant); 21-3496 (Michon). Nor are they given instructions about what to look for in making hiring decisions. R.T. at 1-144 (Gill); 2-345 (Grant). However, the appearance of Lucky employees is important because they are handling food and they must seem sanitary. R.T. at 4-629 (Grant). Except for being over sixteen years of age, there are no minimum qualifications or experience for being hired into Courtesy Clerk or Apprentice Clerk positions. Ex. A-19 at 13; R.T. at 1-64-65 (Gill).
51. Store Managers have absolute discretion in making decisions about hiring. R.T. at 4-693-94 (Grant). Goodspeed believes that Store Managers are in the best position to make hiring, promotion and termination decisions based on the needs of their individual store. R.T. at XX-XXXX-XX (Goodspeed).
52. New employees are hired for specific positions in specific departments. R.T. at 4-693-94 (Grant); 31-5155 (Hardy); 6-983 (Hoffman); 20-3333 (Michon). A new hire to Lucky is usually assigned to a position as a Courtesy Clerk, as a Utility Clerk, or an Apprentice Clerk in any one of the four departments. Courtesy Clerk is the most common entry-level position. Joint Statement ¶ 30. Because there is a high attrition rate in Deli/Bakery, new hires are often assigned to Apprentice positions in that department. R.T. at 4-693-94 (Grant).
53. In deciding the department into which to place a new employee, Store Managers are not constrained by the UFCW Contract or by Lucky policy. R.T. at 1-70-71 (Gill); 31-5153 (Hardy). In fact, Store Managers have complete discretion in making these decisions and make them based on the needs of the store. R.T. at 2-346 (Grant); 6-984 (Hoffman). If there is an opening in the requested department, Store Managers can take the preferences of new hires into account. R.T. at 4-699 (Grant). The decisions of Store Managers regarding placement of new hires are not reviewed by upper-level management. R.T. at 6-984 (Hoffman); 20-3341 (Michon). Store Managers must have final approval from the District Manager or the Grocery Merchandiser before making placements. R.T. at 31-5153 (Hardy).
54. Lucky's standard procedure manual says that "married women who desire to work only a few hours a day often make valuable part-time cashiers." Ex. A-19 at 15. Gill understood this to mean that married women are a good source of part-time employees, and not that married women should only be considered for particular jobs. R.T. at 1-145, 2-215 (Gill).
55. During his tenure as District Manager, Hoffman never received a complaint from a woman about the fact that she was initially placed in the Deli/Bakery or General Merchandise departments. R.T. at 6-991 (Hoffman).
56. The District Manager receives a list of new hires. R.T. at 2-344 (Grant). The Union is not given information about the placement of new hires. R.T. at 2-262 (Gill). UFCW Contract § 2.5.3 requires that Lucky direct new hires to report to the Union within seven days of hire in order that the Union can inform the new hire of the provisions of the UFCW Contract. R.T. at 2-199 (Gill). New hires must join *273 the Retail Clerks Union within thirty days of hire. R.T. at 1-66 (Gill).
A. Lucky's Job Ladders
57. An employee can be promoted from the position of Courtesy Clerk to Apprentice Clerk in either the Grocery, Produce, General Merchandise, or Deli/Bakery departments. Joint Statement ¶ 35.
58. Upon completion of a certain number of work hours set forth in the UFCW Contract, an Apprentice Clerk automatically becomes a Journey Clerk. During the Apprentice Clerk period there are several automatic wage increases after specified numbers of hours worked. Joint Statement ¶ 36.
59. The promotional level above Journey Clerk is Department Head. Joint Statement ¶ 37.
60. The pool of employees considered for promotion to Deli/Bakery Department Head and General Merchandise Department Head is primarily comprised of Deli/Bakery Journey Clerks and General Merchandise Journey Clerks, respectively. Joint Statement ¶ 37.
61. The pool of employees considered for promotion to Produce Manager is primarily comprised of Produce Journey Clerks. Joint Statement ¶ 37.
62. The pool of employees considered for promotion to entry-level department head jobs in the Grocery Department is primarily comprised of Grocery Journey Clerks. The primary entry-level management job classification in the Grocery Department is Fourth Person or, in some stores, Fifth Person. Depending on the size of the store, other Grocery Department Head level job classifications include, Front-End Clerk, Night Crew Manager, and Receiving Clerk. The District Manager, with input from the Store Managers and the appropriate Merchandiser, makes Department Head promotion decisions for Front-End Clerk, Night Crew Manager, and Receiving Clerk. In the case of Produce Manager, Fourth Person, Deli/Bakery Department Head and General Merchandise Department Head promotions, the District Manager's recommendations must be approved by the Vice President of Operations. Joint Statement ¶ 37.
63. The next level of promotion in the Grocery Department is from entry-level Department Head jobs to Third Person. Most of the employees considered for promotion to Third Person come from Fourth Person. Lucky considers movement from Fourth Person to Third Person to be a promotion although both positions receive the same hourly wage rate. Third Person promotions are made by the same decision makers as are Fourth Person promotions. Joint Statement ¶ 38.
64. The next level of promotion in the Grocery Department is from Third Person to Assistant Store Manager. Most of the employees considered for promotion to Assistant Store Manager come from Third Person. Until 1989, the District Managers and the Vice President of Operations made recommendations for promotion to the Assistant Store Manager position to the Senior Vice President of Operations. The Senior Vice President of Operations was responsible for approval. Joint Statement ¶ 39.
65. The final level of promotion in the Grocery Department is to Store Manager. Most of the employees considered for promotion to Store Manager come from Assistant Store Manager. Until 1989, the District Managers and the Vice President of Operations made recommendations for promotion to Store Manager to the Senior Vice President of Operations. The Senior Vice President of Operations was responsible for approval. Joint Statement ¶ 40.
66. No written, oral or hands-on tests are required for assignment or promotion to any retail store position. Joint Statement ¶ 41.
B. The UFCW Contract Provisions on Promotion
67. Section 4.3.1 of the UFCW Contract provides:
Promotion: Determination of which employee is to be promoted will be based upon seniority provided the employee *274 with the highest seniority has the qualifications necessary for the job. Qualifications shall include such factors as experience, job performance, aptitude, attendance, etc. Where merit and ability are approximately equal, seniority shall control.
Joint Statement ¶ 42. This is the only provision in the UFCW Contract regarding promotion criteria. R.T. at 1-96 (Gill); 2-352 (Grant).
68. Section 4.3.2 of the current UFCW Contracts for Locals 588 and 1288 creates a job posting program. That program was implemented in 1989. Section 4.3.2 provides:
All permanent job vacancies above the Journeyman Food Clerk classification shall be posted at each store of the Employer within the seniority area specified herein for a period of five (5) days. The job posting shall specify the job classification and location of the store where the permanent job vacancy exists. Any employee interested in [the] permanent job vacancy must complete a job bid form and return it to the Store Manager on or before the expiration of the posting period. In the event the Employer decides to promote an existing employee to fill the permanent job vacancy, then in that event the selection of the employee to be promoted shall be in accordance with the provision set forth herein.
Any successful bidder who thereafter declines the promotion or is unable to perform the duties of the job shall be ineligible for any subsequent promotional bid for a period of six (6) months.
Joint Statement ¶ 43.
69. Section 4.3.2 of the current UFCW contracts for Locals 588 and 1288 also creates a process by which Non-Food and General Merchandise Clerks may submit a written request for promotion to jobs below Journey Food Clerk. That section provides:
All permanent job vacancies below the Journeyman Food Clerk classification shall be handled on a store-by-store basis. Non-Food and General Merchandise Clerks in each store who have worked in the Non-Food and General Merchandise Clerk classification for a period of two (2) full years and are desirous of promotion and are otherwise qualified for a promotional opportunity in accordance with this provision must file a semi-annual written request for promotion with their Store Manager during the first two (2) working weeks in January and July. In the event the Employer decides to promote an eligible Non-Food or General Merchandise Clerk within the store to fill the permanent job vacancy, then the selection of the employee to be promoted shall be in accordance with this provision.
Joint Statement ¶ 44.
C. Promotion Practices
70. Lucky's general policy is to fill open positions by promoting current employees rather that by hiring people from outside of the stores. R.T. at 1-146 (Gill). Defendant asserts that job posting is not necessary at Lucky because Store Managers know which employees are interested in promotion, R.T. at 6-1063 (Hoffman), and because employees who excel have the ability to communicate their desire for advancement to management, R.T. at 18-3164 (Goodspeed). President Goodspeed believes that job posting would not be helpful. R.T. at 18-3056 (Goodspeed). Grant assumed that Lucky could not institute job posting because it is not provided in the UFCW Contract. R.T. at 4-668 (Grant). However, Grant had not been told this directly by anyone. R.T. at 5-860 (Grant). Defendant has not presented sufficient evidence to convince the court that job posting is proscribed by the UFCW Contract.
71. Lucky does not require that prospective promotees to all positions be interviewed. However, in the past few years, prospective Store Managers have been interviewed. R.T. at 3-438 (Grant).
72. Under UFCW Contract § 4.3.1, if one employee is more qualified than another, the Store Manager does not have to consider seniority in making a promotion decision. R.T. at 1-108 (Gill); 2-353-54 (Grant). Defendant asserts that qualifications are more important than seniority in *275 making promotion decisions. R.T. at 12-2100 (Herkal). Defendant also asserts that it is rare for two employees to have equal qualifications, so that seniority is seldom considered. R.T. at 32-5270 (Hardy). Moreover, because the seniority list is not posted an employee cannot check to see if the promoted employee has more seniority than they do. R.T. at 1-114 (Gill).
73. Lucky has no official or uniform definition of merit, qualifications or ability. R.T. at 1-109-10 (Gill); 2-254, 2-355 (Grant). In addition, Store Managers are not systematically trained about how to define merit, qualifications or ability. Although interpreting the UFCW Contract is part of his job, no one had ever asked Gill how to interpret the term "qualifications" in that document. R.T. at 1-110 (Gill). There is no uniform definition of the term "job experience" in UFCW Contract § 4.3.1. R.T. at 13-2274 (Herkal). An employee's attitude, appearance, career-mindedness and intelligence can also be considered in promotion decisions. R.T. at XX-XXXX-XX (Herkal).
74. The practices and policies of Store Managers are based on their personal judgment and experience. Unless they have worked at another store, they generally have no knowledge of the policies of other Store Managers. R.T. at 2-300 (Grant); 21-3540 (Michon). President Goodspeed believes that written promotion criteria are not necessary because "[t]his is not a big sophisticated business. This is a store that has 50 to 100 employees. They are all in the same building. They are not off working in a vacuum. They talk to each other. They socialize together, go on breaks together, go to lunch together, see each other, see what the other person is doing." R.T. at 18-3172 (Goodspeed). In addition, defendant believes that Store Managers are best qualified to make hiring, promotion and termination decisions based on the needs of the individual stores, R.T. at XX-XXXX-XX (Goodspeed), even though these decisions may be subjective and may vary from store to store.
75. Store Managers do not look at what employees wrote on their initial applications when deciding whom they should promote to what position. R.T. at XX-XXXX-XX (Michon). In fact, job applications are neither available nor reviewed at the time of promotion.
76. There are no documents that reflect who was considered for a promotional opportunity, that reflect an employee's interest in working a different shift, that reflect the qualifications of employees for promotions, that reflect whether a Store Manager considered more than one person for a promotion, or that inform employees how they should express interest in promotion. R.T. at XX-XXXX-XX (Javier).
77. In 1988, Russell Specter told Lucky's District Managers that without job descriptions they would have no basis on which to justify their promotion standards. R.T. at 7-1173 (Specter). Noriega-Ailor suggested that Lucky create formal job descriptions because they would help management focus on whether a prospective promotee has the skills necessary to perform the job. R.T. at 8-1422 (Noriega-Ailor). However, other than what is written in the UFCW Contract, there are no job descriptions for Courtesy, Apprentice or Journey Clerk positions in any department. R.T. at 2-293, 4-656-58 (Grant); 10-1661 (Javier). There are no written promotion criteria for any position below Assistant Store Manager. R.T. at 10-1661 (Javier). The Union did not object when Lucky adopted job descriptions for Night Crew Manager and Third Person in 1989. R.T. at XX-XXXX-XX (Hardy); 13-2268 (Herkal). Lucky has failed to develop job descriptions that would provide guidance and objective standards for evaluation and promotion decisions, although the UFCW Contract does not bar the provision of job descriptions.
78. Lucky's affirmative action policy says that "all promotional standards shall be realistically related to job requirements and shall be periodically re-examined to determine that they are genuinely job-related." Ex. A-67; R.T. at 8-1428 (Noriega-Ailor).
79. In late 1983, Noriega-Ailor organized a performance appraisal system for Lucky. All of the Store Managers were *276 trained in the system which required that performance appraisal forms be completed and sent to the personnel office. However, in late 1984 Noriega-Ailor found that the appraisals forms were not being completed; eighty-one out of 126 stores had not filled out any forms. R.T. at X-XXXX-XX (Noriega-Ailor). Compliance with the system picked up by 1988. R.T. at 9-1605 (Noriega-Ailor). Lucky conducted one appraisal training program for Store Managers in 1987 and another for Store Managers and District Managers in 1989. R.T. at 10-1732, 10-1734 (Javier).
80. Store Managers are not required to review performance appraisal forms before making promotion decisions. R.T. at 9-1624 (Noriega-Ailor). Grant never asks for employees' performance evaluations in deciding whether to promote them. Moreover, despite the rule that employees with over a certain number of disciplinary notices cannot be promoted, it does not appear that anyone checks disciplinary notices. R.T. at 3-448-49 (Grant).
81. If a Store Manager makes an improper promotion decision, the aggrieved employee can bring a grievance. R.T. at XX-XXXX-XX (Herkal).
a. Apprentice Clerks
82. There is no bid form or application process for promotion to Apprentice Clerk positions. Courtesy Clerks have no formal way of knowing who is being considered for promotion. Store Managers need not document the reasons for their promotion decisions and those decisions are not subject to review except through the grievance process. R.T. at 1-104-05 (Gill). Thus, promotion decisions are extremely subjective, vary from store to store, and allow stereotypes and other improper considerations to enter into the decision making process.
83. Courtesy Clerks are equally qualified for Apprentice Clerk positions in each of the four departments. R.T. at 2-359 (Grant). Store Managers consider whether an employee is available when deciding who to promote to Apprentice. They do not consider the experience or education of the potential promotee. R.T. at 1-104-05 (Gill). Michon only promotes Courtesy Clerks to Apprentice positions in Grocery because he "knew their capabilities ... [he] knew what they could do." R.T. at 20-3334 (Michon). Herkal has found that Courtesy Clerks have more relevant experience for promotion to the position of Grocery Apprentice than do Deli/Bakery or General Merchandise Clerks. Deli/Bakery or General Merchandise Clerks usually do not have exposure to the Grocery department. R.T. at 12-2168 (Herkal). Herkal also believes that a less senior Courtesy Clerk can be promoted to a Grocery Apprentice position over a more senior Deli/Bakery or General Merchandise Clerk because there are separate seniority systems for Food and Non-Food employees. R.T. at 12-2167 (Herkal).
84. Store Managers have complete discretion in deciding which department each Apprentice Clerk should be placed in. Store Managers also decide who is to be assigned to night and day work. R.T. at 1-72-75 (Gill); 2-343, 2-357-58 (Grant). They do not have to consider what employees indicated as their department of preference on their initial application. However, Store Managers must consult the Department Head and the District Manager before promoting into a department. R.T. at 2-342, 2-359 (Grant).
b. Journey Clerks
85. An employee must have 2,080 hours of experience to be hired directly into a Journey Clerk position. Only hours spent at the Apprentice level count toward those necessary for promotion to Journey Clerk. Lucky has a system to ensure that this policy is followed. R.T. at 1-76-77 (Gill).
86. Store Managers have complete discretion in deciding which tasks each Apprentice or Journey Clerk should perform (checking, stocking, or receiving merchandise), and which shifts they should work. No records are kept of these decisions and there is no review of these decisions. R.T. at 1-79-80 (Gill); 2-340-41 (Grant); 5-888 (Hoffman).
*277 c. Fifth Person/Front-End Manager
87. The position of Fifth Person only exists in high volume stores, defined as those that take in over a half million dollars a week. R.T. at 2-350, 4-668 (Grant). There are about thirty or forty such stores in the NCD. R.T. at 4-670 (Grant). The duties of a Fifth Person are essentially the same as those of a Fourth Person. R.T. at 4-669 (Grant).
88. The Front-End Manager position was created in 1987. Front-End Managers monitor and supervise the Courtesy Clerks. Ex. A-47; R.T. at 2-332-33 (Grant). Most Front-End Clerks are women. R.T. at 2-336 (Grant). Grant could not remember whether any Front-End Clerks had ever been promoted into management. R.T. at 3-373 (Grant).
89. Michon considered experience, interest, cooperation, availability and job performance in deciding who should be promoted to Front-End Manager. R.T. at 20-3423 (Michon).
d. Department Head/Fourth Person/Receiving Clerk
90. Fourth Person is the principal entry-level management position. R.T. at 1-90 (Gill); 3-382 (Grant). Fourth Person is responsible for light bookkeeping, stocking, customer relations, security matters, building displays, and is often left in charge of the store. R.T. at XX-XXXX-X (Michon). Grant testified that the Fourth Person closes up the store once or twice a week, replaces the Third Person if s/he is on vacation or has stepped up, does some bookkeeping, moves heavy dairy cases, and uses a forklift to build the displays. The Fourth Person must answer customer complaints and ensure the security of the store. R.T. at X-XXXX-XX (Foley). The Fourth Person is required to work late shifts and weekend shifts. R.T. at 4-668-72 (Grant); 6-1031 (Hoffman). However, the Fourth Person also does a lot of office work. R.T. at 5-843 (Grant).
91. Store Managers make the initial recommendation to the Grocery Merchandiser that an employee should be promoted to Fourth Person. R.T. at 3-437 (Grant); 6-1039 (Hoffman). The District Manager has the final say over the promotion. R.T. at 6-1040 (Hoffman). In deciding whom to recommend for promotion to Fourth Person, Store Managers consider whether the employee is interested in the job, is cooperative, and is available to accept the position. R.T. at XX-XXXX-X (Michon). The seniority of the prospective Fourth Person is less important than the person's qualifications, availability and interests. R.T. at 21-3529 (Michon). In recommending employees to be promoted to Third Person and Fourth Person, Hoffman considers whether they have experience in the Grocery department, whether they are a good role model, whether they get a job done, their dependability (punctuality), performance, honesty, ability, intelligence, career-mindedness, and availability. R.T. at X-XXXX-XX, 7-1138 (Hoffman). Hardy considers level of interest, availability, experience, seniority, disciplinary write-ups, and job performance. R.T. at XX-XXXX-X (Hardy). There is no written list of criteria for promotion to Fourth Person. Prior discipline problems do not automatically bar someone from promotion to Fourth Person, unless the disciplinary write-up was recent. R.T. at X-XXXX-XX (Hoffman).
92. Grant testified that in order to be promoted to Fourth Person, a Journey Clerk must have experience on the Night Crew. R.T. at XX-XXXX-X (Michon); 5-799 (Grant). Michon felt that it was necessary for a Night Crew Manager to have night stocking experience, because "they should know the job in order to supervise it." R.T. at 20-3408 (Michon). Hoffman believed that Night Crew experience made an employee more promotable to Fourth Person, although it was not required; nevertheless, he did require that the employee have stocking experience. R.T. at 5-888, 6-1046, 6-1072 (Hoffman). Foley thought that in order to be promoted to Fourth Person, an employee must know how to work the milk box, the beverage aisle, merchandising, and displays. Employees *278 who work in the late evenings have the opportunity to learn these tasks. Some of the duties of Fourth Person, such as stocking merchandise, working the beverage aisle and working the milk box require heavy physical work. R.T. at X-XXXX-XX (Foley). Grant believed that experience in the Grocery department is important for promotion to management, because 65% of Lucky's business comes from the Grocery department. R.T. at 4-653 (Grant).
93. Lucky does not have a written or unwritten policy which prevents Department Heads in the Deli/Bakery or General Merchandise departments from being considered for Fourth Person. However, employees from these departments are rarely promoted to Fourth Person. Generally, Store Managers do not consider employees in the Deli/Bakery or General Merchandise departments for promotion to Fourth Person. R.T. at 21-3530 (Michon). Grant believed that Deli/Bakery and General Merchandise employees would require training before they would be qualified for promotion because they would not have had experience in the Grocery department. R.T. at 3-453, 4-716, 4-721 (Grant). Deli/Bakery and General Merchandise Department Heads would most need training in bookkeeping in order to be qualified for promotion to Fourth Person. R.T. at 5-831 (Grant). Store Managers would be more likely to promote the General Merchandise Department Heads who have Grocery experience that those who do not have Grocery experience to Fourth Person. R.T. at 5-832 (Grant).
94. It is very unlikely that a part-time employee would be promoted to an entry-level management position, R.T. at 3-500 (Grant), although part-time and full-time Journey Clerks are equally eligible for promotion to Fourth Person, R.T. at 1-107 (Gill).
95. Seniority should be the deciding factor in a promotion to Fourth Person only when the potential promotees have the same abilities. Hoffman has never seen a situation where the potential promotees have the same abilities. R.T. at 7-1113 (Hoffman). Herkal testified that seniority is not a significant factor in promotions to positions above Department Head. R.T. at 12-2100 (Herkal).
96. Since 1988 Grant has signed off on recommendations for promotion to Fourth Person. In deciding who to recommend for promotion to Fourth Person or Third Person, the District Manager discusses the qualifications of the candidates with the Store Manager. The District Manager is also familiar with the qualifications of the employees in his or her district. R.T. at 4-641-42 (Grant). Grant acknowledged that Store Managers have their own criteria for making these recommendations, but said that "you have to rely on the Store Manager's past experience. The Store Manager was a Third Person or Fourth Person. The Store Manager would know what is needed for Fourth Person." R.T. at 3-384, 3-420-21 (Grant). Some of the factors Store Managers might consider in making this promotion recommendation are the employee's appearance, attitude, dress, aggressiveness, and family responsibility. R.T. at 3-488 (Grant). However, there is no assurance that these factors are considered in all or most promotion decisions. The District Managers do not have a consistent set of criteria which they check with the Store Managers before approving promotions.
97. Receiving Clerks handle the merchandise that comes into the store and check off inventory. There is no job description or selection criteria for Receiving Clerk. R.T. at 2-305-06 (Grant). Receiving Clerks were replaced by DSD Clerks in every store by April 1989. DSD clerks check the inventory on a computer and do not do as much physical work as did Receiving Clerks. R.T. at 8-1327 (Foley); 2-308 (Grant). The former Receiving Clerks got first choice of the new DSD Clerk jobs. R.T. at 2-315 (Grant). There are no minimum training or experience requirements for DSD clerk, however, there are established selection standards for DSD Clerks. Ex. A-44; R.T. at 2-309-10 (Grant).
e. Third Person/Head Clerk
98. The only job description for Third Person/Head Clerk is in UFCW Contract § 188.8.131.52.3, which says Third Person opens *279 and closes the store and is responsible for the operation of the store in the absence of the Store Manager and Assistant Store Manager. R.T. at 1-91 (Gill).
99. There are no written or unwritten standards for promotions to Third Person. R.T. at 2-356 (Grant). The criteria Store Managers consider in deciding promotions to Third Person are the same as the criteria for promotions to Fourth Person. However, holding the position of Fourth Person is a prerequisite to being a Third Person. R.T. at XX-XXXX-XX (Michon).
100. Most promotions to Third Person are based on qualifications, not seniority. There is no uniform definition of merit, qualifications or ability, and Store Managers need not document their findings as to qualifications. R.T. at 1-109 (Gill).
f. Assistant Store Manager
101. In addition to doing labor scheduling and running the store when the Store Manager is off duty, the Assistant Store Manager performs all of the duties of a Journey Clerk, a Fourth Person and a Third Person. R.T. at 6-1037 (Hoffman).
102. Hoffman testified that there is a written job description for Assistant Store Manager. R.T. at 5-883 (Hoffman). However, Gill testified that there is no job description for Assistant Store Manager. The UFCW Contract does not require that Assistant Store Managers have any minimum of experience, training or education. R.T. at 1-94-95 (Gill).
103. Assistant Store Managers do much of the payroll, help write the work schedules, and oversee the operation of the entire store. R.T. at 4-676 (Grant).
104. Before 1989, Grant signed off on recommendations for promotion to Assistant Store Manager and the Director of Operations made the final approval of the promotion. Grant currently signs off on recommendations for promotion to Assistant Store Manager in his districts, and the President of Lucky has the final approval. R.T. at 3-423 (Grant).
105. There are no set criteria for promotion to Assistant Store Manager. R.T. at 3-424 (Grant). Grant considers the volume and type of store from which the prospective Assistant Store Manager came, and the Store Managers for whom s/he worked. R.T. at 4-679 (Grant). Grant assumes that the District Managers talk to the employees they recommended for promotion. R.T. at 3-440 (Grant).
g. Store Manager
106. The Store Manager is responsible for labor scheduling, for the labor budget, and for the general supervision of the entire store. R.T. at 6-1038 (Hoffman).
107. On average, it takes six to ten years to progress from Fourth Person to Store Manager. R.T. at 6-1038 (Hoffman). From the time an employee is first hired, it takes approximately twelve to fourteen years to become a Store Manager; an employee usually remains in the position of Assistant Store Manager for five or six years before being promoted. R.T. at XX-XXXX-XX (Hardy).
108. There is a written job description for Store Manager. Ex. A-38; R.T. at 2-302 (Grant). The job description includes a summary of duties and a list of primary responsibilities in directing store operations, merchandising, personnel, capital expenditures, and store visits. The special requirements for the position are that the Store Manager must be a high school graduate with a thorough working knowledge of the retail food or drug business, have five to ten years of experience in the retail food or drug business (two years of which must be at the Assistant Store Manager level), have the ability to organize a retail store operation, the ability to train, organize and discipline employees, and the ability to operate a store under a union contract. Ex. A-38.
109. The UFCW Contract contains a job description for Store Manager. The job description says that "a managing clerk is an employee who has charge of and general supervision over not more than one store." UFCW Contract § 9.1.1; R.T. at 1-95 (Gill).
110. There are no written criteria for deciding who should be promoted to Store *280 Manager. R.T. at 1-96 (Gill); 18-3049 (Goodspeed); 5-880 (Hoffman).
111. Although Grant testified that the Store Manager position is open to Deli/Bakery Department Heads as long as they are good workers, R.T. at 5-830 (Grant), Gill acknowledged that he did not know of any Deli/Bakery Department Heads who had become Store Managers. R.T. at 2-249 (Gill). Gill did not think that Deli/Bakery Department Heads would be qualified for the Store Manager job because they would not be familiar with the Grocery and Produce departments. R.T. at 2-268 (Gill).
112. President Goodspeed has the final say in Store Manager promotions. He makes his decisions based on the information provided on the promotion recommendation forms and from conversations, but he does not look at the employee's personnel file. R.T. at XX-XXXX-XX (Goodspeed).
D. The Valley Posting Program
113. Before 1989 there was no job posting, bid process or application system for promotions at Lucky. R.T. at 1-115-16 (Gill). In 1989, Local 588, Sacramento and Local 1288, Fresno set up a job posting program ("Valley Posting Program"), which was rejected by the other retail clerk locals. Ex. A-16 at 13; R.T. at 2-226 (Gill); 12-2134 (Herkal). This is the only job posting program at Lucky. R.T. at 1-122 (Gill). Herkal testified that Lucky cannot implement job posting throughout all of its stores without Union approval, R.T. at 12-2144 (Herkal), however, defendant has not presented sufficient evidence to convince the court that job posting is proscribed by the UFCW Contracts. Goodspeed believes that a job posting program is not necessary because employees find out about openings by word-of-mouth. R.T. at 18-3124 (Goodspeed).
114. Under the Valley Posting Program, openings are posted in all of the stores within the seniority area and bids for the opening are forwarded to the District Manager. The District Manager considers all the bids, interviews candidates for promotion and makes a selection. R.T. at 2-212 (Gill). However, according to Hoffman, the District Manager for the Valley Stores, employees who had not submitted bids would also be considered for promotions. R.T. at 8-1387 (Foley); 6-929 (Hoffman).
115. Foley was responsible for tracking promotion bids. R.T. at 8-1375 (Foley). In 1990, he found that 26% of the employees bidding for Fourth Person were women, while 34% of those eligible to bid were women. R.T. at 8-1384 (Foley).
116. The job posting program allows Journey Clerks in the Non-Food departments who have worked at Lucky for two years to bid for promotion semi-annually. This is the only way that Journey Clerks in the Non-Food departments can move out of those departments. R.T. at 1-122 (Gill); 12-2165 (Herkal). Hoffman was not aware of this provision of the program. R.T. at 6-929 (Hoffman).
117. The Valley Posting Program only applies to promotions to the positions of Produce Department Head, Third Person, Fourth Person/Receiving Clerk, Night Crew Manager and Head Clerk Night. The program does not apply to openings for Deli/Bakery or General Merchandise Head, Assistant Store Manager or Store Manager. R.T. at 1-120-21 (Gill); 6-928 (Hoffman).
118. District Managers have had no problems working with the Valley Posting System, except that it causes a "little administrative hassle." R.T. at 12-2150 (Herkal).
119. President Goodspeed testified that additional training makes an employee more promotable. R.T. at 18-3166 (Goodspeed).
120. The Store Managers and the Department Heads are primarily responsible for training employees. R.T. at 5-760-61 (Grant). However, it is too expensive to offer additional training to all employees. R.T. at 18-3167 (Goodspeed). Store Managers rely on their personal experience to decide who should get training, especially *281 for entry-level management positions. R.T. at XX-XXXX-XX (Gentile); 1-93 (Gill); 3-487 (Grant); X-XXXX-XX (Hoffman). In making these decisions, they might consider appearance, attitude, dress, aggressiveness, and the employee's family responsibility. R.T. at 3-488 (Grant). Store Managers are not required to survey their employees to see who is interested in training and promotion. R.T. at 3-490 (Grant). Nor does Lucky systematically monitor which employees receive on-the-job training. R.T. at 22-3746 (Gentile). Historically, Lucky has relied on on-the-job training. R.T. at 21-3647 (Gentile). There is no formal training program for non-management employees to make them more promotable. R.T. at 3-486 (Grant).
121. Store Managers are more likely to promote an experienced employee than to train an inexperienced employee to fill a position. R.T. at 3-495 (Grant).
122. Section 184.108.40.206 of the UFCW Contract mandates that Grocery Apprentices must get thirteen weeks of checking experience and thirteen weeks of stocking experience during their first year at Lucky. There is no mechanism to ensure that every Grocery Apprentice receives this training. R.T. at 1-66-68 (Gill). However, neither Gill nor Grant knew of any complaints or grievances brought by an employee because they had not received this training. R.T. at 2-212 (Gill); 5-769 (Grant). Neither had Grant received complaints from women that they received less training than did men. R.T. at 5-761 (Grant).
123. Deli/Bakery employees receive several types of formal training: baking school, a baking manual, a deli manual, cake decorating school, a cake decorating manual, and a sixteen hour class on product knowledge and suggestive selling. R.T. at XX-XXXX-XX (Gentile).
124. The Entry-Level Management Training program ("ELMT") was instituted in the NCD in 1988. Ex. B-204; R.T. at 5-857-59 (Grant). ELMT-1 is for Fourth Persons, and ELMT-2 is for Third Persons. R.T. at 8-1371 (Foley); 22-3689 (Gentile). The programs last four weeks and provide hands-on training in all of the departments. Foley monitors which employees are selected to participate in the ELMT. Some of the graduates of the ELMT have been promoted to Assistant Manager. R.T. at XX-XXXX-XX (Javier); 5-762-65 (Grant).
125. Only those employees who are recommended by the District Managers may apply to the program. R.T. at 22-3759 (Gentile). Each District Manager recommends two employees for each program. The selection committee makes the decision of who should participate. Foley testified that over thirty employees have participated in each program, and over 45% of them have been women. R.T. at X-XXXX-XX (Foley); 22-3758 (Gentile).
126. Gentile expressed concern that if the ELMT was made available to Deli/Bakery or General Merchandise Clerks, they would use the program to transfer into the higher paying departments. He suggested that the program should only be open to Journey Clerks and to Department Heads. Ex. B-197 at D0086211; R.T. at XX-XXXX-XX (Gentile). Herkal expressed similar concerns. R.T. at XX-XXXX-XX (Herkal).
127. As of April 9, 1990 the ELMT was not operational in all the Lucky stores. However, it was expanded to all of the stores by the end of 1990. Ex. A-204; R.T. at 5-820 (Grant).
128. There have been seventy-seven graduates from the ELMT program, thirty-five of whom were women. Of the twenty-two graduates who have subsequently been promoted, ten have been women. R.T. at 22-3694 (Gentile).
129. The Retail Management Development Program ("RMDP") was instituted in the NCD in 1987. R.T. at 5-857-59 (Grant). This program is open to employees who have been Assistant Managers for six months to a year, and are being considered for promotion to Store Manager. The program lasts for three months and a large percentage of the graduates have been promoted to Store Manager. Ex. B-189; R.T. at 5-766-68 (Grant). The District Managers also select the participants in this program. R.T. at 21-3662 (Gentile).
*282 130. About sixty people have graduated from the RMDP and women have participated in the program at a higher rate than their incumbency rate among Assistant Store Managers. R.T. at 8-1374 (Foley). Gentile believed that of the sixty participants in the RMDP, seventeen were women. In addition, half of the graduates of the program have been promoted, six women among them. R.T. at XX-XXXX-XX (Gentile).
131. Lucky does not post notices of openings in the ELMT or RMDP programs. Nor does it post the procedures employees must follow to get into the programs. R.T. at 22-3757 (Gentile). Neither the ELMT nor the RMDP is open to General Merchandise or Deli/Bakery Department Heads. R.T. at 22-3739, 22-3755 (Gentile).
132. In 1987 Lucky held two career days to teach their employees about career opportunities in the stores, and to give them a better understanding of Lucky's total operation. Ex. B-236; R.T. at 5-770-72 (Grant); 10-1663, 11-1920 (Javier). The career days were held because the majority of Lucky employees had little knowledge of the jobs outside of their immediate operating area. R.T. at 5-846 (Grant). A pamphlet was handed out which gave a history of Lucky and vague job descriptions. Ex. A-30; R.T. at 5-829 (Grant). However, fewer than 200 of the 15,000 Lucky employees in the retail stores in the NCD attended the 1987 career days. R.T. at 5-825 (Grant). Javier testified that over 500 people attended. R.T. at 11-1921 (Javier).
F. Movement between Departments
133. Lucky does not have a written or unwritten policy that prohibits transfers between departments, but they do not happen often. Most transfers are between the Grocery and Produce departments. Employees transfer into and out of the Deli/Bakery department infrequently. R.T. at 8-1306 (Foley); 1-124-25 (Gill); 4-704-07 (Grant); 6-986 (Hoffman); XX-XXXX-XX (Javier); 20-3343 (Michon). Employees lose their seniority when they transfer between departments. R.T. at 4-705 (Grant).
134. Store Managers have complete discretion in deciding whether to allow an employee to move between departments. In making this decision, the Store Manager should consider whether there is an opening in the department, and whether the prospective transferee has good work ethics, is a productive individual, and is neat and clean in appearance. R.T. at 18-3115 (Goodspeed); 4-704 (Grant). However, Store Managers are reluctant to allow employees to transfer out of the Deli/Bakery department because it is expensive to train new people. R.T. at XX-XXXX-XX (Goodspeed); 31-5166 (Hardy).
135. Grant was not aware of any effort by Lucky management to tell Store Managers and District Managers that they should consider General Merchandise and Deli/Bakery employees for positions as Grocery and Produce Apprentices. R.T. at 3-459 (Grant). Goodspeed admitted that no steps had been taken to facilitate movement from General Merchandise and Deli/Bakery to Grocery and Produce. R.T. at 18-3118 (Goodspeed).
136. Bob Beckerman, Vice President of Deli/Bakery, told Grant that Deli/Bakery employees should be considered for entry-level management positions lest employees consider Deli/Bakery to be a dead-end job. R.T. at 3-454 (Grant). Goodspeed testified that he agreed with Beckerman. R.T. at 18-3119 (Goodspeed). However, Grant did not think that jobs in Deli/Bakery were dead-end jobs because Store Managers could promote outstanding people from any department. R.T. at 4-719 (Grant). Grant received a letter from a Merchandiser in the General Merchandise department which listed "very limited opportunities for advancement" and "locked in to a dead-end department" among the reasons for her resignation. Ex. A-158; R.T. at 5-836 (Grant). The evidence is overwhelming that transfers from the Deli/Bakery and General Merchandise departments to the Grocery and Produce departments are rare, and that Lucky's general practices stand as an impediment to such transfers. Thus, employees who are hired into the Deli/Bakery and General Merchandise departments *283 are generally locked into those departments and face substantial barriers in promotion to management positions.
Movement from Part-Time to Full-Time
A. The UFCW Contract Provisions on Movement to Full-Time
137. Journey Clerks may be assigned either to part-time status (16-39 hours a week) or to full-time status (at least 40 hours a week). Joint Statement ¶ 32. The UFCW Contract guarantees part-time workers sixteen hours of work a week, and full-time workers forty hours of work a week. R.T. at XX-XXXX-XX (Michon).
138. The UFCW Contract permits Journey Clerks to submit bids for more work hours in February and August of each year. R.T. at 1-126 (Gill). Section 4.10.1 of the UFCW Contract provides:
Part-time Journeyman Clerks may bid for full-time forty (40) hour job openings or part-time job openings with more hours excluding relief for vacations, illness, or other authorized absences within the employee's assigned store, based upon said employee's seniority provided that he makes his desire for such work known, in writing, concurrently to the Union and to the store manager. Written requests must be made every six (6) months. Written requests may not be submitted outside the specified period. The time periods for requests shall be the first two (2) full working weeks in February and August. Lists are effective the first shift of the month following the request period.
. . . . .
The Employer shall thereupon place the name of the employee on a list maintained by the Company for such purpose. The names of the employees shall be placed upon the list according to seniority. A copy of that list shall be forwarded at the end of each request period to the Union.
Provided the Journeyman Clerk possesses the necessary qualifications and has complied with the requirement above, he shall be offered any job opening except as restricted by the above, which might occur within the employee's assigned store before any employee is hired into said store. [Note: prior to the 1986-1989 contract, this paragraph provided that the Journeyman Clerk would be offered any opening within the geographic seniority area of his or her union, provided he or she possessed the necessary qualifications and complied with the bid procedure.]
. . . . .
In the event a full-time forty (40) hour job becomes available in a store in which no employee is on the bid list, the most senior employee on the bid list with the necessary qualifications within the geographical jurisdiction of the Union shall be offered the job before any employee is hired into said store.
Joint Statement ¶ 33.
139. Section 4.10.3 of the UFCW Contract provides:
Employees refusing an offer of full-time work, requesting part-time work after having been selected for full-time work, indicating their unavailability for continued full-time work or refusing a job opening with more hours shall not be entitled to exercise rights set forth above until the next bid period.
Joint Statement ¶ 34.
140. The UFCW Contract does not require that full-time clerks have different skills than part-time clerks. R.T. at 1-78 (Gill).
141. Herkal testified that the Union would not recognize any additional rules which Lucky could make regarding movement to full-time positions. R.T. at XX-XXXX-XX (Herkal). However, defendant has not presented sufficient evidence to convince the court that this is true; nor does Lucky consistently follow its established rules for awarding movement to full-time status.
B. Movement Practices
142. Most part-time employees work in the Grocery department. R.T. at 6-1019 (Hoffman).
*284 143. Opportunities for full-time work are not posted. R.T. at 1-127 (Gill). Grant did not remember anyone suggesting that full-time job opportunities should be posted. R.T. at 3-479 (Grant). Grant believes that because there are going away parties for departing employees, other employees in the store are aware when there was an opening for full-time work. R.T. at 3-378 (Grant). However, Grant conceded that as employees are sometimes replaced before they depart, an employee who learns about an opening at a going away party might learn too late to apply for the position. R.T. at 5-861 (Grant).
144. Only Journey Clerks are eligible to bid for movement from part-time to full-time work. Bid forms are available from the Union locals and from Lucky. The bids must be submitted during the first two full weeks in February or in August. Bids are then forwarded to Lucky's San Leandro office where they are compiled into lists in order of seniority, by department (Grocery, Produce and Non-Food), and store. R.T. at 21-3599 (Fumero); 13-2191 (Herkal). The compiled lists are then sent to the Union locals and the District Managers. R.T. at 2-205-06 (Gill).
145. Fumero generates a seniority list on a monthly basis, it lists all of Lucky's employees in the order specified by the UFCW Contract. These lists are distributed to district supervision and a copy is kept in the Payroll and Benefits office. R.T. at 21-3597 (Fumero).
146. Bids are only good for six months and there is no uniform Lucky policy for keeping old bid lists. R.T. at 1-132-33 (Gill). Despite the fact that E.E.O.C. charges had been pending against Lucky with respect to bids for movement to full-time status since 1985, Lucky did not keep all of its bid lists. R.T. at XX-XXXX-XX (Herkal). Prior to 1987, bid lists were only kept for the current year and the previous year. R.T. at 21-3616 (Fumero). However, since 1987 all of the bid list have been maintained. R.T. at 21-3601 (Fumero). Fumero believes that she has a complete set of the bid lists since 1984, but she has not reviewed them to see if they are complete. R.T. at 21-3617 (Fumero). She is certain that she has a complete set of the bid lists from the years 1987, 1988 and 1989. R.T. at XX-XXXX-XX (Fumero).
147. Fumero does not recognize requests for additional hours which were on bid forms. R.T. at XX-XXXX-XX (Fumero).
148. The selection of employees to move from part-time to full-time status is up to the discretion of the Store Manager. The Store Managers receive no training or instructions about how to make these decisions. However, Store Managers must follow seniority in moving employees to full-time. R.T. at 3-502, 5-755 (Grant); XX-XXXX-XXX (Hardy).
149. The District Manager makes the final determination about which employees should move to full-time status. The District Manager considers seniority, the needs of the store, the qualifications of the employee, and the constraints of the labor budget in making this decision. R.T. at 31-5192 (Hardy); X-XXXX-XX (Hoffman). Of the employees who bid for full-time work, the one in the store with the most overall seniority should be promoted to full-time. R.T. at 1-128 (Gill). The UFCW Contract does not define "qualifications" with respect to movement to full-time. However, Gill has interpreted "qualifications" to mean experience in the department. R.T. at 1-129-30 (Gill). Javier testified that seniority and qualifications must be taken into consideration in moving an employee to full-time status. R.T. at 10-1688 (Javier). However, in a December 2, 1987 letter to the E.E.O.C., Javier wrote "[t]he criteria considered, in addition to a person's seniority, when determining assignment to a full-time position is possession of necessary qualifications to perform the job, including experience, job performance, attendance, attitude, and availability to work the hours." Ex. A-113B, SP000076; R.T. at 10-1690 (Javier). Seniority is the tie breaking factor in making promotions to full-time. R.T. at 31-5199 (Hardy).
150. If there is no one on the bid list in the store that has the full-time opening, the most senior person in the geographical *285 area of the Union local should be offered the position. R.T. at 2-207 (Gill).
151. When employees bid for full-time work they are not required to specify whether they desire day or night shifts. However, most employees do specify. R.T. at 1-130 (Gill).
152. There is more full-time work available in Grocery and Night Crew. R.T. at 5-755 (Grant). Part-time employees in the Deli/Bakery and General Merchandise departments may bid for full-time Grocery positions. R.T. at 2-208 (Gill).
153. Lucky needs to employ a large number of part-time clerks in order to give the stores flexibility in scheduling hours. R.T. at 31-5190 (Hardy); 13-2190 (Herkal). Therefore, more employees want to be assigned to full-time work than Lucky can accommodate. R.T. at 5-753 (Grant). Some Lucky employees who have eight or nine years of part-time experience cannot get full-time status. R.T. at 5-754-55 (Grant).
154. Some employees are promoted to full-time status without having submitted a bid, even if other employees have submitted bids. R.T. at 1-131 (Gill); XX-XXXX-XX (Hardy). This happens principally when the bidders lack the experience necessary for the job, or when all the employees on the bid list turn down the offer for full-time status, or when a new store opens. R.T. at XX-XXXX-X (Hardy).
Allocation of Additional Hours/Step-Ups
A. The UFCW Contract Provisions on Allocation of Hours
155. Part-time employees may request additional hours of work. Section 4.10.2 of the UFCW Contract provides:
Part-time employees may request additional available hours within their classification on a store-by-store basis, provided they have the previously mentioned qualifications, are available for the hours, and have notified their Store Manager, in writing, of their desire for more hours, and they shall be afforded such hours by seniority.
Joint Statement ¶ 31.
156. This section of the UFCW Contract is the only written policy that addresses the allocation of additional hours. R.T. at 1-134 (Gill).
B. Allocation of Hours Practices
157. There is no bid policy for the allocation of additional hours. An employee may request additional hours at any time, and the request will not expire as long as the person is employed at the same store. R.T. at 1-135 (Gill); 6-1009 (Hoffman). Openings for additional hours are not posted; employees learn about the availability of additional hours by word of mouth. R.T. at 31-5187 (Hardy). There is a standard form to request additional hours, although requests are not always made in writing. R.T. at 2-202 (Gill). There is no centralized list of additional hours requests, nor are Store Managers required to keep a list of the requests made in their stores. R.T. at 1-136 (Gill). When he was a Journey Clerk, Foley would call around to other Lucky stores if he could not get any additional hours in his store. R.T. at X-XXXX-XX (Foley).
158. Store Managers do not receive any training or instruction from Lucky's upper management about procedures to follow for allocating additional hours, nor are there any documents that inform Store Managers about how to define qualifications, job performance or availability. R.T. at 10-1692 (Javier); 20-3478 (Michon). Decisions of Store Managers are not reviewed by upper management. R.T. at 6-1008 (Hoffman).
159. An employee desiring additional hours must be qualified and must have the ability to do the job to which they would be assigned. R.T. at 21-3543 (Michon). Michon did not follow a clear policy when determining who should receive additional hours. Sometimes he gave the additional hours to the first available person to ask for them. R.T. at XX-XXXX-XX (Michon). Although the UFCW Contract mandates that additional hours should be allocated based solely on seniority, R.T. at 2-202 (Gill); 5-887 (Hoffman); 20-3380 (Michon), in allocating additional hours Michon considered seniority and availability equally, *286 R.T. at 21-3541 (Michon). Grant testified that additional hours are to be allocated according to seniority and availability. R.T. at 5-758 (Grant).
160. More additional hours are available on the night shift than on the day shift. R.T. at 5-759 (Grant). Therefore, an employee who is willing to work Night Crew is more likely to get additional hours. R.T. at 6-1008 (Hoffman). Hardy testified that night hours are easier to obtain because they are not in as much demand as are early hours. R.T. at 31-5184 (Hardy).
161. If an employee disagrees with a Store Manager's allocation of additional hours s/he can bring a grievance. R.T. at 2-203 (Gill); 3-503 (Grant). Although, under UFCW Contract § 7.1, Lucky is required to post employee work schedules by the name of the employee on a weekly basis, R.T. at 2-204 (Gill), the Union does not receive a list of the employees who request additional hours and there is no way for one employee to know which other employees requested additional hours, R.T. at 1-138 (Gill).
162. However, Lucky employees often are aware of the seniority of their co-workers because the work schedules are listed in order of seniority. Senior employees sometimes complain if employees with less seniority receive more hours than they do. R.T. at X-XXXX-XX (Foley).
C. Step-Up Practices
163. Step-ups occur when one employee replaces a senior employee who is sick or on vacation. There is no bidding or posting system for step-ups. When an employee steps-up to a higher paid position, they are paid the wage rate of that position. R.T. at 1-87-88 (Gill).
164. The selection of employees to step-up to higher paid positions is entirely within the discretion of the Store Manager. The Store Manager is not required to consider seniority. R.T. at 1-88 (Gill). There are no written or unwritten criteria by which the Store Manager should decide whom to select for a step-up. In Grant's view step-ups should be awarded to the employee "who gets the job done the best and the fastest." Individual Store Managers have their own criteria for making this decision. R.T. at 3-377 (Grant).
165. Lucky does not keep records of step-ups, although they are considered to be on-the-job training. R.T. at 1-89 (Gill). Step-ups make an employee more qualified for management positions. R.T. at 3-378 (Grant).
166. The legality of the wage rates applicable to Lucky employees is not at issue in this suit. R.T. at 2-191-93.
167. Lucky negotiates wage rates as part of the industry multi-employer bargaining group, of which Safeway, Albertson's, Raley's, Savemart, Fry's and P & W Supermarkets are also members. Safeway is the largest employer in the group. UFCW contracts are renegotiated every three years. R.T. at 2-183-84 (Gill); XX-XXXX-XX (Herkal).
168. There are two separate wage systems at Lucky, Food (Grocery and Produce) and Non-Food (General Merchandise and Deli/Bakery). In general, Food department wage rates are higher than comparable Non-Food department wage rates. R.T. at 1-85 (Gill).
169. UFCW Contract § 1.3.2 states that Non-Food Clerks cannot work with food items. Were Lucky to disobey this section of the contract it would have to pay all of its clerks at the higher Food department wage rate. R.T. at 2-190 (Gill); 12-2162 (Herkal).
170. Courtesy Clerk, the entry-level position in the Grocery department, is the lowest paying job in any department. Ex. A-13; R.T. at 1-55 (Gill). Apprentice positions in the Deli/Bakery and General Merchandise departments, the entry-level positions in those departments, are paid at a lower wage rate than the Apprentice positions in the Grocery and Produce departments. R.T. at 1-61 (Gill).
171. Apprentice Clerks in both the Food and Non-Food departments receive increases in their wage rates based on the number of hours they have worked, in increments of 520 hours. However, the wage rates at *287 the different increments are lower for Apprentices in the Non-Food departments than for Apprentices in the Food departments. Ex. A-13, at D0011627-30; R.T. at 1-86-87 (Gill).
172. In 1989, when Apprentice Clerks in the Grocery and Produce departments reached the third step of their wage progression ($10.13 per hour), they were paid more than were Head Clerks in the Deli/Bakery and General Merchandise departments ($9.359 per hour). Ex. A-13; R.T. at 2-229-30 (Gill).
173. When an employee wishes to file a grievance, s/he must contact the Store Manager and the Union. The Union files a written grievance and requests a meeting with Lucky management. The grievance then goes to the Board of Adjustment, and if it is not resolved, it goes to arbitration. R.T. at 1-111-12 (Gill).
174. The UFCW Contract includes two grievance procedures. Section 18.2 covers disciplinary grievances, and section 18.3 covers interpretation or application disputes. Section 18.3 applies to grievances regarding initial placement, allocation of hours, and movement to full-time and positions. Ex. A-263 at 24; R.T. at 2-216 (Gill).
175. There is no systematic way in which the decisions of the Board of Adjustment or the results of arbitration are communicated to Store Managers other than the one involved in the dispute. R.T. at 1-113 (Gill). As most grievances are disciplinary, Gill believes there is no benefit to informing other Store Managers about the results. R.T. at 2-217 (Gill). However, Gill also does not inform Store Managers about the results of race and sex discrimination grievances. R.T. at 1-147 (Gill). Hardy does not always get copies of the written grievances that are filed in his district. Moreover, he does not always see grievances which relate to promotions in his district. R.T. at 32-5320 (Hardy).
176. Gill sees thirty-five to forty grievances a month. Only two or three of those grievances are interpretation or application disputes. Gill testified that he has never seen a grievance about initial placement; that he sees ten or twelve grievances a year about the allocation of hours (not on the basis of sex or race discrimination); that he sees ten grievances a year about movement to full-time (very few on the basis of sex or race discrimination); that he sees one or two grievances about promotion a month (very few on the basis of sex or race discrimination); and that he has seen one grievance regarding training. R.T. at 2-218-22 (Gill).
177. Gill did not remember counseling any employee about race or sex discrimination. R.T. at 1-152 (Gill).
178. Violations of the discrimination provision under UFCW Contract § 2.4.2 are subject to the grievance and arbitration procedure. Ex. B-263 § 2.4.3; 2-195 (Gill). This discrimination provision applies to hiring, promotion, movement to full-time and the allocation of additional hours. R.T. at 2-196 (Gill). A number of grievances have been pursued by the Union on behalf of women for breach of the UFCW Contract. R.T. at 2-258-59 (Gill).
179. There is an internal complaint procedure at Lucky for discrimination claims. Employees who feel they have been discriminated against can talk to their Store Manager, the District Manager, Virginia Javier or Mark Foley. Foley then conducts an investigation of the alleged discrimination. R.T. at 8-1330 (Foley); 11-1960 (Javier). Since September 1989, there have been between eighty and 100 internal discrimination complains, most of which involved allegations of sexual harassment. In that time period, there were no internal complaints regarding the allocation of hours, movement to full-time, or initial placement. However, there were two or three internal complaints regarding promotion. R.T. at 8-1333 (Foley). As a result of these complaints Foley issued verbal warnings, written warnings, fifteen to twenty suspensions, five or six demotions, and seven or eight terminations. R.T. at X-XXXX-XX (Foley). Javier testified that she saw about three internal complaints regarding promotion decisions between 1986 and 1989. R.T. at 11-1961 (Javier).
*288 180. Since September 1989, Foley has seen about thirty E.E.O.C. complaints, one regarding allocation of hours, one regarding movement to full-time, three or four regarding promotion, and none regarding initial placement. R.T. at 8-1334 (Foley). Between 1986 and 1989, Javier saw seventy E.E.O.C. complaints based on race or sex discrimination, fifteen of which related to promotion, five related to movement to full-time, and five related to the allocation of additional hours. R.T. at 11-1963 (Javier).
Comparison of Lucky's NCD and SCD
181. There are 240 stores in Lucky's Southern California Division. 48% of the employees in the SCD are women. R.T. at 21-3555, 21-3579 (Frazier).
182. The Management and Human Resource departments in Lucky's northern and southern divisions are completely autonomous. R.T. at 21-3556 (Frazier); 10-1698 (Javier). Each division has its own policies with respect to promotion, training, work assignment, allocation of additional hours, and movement to full-time status. Each division has its own affirmative action policy and is solely responsible for implementing it. R.T. at XX-XXXX-XX (Javier).
183. When employees are hired in the SCD they designate themselves as either restricted or as available. Employees designating themselves as available are willing to move to full-time work or to accept additional hours. Employees have the opportunity to change their status every six months. This creates a two-tiered seniority list where any employee on the available list functionally has more seniority than the most senior person on the restricted list. R.T. at 21-3548, XX-XXXX-XX (Frazier); XX-XXXX-XX (Herkal).
184. Deli/Bakery employees are classified in the General Merchandise department in the SCD. The SCD classification of Clerk Helper is the equivalent of Courtesy Clerk in the NCD. R.T. at 21-3559 (Frazier); 13-2264 (Herkal). There is a training program for Clerk Helpers in the SCD which prepares them to move to Apprentice positions. R.T. at 21-3560 (Frazier); 13-2264 (Herkal).
185. The UFCW Contract in the SCD specifically provides that Apprentices in General Merchandise may transfer into a Food department while retaining their seniority and without suffering a reduction in pay. Many employees in the SCD follow the job progression from Clerk Helper, to General Merchandise Clerk, to Food Clerk. R.T. at XX-XXXX-XX (Frazier).
186. Journey Clerks are eligible for the ELMT program in the SCD. R.T. at 21-3564 (Frazier).
187. There has never been an affirmative action program in the SCD. However, the consent decree in Ballesteros v. Lucky Stores sets goals and timetables for the promotion of women in the SCD to entry-level management positions. Ex. A-115B; R.T. at XX-XXXX-XX (Frazier).
Lucky's Affirmative Action Efforts
A. Before 1986
188. An affirmative action equal employment opportunity policy was first adopted by Lucky Corporate in 1975. Herkal was responsible for giving advice and preparing statistical reports in connection with the affirmative action plan in the NCD. Exs. A-32, A-66; R.T. at XX-XXXX-XX (Herkal). As of 1983, Herkal believed Lucky was doing all that it needed to do with respect to affirmative action. R.T. at 13-2209 (Herkal).
189. In late 1983, the only data Lucky had on race was based on a visual survey of its employees. At that time Noriega-Ailor was given the task of compiling gender and race data for the NCD. R.T. at X-XXXX-XX (Noriega-Ailor).
190. Noriega-Ailor's December 6, 1983 memorandum to Robert Grant, Jim Scoggins, Walter Herkal and Robert Gill reported that women were 4.2% of management in the NCD. The memorandum also reported that according to the State of California Civilian Labor Force statistics from the 1980 Census, 34.4% of women were available for management positions. Ex. A-79; R.T. at X-XXXX-XX (Noriega-Ailor).
191. Noriega-Ailor's April 16, 1984 memorandum to Jack Hoover, Jim Scoggins, Walter Herkal and Robert Gill stated *289 that women were 4% of management in the NCD. The memorandum reported that out of 1133 managers in NCD, forty-five were women. Noriega-Ailor believed that the decrease by two in the number of women in management positions, might have been due to the fact that Lucky had reclassified some job titles. Noriega-Ailor discussed these statistics with Herkal, Hoover, and Scoggins. Ex. A-83; R.T. at X-XXXX-XX (Noriega-Ailor).
192. A memorandum from Ray Sweeny to Richard Goodspeed, Jack Hoover, and Rosalind Speaker Thompson, dated August 24, 1984, shows that the NCD had the lowest percentage of women in management of the nine Lucky divisions. The report directed the NCD to review its affirmative action policy and to "reemphasize the importance of compliance." Ex. A-67; R.T. at 8-1421 (Noriega-Ailor). Noriega-Ailor discussed these statistics with Grant, Hoover, Herkal, and Scoggins, and suggested that Lucky formalize its job descriptions because they would help management focus on employee skills in making promotion decisions. R.T. at X-XXXX-XX (Noriega-Ailor).
193. Noriega-Ailor understood that Walter Herkal was responsible for meeting Lucky's affirmative action goals in 1984. R.T. at 8-1429 (Noriega-Ailor).
194. In 1985, the number of women in management in the NCD increased from forty-nine to fifty-eight. Ex. A-84; R.T. at 8-1430 (Noriega-Ailor). At that time, almost all of the Deli/Bakery Department Heads and 60% of the General Merchandise Department Heads were women. R.T. at X-XXXX-XX (Noriega-Ailor). Noriega-Ailor felt that the perception at Lucky was that the depth of knowledge and the scope of responsibility required was lesser for the Non-Food Department Head jobs than for the Food Department Head jobs. R.T. at 8-1438 (Noriega-Ailor). In addition, she has found that when jobs are heavily concentrated with one group of people, changing the composition of people in the jobs generally improves movement in and out of those jobs. R.T. at 8-1440 (Noriega-Ailor).
195. Grant was aware that there were more women than men in the Deli/Bakery, General Merchandise and Floral departments; more men than women in positions above Fourth Person; and that more women were Checkers and more men were Stockers. R.T. at 5-782 (Grant). He did not consider these statistics to be problematic because he had not received any complaints about discrimination in the retail stores before 1986. R.T. at 3-537, 5-785 (Grant). At that time, Grant did not believe that the existence of a statistical disparity between the number of men and women in management was an indication that Lucky was discriminating against women. R.T. at 5-786 (Grant).
196. Following the settlement of the Bockman case in 1985-1986, see supra n. 4, Lucky instituted hiring goals in the meat and Produce departments. Herkal informed Grant that Lucky must try to improve the number of women who were promoted. R.T. at 3-533, 5-789 (Grant). At that time, Lucky's General Management Committee began discussing affirmative action. Thompson was put in charge of the development of an affirmative action plan which she was told to "make simple." Thompson Depo. at 176-177.
197. Under the Bockman consent decree, Javier was required to develop an applicant tracking program for all of the class positions under the decree, and to ensure that the job posting requirements were followed. She also was required to conduct outreach recruitment for women. Ex. A-114B; R.T. at XX-XXXX-XX (Javier).
198. Lucky was involved in an industry-wide E.E.O.C. complaint in the early 1980s which alleged that women were segregated into the lower paying jobs in the meat department. R.T. at 13-2282 (Herkal).
199. Lucky has had a sexual harassment policy since at least 1986. Ex. B-289; R.T. at 19-3209 (Goodspeed).
B. The 1986 NCD Affirmative Action Plan
200. Other than Lucky's affirmative action policy statement, Grant had never received *290 materials about affirmative action before 1986. R.T. at 3-536 (Grant).
201. Thompson designed and led a project to institute affirmative action plans in all of Lucky's divisions in 1986. She gave the divisions general outlines of what their plans should look like and designed a progress reporting structure. Thompson Depo. at 36.
202. Goodspeed approved Lucky's 1986 affirmative action plan for the NCD as soon as he became president of the NCD because it was a "good business practice." Ex. A-35; R.T. at 18-3076, 18-3080 (Goodspeed). At that time he agreed that there was a problem with the underrepresentation of women in management at Lucky. R.T. at 18-3088, 18-3095 (Goodspeed). Grant agreed that the number of women in management and Produce should be increased, Hoffman was aware that women were promoted less than men, and Herkal felt that Lucky had to make improvements in affirmative action. R.T. at 3-545, 3-553 (Grant); 13-2281 (Herkal); 6-1058 (Hoffman). Grant believes that Scoggins and Martin were also aware of the problem. R.T. at 3-554 (Grant). Javier felt that Lucky's senior management had been aware of the underutilization of women since 1986. R.T. at 10-1806 (Javier). In fact, both Gill and Hoffman had received complaints from the Union about discriminatory promotion decisions and failure to move women to full-time positions. Gill informed Grant of the complaints and requested that they meet to discuss the matter. R.T. at 1-153-57 (Gill); 6-926-27 (Hoffman).
203. Although Goodspeed had overall responsibility for the affirmative action plan, Tim Martin was in charge of its implementation. R.T. at 18-3083, 18-3184 (Goodspeed). Javier administered the affirmative action plan, monitored compliance, and compiled and reviewed statistical data from the plan. R.T. at 10-1703 (Javier).
204. The 1986 affirmative action plan for the NCD acknowledged that equal employment opportunity is legally required, but also recognized that it is "sound for moral, social, and economic reasons." Ex. A-35 at D0012509. The plan discussed the underutilization of women in management, stated that promotion and training was necessary to better the distribution of women in Lucky's workforce, and set short, intermediate and long-term goals to accomplish this objective. Ex. A-35 at D0012569-72; R.T. at X-XXXX-XX (Noriega-Ailor). The plan also included a supervisory training program and a reporting requirement for all promotions and transfers. Ex. A-35 at D0012578-79, D0012584; R.T. at X-XXXX-XX (Noriega-Ailor).
205. Javier and Noriega-Ailor found Lucky's senior management to be supportive when they presented the plan to them. R.T. at 11-1850 (Javier); 8-1449 (Noriega-Ailor). Noriega-Ailor had "a very strong feeling from senior management that they were interested in having women advance within the workforce, and they were interested in [her] ideas with regards to [women's] advancement and the probable training programs that might be necessary to assist with that." R.T. at 9-1587 (Noriega-Ailor). However, one reason Noriega-Ailor left Lucky was that she did not feel that senior management was ready to have "women in certain positions." R.T. at 9-1590 (Noriega-Ailor).
206. On March 14, 1986, Noriega-Ailor restructured the Affirmative Action Steering Committee into two parts: the Affirmative Action Executive Committee, comprised of Goodspeed, Herkal, Martin, Fahey, Cundiff, herself and Javier, and the Affirmative Action Implementation Committee, which was to develop programs that would further affirmative action. Ex. A-88; R.T. at XX-XXXX-XX (Goodspeed); 9-1574, X-XXXX-XX (Noriega-Ailor). Despite his position as Vice President, Industrial Relations for the NCD, Herkal does not remember serving on the Affirmative Action Executive Committee. R.T. at 12-2058 (Herkal).
207. Goodspeed thought Lucky had a problem identifying people to move into management positions. The strategic plan for the NCD, dated June 12, 1986 stated that Lucky does not "have a formal system to identify and monitor existing employee *291 talent." Ex. A-133 at 25; R.T. at 18-3057 (Goodspeed).
208. On February 9, 1987, Goodspeed personally attended a meeting of the Affirmative Action Implementation Committee in an effort to improve the attendance of managers. Ex. A-94; R.T. at XX-XXXX-XX (Goodspeed).
209. Lucky management claims it was distracted from their affirmative action efforts by the attempted take-over of the company by Asher Edelman and the liquidation of the Gemco stores in 1986, and by American Stores' merger with Lucky and a three week Teamster strike in 1988. These events negatively affected the position of women at Lucky because many employees were forced to step back on the promotional ladder. R.T. at 18-3152 (Goodspeed); 5-791-96 (Grant); XX-XXXX-XX (Herkal); Thompson Depo. at 73, 190-194. The downsizing of the Lucky workforce in 1986 diverted the time and attention of Lucky management away from implementation of the affirmative action plan. R.T. at 9-1618 (Noriega-Ailor). However, no women stepped down from Store Manager in 1988 or 1989, and no women stepped down from Assistant Store Manager in 1988. Ex. A-56; R.T. at 6-932 (Hoffman). In addition, between 1986 and 1989, the RMDP and ELMT were implemented, stores were opened and remodelled, and new and expensive policies, such as "three's a crowd," were implemented, despite the distraction caused by the attempted take-over and the merger. R.T. at 5-857-59 (Grant). Therefore, the court finds that Lucky's stated reasons do not sufficiently explain Lucky's failure to vigorously pursue or to maintain its affirmative action policy.
C. Discontinuation of the 1986 Affirmative Action Plan
210. Javier testified that she, Herkal, Noriega-Ailor and Martin discontinued Lucky's 1986 affirmative action plan in 1989, shortly after this suit was filed, on the advice of Donahue, Gallagher, Thomas & Wood, Lucky's house counsel. They did not announce their decision to Lucky's senior management. R.T. at 12-2043 (Herkal); XX-XXXX-XX, XX-XXXX-XX (Javier). Javier made the ultimate decision to cancel the affirmative action program. She did not discuss the idea with her supervisors. R.T. at XX-XXXX-XX, XX-XXXX-XX (Javier). She continued to track the numbers of women in management, but did not report the results to the managers. R.T. at 11-1988 (Javier). Since the discontinuation of the affirmative action plan, no reference to the plan has been made in meetings between the Human Resources Department and Lucky management. R.T. at 10-1742 (Javier). Foley also was aware that the 1986 affirmative action plan was no longer in effect. R.T. at 8-1306, 8-1308 (Foley).
211. Grant believed that the 1986 affirmative action plan was still in effect because he had not been told anything to the contrary. He did not know if any of the action-oriented programs were still being executed. However, he believed that the Affirmative Action Implementation Committee was still active. R.T. at 3-524-32, 5-797 (Grant). President Goodspeed never had any discussions about terminating the 1986 affirmative action plan, and considered the plan still to be in effect. R.T. at XX-XXXX-XX (Goodspeed). Herkal testified that he believed that the 1986 affirmative action plan is still in effect because "it is a living document." R.T. at 12-2044, XX-XXXX-XX (Herkal).
D. Lucky's 1988 Affirmative Action Efforts
212. In 1988 Grant felt that Lucky was doing its best with respect to the promotion of women. R.T. at 4-589-90 (Grant). He does not remember seeing any affirmative *292 action progress reports before 1988. R.T. at 3-522 (Grant).
213. Grant asked Herkal to collect the names of men and women who had asked for and who had rejected advancement opportunities between January 1, 1982 and May 31, 1987. Exs. A-59 and A-60; R.T. at 6-909 (Hoffman); XX-XXXX-XX (Javier). Grant had previously asked for information about how many women had turned down promotions, but not about how many men had done the same. R.T. at 3-561 (Grant). A substantial number of women indicated that they were interested in advancement. Ex. A-60; R.T. at 6-910, 6-950-56 (Hoffman). Based on this information, Javier put together a report entitled, "Women in Management," dated April 29, 1988, which was sent to Goodspeed, Herkal, and Martin. The report sought to determine Lucky's status with regard to women in management. Ex. A-105; Thompson Depo. at 51. The report showed that there were no women District Managers, twenty-four male and four female Merchandisers, 150 male and ten female Store Managers, and 152 male and twelve female Assistant Store Managers. Ex. A-105 at D0097148; R.T. at XX-XXXX-XX (Javier).
214. Lucky has never conducted a survey to see if women were interested in moving from the Deli/Bakery and General Merchandise departments to the Grocery and Produce departments. R.T. at 10-1677 (Javier). Nor has Lucky surveyed the interest levels of women in the different departments, or in stocking versus checking. R.T. at 1-152-53 (Gill); 3-559 (Grant).
215. One reason the NCD adopted the Specter goals for promotion of women into Fourth Person and Third Person in 1988 was to limit Lucky's liability in this case. R.T. at 18-3139, 19-3210 (Goodspeed).
216. Specter conducted two training programs with groups of Lucky District Managers in 1988. He also had a separate set of meetings with Store Managers, Assistant Store Managers, and Merchandisers and in NCD. His intent was to meet with everyone at Lucky who had decision making authority. R.T. at X-XXXX-XX (Specter). Specter said that the purpose of the meeting was "to scare the pants off of them" and to get Lucky to support affirmative action. R.T. at 7-1162 (Specter).
217. Prior to the Specter meetings Grant issued a memorandum to the District Managers which urged them to give their "full commitment and support" to the program. Ex. A-52; R.T. at 7-1157 (Specter). At the meetings the "Plan for Success" and a copy of Lucky's sexual harassment policy were distributed. Ex. A-53; R.T. at X-XXXX-XX (Specter). Among other things, Specter discussed the legal standards for sexual harassment, the liability standard under Griggs v. Duke Power, 401 U.S. 424, 91 S. Ct. 849, 28 L. Ed. 2d 158 (1971), for statistical disparities in a workforce, and the evaluation of qualifications in making promotion and hiring decisions. R.T. at 7-1182 (Specter). Specter also told the Lucky District Managers that the statistical underrepresentation of women in management might open them to liability for sex discrimination. R.T. at 7-1173 (Specter). Only three of the 364 District Managers, Store Managers, Assistant Store Managers, and Merchandisers in NCD did not attend the Specter meetings. Ex. A-124; R.T. at 5-812 (Grant).
E. Notes from the Specter Meetings
218. Javier took notes at the May 20, 1988 meeting between Specter and Lucky's District Managers. Ex. A-119; R.T. at XX-XXXX-XX (Javier).
219. Hoffman took notes at the June 14 and June 22, 1988 meetings between Specter and the Store Managers, Assistant Store Manager and Merchandisers. Ex. A-62; R.T. at 6-921 (Hoffman). Specter requested that each person at the meetings volunteer a stereotype that they had heard in the workplace. In response to this question Hoffman remembered hearing that "women won't work late shifts because their husbands won't let them;" "the crew won't work for a Black female;" "women are better with customers than men are;" and "women need training and an opportunity to do floor work." R.T. at 6-923-24 (Hoffman). Hoffman did not think that the Store Managers believed the stereotypes *293 they reported, but he thought that they may have heard some of them from other Lucky employees. R.T. at X-XXXX-XX (Hoffman).
220. Foley took notes during Specter's meetings with Store Managers, Assistant Store Managers and Merchandisers between June 7 and July 13, 1988. The Manager's comments which he recorded included the following: "Women do not want to work late shifts;" "How does your crew get used to working with a woman boss?;" "Women are not encouraged to be promoted;" "[You] get used to someone fitting the `picture' and the `picture' is a White male;" "Women's income is the second income in a household;" "Customers might object to seeing a woman in management;" "Women are afraid to work at night;" "Men need to change there [sic] attitudes;" "Women seem to step down a lot after being promoted;" "Women don't go into [management] because they are not accepted. Have to show them they are wanted in [management];" "Women have the fear they won't be able to handle the promotion. The crew will not accept them;" "Some men do not want to work side by side with women in Produce. Need to change attitudes;" "Resentment from males when female is promoted. Promoted women to 5th Person, taking a lot of heat from men;" "Impossible, can not find any qualified women;" "Women have a hard time managing other women;" "Women work neater, strive to so a better job. Men do not want the competition from women;" "Hard to change old habits, still see the `success model' as a White male;" "Women/minorities will turn out to be are [sic] best [managers]. `Team Players.' Black females are aggressive, they would make great [managers];" "The workforce would not perform for a Black/female [manager];" "We have old timers who are set in there [sic] ways. Need to change attitudes;" "Has [sic] promoted women who step down soon after because the only reason they did it was to become full time journey person;" "Females [who] step down `glow' because there are not many in [management] so when it happens everyone is aware of it. On the other hand when a man steps down it is not that big a deal because we have so many men waiting in the wings;" "Two classes of women 1) old school, been around for [years], they don't see the opportunities and don't want them, 2) new school, just starting out in Lucky and want the opportunities. Need to open the door to [management] for this group;" "Women don't have as much drive to get ahead. Women are not the bread winners;" "Women cry more. Women should be stronger;" and "Women are considered the weak sex. Has men do the hard, physical work." Ex. A-121; R.T. at X-XXXX-XX, X-XXXX-XX (Foley). See also Ex. A-122; R.T. at XX-XXXX-XX; R.T. at X-XXXX-XXX (Specter).
221. Foley felt that the comments were representative of the beliefs of some of the Managers, that others were provoked into making the statements, and that some were repeating comments that they had heard other Lucky employees make. R.T. at 7-1260 (Foley). Javier could not distinguish which comments reflected the true beliefs of the Managers who said them. R.T. at 11-1987 (Javier). Nonetheless, the court finds that these notes reflect the views of at least some of Lucky's Store Managers or the views expressed by Lucky employees who the participants in the meetings overheard.
222. Most of the participants in the Specter meetings felt that the program was overdue and that management must be involved in affirmative action. Exs. A-123B and A-124; R.T. at X-XXXX-XX (Foley).
223. Specter conducted two "Generic Success Model Meetings" in August and September of 1988 to develop selection criteria for Third and Fourth Persons. R.T. at X-XXXX-XX (Specter). Both Foley and Javier took notes at these meetings. Exs. A-126 and A-127; R.T. at 8-1295 (Foley); 10-1802 (Javier).
224. In 1989 the Union rejected Lucky's proposal that the following section be added to section 4.3.1. of the UFCW Contracts.
Notwithstanding anything contained in this section to the contrary, seniority shall have no application where promotions *294 are effectuated in order to promote affirmative action goals and/or principles.
Ex. B-253; R.T. at XX-XXXX-XX (Herkal).
F. Discontinuation of the 1988 Affirmative Action Efforts
225. After August 1988, on the advice of Lucky's counsel, there were no more meetings between Specter and Lucky. R.T. at 10-1799 (Javier). Specter testified that his "program was aborted." R.T. at 7-1226 (Specter). However, Herkal was not aware that the 1988 Specter goals had been eliminated, R.T. at XX-XXXX-XX (Herkal), and as far as Grant knew, the Specter goals had not been changed, R.T. at 4-580 (Grant).
G. Attitudes of Senior Management
226. Before 1986, Herkal believed that the distribution of women and men at Lucky reflected "the interest level of the people and that anybody who wanted to get promoted ultimately would." He did not see any reason to make changes at Lucky to increase the pool of women who would be eligible for promotion. R.T. at 12-2105 (Herkal). However, he admitted that prior to the Bockman litigation, he had believed that the absence of women in warehouse jobs reflected their lack of interest in those jobs and he was surprised to discover that women were interested in and had applied for those jobs. R.T. at XX-XXXX-XX (Herkal). Herkal disagrees with the idea of affirmative action and "accelerated promotion programs" because after the "obvious candidates" and the qualified candidates are promoted, unqualified candidates must be promoted. R.T. at XX-XXXX-XX (Herkal).
227. Gill claimed never to have heard anyone at Lucky say that women were less interested in promotion than men, or that women turned down promotions more than men. R.T. at 1-152-53 (Gill).
228. Grant said that when he was a Store Manager in the 1960s he once put a woman to work as a Receiving Clerk unloading trucks, and she almost filed a grievance against him. R.T. at 2-313 (Grant). At that time, Grant found that women were more interested in cash register work, and that men were more interested in floor work because men viewed that work as "more important." R.T. at 3-568 (Grant). He does not have any reason to believe that women's interests are any different than men's today. R.T. at 2-313-14 (Grant). Grant testified that women step down from management positions at a higher rate than do men. However, he has never reviewed the available statistics on step-downs and he admitted that his opinion was based on personal belief and his experience as a Store Manager thirty years ago. R.T. at 3-567-68 (Grant).
229. In 1989 Goodspeed knew that ten of the 150 Store Managers, and that twelve of the 152 Assistant Store Managers in the NCD were women. R.T. at 18-3092 (Goodspeed). However, he did not believe that women were less interested than men in management positions. R.T. at 18-3110 (Goodspeed). Goodspeed does not think that the small number of women in management at Lucky is the result of purposeful discrimination; but due to "lack of communication, lack of understanding." R.T. at 19-3225, 19-3248 (Goodspeed).
III. EXPERT WITNESSES
230. Dr. Richard Drogin received a Ph.D. in statistics from the University of California, Berkeley in 1970. He is currently a professor of statistics at California State University, Hayward. He teaches courses in sampling techniques, data analysis, development of statistical software, probability theory, stochastic processes, and simulation methods. He is also a partner in the statistical consulting firm of Drogin, Kakigi & Associates. Much of his consulting is in the area of employment discrimination, and he has testified as an expert witness in statistics and probability. Ex. A-10 at 1 (Drogin Report); R.T. at XX-XXXX-XX.
231. Dr. Drogin has previously been accepted as an expert in federal court in about twenty cases. He has been retained by the E.E.O.C., the Office of Civil Rights, *295 the California Department of Fair Employment and Housing, and various legal services organizations. R.T. at 15-2575.
232. This court finds that Dr. Drogin is qualified to testify as an expert in the areas of statistics, probability and computer analysis. R.T. at 15-2576.
233. Dr. Drogin was hired by plaintiffs' counsel to review the computerized personnel and employment records provided by defendant through discovery and to analyze the data. R.T. at XX-XXXX-XX. He examined the initial assignment of new hires, promotional patterns at several levels of the workforce, and the distribution of hours. Ex. A-10 at 1.
234. Dr. Drogin's analysis was based primarily on Lucky's quarterly payroll data. The Lucky payroll tapes provided to plaintiffs in discovery are authentic business records pursuant to Fed.R.Evid. 803(6), 901. Joint Statement ¶ 51. The computerized payroll tapes contain personnel and payroll data for all non-administrative employees working in the NCD. Non-administrative data is available for the period from 1979 to the second quarter of 1990. For 1979 and 1980, only year-end data is available. Quarterly data is available since the second quarter of 1981, except that second quarter 1987 data is missing. Administrative data is available only since the second quarter of 1987. Ex. A-10 at 7; R.T. at XX-XXXX-XX. Dr. Drogin modified the original data on the payroll tapes to account for known errors in certain fields. Ex. A-10 at 8; R.T. at 15-2591.
235. The payroll tapes give a snapshot of each employee's status at the date of the tape by: social security number, name, sex, position held, store, date of hire, termination date (if any), and assorted information on hours worked and income. Ex. A-10 at 7.
236. By comparing a person's status on the payroll tapes in consecutive quarters it is possible to identify promotions and other job movements. However, if a person makes two or more job changes during a quarter, the data will only show a change from the first job held at the end of the prior quarter to the job held at quarter end. Ex. A-10 at 7.
237. Jobs were identified on the computerized payroll tapes by department and job classification codes. In some cases there were several codes for the same job. Therefore, Dr. Drogin created twenty-six job groups. Each job group was specified in terms of a set of department and job classification codes. Most job groups consisted of one job classification in one department, although in a few cases a few job classifications were grouped together to form one job group. The job groups were based upon information provided by defendant. Ex. A-10 at 8; R.T. at 15-2583, 15-2585. Dr. Drogin excluded data that did not fit into his job groups. He found that the computer entries he excluded accounted for no more than .5% of any department. R.T. at XX-XXXX-XX.
238. Defendant concedes that Dr. Drogin's coding of applications was consistent with Dr. Haworth's coding. Lucky's Opening Post-Trial Brief at 8 n. 6 (citing R.T. at 34-5701, XX-XXXX-XX, XX-XXXX-XX).
239. Until the fourth quarter of 1989, there was no way to distinguish among the positions of Fourth Person, Fifth Person, Receiving Clerk, Night Clerk, and Front-End Clerk on Lucky's computerized payroll tapes. In the fourth quarter of 1989 Lucky introduced new payroll job classification codes for certain retail store positions. Joint Statement ¶ 48; Ex. A-10 at 8; R.T. at X-XXXX-XX (Foley). Because of this confusion, Dr. Drogin's promotion analyses do not go beyond the third quarter of 1989. Ex. A-10 at 9.
240. Dr. Drogin also worked from computerized bid lists of employees bidding for full-time positions, going back to February 1985. However, that bid list data is incomplete. Ex. A-10 at 7. Lucky did not maintain a complete set of all full-time bid lists created or bids made during the liability period. Joint Statement ¶ 50.
241. Descriptive statistics simply report or summarize counts, percentages, or averages based upon a given data set. Dr. Drogin used inferential statistics in an attempt to measure whether certain disparities, *296 which are calculated from descriptive statistics, are within the range of what might be expected if the data were generated according to a certain method or procedure. Ex. A-10 at 10.
242. Dr. Drogin used a binomial, or two-tailed test. R.T. at 15-2608. Using a two-tailed test, the significance probability is the likelihood that an observed disparity will occur by random fluctuation. The significance probability can be expressed in terms of a Z-value, which gives the size of the disparity in terms of standard deviations. A Z-value of (+ or -) 1.96, or approximately two standard deviations, indicates that a disparity has one chance in twenty of occurring by random fluctuation (5% probability of occurrence). A Z-value of (+ or -) 3.08, or approximately three standard deviations, indicates that a disparity has one chance in 500 of occurring by random fluctuation (.2% probability of occurrence). Ex. A-10 at 10-11; R.T. at XX-XXXX-XX. Dr. Drogin considered a Z-value of (+ or -) 1.96 to be statistically significant. Ex. A-10 at 11; R.T. at 15-2609.
243. Dr. Drogin used the statistical method of regression analysis, controlling for the effects of a variable by disaggregating the data according to the values of a variable. He divided the data into subgroups and obtained the level of disparity between the different subgroups. He then accumulated the disparities across all of the subgroups to obtain an overall measure of disparity. Ex. A-10 at 12-13.
244. Dr. Drogin compared the movements of men and women from the payroll tapes to a model of availability to determine if there was a significant difference from the model. R.T. at 15-2608.
245. Dr. Drogin also took into account the historical pattern of movement between jobs at Lucky by constructing feeder pools for each target job. Ex. A-10 at 13.
A. Initial Placement
246. Nearly all new Lucky employees are hired as Courtesy Clerks or Apprentices in Grocery, or as Apprentices in Produce, Deli/Bakery, or General Merchandise. Ex. A-10 at 2.
247. The overall percentage of women among new hires from 1984 to 1989 was 46.6%. R.T. at 17-2963.
248. Between the second quarter of 1984 and the end of 1989, Dr. Drogin found that women were 35% of the new hires into Grocery and Produce jobs and 84% of the new hires into Deli/Bakery and General Merchandise jobs. These figures are statistically significant when compared to expected rates, with a Z-value of 61.71 (probability of less that 1 in 10 100). Ex. A-10 at 14, Tab 5; R.T. at XX-XXXX-XX.
249. Dr. Drogin compared the movement of male and female employees out of Courtesy Clerk positions and into Apprentice positions in the different departments. He found that, controlling for store and year, women comprised 31% of those promoted into Apprentice jobs in Grocery and Produce, but 75% of those promoted into Apprentice jobs in Deli/Bakery and General Merchandise. This disparity is statistically significant, with a Z-value of - 15.13 (probability of less that 1 in 1050). Ex. A-10 at 16; R.T. at 15-2620.
250. For the second quarter of 1984 through the third quarter of 1989, Dr. Drogin found statistically significant short-falls, in excess of three standard deviations, in promotions of women into positions as Apprentice Produce Clerk from all positions, Z-value: - 5.46; into part-time Journey Produce Clerk, Z-value: - 2.77; into Apprentice Night Crew, Z-value: - 18.00; and into part-time Journey Night Crew, Z-value: - 19.98. Ex. A-10 at 20.
*297 251. For the second quarter of 1984 through the third quarter of 1989, Dr. Drogin found statistically significant shortfalls, in excess of three standard deviations, in promotions of women into Department Head/Receiving Clerk, Z-value: - 6.93; and into Night Crew Manager, Z-value: - 4.01. Ex. A-10 at 21-22.
252. Dr. Drogin did not find statistical disparities for the liability period in promotions of women to Third Person, Z-value: - 2.58; Assistant Store Manager or Store Manager. R.T. at 15-2655.
C. Movement from Part-Time to Full-Time
253. Dr. Drogin found that since 1984, women comprise approximately 55% of part-time Journey Day Clerks in all departments. However, women are only 30% of full-time Journey Day Clerks. Ex. A-10 at 17.
254. Comparing men's and women's moves from part-time to full-time Journey Day positions, Dr. Drogin found that women received fifty-three fewer moves than would be expected. This disparity is statistically significant, with a Z-value of - 6.75. Ex. A-10 at 17-18; R.T. at XX-XXXX-XX. The results by year from 1987 to 1989 are all statistically significant. R.T. at XX-XXXX-XX. The disparity between the movement of men and women into full-time night jobs is also statistically significant, with a Z-value of - 15.83. R.T. at 15-2633.
255. Dr. Drogin's analysis included day and night part-time clerks because under the UFCW Contract they are equally eligible for promotion to full-time jobs. Ex. A-10 at 18. However, when workers are promoted, day workers generally stay on the day shift and night workers generally stay on the night shift. Ex. A-10 at 3.
D. Movement off the Bid Lists
256. Dr. Drogin confirmed Lucky's practice of awarding full-time positions not taken from the bid list. He also determined that many bid lists were missing; particularly the bid lists for 1984. Since 1985, only half of the movements to full-time can be matched up with a bid. Ex. A-10 at 17.
257. There is no information on promotions in the bid list data, and there is no information on bid lists in the payroll data. Therefore, there is no way to match a bid to each promotion and to be certain that any particular promotion was off the bid list. Dr. Drogin assumed that a person was promoted off the bid list if they were promoted during consecutive quarters, and they were also on a bid list during the corresponding time period. Likewise, if a person was not on an available bid list during the time corresponding to their promotion, Dr. Drogin assumed that they were not promoted off the bid list. Ex. A-10 at 19; R.T. at XX-XXXX-XX.
258. Dr. Drogin compared movements from part-time to full-time taken from the bid list with movements not taken from the bid list. He found that when promotions are taken from the bid list, women are moved up to full-time at a rate consistent with their representation on the bid list (women comprise 52.8% of those taken from a bid list, and 50-55% of bidders), Z-value of - 3.21. However, women receive significantly fewer moves from part-time to full-time among moves not taken from the bid list than would be expected from their representation among the "feeder pool" of part-time clerks (women comprise 31.3% of those promoted who are not taken from a bid list), Z-value of - 6.59. Ex. A-10 at 19, Table 21; R.T. at XX-XXXX-XX.
259. Dr. Drogin performed the same analysis only counting second and fourth quarter moves off the bid list because these promotions can be directly traced to the bid list. The results of this analysis confirmed Dr. Drogin's other results. Ex. 172; R.T. at XX-XXXX-XX.
E. Allocation of Hours
260. Dr. Drogin analyzed the distribution of hours for men and women in Journey Day Clerk and Grocery Apprentice positions, *298 using the year-to-date regular hours field on the payroll tapes. This field contains all hours worked and paid for during the year, including hours worked at the regular rate, overtime hours, and alternate hours. It also includes hours not worked but credited to an employee for pay purposes, such as vacation pay. Ex. A-10 at 23.
261. Dr. Drogin determined that there were statistically significant disparities in the distribution of hours to women and men. Ex. A-10 at 23, Table 19. These disparities occurred with respect to women in part-time Journey Day Clerk, Z-value: - 18.16; full-time Journey Day Clerk, Z-value: - 8.77; and Apprentice Day Grocery, Z-value: - 3.36. Ex. A-10 at 23-24; R.T. at XX-XXXX-XX.
262. Dr. Drogin did not find statistically significant disparities in the allocation of hours to females in Courtesy Clerk positions. Ex. B-407; R.T. at 17-2882. Dr. Drogin did not analyze the allocation of hours in Utility Clerk positions because there are very few women in those positions. R.T. at 17-2882.
263. Dr. Drogin also found a significant disparity in the assignment of alternate or "step-up" hours to women. Controlling for year, store and seniority, he found that women received fewer alternate hours relative to their proportion in the relevant pool, Z-value: - 22.53. Ex. A-10 at 24-25, Table 20.
264. Dr. Drogin performed an analysis of the allocation of overtime hours. He found that for the years of 1987 and 1989, the disparity between the allocation of overtime hours between men and women was statistically significant. R.T. at 17-2978.
265. Dr. Drogin found that there is greater opportunity for higher salary in the Grocery and Produce departments than in the General Merchandise and Deli/Bakery departments. In addition, men earn more than women in the same job group in nearly every year and job group. Ex. A-10 at 5, Table 3.
266. Dr. Drogin notes that Courtesy Clerks earn less than Deli/Bakery or General Merchandise Apprentices. However, he believes that it is inappropriate to compare Deli/Bakery or General Merchandise Apprentices to Courtesy Clerks because this comparison does not take seniority into account. R.T. at 17-2967. Once an employee reaches the Apprentice level in either the Grocery or the Produce departments, their wage rate is higher than the wage rate for most of the positions in the Deli/Bakery or General Merchandise job ladders. R.T. at XX-XXXX-XX.
G. Dr. Drogin's Conclusions
267. Dr. Drogin concluded that there is a consistent pattern of statistical disparities in Lucky's employment process which have perpetuated workforce stratification by sex at Lucky in the past decade. As a result of this pattern, Dr. Drogin concluded that women were effectively blocked from advancement and higher earning power. Ex. A-10 at 26.
268. Dr. John Pencavel received a Ph.D. in economics from Princeton University in 1969. He is currently a professor of economics at Stanford, where he teaches graduate and undergraduate courses in economic theory, labor economics, and econometrics. His specialty is the field of labor economics and he has published papers on issues pertaining to employment, hours of work, and wages. Ex. A-9 at 1 (Pencavel Report).
269. He has testified as an expert witness in other courts in the Northern District of California. R.T. at X-XXXX-XX.
270. This court finds that Dr. Pencavel is qualified to testify as an expert in the areas of labor economics and econometrics. R.T. at 9-1472.
271. Dr. Pencavel was hired by plaintiffs' counsel to address the question of the relative economic status of female and male employees at Lucky Stores, in particular earnings differentials between the sexes. Ex. A-9 at 1; R.T. at 9-1473.
*299 272. Dr. Pencavel worked from a computer file of information on hires in 1979, 1984, and 1986 prepared by Dr. Drogin. R.T. at 9-1474. For each hire he had information on jobs held, hours worked, and earnings at different points in time. R.T. at 9-1475. For 1979 hires, year end information for 1982, 1985, and 1988 was provided. For 1984 hires, year end information for 1987 and 1989 was provided, For 1986 hires, year end information for 1989 was provided. Ex. A-10 at 9. The payroll data was incomplete. R.T. at 9-1477. In particular, before 1987 there were no data for Store Managers and incomplete data for Assistant Store Managers. Ex. A-9 at 2, n. 1.
273. Dr. Pencavel did not have information about the employees' prior experience, health, educational background, shifts worked, or interest. R.T. at 9-1476.
274. Dr. Pencavel performed regression analyses, examining the association between a set of independent and dependent variables. The dependent variables were the standard wage rate, the annual work hours, and annual earnings. The independent variables were gender, ethnicity, age, marital status, and seniority. R.T. at 9-1480. Dr. Pencavel did not control for shift assignments, although such assignments can affect hours, wages and earnings. R.T. at 9-1527.
275. Dr. Pencavel analyzed his data in terms of three cohorts. He looked at the people hired in 1979 and charted their economic status at the end of 1982, 1985 and 1988. He looked at the people hired in 1984 and charted their economic status at the end of 1987 and 1989. Finally he looked at the people hired in 1986 and charted their economic status at the end of 1989. Ex. A-9 at 2.
276. Dr. Pencavel looked at the data for each of his cohorts after three years. At that time, approximately 80% of each of the cohorts had resigned or been fired from Lucky. R.T. at 26-4389. In 1982 only 26.7% of the 1979 cohort remained at Lucky. In 1987 only 18.2% of the 1984 cohort remained, and in 1989 only 13.4% of the 1986 cohort remained. Ex. B-369; R.T. at X-XXXX-XX. Moreover, 85.9% of those hired into Grocery and Produce and 86.5% of those hired into Deli/Bakery and General Merchandise in 1986 had left Lucky after three years. Ex. B-372; R.T. at 9-1547. For the 1984 cohort, by 1987 only eight employees (six men and two women) had been promoted to Department Head/Receiving Clerk; of those hired into Grocery or Produce, fourteen employees had been promoted to entry level management positions (Night Crew Manager, Department Head/Receiving Clerk, or Produce Department Head); of those hired into Deli/Bakery or General Merchandise, thirty-one had reached Department Head. R.T. at X-XXXX-XX.
277. The average tenure of a Courtesy Clerk or of an Apprentice in Deli/Bakery is nine months. R.T. at 26-4391.
278. Dr. Pencavel found that the attrition rate was higher for men than it was for women in each of his cohorts. Of the 1979 cohort, 24.2% of the men and 32.8% of the women were employed at Lucky at the end of 1982; 17% of the men and 23.7% of the women were employed at the end of 1985; and 12.8% of the men and 18.4% of the women were employed at the end of 1988. Of the 1984 cohort, 16.3% of the men and 20.8% of the women were employed at the end of 1987; and 12% of the men and 15.6% of the women were employed at the end of 1989. Of the 1986 cohort, 12.2% of the men and 14.8% of the women were employed at Lucky at the end of 1989. Ex. A-9 at 3.
279. Dr. Pencavel calculated the average hourly wage rates by sex for each cohort using a one-tailed test. R.T. at 9-1531. He considered a 5% level of significance to be statistically significant; and a 1% level of significance to he highly significant. Ex. A-9 at 5 n. 4; R.T. at X-XXXX-XX.
*300 280. He found that for the 1979 cohort the difference between the hourly wage rate of men and women was 44 cents at the end of 1982, a .33% level of significance (33 in 1000 chance of generating this wage difference when there is in fact no difference, the standard error is 16 cents); 35 cents at the end of 1985, a .15% level of significance; and 40 cents at the end of 1988, a .05% level of significance. For the 1984 cohort the difference between the hourly wage rate of men and women was $1.93 at the end of 1987, and $1.36 at the end of 1989. The percentage level of significance for the 1984 cohort was less than .01%. For the 1986 cohort the difference between the hourly wage rate of men and women was $1.85 at the end of 1989. Ex. A-9 at 4; R.T. at X-XXXX-XX. After adjusting for seniority within the year of hire and store, the wage differential between men and women was still statistically significant. Ex. A-9 at 6; R.T. at 9-1485.
281. Dr. Pencavel found that there was little difference between male and female wage rates within job categories, as the UFCW Contract dictates wage rates by job category. Therefore, he concluded that men and women must be distributed differently across job categories. Ex. A-9 at 5, Table 1-2; R.T. at 9-1487.
282. He also found that women are more likely to work in the General Merchandise and Deli/Bakery departments, which pay a relatively lower wage than the Grocery and Produce departments. Ex. A-9 at 5, Table 3-4. In 1982, no men from the 1979 cohort were working in Deli/Bakery or General Merchandise, but 10% of the women were. In 1989, .5% of the men from the 1986 cohort and 49% of the women were in Deli/Bakery or General Merchandise. Ex. A-9 at 5, R.T. at 9-1488. In 1988, the 1979 cohort was 37% female and 63% male. However, part-time Journey Day Clerks from the 1979 cohort were 68% female and 32% male, while full-time Journey Day Clerks were 33% female and 67% male. Ex. A-9, Table 3; R.T. at 9-1518.
283. Dr. Pencavel calculated the average annual hours of work for each cohort. He found that for the 1979 cohort the difference between the annual hours worked of men and women was 201 hours at the end of 1982; 221 hours at the end of 1985; and 209 hours at the end of 1988. For the 1984 cohort the difference between the annual hours worked of men and women was forty-one hours at the end of 1987, and seventeen hours at the end of 1989. For the 1986 cohort the difference between the annual hours worked of men and women was seven hours at the end of 1989. The difference in hours for the 1979 cohort in all years is statistically significant. However, the hours difference for the 1984 and 1986 cohorts was not statistically significant. Ex. A-9 at 7, Table 5-6; R.T. at X-XXXX-XX.
284. When controlling for seniority and store, Dr. Pencavel found that women worked significantly fewer hours than men in all years studied except for 1989. Ex. A-9 at 9.
285. Dr. Pencavel found that for each cohort in every year, the simple correlation between hourly wage rates and annual hours worked is positive for male employees and negative for female employees. He concluded that because men who work relatively long hours tend to be paid at high hourly wage rates (including premium pay) while women who work relatively long hours tend to be paid at low hourly wage rates, the annual earnings differential between men and women exceeds the differential in hourly wage rate and the differential in annual hours. Ex. A-9 at 10; R.T. at 9-1495.
286. Dr. Pencavel found that in each cohort in every year the average earnings of women is below that of men. He found that for the 1979 cohort the difference between the annual earnings of men and women was $3,410 at the end of 1982; $4,414 at the end of 1985; and $6,938 at the end of 1988. For the 1984 cohort the difference between the annual earnings of men and women was $4,466 at the end of 1987, and $5,433 at the end of 1989. For the 1986 cohort the difference between the annual earnings of men and women was $14,277 at the end of 1989. The percentage level of significance for all of the cohorts *301 in every year was less than .01% This shortfall is highly statistically significant. Ex. A-9 at 10; R.T. at 9-1496-7.
287. Based on his finding that the annual earnings of women at Lucky are between 76% and 82% of the annual earnings of men, and the existence of a significant difference between the annual earnings of male and female employees who were single, Dr. Pencavel concluded that the economic status of women at Lucky is inferior to that of men. He also concluded that his findings are robust and consistent as they are unaffected by adjustments for seniority, store, marital status, age and ethnic background. Ex. A-9 at 12; R.T. at X-XXXX-XX, 9-1599.
288. Dr. William T. Bielby received a Ph.D. in sociology from the University of Wisconsin, Madison in 1976. He is currently a professor of sociology at the University of California, Santa Barbara, where he teaches graduate and undergraduate courses in organizational behavior, labor markets, and quantitative methods. His research specialties include the organization of work, quantitative models and methods, and social inequality. In the past eight years, his work has focused on the issue of gender and work. Ex. A-8 at 1.
289. Dr. Bielby has not testified as an expert in any previous lawsuits or served as an expert consultant in any prior litigation. R.T. at 14-2361.
290. This court finds that Dr. Bielby is qualified to testify as an expert in the areas of the organization of work and its relationship to gender. R.T. at 14-2361.
291. Dr. Bielby was hired to examine Lucky's employment policies regarding job assignment and promotion, and to use his expertise in the areas of the organization of work and gender in work to explain the kinds of gender based job assignment and promotion patterns that he observed. Ex. A-8 at 1; R.T. at 14-2361.
292. He worked from workforce statistics compiled by Dr. Drogin from the employment tapes provided by Lucky. That data included incumbency by sex, by year, and job moves by sex and year. He also worked from the UFCW Contract, the E.E.O.C. determination, and the depositions of Store Managers. R.T. at 14-2362, 14-2365.
293. Dr. Bielby found that differences between the entry level jobs of men and women at Lucky can contribute substantially to subsequent differences in their pay and the advancement. Ex. A-8 at 2.
294. Based on data collected by the U.S. Employment Service during the 1960s and early 1970s on 400 California firms, Bielby and Baron concluded that some employers practice "statistical discrimination," where some job titles are reserved for men and others are reserved for women. In statistical discrimination, the assignment of some jobs exclusively to men and others exclusively to women is based on employers' perceptions of sex differences in job qualifications. Ex. A-8 at 3.
295. Organizations can have short job ladders, which end after one or two promotions, and long job ladders, which eventually lead to top management positions. If job ladders are segregated by sex, with women concentrated in the shorter ladders, women will be promoted less frequently and will be underrepresented in higher level positions. Ex. A-8 at 1-2.
296. Dr. Bielby found that organizational arrangements based on perceptions of sex differences rarely changed unless top management took explicit steps to change the policies and practices that exclude women from some jobs. Ex. A-8 at 3; R.T. at 15-2566.
297. Dr. Bielby found that employment practices in Lucky's NCD stores are similar in many ways to those in the highly segregated firms which he examined in 1960s and 1970s. He concluded that Lucky's job ladders are segregated by department: men are rarely assigned to the Deli/Bakery and General Merchandise departments, while women are largely excluded from the Produce department. In addition, movement out of the Deli/Bakery and General Merchandise department and into the Grocery department, where the opportunities for promotion to management are much *302 greater, is rare. Ex. A-8 at 3-4; R.T. at XX-XXXX-XX.
298. At Lucky, some jobs (e.g. Night Crew and Produce) have come to be perceived as "men's work," while other jobs (e.g. Deli/Bakery and General Merchandise) come to be perceived as "women's work." Department-specific job ladders such as those at Lucky have been identified as an important factor in sustaining segregation. Ex. A-8 at 4.
299. Gender stereotypes are beliefs about personality traits, role behaviors, physical characteristics and occupational preferences that differentiate men and women. Ex. A-8 at 5. Stereotypical perceptions sustain sex discrimination in higher level jobs, as individual women are evaluated according to Store Managers' perceptions of women as a group rather than as individuals. Ex. A-8 at 4.
300. Dr. Bielby stated that sex segregation is part of the collective memory of Lucky managers. R.T. at 14-2403. He received the notes taken by Mark Foley at the series of meetings with Store Managers during the summer of 1988 after he completed his report. Ex. A-121; R.T. at 14-2404. However, Dr. Bielby believed that the notes provide numerous examples of stereotypical gender beliefs regarding the suitability of women for management positions, night shift work, work perceived to be physically demanding, and full-time work. Ex. A-8 at 7-8; R.T. at 14-2406.
301. Stereotypes are most consequential in situations where evaluative criteria are ambiguous. Dr. Bielby found that when the required qualifications for a job are ambiguous and information on the decision making process is unavailable to candidates, subjects were likely to recommend candidates of their own race and sex. R.T. at 14-2395. In contrast, when evaluative criteria are clear and the decision making process is public, race and sex were less likely to factor into choices. Ex. A-8 at 6; R.T. at 14-2372.
302. Dr. Bielby found that there was no systematic procedure for evaluating employees at Lucky, so that the information Store Managers had about individual employees varied depending, inter alia, on the size of the store, and the shift the employee worked. R.T. at XX-XXXX-XX. Some factors that some managers claimed to consider in hiring and promotion decisions were performance, ability, appearance, dependability, willingness to accept responsibility, career-mindedness, and stability of home life. R.T. at 14-2385.
303. When decision makers know in advance that they must justify their decisions with respect to unambiguous criteria, stereotypical beliefs are less likely to influence their actions. Ex. A-8 at 7; R.T. at 15-2561. Dr. Bielby stated that Lucky's grievance system only creates limited accountability because it is most often used in discharge and discipline cases and because it cannot be invoked until after the fact of a promotion decision. R.T. at 14-2400, 15-2562.
304. Generally speaking, among men and women who hold similar jobs and face similar advancement opportunities, Dr. Bielby found that gender differences in attitudes towards promotion are trivial or non-existent. Ex. A-8 at 9-10.
305. Dr. Bielby noted that many studies show that individuals who face limited opportunities for advancement have lower aspirations and a weaker commitment to their work compared to others with more promising career prospects. Ex. A-8 at 9; R.T. at XX-XXXX-XX. Dr. Bielby's research on promotion interests suggests that the lack of advancement opportunities for women at Lucky has almost certainly reduced their expectations regarding promotion. In addition, the lack of opportunity has probably discouraged some women from acting upon their aspirations for job advancement and engaging in promotion-seeking behaviors. Ex. A-8 at 10.
306. Dr. Bielby believed that the implementation of a posting system for advancement opportunities and a formal system for applying for new openings along with the clear, written specification of criteria for promotion would almost certainly lead to a substantial increase in qualified employees' pursuit of job advancement. Ex. A-8 at 10-11.
*303 307. Dr. Bielby concluded that Lucky's NCD workforce is highly segregated by sex in the assignment of employees to jobs, departments, shifts, and hours. Vague and ambiguous criteria and limited accountability reinforces the influence of gender stereotypes on manager's decisions about assignment of women to jobs, departments, shifts, hours, and training opportunities and manager's decisions about promotion of women into management level positions. Given the high level of ambiguity and individual discretion involved in making such decisions, it is inevitable that personnel practices will be influenced by stereotypes regarding gender and race. Ex. A-8 at 4; R.T. at 14-2365.
308. In addition, Dr. Bielby found that in the absence of a formal system for posting promotion opportunities and eliciting information on the qualifications and interests of individual employees, managers are likely to judge individual female employees on the basis of their stereotypical beliefs of women as a group. Such judgments lead to personnel decisions that sustain and reinforce a sex segregated job structure which limits women employees' opportunities for advancement. Ex. A-8 at 8. As a result, women are disproportionately assigned to departments with limited promotion opportunities at Lucky and have limited access to job duties, hours, and shifts that increase an employee's eligibility for management positions. Ex. A-8 at 11.
309. Dr. Bielby also concluded that there is no evidence that male and female employees differ in the features they see as important in a job. However, in the absence of accountability and formal criteria for personnel decisions, the interests and qualifications of individual women employees are ignored and the differences between male and female employees are greatly exaggerated. Ex. A-8 at 11; R.T. at XX-XXXX-XX, 15-2557.
310. Dr. Bielby suggested that in order to minimize the effect of gender stereotypes, employers should use employment practices that specify criteria for decision making, that provide accurate information about applicants, and that hold decision makers responsible. Under those circumstances, decision makers have a specific set of guidelines to follow and the person being considered for the job or promotion has information about how the decision is to be made, so the person can take efforts to make his or her qualities known to the decision maker. R.T. at XX-XXXX-XX.
311. Regarding the results of Dr. Drogin's study, Dr. Bielby said that "it is difficult to reconcile segregation that dramatic with what we know about the interests of men and women in the kinds of work they seek." However, he conceded that the statistical disparities could be the result of a combination of discrimination against women and a lack of interest on the part of women. R.T. at XX-XXXX-XX.
312. Defendant's expert, Dr. Daum, claims that Dr. Bielby's conclusions are not applicable to Lucky because the studies on which he based his report involve different populations, because most of the work at Lucky is part-time, because some jobs at Lucky can only be done at night, and because none of the studies involved retail food stores. R.T. at XX-XXXX-XX. However, Dr. Bielby insists that the studies he cites in his report have been found to have external validity and are accepted throughout the scientific community. R.T. at 15-2555.
313. Dr. Jeffrey W. Daum received a Ph.D. in Social-Industrial Psychology from Louisiana State University in 1972. He is currently Vice President of HRStrategies, and President of Confer, Inc., an HRStrategies subsidiary. His consulting experience spans public and private sector organizations. He has designed, implemented and analyzed surveys over the past twenty years for organizations such as General Motors, JCPenney, Post-Newsweek, banking institutions, supermarket chains (Albertson's), state university systems, and the chemical industry. Ex. B-18. Dr. Daum has not previously testified as an expert. R.T. at 22-3802.
314. This court finds that Dr. Daum is qualified to testify as an expert in the *304 areas of employment-related surveys, employee job interest, work hour preferences and employee selection. R.T. at 22-3803.
A. Dr. Daum's Work Interest Survey
315. Dr. Daum was hired by defendant's counsel to design a survey instrument which would assess interest patterns in the Lucky workforce. His task was to see if job interest patterns were genderspecific and if they were related to the distribution of employees at Lucky. R.T. at 22-3803.
316. Dr. Daum's job interest survey was divided into three sections. In Part 1 subjects were asked to relate their job history at Lucky by checking off, from a list of twenty-six jobs, their current position, all the previous positions they have held, the first job they applied for, and the first job they held at Lucky. In Part 2 subjects were asked to rank a list of eleven job features; and in Part 3 subjects are asked to rank a list of nine jobs with respect to each of the eleven job features. Ex. B-20; R.T. at 22-3828. Dr. Daum designed three different versions of the survey in which the columns of paired comparisons were placed in different orders. R.T. at 22-3830. The job features and the job titles were listed in alphabetical order in all three surveys. R.T. at 22-3825.
317. The job titles listed in Part 1 of the survey were not official Lucky job titles. For example, Checker and Stocker are officially the same job, but Lucky employees consider themselves to be one or the other and were not confused by having both job titles in the survey. Apprentice Clerk positions also were not included in the survey. However, Dr. Daum believed that each subject was able place himself or herself within one of the twenty-six job titles. R.T. at XX-XXXX-XX.
318. Part 2 of Dr. Daum's job interest survey identified eleven job features: 1) advancement opportunity, 2) co-worker contact, 3) level of customer contact, 4) degree of flexibility in scheduling hours, 5) average number of hours available, 6) rate of pay (dollars per hour), 7) level of physical effort, 8) degree of responsibility, 9) shift requirements, 10) stability of hours, and 11) level of stress. Ex. B-20 at 3.
319. Dr. Daum drew up this list of features based on what Lucky employees in the SCD and Lucky management in NCD had indicated was important to them in making career choices or which they considered when they thought about progressing within the Lucky system. R.T. at XX-XXXX-XX. Dr. Daum did not conduct a job analysis to determine which job features or jobs should be included in his survey as he did when he designed the job interest survey for Albertson's supermarket. R.T. at 24-4108, 24-4111.
320. Dr. Daum included the job features "shift requirements" and "stability of hours" in order to emphasize the issues in this case. He did not include those two job features in his study for Albertson's. R.T. at XX-XXXX-XX. Dr. Daum does not believe that his selection of job features caused a gender bias in the survey because there were low correlations between the responses to the features and to the jobs. In addition, a manova analysis showed no consistent relationship between the responses to the features. R.T. at XX-XXXX-XX.
321. Dr. Daum's job features could be interpreted in multiple ways. For example, if a subject responded that a job feature was important to him or her in selecting a job, that could mean that it is important to have either a high or a low degree of that job feature. R.T. at 25-4153, 25-4155. However, Dr. Daum believed that as long as subjects were consistent in their interpretation of the job features, the results of the survey would be valid. The test of intransitivity which Dr. Daum ran showed that subjects responded to the paired comparisons in a consistent manner. Ex. B-22; R.T. at 26-4325.
322. Part 3 of Dr. Daum's job interest survey identified nine jobs: 1) Checker, 2) Deli/Bakery Clerk, 3) Fourth Person, 4) General Merchandise Department Head, 5) Night Crew, 6) Night Crew Manager, 7) Produce Clerk, 8) Receiving Clerk, and 9) Store Manager. Ex. B-20 at 6.
*305 323. Dr. Daum assumed that each subject was equally familiar with the jobs on the survey and that there was no gender difference in the subjects' job familiarity. R.T. at 25-4158. However, subjects who worked in stores which did not have Deli/Bakeries would have less occasion to know about Deli/Bakery positions. R.T. at 25-4160. In addition, about 20% of the subjects had worked at Lucky for less that a year and would not have had a lot of time to become familiar with the different jobs. R.T. at 25-4161.
324. Dr. Daum assumed that his subjects knew exactly what they want at Lucky, and would do what they had to do to get it. R.T. at XX-XXXX-XX.
325. Dr. Daum's survey was administered on a Wednesday and a Thursday in March in sixteen stores in the NCD. R.T. at 22-3847. 417 women and 505 men took the survey. R.T. at 23-3896. Dr. Daum chose which stores to survey by placing all of the stores in the NCD on a matrix of volume and geographical location. R.T. at 25-4180. For each of the sixteen cells in the matrix, he chose the store that most closely approximated the average gender and race composition of the entire cell. R.T. at XX-XXXX-XX. Dr. Daum found that there was no statistically significant difference in the survey responses that corresponded to the percentage of women working in each store. Ex. B-25; XX-XXXX-XX. Although the number of stores in each cell was not equal, Dr. Daum did not weight the survey responses based on the number of employees represented by each cell. R.T. at 25-4176. Dr. Daum did not believe that it was possible to do a random sample in a real environment, and did not think that there were any systematic sampling errors in his stratified sample. R.T. at XX-XXXX-XX.
326. Dr. Daum's "universe," or population to which his data can be generalized, consists of those people who are currently employed in the retail stores in Lucky's NCD, with the exception of pharmacists, former Alpha Bets store employees, and administrative and warehouse employees. R.T. at XX-XXXX-XX. Dr. Daum ran a test of proportions and found that, with the exception of the overrepresentation of General Merchandise Department Head, he had captured employees in each job title in approximately the proportion that the job was represented in the NCD. Ex. B-24; R.T. at 25-4207.
327. Dr. Daum's survey employed paired comparisons. This methodology assumes that if a job feature is important to a subject, the subject will choose it consistently over other job features, and that if a job feature is less important to a subject, the subject will flip-flop between choosing it or other job features. R.T. at 25-4157.
328. Dr. Daum employed paired comparisons in an attempt to avoid having subjects give "socially desirable" responses to his survey questions. R.T. at XX-XXXX-XX, 24-4093. Dr. Daum assumed that forcing his subjects to focus on the importance of the job features would screen out their understanding of which jobs were more socially desirable or were more traditionally gender appropriate. R.T. at 24-4097. In addition, Dr. Daum believed that paired comparison methodology controls for the expectations of subjects because they are not asked to rate the jobs in terms of the likelihood that they will get the job they want. R.T. at 23-3909. However, paired comparisons cannot provide information about how subjects come to have the preferences they have, or if those preferences have changed since they have been employed at Lucky. R.T. at 24-4090. Moreover, this methodology cannot ascertain the trade-offs that people make in choosing to seek different jobs. R.T. at XX-XXXX-XXX, XX-XXXX-XX.
329. Dr. Daum did not ask "Have you been discriminated against at Lucky?" He believed that the difficulty of interpreting that type of information outweighed the benefit of obtaining it. R.T. at 24-4092, 26-4326. Dr. Daum also did not ask "Are *306 you interested in a promotion, if so what job?" Dr. Daum said that while he would have been interested in tracking the subjects to see whether their behavior was consistent with their survey responses, he was not concerned about their professed interests. R.T. at XX-XXXX-XX.
330. Based on the responses to his job interest survey, Dr. Daum computed standardized preference scores from a standard table which converts frequency counts, the number of times one item in a pair is chosen over all of the other items, into standardized scores. The standardized scores have "a normal distribution with an average or mean of 50 and a standard deviation of 10." R.T. at 23-3898. See infra n. 27.
331. Based on his calculations of randomness and intransitivity in the responses, Dr. Daum found that the subjects understood the survey and that they responded to it consistently. Ex. B-22; R.T. at XX-XXXX-XX. There is no published standard for what is acceptable. However, Dr. Daum believed that, in addition to being internally consistent, the survey was consistent with the choices that people actually make in deciding whether to pursue a job because the elements of the survey were personal and realistic. R.T. at XX-XXXX-XX.
332. Dr. Daum found that 83.3% of the women and 84.8% of the men surveyed said that they got the job for which they initially applied. Ex. B-28; R.T. at 23-3890. Dr. Daum made no provision for subjects who applied for more than one job. R.T. at XX-XXXX-XX. About 100 of the 900 employees sampled said that they had applied for "any job." R.T. at 23-3891. Dr. Daum compared the results of this survey question to Dr. Haworth's initial placement database sample and found that the results were consistent. Ex. B-27; R.T. at XX-XXXX-XX. Dr. Daum did not know how Dr. Haworth classified the applications in the initial placement data base, he did not know how she calculated her percentages, and he did not look at the application or bid sheet data himself. R.T. at 24-4106, 25-4146.
333. The results of Dr. Daum's survey showed a disparity between the jobs men and women initially applied for at Lucky. Ex. B-26; 23-3893. Dr. Daum found that while men's and women's standardized preference scores for each of the nine jobs across all of the job features followed the same trend, Ex. B-29; R.T. at XX-XXXX-XX, certain jobs were statistically significantly preferred by one gender across all of the job features. Checker, Deli/Bakery Clerk, and General Merchandise Department Head were statistically significantly preferred by women across all of the job features; and Night Crew, Night Crew Manager, Produce Clerk, Receiving Clerk and Store Manager were statistically significantly preferred by men across most job features. Ex. B-31 and B-30; R.T. at 23-3904.
334. Based on all of the job features and weighting for the importance of each job feature, Dr. Daum ranked the standardized preference scores for all of the jobs. He found that women most preferred Checker (score 55.39) and least preferred Night Crew (score 42.61). Men most preferred Store Manager (score 55.58) and least preferred Deli/Bakery Clerk (score 38.09). While Dr. Daum found that men's and women's four most preferred and four least preferred jobs were the same, he also found that the difference between men's and women's preference scores was statistically significant for all of the jobs with the exception of Fourth Person. Ex. B-32; R.T. at XX-XXXX-XX.
335. Dr. Daum found that subjects tended to rate their incumbent positions higher than they rated other jobs. R.T. at *307 25-4163. However, he contended that the interest patterns he observed were consistent even when incumbents were not counted. R.T. at XX-XXXX-XX.
336. Dr. Daum found that while women ranked Deli/Bakery Clerk eighth of the nine jobs, 87.3% of those who listed Deli/Bakery Clerk as the first job for which they applied were women. Ex. B-32; Ex. B-27; 23-3914. In addition, 7% of women chose Deli/Bakery Clerk as their most preferred job, while only 2% of men did so. Ex. B-33. Therefore, Dr. Daum explained, he would expect three times as many women as men to be Deli/Bakery Clerks. R.T. at 23-3917.
337. Dr. Daum found that men rated Deli/Bakery Clerk last with respect to the job feature of stress, while women rated Store Manager last with respect to that job feature. R.T. at 25-4215. Men rated Night Crew eighth of the nine jobs with respect to the job feature of physical effort, and rated Deli/Bakery Clerk ninth. Daum did not find these results to be inconsistent with his contention that the subjects understood the task of the survey. R.T. at XX-XXXX-XX.
338. Dr. Daum attempted to chart the job preferences of employees whose current job feeds into Fourth Person. Ex. B-36; R.T. at 23-3929. Dr. Daum charted the standardized preference scores of employees in Night Crew, Fifth Person, Produce Manager, Checker, Produce Clerk, Stocker, Receiving Clerk, Night Crew Manager, and Front-End Clerk, but did not test to see if those jobs actually fed into Fourth Person. Ex. B-36; R.T. at 25-4241. He found that out of the 435 employees in his feeder pool, only eighteen indicated that Fourth Person was their most preferred job. Ex. B-36; R.T. at 23-3931. However, while women constituted 32.9% of his feeder pool, they only accounted for 16.7% of those who chose Fourth Person as their first preference. Ex. B-36; R.T. at 23-3932. This result is not statistically significant. R.T. at 25-4242.
339. Dr. Daum then considered the standardized preference scores of those employees in his Fourth Person feeder pool who chose Fourth Person as their most preferred job, and those who chose Store Manager as their most preferred job and chose Fourth Person second. Ex. B-37; R.T. at 23-3934. He found that women constituted 32.9% of this feeder pool and accounted for 28.4% of those who chose Fourth Person as their first or second preference. Ex. B-37. This result is not statistically significant. R.T. at 25-4245.
340. Dr. Daum also considered the standardized preference scores of Checkers and Stockers who chose Fourth Person as their most preferred job, and those who chose Store Manager as their most preferred job and chose Fourth Person second. Ex. B-38; R.T. at 23-3937. He found that women constituted 47.8% of this feeder pool and accounted for 35.4% of those who chose Fourth Person as their first or second preference. Ex. B-38. This result is statistically significant. R.T. at 23-3940.
341. Dr. Daum concluded that, when one looks at the appropriate feeder pool, there are consistent differences in the level of interest of men and women in the position of Fourth Person. R.T. at 23-3944. However, Dr. Daum admitted that regardless of how an employee rated Fourth Person, if the employee has no interest in the jobs through which they would have to progress in order to reach Fourth Person, it is unclear that they would tolerate what it takes to become a Store Manager. R.T. at 25-4221.
342. Dr. Daum also admitted that a person who rated Receiving Clerk as their first choice and Fourth Person as their second choice would not appear in his feeder pools, while a person who gave a lower preference score to the Fourth Person job but rated it as their first choice would appear in his feeder pools. R.T. at XX-XXXX-XX. However, Dr. Daum believed that it was unlikely that this had occurred because he found the highest positive correlation across all the job features between Store Manager and Fourth Person. Thus, he thought it was improbable that an individual would express great interest in Store Manager, and little interest in Fourth Person. R.T. at 26-4347.
*308 343. Dr. Daum also attempted to chart the job preferences of employees whose current job feeds into Night Crew Manager. Ex. B-39; R.T. at 23-3945. Dr. Daum charted the standardized preference scores of employees in Night Crew, Checker, Stocker, and Produce Clerk. Ex. B-39; R.T. at 23-3945. He found that out of the 350 employees in his feeder pool, twenty-two indicated that Night Crew Manager was their most preferred job. Ex. B-39. However, while women constituted 34.4% of this feeder pool, they did not account for any of those who chose Night Crew Manager as their first preference. Ex. B-39.
344. Dr. Daum then considered the standardized preference scores of those employees in his Night Crew Manager feeder pool who chose Night Crew Manager as their most preferred job, or as their second or third most preferred job after Store Manager or Fourth Person. Ex. B-40; R.T. at 23-3947. He found that women constituted 34.4% of this feeder pool, and accounted for 7.5% of those who chose Night Crew Manager as their most preferred job, or as their second or third most preferred job after Store Manager or Fourth Person. Ex. B-40. Dr. Daum did not indicate whether this result was statistically significant.
345. Dr. Daum charted the standardized preference scores for each of the job features. He found that the ranking of the job features was similar, but that some of the job features were more important to one gender than to the other. Ex. B-42; R.T. at 23-3952. The only job features for which there was a statistically significant difference between the standardized preference scores of men and women were "level of customer contact" and "level of physical effort." Ex. B-42. Dr. Daum concluded from these results that management did not differentially influence the job interests of male and female employees. If it had, the interests of the women employees who were blocked from attaining the positions which they most desired would be different than the interests of the men who had not faced blocked opportunities. R.T. at 23-3953. Dr. Daum explained that he would expect blocked opportunity to be reflected in the subjects' preference of job features but not in their preference of jobs. R.T. at 23-3954. However, Dr. Daum did not know of any scholarly research on the subject of the effect of management conduct on employee job preferences. R.T. at 24-4089.
346. Dr. Daum assumed that if management has an impact on the job interests of employees, that impact should increase with the employees' interaction with management. R.T. at 24-3974. Dr. Daum concluded that Lucky management did not differentially influence the job interests of female employees because female employees who had worked at Lucky for less than one year did not have standardized preference scores for Fourth Person or Store Manager which differed statistically significantly from female employees who had worked at Lucky for two years or more. Ex. B-46; XX-XXXX-XX. The only job with respect to which women's standardized preference scores changed statistically significantly over time was Night Crew Manager. R.T. at 25-4263. Dr. Daum found that men's standardized preference scores for Fourth Person and Store Manager do not change statistically significantly in their first year at Lucky. However, men's standardized preference scores for Fourth Person drop statistically significantly after two years at Lucky. B-426; XX-XXXX-XX.
347. Dr. Daum did not survey the interest levels of terminated Lucky employees. R.T. at 25-4169. Rather, he assumed that the job preference scores of the employees who had been at Lucky for less than one year would be about the same as those of employees who had resigned or were terminated. R.T. at 24-3977, XX-XXXX-XX.
348. Dr. Daum compared the results of this job interest survey to the results of the job interest survey that he conducted in the SCD and the job interest survey that he *309 conducted for Albertson's. He concluded that the results of this job interest survey reflect interest patterns which are related to the subjects' understanding of the different jobs; in addition, the survey reflects actual interest patterns which have not been moderated by the influence of management. R.T. at XX-XXXX-XX.
B. Dr. Daum's Work Hours Interest Survey
349. Dr. Daum was hired by defendant's counsel to design a survey instrument which would address the issue of men's and women's different preferences for working different hours and different days during the week. R.T. at 24-4053, 24-4062. Dr. Daum's goal was to measure both whether the subjects wanted to work more hours and whether they were available to work more hours. R.T. at XX-XXXX-XX.
350. The survey was conducted on October 20, 1990 in Lucky's SCD. Dr. Daum believed that there were enough parallels between the SCD and the NCD for the results of the survey to be applicable to both divisions. R.T. at XX-XXXX-XX, XX-XXXX-XX. Dr. Daum chose which stores to survey by placing all of the stores in the SCD on a matrix of volume and geographical location. For each of the four cells in the matrix, he chose to survey the store that most closely approximated the average gender and race composition of the entire cell. R.T. at XX-XXXX-XX. 184 employees, excluding pharmacists, took the survey. R.T. at 24-4066. Dr. Daum surveyed all but six of the employees who worked in the four stores on the day of the survey. R.T. at 26-4280. However, there were only three women who regularly worked night jobs in the sample. R.T. at 26-4299. Dr. Daum did not know if the percentage of part-time to full-time workers in the survey was representative of the percentage in the NCD. R.T. at 26-4300.
351. In Dr. Daum's work hours interest survey subjects were asked to select their current job title out of a list of twenty-six job titles, to identify the number of hours they work per week, and to chart their actual work schedule. The subjects were also asked how many additional hours they would like to work and would be available to work. Finally, the subjects were asked to design their ideal work schedule. Ex. B-424.
352. Some of the questions on the survey were open to multiple interpretations. For example, on question eight a subject could have marked just the particular four additional hours they wanted to work, or they could have marked all of the time in which they were available to work the four additional hours. Dr. Daum found that the blocks of time which the subjects marked consistently exceeded the number of additional hours which they said they wanted to work. R.T. at 26-4286. Question nine asked if the subject was willing and available to work the schedule they marked in questions six, seven, and eight. It is unclear how the subjects interpreted that question. R.T. at XX-XXXX-XX.
353. Question ten asked subjects to indicate what time periods they both wanted and would be available to work consistently during the upcoming six month period. Ex. B-424. Dr. Daum found that of the women who worked part-time, 70.3% wanted *310 a day schedule, 7.8% wanted a night schedule, and 21.9% wanted a mixed schedule. Of men who worked part-time, 38.5% wanted a day schedule, 26.9% want a night schedule, and 34.6% wanted a mixed schedule. Ex. B-47; XX-XXXX-XX. In addition, Dr. Daum found that of the women who worked full-time, 84.2% wanted a day schedule, 5.3% wanted a night schedule, and 10.5% wanted a mixed schedule; and of men who worked full-time, 75.5% wanted a day schedule, 14.3% wanted a night schedule, and 10.2% wanted a mixed schedule. Ex. B-47. Dr. Daum characterized someone as preferring a day schedule if they checked off more hours between 7:00 a.m. and 7:00 p.m. than between 7:00 p.m. and 7:00 a.m.; he characterized someone as preferring a night schedule if they checked off more hours between 7:00 p.m. and 7:00 a.m. than between 7:00 a.m. and 7:00 p.m.; and he characterized someone as preferring a mixed schedule if they checked off an equal number of blocks in each of the twelve hour time periods. R.T. at XX-XXXX-XX.
354. Dr. Daum concluded that although night schedules were generally unpopular, three times as many men wanted to work nights as women. However, he also found that about half of the subjects wanted to keep their current schedule. R.T. at 26-4306. In addition, he did not ask if the subjects would be willing to work a different schedule if they were promoted. R.T. at XX-XXXX-XX.
355. Dr. Daum found that the proportion of women who prefer to work a 7:00 a.m. to 11:00 a.m. shift was statistically significantly greater than the proportion of men who prefer to work that shift. The proportion of women who prefer to work an 11:00 a.m. to 3:00 p.m. shift was highly statistically significantly greater than the proportion of men who prefer to work that shift. Dr. Daum found no significant difference in men's and women's preference for the 3:00 p.m. to 7:00 p.m. shift or the 3:00 a.m. to 7:00 a.m. shift. Dr. Daum found that the proportion of men who prefer to work a 7:00 p.m. to 11:00 p.m. shift was highly statistically significantly greater than the proportion of women who prefer to work that shift. The proportion of men who prefer to work an 11:00 p.m. to 3:00 a.m. shift was highly statistically significantly greater than the proportion of women who prefer to work that shift. Ex. B-48; R.T. at 24-4077.
356. Dr. Daum found that 61% of women who worked part-time wanted to work full-time, while only 48% of men who worked part-time wanted to work full-time. He addition, he found that women who worked part-time said that they wanted more additional hours than did men who worked part-time. R.T. at XX-XXXX-XX.
357. Dr. Daum concluded that the work environment at Lucky is different than that in most businesses because work takes place twenty-four hours a day. R.T. at 24-4081. In addition, Store Managers make job-related decisions based on their knowledge of employees work histories. Therefore, the fact that there are no written job descriptions or promotion criteria at Lucky does not indicate that there is either stereotypical thinking or discrimination. R.T. at XX-XXXX-XX.
358. Dr. Joan Gustafson Haworth received a Ph.D. in economics from the University of Oregon in 1970. She was a tenured faculty member at Florida State University until 1989 when she retired. Dr. Haworth is the president of Economic Research Services Inc., a consulting firm. Ex. B-49; R.T. at 26-4374.
359. Dr. Haworth has previously been qualified to testify as an expert in statistics and labor economics in federal court on behalf of plaintiffs, such as the Lawyer's Committee and the Massachusetts African-American Police Association, and defendants, such as Sears, General Motors, and the system of higher education of Oregon. R.T. at 26-4379.
360. This court finds that Dr. Haworth is qualified to testify as an expert in the *311 areas of statistics and labor economics. R.T. at XX-XXXX-XX.
361. Dr. Haworth was hired by defendant's counsel to review Lucky's employment data and to determine whether the data was consistent with plaintiffs' allegations of discrimination. R.T. at 26-4380.
362. Dr. Haworth based her testimony on her review of Lucky's payroll data, the UFCW Contracts, the available bid lists for full-time positions, employee personnel folders, sample weekly schedules, data from the Valley Posting Program, available job applications, data about labor force participation rates outside of Lucky, Dr. Daum's conclusions, plaintiffs' experts' reports, and the Foley affirmative action progress reports. R.T. at XX-XXXX-XX.
A. Dr. Haworth's Initial Placement Database
363. Dr. Haworth's initial placement database was drawn from the set of approximately 28,000 Lucky applications from May 1984 through 1990. As reviewing all of the applications would have been burdensome, the parties stipulated to construct a database which would be representative of the population of employees hired by Lucky between 1984 and 1989. Ex. A-203; R.T. at 27-4435.
364. Dr. Haworth's initial placement database did not conform to the sampling plan to which the parties stipulated. Ex. A-203; R.T. at XX-XXXX-XX. In fact, Dr. Haworth's actual sample of 5145 applications was significantly smaller than the 7164 applications that was that was called for in the stipulation. R.T. at XX-XXXX-XX. In addition, Dr. Haworth disregarded the provision of the stipulation that required the replacement of missing applications in the case that more than 2% of the applications sought were not obtained. Ex. 203, ¶ 7. Dr. Haworth was missing more than 2% of the applications which she sought for Courtesy Clerk, Grocery Apprentice, Deli/Bakery Apprentice, and General Merchandise Apprentice. R.T. at XX-XXXX-XX. The sample which Dr. Haworth obtained contained about 24% fewer applications than was prescribed in the original sample plan. R.T. at 34-5660.
365. The numbers of Courtesy Clerks, Grocery Apprentices, and Deli/Bakery Apprentices which were to have been sampled under the stipulation would have ensured a margin of error of less than 3%. R.T. at 27-4448. Dr. Haworth testified that a margin of error of 3% was reliable, although she would have preferred a smaller margin of error. R.T. at 27-4450. Dr. Haworth's sample did not meet the 3% margin of error for hires of Deli/Bakery, Produce and Grocery Apprentices. R.T. at XX-XXXX-XX. In addition, the margin of error for Night Crew Apprentices exceeded 5%. R.T. at XX-XXXX-XX.
366. The majority of the missing applications were of terminated employees. R.T. at 27-4466. Dr. Haworth believed that many of the missing applications belonged to strike replacements who were not supposed to be included in the database. R.T. at 27-4508. In 1989, over 600 people were temporarily hired by Lucky as potential strike replacements and were classified as Grocery Day apprentices. R.T. at XX-XXXX-XX; XX-XXXX-XX (Herkal). These strike hires lacked race or sex codes in the payroll data, and their applications did not indicate the job for which they applied because they were not confined by the UFCW Contract to a particular job. R.T. at XX-XXXX-XX. Dr. Haworth estimated that there were about 600 strike hires. R.T. at XX-XXXX-XX.
367. Dr. Haworth coded all Grocery department applications, including Stocker applications, as applications for day work. She only counted an application as a night application if the word "night" was specifically used on the form. R.T. at 27-4491, XX-XXXX-XX. If an applicant wrote "clerk," Dr. Haworth counted them as having applied for a Grocery Day Apprentice job. 27-4488. However, she admitted that night workers do more stocking than do day workers, R.T. at XX-XXXX-XX, and "stocker" applications were a larger proportion of night hires than were "night" applicants. Ex. A-250. When "stocker" applicants are added to "night" applicants *312 the female application rate triples to about 25%. Ex. B-59; R.T. at XX-XXXX-XX.
368. Dr. Haworth counted Floral Clerks as Produce hires when looking at the promotion data, but considered them to be General Merchandise Applicants when she looked at the initial placement data. R.T. at 27-4485, 29-4813, XX-XXXX-XX.
369. Dr. Haworth disregarded applications that were for jobs other than those at issue in this litigation (e.g. applications for management, maintenance, security, or bookkeeping). R.T. at 29-4802. 55.18% of the applications for "any job" were from women. Ex. B-59. If an applicant listed more than one job, Dr. Haworth counted them as having applied for each job. R.T. at 27-4433. Dr. Haworth disregarded the more than 200 applications which left blank the question "for what type of work are you applying." R.T. at 29-4796.
370. Dr. Haworth admitted that many factors could affect which position a person would indicate they wanted on their initial application, including his or her knowledge of the jobs available, and what the person was told by store management. R.T. at XX-XXXX-XX. However, it is clear from the overwhelming evidence in this case that applicants were given little information and the information they were given varied from store to store.
371. Dr. Haworth testified in E.E.O.C. v. Sears, 628 F. Supp. 1264 (N.D.Ill.1986), aff'd, 839 F.2d 302 (7th Cir. 1988), that "[c]ounts of applicants are unreliable indicators of interest when, as in this case, there are many different types of jobs for which the same application form is submitted, and most forms contain insufficient information from which to determine whether the applicant is interested, qualified, and available for the specific position at issue." Ex. B-209 at 147-8.
372. Dr. Haworth found that of those she considered to have applied for night jobs, less than 30% received night positions; likewise, of those she considered to have applied for Produce jobs, about 33% received Produce positions. R.T. at XX-XXXX-XX, XX-XXXX-XX.
373. Dr. Haworth testified that Store Managers do not consider an employee's initial application in making promotion decisions. R.T. at XX-XXXX-XX. In addition, Dr. Haworth believed that employee job preferences are likely to change after their initial placement. R.T. at XX-XXXX-XX.
374. Dr. Haworth found that only fourteen out of the 2302 people who were hired into Courtesy Clerk positions initially applied for night work; and only twenty-four applied for positions in the Produce department. Ex. A-217; R.T. at XX-XXXX-XX. However, of the sixty Courtesy Clerks who were promoted to night positions, only one had applied for night work. Of the thirty-four Courtesy Clerks who were promoted to Produce Clerk positions, only two had applied for Produce positions. Ex. A-250. Therefore, what applicants wrote on their job applications has very little effect on the job to which they were assigned by the Store Manager.
375. Dr. Haworth admitted that the 105 "night" applications, 206 Produce applications, and twenty-one janitor applications upon which her initial placement database relied were either insignificant sample sizes or had more than a 5% margin of error. Ex. A-212; XX-XXXX-XX, XX-XXXX-XX.
376. Dr. Haworth testified that the results of Dr. Daum's job interest survey for the positions of Courtesy Clerk, Deli/Bakery and General Merchandise Clerk, and Night Crew, confirmed the reliability of her initial placement database sample. Ex. B-27; R.T. at XX-XXXX-XX; see supra n. 12. However, their results for the percentage of applicants for other job titles were not very close. Ex. B-26; R.T. at XX-XXXX-XX.
377. Dr. Haworth testified that the initial placement database sample was reliable *313 for the purposes for which it was used. R.T. at 27-4503.
378. Dr. Haworth looked at the proportion at which men and women applied for different jobs at Lucky, as indicated by the initial placement database. Based on those numbers, she calculated a preference adjustment ratio. Dr. Haworth then applied her preference adjustment to the lottery model which Dr. Drogin used to determine the number of men and women he predicted would be placed into the different jobs at Lucky. Finally, Dr. Haworth derived an adjusted predicted figure which she believed more accurately reflects the interests of Lucky employees. Exs. B-52 and B-59; R.T. at XX-XXXX-XX.
B. Dr. Haworth's Conclusions
379. Dr. Haworth concluded that although not all women have the same interests as the "average women" and not all men have the same interests as the "average man," there are average differences in the work preferences of men and women. R.T. at 26-4395.
380. Dr. Haworth assumed that the relative level of interest of male and female employees in entry-level jobs was reflected in the relative proportion of male and female employees applying for particular jobs. R.T. at XX-XXXX-XX, XX-XXXX-XX, XX-XXXX-XX, 31-5115. Therefore, she applied preference adjustment factors based on the proportion of men and women applying for particular jobs to all employees hired into entry-level jobs. Ex. B-53; R.T. at 27-4534, 28-4720, 28-4721.
381. Dr. Haworth's preference adjusted model closely approximated the gender composition of actual hires into entry-level jobs. Exs. B-52 and B-53; R.T. at 27-4541. When Dr. Haworth revised her analysis at trial to ensure the consistent coding of Floral Clerks as General Merchandise employees, the disparity between actual and predicted placements into the Produce department was virtually eliminated. R.T. at 27-4430, XX-XXXX-XX, 30-4916, XX-XXXX-XX. The differences in the gender compositions of actual and predicted placements into other entry-level positions were within a 2-5% margin of error. R.T. at XX-XXXX-XX, XX-XXXX-XX, XX-XXXX-XX, XX-XXXX-XX, XX-XXXX-XX, 29-4873, XX-XXXX-XX.
382. Haworth concluded that female Grocery clerks use their seniority to avoid working after 7:00 p.m. Dr. Haworth concluded that men prefer night work more than do women. R.T. at XX-XXXX-XX. Therefore, she assumed that women would be less inclined than men to assume Fourth Person duties or to work the shifts that are necessary preparation for Fourth Person. R.T. at 28-4660. However, the overwhelming majority of Fourth Persons are promoted from day positions: 47.3% from full-time day positions, and 37.1% from part-time day positions. Ex. B-146.
383. Using a preference adjustment factor which was based on Dr. Daum's findings about men's and women's preferences for Fourth Person, Dr. Haworth concluded that the number of female moves to Fourth Person would be 70-90% of the number predicted by Dr. Drogin's lottery model. Ex. B-37, B-157; R.T. at XX-XXXX-XXXX.
384. Dr. Haworth believed that disparities in the promotions of men and women that were not related to attitudes towards night work, those from Courtesy Clerk into Deli/Bakery and Produce Apprentice, were explained by differences in interest. Ex. B-31, B-33, B-35, B-125, B-126; R.T. at 28-4638, 30-5057.
385. Dr. Haworth analyzed all moves to full-time using a preference adjustment factor that adjusted the gender composition of the relevant feeder pools. Dr. Haworth found that there was no statistically significant disparity in moves to full-time. Exs. B-123 and B-124; R.T. at XX-XXXX-XX. However, she conceded that as many as 239 moves (41.78% of all moves) could not be tied to a bid list, and that of those moves only 19.2% went to women. Ex. A-219; R.T. at XX-XXXX-XX.
386. Dr. Haworth asserted that there is a group of primarily female senior part-time employees who work less than average regular hours, very few premium hours, and never bid for full-time work. R.T. at 27-4578, XX-XXXX-XX, 31-5087. *314 However, Dr. Haworth was not able to quantify the size of this group. R.T. at XX-XXXX-XX. Dr. Haworth concluded that in an environment where additional hours tend to be available in the evenings or at night the gender differences in the average hours worked and earnings of part and full-time employees at Lucky were entirely consistent with gender neutral employment practices. R.T. at 27-4553, 30-4987, XX-XXXX-XX, XX-XXXX-XX, 31-5129.
387. Dr. Haworth concluded that the differences in average premium pay earned by male and female employees could be explained by the greater availability of male employees for night work. Ex. A-224; R.T. at XX-XXXX-XX, 28-4597, XX-XXXX-XX.
388. Dr. Carl C. Hoffman received a Ph.D. in sociology from the University of North Carolina at Chapel Hill in 1977. He is currently the president and director of Hoffman Research Associates, a social research firm specializing in consulting in human resource research and development. Ex. A-242 at 3.
389. Dr. Hoffman has testified as an expert in federal court on about twenty previous occasions. He has testified in sex discrimination, race discrimination, securities fraud, and libel cases, in the areas of statistics, labor force economics and survey research. R.T. at 32-5342.
390. This court finds that Dr. Hoffman is qualified to testify as an expert in the areas of the design, administration and analysis of employee job interest surveys, and survey research and statistics. R.T. at 32-5347.
391. Dr. Hoffman was hired by counsel for plaintiff to review the job interest and work hours interest surveys performed by Dr. Daum. Ex. A-242 at 3. In addition, he performed his own cluster analysis and multidimensional scaling on Dr. Daum's data. Ex. A-242 at 31-38.
A. Critique of Dr. Daum's Job Interest Survey
392. Dr. Hoffman has conducted interest surveys for Delta Airlines (1978), Bristol-Myers (1981) and Purolator Courier (1983). R.T. at 32-5343. He believes that job interest surveys can be used in discrimination cases, but he does not believe that Dr. Daum's survey was well executed, nor does he think Dr. Daum's survey answered the relevant questions in this case. R.T. at XX-XXXX-XX. Moreover, Dr. Hoffman did not feel confident relying on the results of an interest survey once he had found that management engaged in disparate treatment. R.T. at 33-5477.
393. Dr. Hoffman concluded that Dr. Daum's job interest survey was flawed because the universe was not defined to include all employees of Lucky Stores or the employment practices they experienced. Ex. A-242 at 4. Dr. Daum surveyed active Lucky employees on a Wednesday and a Thursday. Although Dr. Hoffman did not know if the responses of terminated employees would have changed the results of Dr. Daum's survey, those responses might have shown a pattern in the job interests of employees who were dissatisfied at Lucky. R.T. at XX-XXXX-XX. However, Dr. Hoffman did not survey terminated employees in the job interest surveys he conducted for Delta Airlines or Bristol-Myers, and he found that the responses of terminated employees did not change the results of his Purolator Courier job interest survey. R.T. at 33-5494.
394. Dr. Hoffman concluded that Dr. Daum's job interest survey was flawed because the sample was not a probability sample, and as a consequence, the analysis done by Dr. Daum cannot be generalized to the population of employees that works at Lucky Stores. Ex. A-242 at 4. Dr. Hoffman's test of lack of fit showed that Dr. Daum's sample was not representative of the stores from which he drew it. In fact, the distribution of jobs in the sample more closely represented the NCD than it did the stores from which Dr. Daum drew it. R.T. at 32-5406. Dr. Hoffman did not know if any of the jobs were statistically significantly overrepresented or underrepresented in the sample. R.T. at XX-XXXX-XX.
*315 395. Dr. Hoffman concluded that Dr. Daum's job interest survey was flawed because the sample Dr. Daum described was a stratified sample. Yet, his statistics did not weight the strata according to their representation in the sample nor did it appear that the standard errors or variances were computed according to his stratified design. Ex. A-242 at 4; R.T. at XX-XXXX-XX. Dr. Hoffman said that by picking the sixteen stores in which the proportion of Black and female employees was closest to average, Dr. Daum made his survey insensitive to any job interest patterns that might be associated with the male/female population in a store. R.T. at 32-5400. In addition, Dr. Daum overweighted the cells that only had a few stores in them. R.T. at 32-5402. When Dr. Hoffman weighted the store populations according to their representation in the strata, he found that the difference between men's and women's endorsement scores for Store Manager increased. Exs. B-32 and A-233; R.T. at XX-XXXX-XX. However, when he weighted the endorsement scores for Store Manager of just Courtesy Clerks, Checkers and Stockers, the difference between men's and women's scores decreased. Exs. A-232 and A-234; R.T. at 32-5412.
396. Dr. Hoffman concluded that Dr. Daum's job interest survey was flawed because his sample was not representative of the occupations employed at Lucky Stores or of the times that people work at Lucky Stores. Ex. A-242 at 4; R.T. at XX-XXXX-XX. Moreover, while four of the eleven job features on the survey relate to work hours and two features relate to the social aspects of a job, only one feature relates to opportunity for advancement and pay respectively. Ex. B-20 at 3; R.T. at 32-5380, XX-XXXX-XX. Dr. Hoffman suggested that because women tend to have stronger familial obligations than do men, Dr. Daum may have given the survey a gender bias by having more work hours oriented job features. R.T. at 32-5381. Dr. Daum responded that the responses of the subjects to the work hours oriented job features did not correlate, therefore he did not believe that the inclusion of those job features biased the survey in any way. R.T. at 33-5554.
397. Dr. Hoffman concluded that Dr. Daum's job interest survey was flawed because his survey never asked employees how they had been treated, what jobs they wanted, what they are willing to do to get those positions, or what they believed to be the likelihood of their obtaining those jobs if they wanted to pursue them. Ex. A-242 at 4; R.T. at XX-XXXX-XX. Rather than asking the subjects to make choices based on the context of their real lives (family, interests outside of work, etc.), the survey substituted an abstract mathematical model. R.T. at XX-XXXX-XX. In addition, while people make job decisions based on concrete information, the survey did not provide concrete information to the subjects (e.g. it was unclear from the information presented whether the appropriate response was that it was important to have a high or a low stress job). R.T. at 32-5282.
398. Dr. Hoffman concluded that Dr. Daum's job interest survey was flawed because Dr. Daum's survey was superficial; it did not focus on change or the components of change. The subjects' responses were clearly associated with the jobs that they had held and with which they were familiar and were not based on their knowledge of individual jobs. Ex. A-242 at 4. In addition, Dr. Hoffman did not think that the fact that Dr. Daum's results were consistent showed that subjects were able to respond to the job features independent of their knowledge about the different jobs. For example, it is unclear why men would rate Deli/Bakery Clerk as their least preferred job with respect to stress. Ex. A-198; R.T. at 32-5386. Dr. Hoffman did not believe that men really considered Deli/Bakery Clerk to be either the most or the least stressful job at Lucky. R.T. at 32-5387.
399. Dr. Hoffman concluded that Dr. Daum's job interest survey was flawed because if the survey results were analyzed using measurement models consistent with the techniques used in the survey, the method of paired comparisons, there was little or no difference between men and *316 women in their desire to obtain management positions. Ex. A-242 at 4. In fact, Dr. Hoffman found the results of Dr. Daum's survey to be somewhat surprising, as he has generally found that men and women rank job features differently. Ex. B-43; R.T. at 32-5384.
400. Dr. Hoffman studied the differences in the job interests of employees who had been with Lucky for long and short periods of time. Ex. A-242 at 7. Looking at the Grocery line of progression, Dr. Hoffman found that women with more than two years of experience at Lucky are less interested in advancement opportunity than were women with less than two years of experience at Lucky. In addition, he found that what had been an insignificant difference between men's and women's interests became statistically significant. Id.; R.T. at XX-XXXX-XX. Only the advancement opportunity job feature showed a significant relative change over time. R.T. at 33-5499. From the survey there is no way of knowing if these changes in attitude are a result of discouragement, of learning what the jobs entail, or some other factor or factors. R.T. at 33-5451.
401. Dr. Hoffman found that most of the people who said that they did not get the first job for which they applied were assigned to positions as Courtesy Clerks (44% women, 56% men). However, 77.5% of the people who did not get the first job for which they applied and were assigned to Deli/Bakery positions were women (thirty-one of forty). 85.7% of the people who did not get the first job for which they applied and were assigned to General Merchandise positions were women (twelve of fourteen). R.T. at 33-5417.
402. Dr. Hoffman found that there was no statistically significant difference between the preference scores for Store Manager of the men and women who worked in the three stores that did not have Deli/Bakery departments. The same was true in the six stores that had had a female Store Manager or Assistant Store Manager in the past three years. R.T. at 33-5422. In seven of the ten stores where there was no significant difference between men's and women's preference scores for Store Manager, there was either no Deli/Bakery department or a woman had been Store Manager or Assistant Store Manager in the recent past. However, the other three stores were large, had Deli/Bakery departments, and had not had a female Store Manager or Assistant Store Manager in the recent past. R.T. at 33-5423, 33-5425. Dr. Hoffman concluded that the treatment of women at Lucky affected their job preferences. R.T. at 33-5423.
403. Dr. Daum standardized the preference scores on a scale between twenty-five and seventy-five. Dr. Hoffman does not believe that the standardization resulted in valid preference scores based on individual responses. R.T. at 33-5426. Because paired comparisons only produce an ordered statistic, Dr. Hoffman believes that Dr. Daum's standardization added a level of complexity to the data that was not justified by the measurement technique. R.T. at XX-XXXX-XX. In his report, Dr. Hoffman averaged the rankings of jobs by job feature in determining endorsement scores instead of weighting the job features based on their rankings. R.T. at 34-5576.
404. Dr. Hoffman did not think that Dr. Daum's calculation of the largest potential management feeder pool, those who chose Fourth Person as their most preferred job, accurately assesses the relative interest of men and women in the feeder jobs to Fourth Person. First, Fourth Person is an entry-level job, not a job to seek in itself. Second, Dr. Daum did not weight the feeder jobs based on the likelihood someone would be promoted to Fourth Person from that job. Ex. B-36; R.T. at XX-XXXX-XX. Dr. Hoffman had similar concerns about Dr. Daum's calculations for exhibits B-37 and B-39. R.T. at 33-5446.
405. Dr. Hoffman found that in the SCD both men and women rated Deli/Bakery *317 and General Merchandise positions statistically significantly higher with regard to the job feature of advancement opportunity than did employees in the NCD. R.T. at 33-5452.
B. Cluster Analysis
406. Dr. Hoffman's cluster analysis was based on the principle that he could group individuals based on the similarity of their responses to Dr. Daum's 396 paired comparisons. Ex. A-242 at 31. He preferred this methodology to Dr. Daum's method of comparing the subject's first preferences because it used all of the information in Dr. Daum's survey. R.T. at 34-5615.
407. Dr. Hoffman ran separate cluster analyses for Courtesy Clerks, and for Checkers/Stockers. Ex. A-242 at 31. He also grouped all of the subjects who were Fourth Person or above (not including Fifth Person) into a management group. R.T. at 34-5582.
408. Dr. Hoffman constructed a similarities matrix on which individuals who responded similarly would be grouped together and those who responded differently would be further away. His object was to evaluate the meaning of the resulting clusters by examining the characteristics of the members of the clusters and correlating those characteristics with those of individuals in the management group. Ex. A-242 at 37.
409. Dr. Hoffman found that the Courtesy Clerks fell into four clusters. The largest cluster was composed of employees who were interested in management positions. 48% of the employees in that cluster were female, while 52% of Courtesy Clerks were female. Dr. Hoffman also found that 42% of the Checkers and Stockers who were most interested in management positions were women. Ex. A-242 at 37.
410. According to the computer program's analysis, the best solution for Courtesy Clerks occurred when there were seven clusters. Ex. B-437, SP006003; R.T. at 34-5592. However, Dr. Hoffman stopped at a four cluster solution based on his independent analysis of the clusters. R.T. at XX-XXXX-XX, XX-XXXX-XX. Regardless, Dr. Hoffman testified that if he had stopped at the seven cluster solution the gender composition in the group that was interested in management would have been the same as it was in the four cluster solution. R.T. at 34-5610.
411. Based on his cluster analysis, Dr. Hoffman concluded that there is a great deal of interest in management on the part of women in Courtesy Clerk, Checker and Stocker positions, and that women are available in numbers that are roughly equal to their representation in those positions. Ex. A-242 at 38.
C. Critique of Dr. Daum's Work Hours Interest Survey
412. Dr. Hoffman found that Dr. Daum's work hours interest survey had many of the same methodological problems as his job interest survey. He said that the survey was not detailed enough and the questions were ambiguous. R.T. at 33-5453. In addition, as the sample for the survey was drawn on one day from four stores in the SCD, Dr. Hoffman could not generalize the results to the entire NCD or SCD. R.T. at 33-5454. Moreover, because the sample was small, the differences between the responses of men and women must be quite large in order to be statistically significant. R.T. at 34-5570.
413. Dr. Hoffman thought that the work hours interest survey was poorly designed. As a result, a significant number of the subjects filled out schedules that provided them with more than forty hours of work a week. Ex. A-242 at 39; R.T. at XX-XXXX-XX. Dr. Daum believed that the fact that some subjects filled out schedules that provided more than forty hours of work indicated that they were able to work forty of the hours which they indicated. R.T. at 34-5570.
414. Dr. Daum calculated the difference between male and female preferences for shifts incorrectly on exhibit B-48. His calculations reflect the actual hours the subjects worked instead of the hours the subjects would ideally have liked to work. *318 R.T. at 33-5462. Even if Dr. Daum had made the correct calculations, Dr. Hoffman did not think that he could conclude that women preferred a day schedule more than men. R.T. at 33-5471.
415. Dr. Hoffman concluded from Dr. Daum's work hours survey that women who worked part-time worked significantly more hours than men, and women and men who worked part-time wanted more hours to the same extent. Ex. A-242 at 39.
CONCLUSIONS OF LAW
I. JURISDICTION AND VENUE
416. This court has jurisdiction in this case pursuant to 28 U.S.C. § 1331, 42 U.S.C. § 2000e, et seq., and 42 U.S.C. § 1981.
417. This court has pendent jurisdiction to hear plaintiffs' claims under the California Fair Employment and Housing Act ("FEHA"). Ackerman v. Western Elec. Co., Inc., 643 F. Supp. 836, 841 n. 1 (N.D.Cal.1986), aff'd, 860 F.2d 1514 (9th Cir.1988).
418. Venue properly lies in the Northern District of California.
II. LIABILITY PERIOD
419. The liability period runs from January 14, 1986 to the present for plaintiff's race claims. Joint Statement ¶ 1.
420. The liability period runs from June 26, 1983 to the present for plaintiff's claims regarding reclassification from part-time to full-time based on sex. Joint Statement ¶ 1.
421. The liability period runs from May 2, 1984 to the present for plaintiff's other sex-based claims. Joint Statement ¶ 1.
III. THE PLAINTIFF CLASS
422. The plaintiff class consists of all past, present, and future Black and female employees of Lucky who have worked, currently work, or may in the future work in the retail food stores within Lucky's NCD. Stipulation and Order Regarding Class Certification at 2 (filed October 11, 1989).
423. At this time the court will not resolve any issues that were left open in the parties' partial class definition.
IV. LEGAL STANDARDS
424. All of plaintiff's race claims under section 1981 have been dismissed or have settled. See September 11, 1991 Order at 35-36; see also Order Granting Joint Motion for Preliminary Approval of Settlement Agreement, Certifying Settlement Class, Approving Form and Method of Class Action Notice, and Granting Scheduling Order (filed July 29, 1992).
Title VII Generally
425. The 1991 Civil Rights Act ("1991 Act") applies to plaintiffs' claims. See Stender v. Lucky Stores, Inc., 780 F. Supp. 1302, 1308 (N.D.Cal.1992).
426. Title VII allows plaintiffs to proceed under either the theory of disparate treatment or the theory of disparate impact. See International B'hood. of Teamsters v. United States, 431 U.S. 324, 335 n. 15, 97 S. Ct. 1843, 1854 n. 15, 52 L. Ed. 2d 396 (1977).
427. Title VII and the regulations promulgated thereunder, require that an employer retain personnel and employment records. 42 U.S.C. § 2000e-8(c) and 29 C.F.R. § 1602.1 et seq.
428. Where an employer has failed to retain records, plaintiff is entitled to an inference that the documents would have supported his or her case. Hicks v. Gates Rubber Co., 833 F.2d 1406, 1418-19 (10th Cir.1987); E.E.O.C. v. Jacksonville Shipyards, Inc., 690 F. Supp. 995 (M.D.Fla. 1988).
Title VII Disparate Treatment
429. A showing of disparate treatment requires proof that plaintiffs were intentionally treated less favorably because of race, color, religion, sex or national origin and proof of discriminatory motive. In a discriminatory treatment class action, plaintiffs must prove by a preponderance of the evidence that sex or race discrimination was the defendant's "standard operating *319 procedure," and was maintained through systematic intentional discrimination. Bazemore v. Friday, 478 U.S. 385, 398, 106 S. Ct. 3000, 3007, 92 L. Ed. 2d 315 (1986) (Brennan, J. concurring); Penk v. Oregon State Bd. of Higher Educ., 816 F.2d 458, 463 (9th Cir.1987), cert. denied, 484 U.S. 853, 108 S. Ct. 158, 98 L. Ed. 2d 113 (1986).
430. To establish a prima facie case of purposeful discrimination a plaintiff must offer evidence that "give[s] rise to an inference of intentional discrimination." Texas Dept. of Community Affairs v. Burdine, 450 U.S. 248, 253, 101 S. Ct. 1089, 1094, 67 L. Ed. 2d 207 (1981). In McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S. Ct. 1817, 36 L. Ed. 2d 668 (1973), the Supreme Court identified one model for establishing a prima facie case of disparate treatment. However, the facts a plaintiff must assert to raise an inference of discrimination will necessarily vary depending upon the situation. Foster v. Arcata Assoc., Inc., 772 F.2d 1453, 1460 (9th Cir.1985), cert. denied, 475 U.S. 1048, 106 S. Ct. 1267, 89 L. Ed. 2d 576 (1986). The principal requirement is simply that the plaintiff "must carry the initial burden of offering evidence adequate to create an inference that an employment decision was based on a discriminatory criterion illegal under the Act." Teamsters, 431 U.S. at 358, 97 S. Ct. at 1866.
431. Proof of discriminatory intent may be direct, circumstantial, or may be inferred from statistical evidence. See Green v. USX Corp., 896 F.2d 801, 807 (3d Cir.1990) (disparate impact of employer's challenged actions may be highly relevant in evaluation of disparate treatment claim), cert. denied, ___ U.S. ___, 111 S. Ct. 53, 112 L. Ed. 2d 29 (1990); see also Village of Arlington Heights v. Metropolitan Housing Dev. Corp., 429 U.S. 252, 97 S. Ct. 555, 50 L. Ed. 2d 450 (1977) (statistical evidence of racial disparity may be probative of purposeful discrimination). "All evidence that a plaintiff presents can contribute to this inference, and should therefore be considered as cumulative." Segar v. Smith, 738 F.2d 1249, 1278 (D.C.Cir.1984), cert. denied, 471 U.S. 1115, 105 S. Ct. 2357, 86 L. Ed. 2d 258 (1985).
432. Inferences drawn from statistical proof are strongest when they are substantiated by other evidence in the case which brings "the cold numbers convincingly to life." Teamsters, 431 U.S. at 339, 97 S. Ct. at 1356 (plaintiffs succeeded in establishing prima facie case by using statistical data bolstered by anecdotal evidence of specific instances of discriminatory treatment). The relevance and usefulness of statistical proof depends upon all of the surrounding facts and circumstances. Hazelwood School Dist. v. United States, 433 U.S. 299, 312, 97 S. Ct. 2736, 2744, 53 L. Ed. 2d 768 (1977).
433. The use of subjective factors to evaluate employees for placement, promotion, or training is not per se prohibited by Title VII. Ward v. Westland Plastics, Inc., 651 F.2d 1266, 1270 (9th Cir.1980). However, "`[i]n forbidding employers to discriminate against individuals because of their sex, Congress intended to strike at the entire spectrum of disparate treatment of men and women resulting from stereotypes.'" Price Waterhouse v. Hopkins, 490 U.S. 228, 251, 109 S. Ct. 1775, 1791, 104 L. Ed. 2d 268 (1989) (quoting Los Angeles Dept. of Water & Power v. Manhart, 435 U.S. 702, 707, n. 13, 98 S. Ct. 1370, 1375, n. 13, 55 L. Ed. 2d 657 (1978)).
434. In addition, the Ninth Circuit has recognized that "`subjective practices are particularly susceptible to discriminatory abuse and should be closely scrutinized.'" Jauregui v. City of Glendale, 852 F.2d 1128, 1136 (9th Cir.1988) (quoting Atonio v. Wards Cove Packing Co., Inc., 810 F.2d 1477, 1481 (9th Cir.1987), rev'd on other grounds, 490 U.S. 642, 109 S. Ct. 2115, 104 L. Ed. 2d 733 (1989) (the Supreme Court did *320 not consider the issue of subjective criteria)); see also 1991 Act § 105(a). Thus, the court in Jauregui held that where the potential for manipulation inherent in the use of subjective evaluations is high, courts can infer the intent necessary to establish a claim for disparate treatment. See also Domingo v. New England Fish Co., 445 F. Supp. 421, 436 (W.D.Wash.1977), aff'd, 727 F.2d 1429 (9th Cir.1984).
435. If the plaintiff succeeds in establishing a prima facie case, the burden of production shifts to the defendant to articulate some legitimate, nondiscriminatory reason for the adverse employment decision. Should the defendant succeed in articulating such a reason, the plaintiff can still prevail by demonstrating that the proffered reason is a pretext for a discriminatory motive. Lowe v. City of Monrovia, 775 F.2d 998, 1005 (9th Cir.1985), as amended, 784 F.2d 1407 (9th Cir.1986).
Title VII Disparate Impact
436. A disparate impact plaintiff, unlike a plaintiff proceeding under the disparate treatment theory, may prevail without proof of intentional discrimination by proving that employment practices that are fair in form are discriminatory in practice. Griggs v. Duke Power Co., 401 U.S. 424, 431, 91 S. Ct. 849, 853, 28 L. Ed. 2d 158 (1971). The plaintiff may establish a prima facie case of disparate impact by: 1) identifying the specific employment practices being challenged; 2) establishing disparate impact on a protected group; and 3) demonstrating that the disparity is the causal result of one or more of the employment practices identified. Wards Cove Packing Co., Inc. v. Atonio, 490 U.S. 642, 656-57, 109 S. Ct. 2115, 2124-25, 104 L. Ed. 2d 733 (1989); Watson v. Forth Worth Bank & Trust, 487 U.S. 977, 994-95, 108 S. Ct. 2777, 2788-89, 101 L. Ed. 2d 827 (1988). In addition, the 1991 Civil Rights Act provides that "if the complaining party can demonstrate to the court that the elements of a respondent's decisionmaking process are not capable of separation for analysis, the decisionmaking process may be analyzed as one employment practice." 1991 Act § 105(a).
437. Even before the enactment of the 1991 Civil Rights Act, the Supreme Court had held that disparate impact analysis may be applied to subjective decision making and to subjective employment criteria as well as to objective or standardized tests. Watson, 487 U.S. at 990, 108 S. Ct. at 2786-87. "If an employer's undisciplined system of subjective decisionmaking has precisely the same effects as a system pervaded by impermissible intentional discrimination, it is difficult to see why Title VII's proscription against discriminatory actions should not apply." Id. at 990-91, 108 S. Ct. at 2787.
438. In Allen v. Seidman, 881 F.2d 375 (7th Cir.1989), plaintiffs challenged the disparate impact of a bank examiner test in which testers based a portion of the evaluation on an unstructured personal interview. The Seventh Circuit upheld the district court's finding that the plaintiff class had established a prima facie case of disparate impact based upon the evaluation procedure. Id. at 381. The court reasoned that
[i]n a test notably devoid of objective standards, where far from using blind grading the testers based an unknown part of the grade on the results of an unstructured personal interview, the danger is acute that racial bias of which the testers may well be unconscious will influence the grade.... The subjectivity of [the test] deprived the testers of better information and may have inclined them to fall back on race and on vocationally irrelevant cultural factors correlated with race; if so the test was discriminatory in an uncontroversial sense.
439. Courts have found the causal requirement with respect to discretionary and subjective hiring practices to be satisfied where the employer had no written or otherwise established selection guidelines and the office manager was unable to identify the significant factors in evaluation and selection of applicants, E.E.O.C. v. Rath Packing Co., 787 F.2d 318, 328 (8th Cir.), cert. denied, 479 U.S. 910, 107 S. Ct. 307, 93 L. Ed. 2d 282 (1986); where subjective promotion practices were used, vacancies were *321 not posted and a recently adopted system to communicate vacancies was found to be incomplete and untimely, Paxton v. Union Nat'l. Bank, 688 F.2d 552, 563-64 (8th Cir.1982), cert. denied, 460 U.S. 1083, 103 S. Ct. 1772, 76 L. Ed. 2d 345 (1983); where there was no official application form and exclusive reliance on word-of-mouth recruiting, Davis v. Richmond, Fredericksburg & Potomac R.R. Co., 593 F. Supp. 271, 278 (E.D.Va.1984), aff'd in part, rev'd and vacated in part on other grounds, 803 F.2d 1322 (4th Cir.1986); and where the employer lacked objective criteria or guidelines for promotion to position of locomotive engineer, and the decision was left to the discretion of a small number of male supervisors. Davis, 593 F. Supp. at 279.
440. In such cases, the courts have emphasized that an employer's failure to articulate and apply objective criteria "could only reinforce the prejudices, unconscious or not, which Congress in Title VII sought to eradicate as a basis for employment." Rath Packing, 787 F.2d at 328. Moreover, a system in which promotional opportunities are not posted but rather publicized by word-of-mouth is a discriminatory practice because of its "tendency to perpetuate the all-male composition of higher prestige, better paying jobs." Davis, 593 F. Supp. at 278-79.
441. An employer's policy of leaving initial placement, promotion and training decisions to the discretion of lower level supervisors does not, by itself, raise an inference of discriminatory conduct. Watson, 487 U.S. at 990, 108 S. Ct. at 2786-87. However, in Watson, the Supreme Court found that evidence of discretionary and subjective decision making was supplemented by evidence that subconscious stereotypes and prejudices operated in the vacuum created by the absence of objective, publicized criteria. In Watson, a Black plaintiff had been told that the teller position involved major responsibility with "a lot of money ... for blacks to have to count." Id. Although the Court found that such remarks might not in themselves be sufficient to prove discriminatory intent, they do suggest "a lingering form of the problem that Title VII was enacted to combat." Id.
442. Once the plaintiff has established a prima facie case of disparate impact, the 1991 Civil Rights Act shifts both the burden of production and the burden of proof to the employer to show that a challenged employment practice is job related and is consistent with business necessity. See 1991 Act § 105. This allocation of the burden of proof is consistent with case law prior to the Supreme Court's decision in Wards Cove.See Albemarle Paper Co. v. Moody, 422 U.S. 405, 431, 95 S. Ct. 2362, 2378, 45 L. Ed. 2d 280 (1975); Griggs v. Duke Power Co., 401 U.S. at 432, 91 S. Ct. at 854.
443. Section 105(b) of the 1991 Civil Rights Act defines "business necessity" by reference to an interpretive memorandum. That memorandum states:
The terms "business necessity" and "job related" are intended to reflect the concepts enunciated by the Supreme Court in Griggs v. Duke Power Co., 401 U.S. 424 [91 S. Ct. 849, 28 L. Ed. 2d 158] (1971), and the other Supreme Court decisions prior to Wards Cove Packing Co. v. Atonio, 490 U.S. 642 [109 S. Ct. 2115, 104 L. Ed. 2d 733] (1989).
137 Cong.Rec. S 15276 (daily ed. Oct. 25, 1991). Under Griggs, in order to prove business necessity, an employer must show that its selection criteria bear "a manifest relationship to the employment in question."Griggs, 401 U.S. at 432, 91 S.Ct. *322 at 854. The employer must also demonstrate that the employment practice significantly serves legitimate employment goals. See New York City Transit Authority v. Beazer, 440 U.S. 568, 587 n. 31, 99 S. Ct. 1355, 1366 n. 31, 59 L. Ed. 2d 587 (1979). The employer is not required to show that those employment goals "require" the employment practice. See id.
444. Under the 1991 Civil Rights Act, if a defendant successfully mounts a defense of business necessity a plaintiff may rebut that defense by demonstrating that there exists an alternate employment practice which serves the employer's business necessity, but does so without causing a disparate impact, and that the employer refuses to adopt that alternative employment practice. 1991 Act § 105(a). "Alternative employment practice" is defined in the Act as that practice consistent with "the law as it existed on June 4, 1989," 1991 Act § 105(a), the day before the Supreme Court issued its opinion in Wards Cove. Before Words Cove the Supreme Court had described the "alternate business practice" rebuttal:
[T]he complaining party [may] show that other tests or selection devices, without a similarly undesirable [discriminatory] effect, would also serve the employment's legitimate interest in "efficient and trustworthy workmanship."
Albemarle, 422 U.S. at 425, 95 S. Ct. at 2375 (quoting McDonnell Douglas, 411 U.S. at 801, 93 S.Ct. at 1823).
445. The 1991 Civil Rights Act does not alter plaintiffs' burden of proving that an employment practice has a disparate impact on a protected group. See 1991 Act § 105(a). Nor does the Act change the rule that a statistical analysis comparing segments of an employer's workforce is inadequate to carry a plaintiff's burden of proof. See Wards Cove, 490 U.S. at 653-55, 109 S. Ct. at 2122-24.
The Use of Statistical Evidence in Title VII Cases
446. Both the Supreme Court and the Ninth Circuit have approved the use of statistical evidence, specifically regression analyses, to prove discrimination in Title VII cases. Bazemore v. Friday, 478 U.S. at 400, 106 S.Ct. at 3008-09; E.E.O.C. v. General Telephone Co. of Northwest, Inc., 885 F.2d 575, 579-82 (9th Cir.1989), cert. denied, ___ U.S. ___, 111 S. Ct. 370, 112 L. Ed. 2d 332 (1990).
447. As a threshold matter, "[t]he proponent of [a] survey bears the burden of establishing its admissibility." Keith v. Volpe, 858 F.2d 467, 480 (9th Cir.1988) (citing Toys "R" Us, Inc. v. Canarsie Kiddie Shop, 559 F. Supp. 1189, 1205 (E.D.N.Y.)). That party must show that "the survey was conducted in accordance with generally accepted survey principles and that the results were used in a statistically correct manner." Id. (citing Baumholser v. Amax Coal Co., 630 F.2d 550, 552 (7th Cir.1980)).
448. "Technical inadequacies in the survey, including the format of the questions or the manner in which it was taken, bear on the weight of the evidence, not its admissibility." Id. (citing Prudential Insurance Co. v. Gibraltar Financial Corp., 694 F.2d 1150, 1156 (9th Cir.1982), cert. denied, 463 U.S. 1208, 103 S. Ct. 3538, 77 L. Ed. 2d 1389 (1983)).
449. Statistical evidence must indicate a "longlasting and gross" disparity before the trier of fact may infer that the disparity between the defendant's workforce and the relevant labor market was due to a pattern or practice of intentional discrimination. Teamsters, 431 U.S. at 339 n. 20, 97 S. Ct. at 339 n. 20; Gay v. Waiters' and Dairy Lunchmen's Union, 694 F.2d 531, 551 (9th Cir.1982).
450. A defendant may not rebut an inference of discrimination "by merely pointing to flaws in the plaintiff's statistics." General Telephone, 885 F.2d at 581. Rather, the defendant must introduce "`evidence to support the contention that the missing factor can explain the disparities as a product of a legitimate, nondiscriminatory selection criterion.'" Id. at 580 (quoting Palmer v. Shultz, 815 F.2d 84, 101 (D.C.Cir.1987)).
451. A disparity between the number of men and women in certain job classifications *323 may be attributed to one of three possible causes: the effect of legitimate non-discriminatory selection criteria, unlawful discriminatory animus, or chance. The question is at what point is a disparity sufficiently large, and the probability that the disparity was caused by chance sufficiently low, that an inference of discrimination can be drawn solely from the fact of the disparity. Many courts have followed the social science convention which holds that for disparities below a 5% probability level ("P-value"), chance explanations become suspect. D. Baldus & J. Cole, Statistical Proof of Discrimination at 291 (1980); see also Castaneda v. Partida, 430 U.S. 482, 496 n. 17, 97 S. Ct. 1272, 1281 n. 17, 51 L. Ed. 2d 498 (1977); Palmer v. Shultz, 815 F.2d at 92.
452. A P-value indicates the statistical significance of a disparity on a scale ranging from 0 to 1. The level of significance rises as the P-value declines; thus, a difference in selection rates which is significant at the .01 or 1% level is of greater statistical significance than a difference which is significant at the .05 or 5% level. Statistical Proof at 308. For large samples, there is a 1 in 20 chance that a disparity which is significant at the 5% level will occur by chance and there is a 1 in 100 chance that a disparity which is significant at the 1% level will occur by chance. Id. at 297.
453. In Castaneda, 430 U.S. at 496 n. 17, 97 S. Ct. at 1281 n. 17 and Hazelwood, 433 U.S. at 308 n. 14, 311 n. 17, 97 S. Ct. at 2742 n. 14, 2743 n. 17, the Supreme Court approved the rule that a disparity of two or three standard deviations was sufficient to establish a prima facie case of disparate treatment. For large samples, a disparity of two standard deviations is usually significant at the .05 or 5% level and a disparity of three standard deviations is significant at the .01 or 1% level. Statistical Proof at 297.
454. A Z-value is the number of standard deviations that an observed value is from an expected value. A positive Z-value indicates that an observed value is greater than was expected and a negative Z-value indicates that an observed value is less than was expected. Barnes, Statistics as Proof at 198-99. Z-values of five to six and twelve have been held to be sufficient to establish an inference of intentional discrimination. Gay, 694 F.2d at 551. A disparity of just over three standard deviations has been held to support an inference of discrimination. Hillery v. Pulley, 563 F. Supp. 1228, 1245 (E.D.Cal.1983). Two standard deviations have also been held to be sufficient to establish liability. Palmer v. Shultz, 815 F.2d at 92. However, the Ninth Circuit has advised that courts must be "extremely cautious" about drawing inferences from statistical deviations in the range of one to three. Gay, 694 F.2d at 551.
455. Statisticians can employ either one or two-tailed tests in measuring significance levels. The terms one-tailed and two-tailed indicate whether the significance levels are calculated from one or two tails of a sampling distribution. Two-tailed tests are appropriate when there is a possibility of both overselection and underselection in the populations that are being compared. One-tailed tests are most appropriate when one population is consistently overselected over another. The practical difference between one and two tailed tests is that the P-value produced by a two-tailed test is usually twice as great as that produced by a one-tailed test. Statistical Proof at 307.
*324 Punitive Damages under Title VII
456. In an order dated July 14, 1992, this court held that it would not result in manifest injustice to subject defendant to the punitive damages provision established by the 1991 Civil Rights Act. The court reasoned that defendant has always been subject to compensatory and punitive damages under FEHA and 42 U.S.C. § 1981 for the same conduct; and should have expected to be subject to much greater liability for punitive damages under those statutes. While defendant was exposed to unlimited punitive damages liability under FEHA and section 1981, it can only be held liable for $300,000 in punitive damages under the 1991 Civil Rights Act. 1991 Act § 102(b) (3) (D). Moreover, since a showing of unlawful intentional discrimination is a prerequisite to the award of punitive damages, defendant can hardly claim that although it was aware of the wrongfulness of its conduct it relied on the fact that its violations of plaintiffs' civil rights would be inexpensive. Title VII prohibits discrimination, it does not impose a licensing fee for the privilege of continued discrimination. Accordingly, the court held that the retroactive application of the new punitive damages provision of Title VII would not result in "manifest injustice." July 14, 1992 Order at 10-11 (quoting Bradley v. School Board of the City of Richmond, 416 U.S. 696, 711, 94 S. Ct. 2006, 2016, 40 L. Ed. 2d 476 (1974)).
457. The standard for proving punitive damages under the 1991 Civil Rights Act is easier to satisfy than the standard under FEHA or section 1981. The Title VII punitive damages provision allows the award of punitive damages for "reckless indifference." 1991 Act § 102(b) (1). This evidentiary standard is more lenient than that under either FEHA or 42 U.S.C. § 1981; under FEHA only "oppression, fraud, or malice" are actionable, Cal.Civ.Code § 3294(a), (c), and under section 1981, a defendant may be liable for conduct "motivated by evil motive or intent, or when it involves reckless or callous indifference to the federally protected rights of others," Smith v. Wade, 461 U.S. 30, 56, 103 S. Ct. 1625, 1640, 75 L. Ed. 2d 632 (1983).
458. In addition, although the Act contains no explicit burden of proof for punitive damages claims, 1991 Act § 102, "[c]onventional rules of civil litigation generally apply in Title VII cases ... and one of these rules is that parties to civil litigation need only prove their case by a preponderance of the evidence." Price Waterhouse v. Hopkins, 490 U.S. at 253, 109 S. Ct. at 1792.
459. There appears to be no distinction between the standard for establishing a right to punitive damages and the standard for establishing liability for disparate treatment. Congress could have made such a distinction, but chose not to. The burden of proof, preponderance of the evidence, is the same for both claims. In addition, if a plaintiff were able to establish the intentional discriminatory conduct required to prove disparate treatment, s/he would by definition have satisfied the requirement of showing the "reckless indifference" required for an award of punitive damages. Although the 1991 Civil Rights Act does not define "reckless indifference," punitive damages are available under both section 1981 and section 1983 upon a showing of intentional violation of federal law. Wade, 461 U.S. at 51, 103 S.Ct. at 1637-38; See also Yarbrough v. Tower Oldsmobile, Inc., 789 F.2d 508, 514 (7th Cir.1986). In fact, *325 the Wade court equated reckless or callous disregard with intentional discrimination and held that either was "sufficient to trigger a jury's consideration of the appropriateness of punitive damages." Wade, 461 U.S. at 51, 103 S. Ct. at 1637-38. In addition, the Wade court noted that "[t]here has never been any general common-law rule that the threshold for punitive damages must always be higher than that for compensatory liability." Id. at 53, 103 S. Ct. at 1638.
California Fair Employment and Housing Act
460. The California Fair Employment and Housing Act ("FEHA") provides:
It shall be an unlawful employment practice, unless based upon an bona fide occupational qualification, or, except where based upon applicable security regulations established by the United States or the State of California:
(a) For an employer, because of the race, religion, creed, color, national origin, ancestry, physical handicap, medical condition, marital status, or sex of any person, to refuse to hire or employ the person or to refuse to select the person for a training program leading to employment, or to bar or to discharge the person from employment or from a training program leading to employment, or to discriminate against the person in compensation or in terms, conditions or privileges of employment.
California Government Code § 12940.
461. A plaintiff may establish a prima facie case under FEHA by showing that the defendant's practice has had a disparate impact on a protected group. The burden of proof then shifts to the employer to demonstrate that the practice is a business necessity, which is valid and job-related. Plaintiff may still prevail if s/he can show that the employment practice was merely a pretext for discrimination. Title 2, California Code of Regulations, Section 7286.7(b); F.E.H.C. v. City and County of San Francisco, FEHC Dec. No. 82-11, at 25 (1982), decision aff'd, City and County of San Francisco v. F.E.H.C., 191 Cal. App. 3d 976, 236 Cal. Rptr. 716 (1987).
462. In the course of ruling on the parties' motions in limine in this case, this court stated that it would hear plaintiffs' pendent state claims under FEHA only if those claims would be tried under Title VII standards and burdens of proof. September 11, 1991 Order at n. 2; R.T. at 1-4. Thus, the court's rulings under Title VII apply with equal force to plaintiffs' claims under FEHA, except insofar as they relate to punitive damages.
V. THE USE OF INTEREST SURVEYS
463. Defendant offers the results of a job interest survey in response to plaintiffs' evidence showing statistically significant disparities between male and female Lucky employees in initial placement, promotion, movement from part-time to full-time and the assignment of additional hours. Defendant argues that its job interest survey proves that plaintiffs' statistics measure the impact of average gender differences in job interest, not the impact of discrimination. Moreover, defendant claims to have presented credible evidence of differing job interests between its male and female employees which explains the statistical disparities which plaintiffs' allege.
464. The Ninth Circuit has held that "[w]hile failure to include ... employment interests may render the regression analysis less precise, merely pointing to such an imperfection does not, without more, defeat a showing of intentional discrimination established by the regression analyses." General Telephone, 885 F.2d at 582. Rather, the defendant's burden of proof is "to produce credible evidence that curing the alleged flaws would also cure the statistical disparity." Id. at 583. There is an especially heavy burden on defendant "if the survey purports to represent subjective data, such as attitudes or beliefs, because of the inherent difficulty of accurately measuring such data." Richardson v. Quick Trip Corp., 591 F. Supp. 1151, 1153 (S.D.Iowa 1984) (citing United States Handbook of Recommended Procedures *326 for the Trial of Protracted Cases, 25 F.R.D. 351, 428-29 (1960)).
465. As defendant points out, evidence of interest in the job has always been a part of Title VII cases. The second prong of the four-part test in McDonnell Douglas, 411 U.S. at 802, 93 S. Ct. at 1824, requires proof that the plaintiff applied for the position in question. The Ninth Circuit has stated that the proper measure of the pool of eligible employees should always be the actual applicants unless there is proof of a selection practice that makes use of the actual pool of applicants inappropriate. Moore v. Hughes Helicopters, Inc., 708 F.2d 475, 482 (9th Cir.1983).
466. Therefore, job interest surveys should only be used when it is impossible to establish the make up of the actual or relevant applicant pool. In this case, the database of job applications which was available to the parties was faulty, see infra ¶ 482, and the available bid lists were incomplete, Joint Statement ¶ 50. Defendant has, therefore, replaced this actual application data with the results of job interest and work hours interest surveys.
467. The use of interest surveys to reconstruct missing applicant pool data allows generalizations about the interests of men and women to be assumed in the absence of an actual applicant pool. See E.E.O.C. v. Sears, Roebuck & Co., 839 F.2d 302, 361 (7th Cir.1988) (Cudahy, J. concurring in part and dissenting in part). That fact is particularly relevant in this case because Dr. Daum's survey did not include subjects who left or were terminated by Lucky. Therefore, the survey could not adequately account for the possibility that men and women's work preferences may be influenced differently by their experiences on the job.
468. Of course, an employer cannot be held liable for the fact that women's job interests may be tainted by societal values and influences. Davis v. City of Dallas, 483 F. Supp. 54, 61 (N.D.Tex.1979). By the same token, however, an employer cannot justify the disparate treatment of its male and female employees simply by claiming that the employees wish to be treated differently.
469. To accept the use of interest surveys as a defense in disparate treatment and disparate impact cases is to accept that an employer is permitted to discriminate against the minority of women who are interested in seeking non-traditional employment as long as the majority of women are not interested in such work. Even in a situation where gender stereotypes about work interest patterns reflect reality, it is unlawful for an employer to discriminate against those whose work interests deviate from the stereotype. Therefore, the court holds that job interest surveys cannot be used as a defense in disparate treatment cases. Defendant's survey only has evidentiary weight as a rebuttal to plaintiff's statistical argument in its disparate impact claim.
470. In rebuttal, interest surveys may be used to explain the statistical disparities between men and women which plaintiff alleges. However, anecdotal evidence of disparate treatment or of disparate impact cannot be rebutted by job interest surveys. Generalizations about women's job interests cannot be used to trump the testimony of individual women about their job interests.
471. One purpose of Title VII is to promote gender-equal opportunity in the work force. See Alexander v. Gardner-Denver Co., 415 U.S. 36, 44, 94 S. Ct. 1011, 1017-18, 39 L. Ed. 2d 147 (1974). Insofar as women's attitudes towards work are shaped in part by the past discrimination of their employer, the results of job interest surveys must reflect that past discrimination. To give much credence to proof which is tainted by the lingering effects of *327 the past discrimination of an employer would be to prevent courts from recognizing and taking into consideration the very evolution in women's work interests which Title VII was enacted to encourage.
472. For the reasons stated above, the court has grave reservations about the propriety of admitting job interest surveys into evidence in Title VII cases. Those courts which have admitted interest surveys into evidence have assumed their admissibility without discussion. See E.E.O.C. v. Sears, Roebuck & Co., 628 F. Supp. 1264 (N.D.Ill.1986); E.E.O.C. v. Mead Foods, Inc., 466 F. Supp. 1, 3-4 (W.D.Okla.1977); but see E.E.O.C. v. Sears, Roebuck & Co., 839 F.2d at 360-66 (Cudahy, J. concurring in part and dissenting in part). Despite this court's reservations, and in the absence of any countervailing authority, this court will give defendant the benefit of the doubt and allow the above considerations to go only to weight. Thus, the court will assess the weight to be given to defendant's rebuttal argument that the gender-based disparities which plaintiff alleges are accounted for when adjustments are made to control for differences in men's and women's work interests.
VI. FINDINGS ON EXPERT TESTIMONY
473. The sociological evidence which Dr. Bielby presented to this court is consistent with evidence that has been accepted by other courts. See e.g. Hopkins v. Price Waterhouse, 618 F. Supp. 1109, 1117-18 (D.C.D.C.1985), aff'd after remand, 920 F.2d 967 (D.C.Cir.1990). Dr. Daum criticized Dr. Bielby's reliance on studies that involved populations unlike that at Lucky. However, the studies on which Dr. Bielby relied have been found to have external validity, supra Finding of Fact ("FF") ¶ 312, and Dr. Bielby's conclusions were consistent with Jane't Noriega-Ailor's experiences in the Grocery industry, supra FF ¶ 194. Therefore, the court finds Dr. Bielby's testimony to be persuasive.
474. The court finds Dr. Pencavel's report to be persuasive. Defendant did not make a significant effort to disprove his conclusion that the economic status of women at Lucky is inferior to that of men. Supra FF ¶ 287; but see supra FF ¶¶ 386, 387. Moreover, the statistical disparities which Dr. Pencavel found were unexplained by any other evidence presented in this case.
475. The court finds Dr. Drogin's statistics to be reliable and finds his explanation for the statistical disparities between the placement of men and women at Lucky to be more persuasive than Dr. Daum's.
476. The court finds the results of Dr. Daum's job interest survey to be unpersuasive for the following reasons:
a. Dr. Daum did not survey any Lucky employees who had resigned or been terminated, supra FF ¶ 347, although the responses of those employees might have shown a pattern in the job interests of employees who were dissatisfied at Lucky.
b. Dr. Daum's sample was stratified. He surveyed employees at the sixteen stores in which the proportion of Black and female employees was closest to average. Supra FF ¶ 325. Thus, Dr. Daum made his survey insensitive to any job interest patterns that might be associated with the male/female population in a store and overweighted the cells that only had a few stores in them. Supra FF ¶ 395.
c. Four of the eleven job features on Dr. Daum's job interest survey related to work hours, while only two features related to the social aspects of a job, and only one feature related to opportunity for advancement and pay respectively. Supra FF ¶ 318. Thus, the survey exaggerated any difference between men's and women's work interest that was based on work hour preferences. Supra FF ¶ 396.
d. The job interest survey did not directly ask the subjects how they had been treated at Lucky, what jobs they wanted, what they are willing to do to get those jobs, or what they believed to be the likelihood of their obtaining those jobs if they were to pursue them. Supra FF ¶ 329. In addition, the survey presented abstract choices to the subjects instead of presenting the type of concrete information on *328 which people rely in making job choices. Supra FF ¶ 397.
e. Some of the responses to the job interest survey were inconsistent with Dr. Daum's assertion that the subjects understood the purpose of, and terms used in, the survey. For example, it is unclear why men would rate Deli/Bakery Clerk as their least preferred job with respect to stress. Supra FF ¶ 337. In addition, the fact that subjects tended to rate their incumbent positions higher than they rated other jobs shows that the survey responses were tainted by existing job patterns. Supra FF ¶ 335.
f. The court also questions the method which Dr. Daum used to standardize the subjects' preference scores.Supra FF ¶ 330. As paired comparisons only produce an ordered statistic, Dr. Daum's standardization added a level of complexity to the data that was not justified by his measurement technique. Supra FF ¶ 403.
g. Dr. Daum's feeder pool for management positions was constructed so that it was possible that a person who rated Receiving Clerk as their first choice and Fourth Person as their second choice would not appear in the feeder pool, while a person who gave a lower preference score to the Fourth Person job but rated it as their first choice would appear in the feeder pool. Supra FF ¶ 342.
h. Dr. Daum concluded that management did not differentially influence the job interests of male and female Lucky employees. He concluded that if it had, the job feature preferences of the women employees who were blocked from attaining the positions which they most desired would be different from the job feature preferences of the men who had not faced blocked opportunities. Supra FF ¶ 345. However, it is unclear to the court why Dr. Daum expected blocked opportunity to be reflected in the subjects' preference of job features but not in their preference of jobs, as the subjects based their job preferences on an analysis of job features. The court also notes that Dr. Daum's finding that men's standardized preference scores for Fourth Person and Store Manager drop statistically significantly after two years at Lucky while the standardized preference scores for Fourth Person and Store Manager of women remain unchanged, supra FF ¶ 346, is inconsistent with his conclusion that men are more interested in management positions than are women.
477. The court finds that Dr. Daum did not validate the results from his NCD job interest survey by comparing them to the results from his Albertson's survey or the survey he took in the SCD. The court finds that there were significant differences between the survey instrument which Dr. Daum used for the Albertson's study and the one he used for Lucky's NCD. Supra n. 15. Moreover, there are significant differences between Lucky's northern and southern divisions that make any comparison between the results of the surveys taken in those two divisions misleading. Supra FF ¶¶ 181-187.
478. Most importantly, the court finds that the results of Dr. Daum's survey do not sufficiently explain the gross disparities between Lucky's male and female employees found by plaintiffs. Accordingly, the court will exercise caution in attaching significant weight to Dr. Daum's job interest survey.
479. The court finds the results of Dr. Daum's work hours interest survey to be unpersuasive for the following reasons:
a. Dr. Daum's work hours interest survey had many of the same methodological problems as his job interest survey. For example, the survey was not detailed enough and the questions asked were ambiguous. In addition, the sample for the survey was small and was drawn from Lucky's SCD, supra FF ¶ 350; the court received much testimony that the NCD and the SCD were significantly different, supra FF ¶¶ 181-187, so as to make the responses *329 of employees from one division ungeneralizable to employees from the other division.
b. Dr. Daum's work hours interest survey was poorly designed, so that a significant number of subjects filled out schedules that provided them with more than forty hours of work a week. Supra FF ¶ 352.
c. Defendant failed to reconcile its contention that women and men have different levels of interest in night work with Dr. Daum's finding that there is no statistically significant difference between men's and women's preference for Fourth Person, which is a night job. Ex. B-31.
480. Accordingly, the court will exercise much caution in attaching significant weight to Dr. Daum's work hours interest survey.
481. The court finds that the initial placement database on which Dr. Haworth based her analysis was unreliable for the following reasons:
a. Dr. Haworth's initial placement database did not conform to the sampling plan to which the parties stipulated. The sample which Dr. Haworth obtained contained about 24% fewer applications than prescribed in the original sample plan. Supra FF ¶ 364.
b. Dr. Drogin found that the sample which Dr. Haworth obtained had a statistically significantly unequal distribution of applications by job category, by year, and by age, compared to actual hiring rates. Exs. A-167, A-245, A-243 and A-246; R.T. at XX-XXXX-XX, XX-XXXX-XX, and XX-XXXX-XX.
c. Dr. Haworth's sample did not meet the 3% margin of allowable error for hires of Deli/Bakery, Produce and Grocery Apprentices. The margin of error for Night Crew Apprentices exceeded 5%. Supra FF ¶ 365.
d. Dr. Haworth only coded an application as a night application if the word "night" was specifically used on the form. However, she admitted that night workers do more stocking than do day workers, and that "stocker" applications were a larger proportion of night hires than were "night" applicants. In addition, Dr. Haworth found that of those she considered to have applied for night jobs, less than 30% received night positions. Supra FF ¶¶ 367, 372.
e. Dr. Haworth testified that the results of Dr. Daum's job interest survey for the positions of Courtesy Clerk, Deli/Bakery and General Merchandise Clerk, and Night Crew, confirmed the reliability of her initial placement database sample. See supra n. 12. However, their results for the percentage of applicants for other job titles were not very close: 11.03% of the women in the initial placement database sample said that they first applied for a Produce Clerk position, compared to 0% of the women in Dr. Daum's survey. Likewise, 9.5% of the women in the initial placement database sample said that they first applied for a Janitor position, compared to 0% of the women in Dr. Daum's survey. See supra n. 17.
482. Thus, the court finds that Dr. Haworth's initial placement database is unreliable, and that the conclusions which she drew from this database are also unreliable.
483. Irrespective of the reliability of Dr. Haworth's initial placement database sample, Dr. Haworth testified that Store Managers do not consider an employee's initial application in making promotion decisions. In addition, she believed that employee job preferences are likely to change after their initial placement. Therefore, it is clear to the court that the job interests of Lucky employees are not seriously considered in placement and promotion decisions.
484. Irrespective of the reliability of Dr. Haworth's initial placement database sample, Dr. Haworth had previously testified in E.E.O.C. v. Sears Roebuck & Co., 628 F. Supp. 1264 (N.D.Ill.1986), aff'd, 839 F.2d 302 (7th Cir.1988), that the number of applications for certain jobs is an unreliable indicator of interest when the same application form is submitted for different types of jobs and when most forms contain insufficient information from which to determine whether the applicant is interested, *330 qualified, and available for the position. Supra FF ¶ 371.
485. Moreover, the court finds that differences in the job interests of men and women are not the sole cause of the statistical disparities which plaintiffs found. When Dr. Drogin compared men and women who had filled out their applications in a similar fashion, he still found a statistically significant disparity in their hire rates, Z-value of 18.96. Ex. A-248; R.T. at XX-XXXX-XX. Accordingly, the court will exercise much caution in attaching significant weight to the findings of Dr. Haworth.
486. While the court finds Dr. Hoffman's critique of Dr. Daum's report to be persuasive, the court is not persuaded by Dr. Hoffman's cluster analysis. Dr. Hoffman's cluster analysis was based on the data which Dr. Daum obtained from his job interest survey. Thus, the cluster analysis must be tainted with many of the same flaws as are the job interest survey results.
VII. DISPARATE TREATMENT
Evidence of Intent
A. Evidentiary Rulings
487. On July 22, 1991, defendant made a motion under Rule 407 of the Federal Rules of Evidence to exclude all documentary and testimonial evidence which referred to Lucky's affirmative action efforts. Plaintiffs' filed their objection to defendant's motion on August 30, 1991.
488. Although the failure to follow an affirmative action plan is not per se a violation of Title VII, evidence that an employer violated its own affirmative action plan may be relevant to the question of discriminatory intent. Gonzales v. Police Department, City of San Jose, 901 F.2d 758, 761 (9th Cir.1990) (affirmative action plan pursuant to a consent decree). See also Craik v. Minnesota State Univ. Board, 731 F.2d 465, 472 (8th Cir.1984) ("evidence that an employer has failed to live up to an affirmative action plan is relevant to the question of discriminatory intent"); Yatvin v. Madison Metro. School District, 840 F.2d 412, 415-416 (7th Cir. 1988) (repeated violation of an affirmative action plan may be used as evidence in support of a discrimination claim); Taylor v. Teletype Corp., 648 F.2d 1129, 1135 n. 14 (8th Cir.), cert. denied, 454 U.S. 969, 102 S. Ct. 515, 70 L. Ed. 2d 386 (1981) (evidence of violation of affirmative action program and failure to support the affirmative action director is relevant to discern attitude towards race). In addition, the qualified evidentiary privilege to which self-critical affirmative action related materials are entitled must be interpreted narrowly. General Telephone, 885 F.2d at 578.
489. Accordingly, the court holds that the documentary and testimonial evidence which refers to Lucky's affirmative action efforts is admissible. The court will consider that evidence insofar as it shows that Lucky management was aware of the statistical disparities in the placement and promotion of men and women, the attitude of Lucky management towards fulfilling its affirmative action goals, or discriminatory attitudes on the part of Lucky management. The court will not consider evidence of the affirmative action goals which Lucky set or whether Lucky actually satisfied or failed to satisfy those goals.
490. The court previously held that the 1988 meetings that Russell Specter held with Lucky Store and District Managers were not entitled to attorney-client privilege. Attorney-client privilege may not be asserted for meetings which take place for purposes other than facilitating the provision of professional legal services to a client. 8 Wright and Miller, Federal Practice and Procedure § 2017 at 137. Based upon the trial testimony of members of Lucky management and Russell Specter, trial exhibits and material produced in conjunction with the order to compel discovery of notes from the Specter meetings, the court held that the contract between Lucky and Specter was a consultancy agreement which did not designate Specter as an attorney. Moreover, the 1988 meetings were conducted for the purpose of training and informing Store and District Managers about Lucky's affirmative action plans. *331 August 27, 1991 Order; R.T. at X-XXXX-XX; supra FF ¶ 39.
B. Subjective Decision Making
491. Plaintiffs have shown that decisions regarding hiring, initial placement, and promotion are left to the sole discretion of Store Managers. Supra FF ¶¶ 49, 51, 82, 84, 86, 134, 158, and 164. In addition, the UFCW Contract provisions governing placement, training, and promotion place no practical constraints on that discretion. Supra FF ¶¶ 53, 72, 95, 100, 122, and 149.
492. Lucky Store Managers use different criteria in making assignments and promotions, and these criteria are typically ambiguous, subjective, or both. Supra FF ¶¶ 50, 73, 74, 76, 80, 83, 89, 91-94, 96, 99, 100, 105, 110, 112, 149, 154, 158, 159, and 164.
493. Lucky's training policy is another aspect of discretionary and subjective decision making. There are no uniform criteria for selecting employees to receive training. The choice is within the Store Manager's complete discretion. Supra FF ¶¶ 120, 125, and 131.
494. The absence at Lucky of personnel policies requiring decision makers to collect accurate information about applicants, apply written selection criteria, and be accountable for their decisions, increases the likelihood that gender stereotypes will influence their decisions. Supra FF ¶¶ 307 and 308.
495. Accordingly, the court finds that the potential for discriminatory abuse inherent in the ambiguous and subjective practices employed by defendant in making placement, promotion and training decisions is sufficiently high to infer the intent necessary to establish a claim for disparate treatment.See Jauregui, 852 F.2d at 1136.
496. Lucky defended two class action sex discrimination suits in the early 1980s, each of which resulted in a consent decree. One suit, the Bockman litigation, alleged failure to hire women in the warehouse; the other, an industry-wide suit, alleged sex-segregation in jobs in the meat department. Supra n. 4, FF ¶¶ 197-198.
497. In December 1983, Lucky's Human Resources Manager, Jane't Noriega-Ailor, informed Lucky management that women were underrepresented in management positions. In 1984 she informed Lucky management that the numbers and percentages of women in management had actually decreased. Supra FF ¶¶ 190, 191.
498. In 1984, the NCD received a report from the Corporate office indicating that the division had the lowest percentage of females in management of all Lucky divisions and directing the division to review Lucky's affirmative action policy and "reemphasize the importance of compliance." Noriega-Ailor discussed the figures and the reports with members of Lucky management, including Walter Herkal, Jim Scoggins and Robert Grant. Supra FF ¶ 192.
499. Noriega-Ailor made a number of recommendations, including improved recruiting and selection procedures and preparation of formal job descriptions. Supra FF ¶ 192. She also discussed the fact that women in management were disproportionately concentrated in the Deli/Bakery and General Merchandise departments, and emphasized the need to increase the number of women in other departments and to make sure that men were added into the Deli/Bakery and General Merchandise Department Head positions. Supra FF ¶ 194. Bob Beckerman, the Vice President of Deli/Bakery, urged Lucky management to consider Deli/Bakery and General Merchandise Department Heads as part of the feeder pool for promotion to Fourth Person. *332 These suggestions were not implemented. Supra FF ¶ 136.
500. Lucky's senior management, including Goodspeed, Grant, Hoffman, and Herkal have been aware of the underrepresentation of women in management positions and the overrepresentation of women in the Deli/Bakery and General Merchandise departments since at least 1986. Supra FF ¶¶ 202, 204, 213, and 229.
501. By 1986, several UFCW locals had complained to both Gill and Hoffman about the fact that women were not being promoted. Supra FF ¶ 202.
D. Discriminatory Attitudes
502. Plaintiffs introduced evidence of discriminatory attitudes and stereotyping of women which was taken from notes of District and Store Manager meetings conducted by Russell Specter in 1988.
503. The notes record such comments as that women do not want to work late shifts, that men do not want to compete with women or have a woman as their boss, that a woman's income is a second income in a household, that men resent the promotion of women, that Black women are aggressive, that women who are promoted frequently step down, and that women do not have the drive to get ahead. Supra FF ¶¶ 219, 220.
504. Although he had never reviewed the available statistics on step-downs and his opinion was based only on personal belief, Grant claimed that women stepped down from management positions at a greater rate than men. Grant also believed that men preferred working on the floor to working at the cash register because they viewed floor work as "more important," and that women preferred working at the cash register. Grant acknowledged that he had no basis for his beliefs about employee interest in departments or shifts or assignments other than his experience as a Store Manager thirty years ago. Supra FF ¶ 228.
505. Herkal disagreed with the idea of affirmative action goals and saw no reason to make changes at Lucky to increase the pool of women eligible for promotion. He felt that the distribution of men and women at Lucky reflected the interest level of the employees and that anybody who wanted to get promoted ultimately would. However, Herkal admitted that, prior to the Bockman litigation, he had believed that the absence of women in warehouse jobs reflected their lack of interest, and he was surprised to discover that women were interested in, and had applied for, those jobs. Supra FF ¶ 226.
506. Even if the court were only to consider the evidence regarding defendant's subjective decision making, the court finds that plaintiffs have proven discriminatory intent beyond a preponderance of the evidence.
507. The evidence of defendant's knowledge of the underrepresentation of women in management and the overrepresentation of women in the Deli/Bakery and General Merchandise departments, and the discriminatory attitudes of Lucky management merely confirm this finding.
508. Dr. Drogin found that women were 35% of the new hires into Grocery and Produce jobs while they were 84% of the new hires into Deli/Bakery and General Merchandise jobs (Z-value of 61.71). Supra FF ¶ 248.
509. This disparity is far greater than those which courts have previously held sufficient to establish an inference of intentional discrimination. Hazelwood, 433 U.S. at 311 n. 17, 97 S. Ct. at 2743 n. 17.
510. Defendant attempts to rebut these statistics by claiming that the relative level of interest of men and women in the entry-level jobs at Lucky is reflected in the proportion of men and women who apply for particular jobs. Supra FF ¶ 380. However, considering the unreliability of defendant's initial placement database, the evidence that the interest of applicants is not considered in deciding placement, and the gross statistical disparities between the placement of men and women established *333 by plaintiffs, the court finds that defendant failed to rebut plaintiffs' evidence.
511. When plaintiffs' unrebutted statistical evidence is viewed in conjunction with the direct and indirect evidence of discriminatory intent and attitudes on the part of Lucky management, their highly subjective decision making, and their knowledge of gender imbalance in the workforce, the cumulative evidence is sufficient to sustain plaintiffs' burden of proving disparate treatment in the initial placement of women.
512. Although courts have extended a "measure of immunity" to seniority systems that distribute advantages in a way that tends to perpetuate the effects of prior discriminatory employment practices, Teamsters, 431 U.S. at 350, 97 S. Ct. at 1862, in this case, seniority is seldom if ever the determining factor in promotions, supra FF ¶¶ 53, 72, 95, 100, 122, and 149.
513. Dr. Drogin compared the movement of male and female employees out of Courtesy Clerk positions and into Apprentice positions in the different departments. He found that women comprised 31% of those promoted into Apprentice jobs in Grocery and Produce, but 75% of those promoted into Apprentice jobs in Deli/Bakery and General Merchandise (Z-value of -15.13). Supra FF ¶ 249.
514. Dr. Drogin found statistically significant shortfalls, in excess of three standard deviations, in promotions of women into positions as Apprentice Produce Clerk from all positions (Z-value of -5.46); into part-time Journey Produce Clerk (Z-value of -2.77); into Apprentice Night Crew (Z-value of -18.00); and into part-time Journey Night Crew (Z-value of -19.98). Supra FF ¶ 250.
515. Dr. Drogin found statistically significant shortfalls, in excess of three standard deviations, in promotions of women into Department Head/Receiving Clerk (Z-value of -6.93); and into Night Crew Manager (Z-value of -4.01). Supra FF ¶ 251.
516. These disparities are far greater than those which courts have previously held to be sufficient to establish an inference of intentional discrimination. Hazelwood, 433 U.S. at 311 n. 17, 97 S. Ct. at 2743 n. 17.
517. Dr. Drogin did not find statistically significant disparities for the liability period in promotions of women to Third Person (Z-value of -2.58), Assistant Store Manager or Store Manager. Supra FF ¶ 252.
518. Defendant attempts to prove that the statistical disparities between the promotions of men and women at Lucky are explained by women's aversion to working night shifts and to differences in the job interests of men and women. Supra FF ¶¶ 382 and 384. However, considering the unreliability of defendant's initial placement database, the evidence that the interest of applicants is not considered in deciding placement, and the gross statistical disparities between the placement of men and women which plaintiffs found, the court finds that defendant has failed to rebut plaintiffs' evidence.
519. The court finds that plaintiffs' unrebutted statistical evidence on promotion is compelling. The court has also considered evidence of Lucky practices and policies that affect promotional opportunities, such as prohibiting transfers out of the Deli/Bakery and General Merchandise departments. Supra FF ¶¶ 126, 133-136. The statistical evidence is significant for some positions, but not for others. However, the court finds that the lack of statistically significant disparities in promotions of women to Third Person, Assistant Store Manager and Store Manager is caused by women being blocked from upper management positions at the lower rungs of the promotional ladder. In addition, the court has considered direct and indirect evidence of discriminatory intent on the part of Lucky managers. In considering plaintiffs' evidence cumulatively, the court concludes that Lucky's promotion process is tainted by discretionary decision making and the use of variable and subjective criteria. Accordingly, the court finds that the evidence is sufficient to sustain plaintiffs' burden of proving disparate treatment in Lucky's promotion practices as a whole.
*334 Movement from Part-Time to Full-Time
520. The court finds that Lucky did not consistently follow the bid list policy, promoted employees who were not on the list, and failed to keep records of many bid lists. Supra FF ¶¶ 146, 154, 256 and 385. Therefore, it is appropriate for the court to consider the overall rate of movement of women to full-time positions and the disparity in movement of women to full-time when compared with their representation in the relevant "feeder pool." See Box v. A & P Tea Co., 772 F.2d 1372, 1376-77 (7th Cir.1985) (plaintiff need not establish that she applied for position at issue where employer did not have a formal systematic procedure for posting openings and processing applications), cert. denied, 478 U.S. 1010, 106 S. Ct. 3311, 92 L. Ed. 2d 724 (1986).
521. Dr. Drogin found that women received fifty-three fewer moves to full-time status than would be expected (Z-value of -6.75). The disparity between the movement of men and women into full-time night jobs was also statistically significant (Z-value of -15.83). Supra FF ¶ 254.
522. These disparities are greater than those which courts have previously held to be sufficient to establish an inference of intentional discrimination. Hazelwood, 433 U.S. at 311 n. 17, 97 S. Ct. at 2743 n. 17.
523. Dr. Drogin confirmed Lucky's practice of awarding full-time positions not taken from the bid list. Supra FF ¶ 256. He found that when promotions were taken from the bid list, women are moved up to full-time at a rate consistent with their representation on the bid list (women comprise 52.8% of those taken from a bid list, and 50-55% of bidders). However, among moves not taken from the bid list, women receive significantly fewer moves from part-time to full-time than would be expected from their representation among the "feeder pool" of part-time Clerks (women comprise 31.3% of those promoted who are not taken from a bid list). Supra FF ¶ 258.
524. Defendant attempted to show that there was no statistical disparity between the moves of men and women to full-time positions when the gender composition of the feeder pools was adjusted to account for differences in men's and women's work preferences. Supra FF ¶ 385. However, considering the unreliability of defendant's initial placement database, the evidence that the interest of applicants is not considered in deciding placement, and the gross statistical disparities between the placement of men and women which plaintiffs found, the court finds that defendant has failed to rebut plaintiffs' evidence.
525. The court holds, on the basis of the unrebutted statistical evidence concerning movement of women from part-time to full-time status and nonstatistical evidence of discriminatory intent, that plaintiffs have sustained their burden of proving disparate treatment with respect to the movement of women from part-time to full-time positions.
Allocation of Hours/Step-Ups
526. Dr. Drogin found that there were statistically significant disparities in the distribution of hours to women and men. These disparities occurred with respect to women in part-time Journey Clerk Day (Z-value of -18.16); full-time Journey Clerk Day (Z-value of -8.77); and Apprentice Day Grocery (Z-value of -3.36). Supra FF ¶ 261. However, Dr. Drogin did not find statistically significant disparities in the allocation of hours to females in Courtesy Clerk positions. Supra FF ¶ 262.
527. Dr. Drogin found a significant disparity in the assignment of step-up hours to women. He found that women received fewer step-up hours relative to their proportions in the relevant pool (Z-value of - 22.53). Supra FF ¶ 263.
528. These disparities are greater than those which courts have previously held to be sufficient to establish an inference of intentional discrimination. Hazelwood, 433 U.S. at 311 n. 17, 97 S. Ct. at 2743 n. 17.
529. Defendant tried to show that the gender difference in the allocation of hours and step-ups of Lucky employees was due to women's aversion to night work and was consistent with gender neutral employment practices. Supra FF ¶ 386. However, considering the unreliability of defendant's initial *335 placement database, the evidence that the interest of applicants is not considered in deciding placement, and the gross statistical disparities between the placement of men and women which plaintiffs found, the court finds that defendant has failed to rebut plaintiffs' evidence.
530. The court finds that these statistical disparities, which have been established by substantially more than the preponderance of the evidence, in conjunction with the evidence regarding Lucky management's knowledge of problems of underrepresentation of women in management, and evidence of discriminatory attitudes and stereotyping by Lucky managers, are sufficient to sustain plaintiff's burden of proving disparate treatment with respect the assignment of additional hours to women.
VIII. DISPARATE IMPACT
Subjective Decision Making
531. Plaintiffs have presented uncontroverted evidence of defendant's ambiguous and subjective decision making practices. In fact, the court has found that those practices are wholly lacking in standards. See VII, Evidence of Intent, B.
532. Plaintiffs have also presented persuasive statistical evidence that those ambiguous and subjective decision making practices have a disparate impact on Lucky's female employees. See VII, Initial Placement, Promotion, Movement to Full-Time, and Allocation of Additional Hours.
533. Where the system of promotion is pervaded by a lack of uniform criteria, criteria that are subjective as well as variable, discretionary placements and promotions, the failure to follow set procedures and the absence of written policies or justifications for promotional decisions, the court is not required to "pinpoint particular aspects of [the system]" that are unfavorable to women. Allen v. Seidman, 881 F.2d at 381.
534. Moreover, the court finds that the elements of Lucky's subjective and ambiguous decision making processes are not separable for the purposes of analysis, and therefore may be analyzed as one employment practice. 1991 Act § 105(a).
535. The court holds that Lucky's highly discretionary and subjective decision making with respect to initial placement, promotion, and selection of employees for additional training is a protracted and even more amorphous version of the "unstructured oral interview" which was found to have a disparate impact on the promotional opportunities of Black employees in Allen v. Seidman, 881 F.2d at 381.
536. In addition, the court finds that Lucky's policy of leaving initial placement, promotion and training decisions to the sole discretion of lower level supervisors whose conscious and subconscious prejudices are unchecked by objective and publicized decision making criteria, has had a disparate impact on female employees at Lucky. See Watson, 487 U.S. at 990, 108 S. Ct. at 2786-87.
537. Thus, plaintiffs have established a prima facie case of disparate impact based on Lucky's amorphous and subjective decision making policies, under both 1991 Civil Rights Act and pre-1991 Civil Rights Act standards.
538. Defendant argued that its decision making policies are necessitated by the nature of the grocery business. Supra FF ¶¶ 70, 74, and 149. However, plaintiffs have offered proof that defendant does not consistently follow the policies that it has established. Supra FF ¶¶ 73, 80, 82, 96, 100, 114, 122, 149, and 154. Moreover, defendant has not shown that Lucky's subjective decision making policies are a business necessity which bear "a manifest relationship to the employment in question," Griggs, 401 U.S. at 432, 91 S. Ct. at 854, or that the policies serve, "in a significant way, the legitimate employment goals of the employer," Wards Cove, 490 U.S. at 659, 109 S. Ct. at 2125.
539. Therefore, plaintiffs have sustained their burden of proving that Lucky's standard policy of discretionary, subjective and frequently unreviewed decision making with respect to initial placement, promotion and training has had a disparate impact on *336 the promotional opportunities of women under both 1991 Civil Rights Act and pre-1991 Civil Rights Act standards.
Failure to Follow Bid Procedures
540. Plaintiffs presented testimony and documentary evidence that Lucky's practice of bidding for movement from part-time to full-time status is not consistently followed. Supra FF ¶ 154. Dr. Drogin confirmed that this practice was not followed consistently. Supra FF ¶ 256.
541. Furthermore, Dr. Drogin testified that although there is a small disparate impact on women when movements from part-time to full-time is taken from the bid list (Z-value of -3.21); there is a statistically significant disparity with respect to women for those movements not taken from the bid list (Z-value of -6.59). Supra FF ¶ 258.
542. The court finds that Lucky's practice of departing from bid procedures in awarding movement to full-time status has had a disparate impact on Lucky's female employees. Moreover, defendant's failure to follow its own policies is excepted by neither the business necessity nor the business justification defenses.
543. Accordingly, the court concludes that plaintiffs have sustained their burden of proving that Lucky's practice of departing from bid-list procedures in awarding movement to full-time status has resulted in a disparate impact on women with respect to opportunities for full-time work.
IX. PUNITIVE DAMAGES
544. Based on Lucky's knowledge of significant problems with the underrepresentation of women in management, its failure to implement the recommendations of the Human Resource Director to promulgate formal job descriptions and promotion criteria, the discriminatory attitudes of some of its Store Managers, and its abandonment of two affirmative action programs despite continued evidence of a gross gender imbalance in the Deli/Bakery and General Merchandise departments, plaintiffs have made a sufficient showing to permit them to go forward on their motion for punitive damages. The court must hear further evidence to determine if an award of punitive damages is justified under Title VII.
X. FEHA CLAIM
545. As plaintiffs have satisfied their burden of proof under Title VII, they have also satisfied their burden of proof under the less stringent standards of FEHA.
546. This court previously denied plaintiffs' request for punitive damages under FEHA with respect to their class claims. September 11, 1991 Order at 34-35.
1. Plaintiffs have met their burden of proving that sex discrimination was the standard operating procedure at Lucky with respect to placement, promotion, movement to full-time positions, and the allocation of additional hours. Bazemore, 478 U.S. at 398, 106 S. Ct. at 3007-08. The court finds that defendant's explanation that the statistical disparities between men and women at Lucky are caused by differences in the work interests of men and women to be unpersuasive. Accordingly, the court finds that plaintiffs have made a sufficient showing to prevail on their claim for disparate treatment.
2. Plaintiffs have made a sufficient showing to support their claim for disparate impact with respect to defendant's subjective decision making policies and defendant's failure to follow bid procedures.
3. If plaintiffs wish to bolster their disparate treatment case with additional anecdotal evidence, the court will entertain their presentation. This matter will be taken up at the trial on damages which is scheduled to begin September 14, 1992.
4. Plaintiffs have made a sufficient showing for the court to consider further evidence on their motion for punitive damages under Title VII.
5. Plaintiffs have satisfied their burden of proof under FEHA.
IT IS SO ORDERED.
 The eleven retail clerks unions are 373 Vallejo, 428 San Jose, 588 Sacramento, 648 San Francisco, 775 San Mateo, 839 Salinas, 870 Alameda County, 1119 Marin County, 1179 Contra Costa, 1288 Fresno, and 1532 Santa Rosa. There is also a building service employees union. R.T. at 1-165-67 (Gill).
 The night premium wage is an additional fifty cents per hour. R.T. at 1-73 (Gill).
 Every Lucky store also has a meat department, and some stores have a pharmacy department. This case does not involve positions in those departments.
 The Bockman litigation involved allegations of sex discrimination in Lucky's warehouse facilities in the San Leandro complex. R.T. at 10-1641 (Javier).
 The milk box comes in crates that weigh forty pounds each and are stacked six high. These must be pulled off a pallet and moved around the store. R.T. at X-XXXX-XX (Foley).
 The court is incredulous that Lucky's 1986 affirmative action plan for the NCD could have been discontinued without the knowledge or approval of any of Lucky's senior management. The court cannot determine whether the variation in the testimony of Lucky's senior management concerning the termination of the affirmative action plan is the result of subterfuge, carelessness, or lack of interest in the plan. Regardless, as the employees entrusted with the implementation and administration of the plan carried on this charade with no intervention by senior management, the consequences must be borne by Lucky.
 Dr. Drogin's expected rates are based on the assumption that job assignments of new hires in the same store and year are made independent of sex. Ex. A-10 at 14.
 In analyzing promotions into various target jobs at Lucky, Dr. Drogin used the following control variables: for promotions to Apprentice and part-time Journey Clerk positions, he controlled for year of promotion, feeder job, and store; for full-time Journey Clerk positions, he controlled for year of promotion, feeder job, store, and year of hire. Ex. A-10 at 12.
 In analyzing promotions into positions above Journey Clerk, Dr. Drogin used the following control variables: year of promotion, feeder job, and district. Ex. A-10 at 12.
 Dr. Pencavel counted as employed those employees on the payroll tapes with positive wage rates, end of year hours of work, end of year earnings, and weeks of seniority who also had admissible values on gender, marital status, job category, and store of employment. Ex. A-9 at 3 n. 2.
 There were two stores where the survey was not completed. Dr. Daum got responses from 50% and 60% of the employees in those stores respectively, and he thought that their responses would be representative of those of the entire store. R.T. at XX-XXXX-XX.
 42% of the women in the initial placement database sample said that they first applied for a Courtesy Clerk position, compared to 40.1% of the women in Dr. Daum's survey. 83.1% of the women in the initial placement database sample said that they first applied for a Deli/Bakery or General Merchandise Clerk position, compared to 87.3% of the women in Dr. Daum's survey. 8.6% of the women in the initial placement database sample said that they first applied for a Night Crew position, compared to 5.7% of the women in Dr. Daum's survey. Ex. B-27.
 Dr. Daum defined statistical significance to be a disparity with a P-value of 1% or less. Ex. B-31.
 Dr. Daum surveyed twelve stores in the SCD. R.T. at 25-4174. Before conducting the survey Dr. Daum was unaware that additional hours were awarded in a different manner in the SCD than in the NCD. R.T. at 24-3991. However, Dr. Daum adjusted his survey to account for the differences between the nature of the Receiving Clerk job in the NCD and SCD. R.T. at XX-XXXX-XX. The survey Dr. Daum conducted in the SCD only included four jobs. While Dr. Daum admitted that it is the relative positions of the jobs that is measured in paired comparison surveys, Dr. Daum argued that the surveys were comparable because in each survey the subjects were to focus on the different job features. R.T. at XX-XXXX-XX. In addition, he did not expect the difference in the gender composition of the jobs in the NCD and the SCD to affect interest patterns. R.T. at 25-4204.
 The job features in this survey were different than those in the Lucky surveys. The features were: advancement opportunity, co-worker contact, level of customer contact, degree of flexibility in scheduling hours, average hours available, rate of pay (dollars per hour), level of physical effort, level of pleasant environment, and degree of responsibility. R.T. at 24-4030. The jobs in this survey were also different: Baker or Deli Clerk, Checker, Head Night Stocker, Produce Clerk, Produce Manager, Receiving Clerk, Scan Coordinator, Store Director, and Third Person. R.T. at 24-4031.
 For the purpose of this analysis Dr. Daum considered a P-value of less that 5% to be statistically significant and a P-value of less that 1% to be highly statistically significant. Ex. B-48.
 11.03% of the women in the initial placement database sample said that they first applied for a Produce Clerk position, compared to 0% of the women in Dr. Daum's survey. Likewise, 9.5% of the women in the initial placement database sample said that they first applied for a Janitor position, compared to 0% of the women in Dr. Daum's survey. Ex. B-26; R.T. at XX-XXXX-XX.
 Dr. Hoffman found that 55% of Fourth Persons are promoted from Checker and Stocker positions, 7% come from Night Crew Manager, but only 4% come from Night Crew. Produce Manager and Produce Clerk do not feed into Fourth Person. R.T. at 33-5445.
 The court required the plaintiff to show 1) that s/he belonged to a protected class; 2) that s/he applied and was qualified for a job for which the employer was seeking applicants; 3) that, despite being qualified, s/he was rejected; and 4) that the position remained open and the employer continued to seek applicants from persons having plaintiff's qualifications. McDonnell Douglas, 411 U.S. at 802, 93 S. Ct. at 1824.
 Under Wards Cove, an employer could rebut a prima facie showing of disparate impact by producing evidence that the "challenged practice serves, in a significant way, the legitimate employment goals of the employer." Wards Cove, 490 U.S. at 659, 109 S. Ct. at 2125. The "business necessity" standard which was originally articulated in Griggs, 401 U.S. at 431, 91 S. Ct. at 853, and which was readopted by the 1991 Civil Rights Act, places higher burdens of production and proof on the defendant than did the Wards Cove "business justification" standard.
 These are the legal standards of which the Store Managers were informed at the 1988 Specter meetings. R.T. at 7-1182 (Specter).
 This application can be misleading for three reasons. First, the two or three standard deviations rule does not focus directly on the probability that the statistical disparity observed in the record would occur if the outcomes of the selection process were unassociated with race or sex. Second, the two or three standard deviations rule has been treated by some courts as a rule of law rather than as an aid to interpretation. The complexity of these calculations can easily mislead the court into believing that the statistical significance rather than the magnitude of a disparity is the primary consideration. Third, the two or three standard deviations analysis may produce results which are mathematically incorrect either because the assumed sampling model is inappropriate or because the conclusions reached may disagree with those that would emerge from a correct application of the assumed model. Statistical Proof, 1987 Supplement at 171-176.
 In Bradley v. School Board of the City of Richmond, 416 U.S. 696, 711, 94 S. Ct. 2006, 2016, 40 L. Ed. 2d 476 (1974), the Supreme Court held that the presumption of the retroactive application of a statute may only be rebutted if it will "result in manifest injustice or [if] there is statutory direction or legislative history to the contrary." The Ninth Circuit recently reaffirmed its endorsement of the Bradley test in FDIC v. New Hampshire Ins. Co., 953 F.2d 478, 486-87 (9th Cir.1992).
 In comparison, both FEHA and section 1981 require plaintiffs to make a showing of "clear and convincing evidence" in order to prevail on a punitive damages claim. See Cal.Civ.Code § 3294(a); Mitchell v. Keith, 752 F.2d 385, 390 (9th Cir.), cert. denied, 472 U.S. 1028, 105 S. Ct. 3502, 87 L. Ed. 2d 633 (1985) (borrowing from California law for punitive damages standard).
 However, the award of punitive damages is ultimately up to the discretion of the court.
 Failure to prove that the plaintiff applied for the job can be excused where the plaintiff proves that s/he made a reasonable attempt to convey his or her interest in the job to the employer, see, e.g., Chambers v. Wynne School Dist., 909 F.2d 1214, 1217 (8th Cir.1990); E.E.O.C. v. Metal Service Co., 892 F.2d 341, 348 (3d Cir.1990), or that s/he was deterred from making any application, Teamsters, 431 U.S. at 366-67, 97 S. Ct. at 1870-71.
 Dr. Daum did not explain his standardization procedure to the court in any detail. In fact, his procedure was most clearly explained by defendant's counsel during the cross-examination of Dr. Hoffman. R.T. at XX-XXXX-XX.
 Although Lucky's ambiguous and subjective decision making practices provide ample evidence to support the court's finding of intent; the court's finding also rests on evidence of Lucky's knowledge of the underrepresentation of women in management and the overrepresentation of women in the Deli/Bakery and General Merchandise departments, and the discriminatory attitudes of some of Lucky's management.