Opinion
No. 0103645.
Memorandum Dated November 7, 2006.
The plaintiffs, a class of hourly associates who worked at defendant Wal-Mart Stores, Inc.'s ("Wal-Mart") Massachusetts stores at any point between August 1995 and December 31, 2005, allege that, as a result of Wal-Mart's cost-cutting policies and pressure placed on store managers to keep payroll costs down, they were not given all of their earned rest breaks and they were not properly compensated for all time worked. The court (Murphy, J.) certified this class in January 2004 and again in December 2004.
Among the time-shaving exercises the plaintiffs allege Wal-Mart encouraged are as follows:
when an associate's payroll record indicated that a meal break was not taken, the manager was to insert a meal period;
when an associate's payroll record indicated an odd number of swipes for a given shift, the manager was to clock the associate out one minute after the associate's last punch-in for that shift; associates were encouraged to miss rest breaks.
Originally, the plaintiffs also claimed that associates were encouraged to miss meal periods. In a decision dated September 18, 2006, this court (Murtagh, J.) allowed Wal-Mart's motion for summary judgment on the plaintiffs' claims seeking compensation for missed meal periods.
The plaintiffs intend to rely on Dr. Marvin Shapiro's ("Shapiro") analysis of Wal-Mart's payroll records to establish liability and damages. This case is before the court on Wal-Mart's motion to exclude Shapiro's testimony as unreliable. See Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 592-93 (1993); Commonwealth v. Lanigan, 419 Mass. 15, 26 (1994). Wal-Mart has also moved to decertify the class, arguing that without Shapiro's testimony, the plaintiffs will not be able to sustain their burden of demonstrating that common issues predominate over individual issues. See Mass.R.Civ.P. 23(b). An evidentiary hearing on these issues took place on July 26, 27, and 28, 2006. See Weld v. Glaxo Wellcome, Inc., 434 Mass. 81, 85 (2001) ("Although it is within a judge's discretion to hold an evidentiary hearing, there is no such requirement"). This court heard further argument at the final pretrial conference on October 11, 2006. For the following reasons, Wal-Mart's motion to exclude Shapiro's testimony is ALLOWED; and WalMart's motion to decertify the class is ALLOWED.
There is a separate motion in limine pending in which Wal-Mart seeks to exclude the testimony of the plaintiffs' expert Dr. Thomas Kochan ("Kochan"). This court has made no decision regarding his testimony but notes that Shapiro's testimony, unlike Kochan's, directly affects certification.
PROCEDURAL HISTORY
The plaintiffs first moved for class certification in November 2003. This court (Murphy, J.) allowed that motion in a decision dated January 5, 2004 ("January 2004 Decision"). In the January 2004 Decision, the court went through each of the requirements for certification of a class action, finding that the plaintiffs satisfied the numerosity, commonality, typicality, adequacy, predominance, and superiority requirements. First, the court held that, given that "the number of so-called `low end retail' associates who have worked for Wal-Mart in Massachusetts during the class period is estimated without dispute to be easily in excess of 50,000[,] individual joinder would be impractical and the `numerosity' requirement of Rule 23 is therefore satisfied." January 2004 Decision, at 3.
Second, the plaintiffs had alleged "common issues of state labor law in terms of compensation requirements for hourly [associates]," and common issues of fact with respect to Wal-Mart's policy to provide earned meal and rest breaks to its associates and Wal-Mart's "systemic business strategy creating inescapable pressure on middle management by dint of the terms and condition of their compensation to understaff Wal-Mart stores, and ineluctably to cajole or coerce [associates] to perform services `off-the-clock' to meet productivity and profit goals." Id. at 4. The court held that the commonality requirement was consequently satisfied. Third, the court held that the typicality requirement was similarly satisfied in that Wal-Mart "acted consistently towards all of its `low end,' retail store, shift [associates] . . . [such that] the legal theories upon which the named [p]laintiffs have proceeded against Wal-Mart are the identical theories with which the class members would also seek redress." Id. at 4-5. Fourth, the court held that "no individual plaintiff would have any realistic opportunity to litigate such a complex matter against such a powerful corporate entity as Wal-Mart[,]" therefore, a class action was the superior method by which to proceed. Id. at 8-9.
Fifth, the court addressed predominance, spending the bulk of its discussion on this requirement. Wal-Mart argued "that each of the plaintiffs' claims will require an individualized inquiry into the subjective mind state of each of the plaintiffs and the particular circumstances surrounding the underlying conduct relative to the particular claim of Wal-Mart's unlawful conduct." Id. at 6. With the Associate Handbook and written policies, PD-07 and PD-43, the plaintiffs had "presented sufficient evidence demonstrating that Wal-Mart had a consistent corporate policy" regarding meal periods, rest breaks, and off-the-clock work; the plaintiffs had also presented evidence demonstrating "that regardless of position or shift all [associates] were subjected to a uniformly implemented . . . policy requiring uncompensated work from store [associates] in the Commonwealth." Id. at 7. The plaintiffs asserted that they intended "to produce expert statistical analysis which will be probative and admissible" in order to "discount individualized non-actionable reasons for Wal-Mart's [alleged] failures to compensate." Id.
The court found that, at that point of the litigation, the plaintiffs' intent to produce expert statistical evidence sufficiently addressed the issue that "there may be — and indeed demonstrably are — numerous innocuous reasons why a particular [associate] did not receive compensation for labor . . . [and that] each instance of uncompensated work is potentially excusable[.]" Id. Additionally, the court held that the fact that "[i]ndividual damages may be determined through numerous methodologies after liability on a case wide basis has been established" did not preclude class certification. Id. at 8.
Finally, the court held that the adequacy requirement was satisfied because "[r]ecovery upon the theories pled, if obtained, will potentially prove adequate for the class, and there is no potential conflict between the named [p]laintiffs and the putative class members on the basis of waiver of a particular legal right." Id. at 5. In so ruling, the court declined to address Wal-Mart's argument that the named plaintiffs cannot be adequate representatives of the class "because the named [p]laintiffs, in their efforts to avoid removal of the action to federal court, . . . waived any claim under the Payment of Wages statute, G.L.c. 149, §§ 148 150 . . ." Id.
Wal-Mart appealed the January 2004 decision certifying the class, and a Single Justice of the Appeals Court (McHugh, J.) ("Appeals Court") issued its decision on June 8, 2004 ("June 2004 Decision"). In its appeal, Wal-Mart did not argue
that the plaintiffs . . . failed to meet the first three requirements of [Massachusetts] Rule [of Civil Procedure] 23(a), the so-called numerosity, commonality and typically [sic] requirements. Instead, it argue[d] that the trial judge abused his discretion when he decided that common questions predominated over individual questions and that the representative would adequately and fairly protect the interests of that class.
June 2004 Decision, at 3-4. Consequently, the June 2004 Decision focused on the predominance and adequacy requirements.
With respect to predominance, the Appeals Court noted that while this court (Murphy, J.) had been "persuaded by the plaintiffs' presentation" and ultimately decided to certify the class, id. at 4, it was "a decision by indecision: to rebut the claim that individual questions will require a separate trial, the plaintiffs offer[ed] statistical evidence and methodologies . . . If, therefore, the class-justifying evidence is inadmissible, the case cannot proceed as a class action." Id. at 5. The January 2004 Decision therefore "effectively left for another day a question that determines whether the action can truly proceed on a class basis." Id. The Appeals Court referenced the plaintiffs' intent to rely on "statistical extracts from records sampling, questionnaires, and extrapolations from data obtained from all of those sources, [all of which] raises very troubling issues both of reliability and of due process." Id. at 5-6. As result of these "troubling issues," "`certification' is of doubtful durability unless the plaintiffs successfully meet their burden of showing that the statistical evidence is reliable enough to be admissible and that its admission would not violate Wal-Mart's due process rights." Id. at 6 (emphasis added). "If the proposed evidence proves to be inadmissible, the class certification may be stricken and the action may thereafter proceed on an individual basis." Id. at 10.
The Appeals Court then reviewed this court's refusal to address Wal-Mart's argument that the named plaintiffs' waiver of their claims under G.L.c. 149, §§ 148 and 150 "might conflict with a desire by other class members to assert those claims. That potential for conflict . . . means that the named parties are not adequate class representatives." Id. at 7. "In Spence v. Reeder, 382 Mass. 398, 412-13 (1981), however, the Supreme Judicial Court described, albeit in dictum, the precision with which a court must approach an inquiry regarding the validity of a representative's waiver of procedural rights possessed by class members." Id. at 8. The Appeals Court held that the motion judge should have determined whether the foreclosure of "a class member's ability to pursue statutory damages . . . comports with due process either because class members have been afforded notice and an opportunity to intervene under Mass.R.Civ.P. 23(d), . . . or because the waived claims are so insubstantial that waiver does not deprive absent class members of an interest significant enough to trigger due process requirements." Id. at 9-10. By certifying the class "without determining that the named parties were adequate class representatives and without considering whether waiver of claims under the Payment of Wages statute required some action under Rule 23(d) to insure protection of absent members' due process rights as the case proceeded[,]" this court abused its discretion. Id. at 10. The Appeals Court therefore vacated the class certification and remanded the matter back to this court. Id.
Before the plaintiffs moved, again for certification consistent with the June 2004 Decision, the plaintiffs allegedly discovered new evidence concerning Wal-Mart's practice of deleting time from or otherwise manipulating associates' payroll records. The plaintiffs filed a motion to amend their complaint to add a claim alleging Wal-Mart violated G.L.c. 149, § 148. This court (Fahey, J.) granted this motion in September 2004, and the plaintiffs filed their Second Amended Complaint in October 2004. In their second motion for certification, the plaintiffs sought certification of their newly added G.L.c. 149, § 148, claim, arguing that the instructions the Appeals Court gave to the motion judge in the June 2004 Decision regarding waiver were moot, given the addition of that statutory claim to their complaint.
Wal-Mart moved for clarification or, alternatively, for reconsideration of this order allowing the plaintiffs to amend their complaint. This court (Fahey, J.) allowed this motion in a decision dated December 6, 2004.
The date stamp on the document indicates that this motion was filed with the court in October 2004, but it does not appear on the docket sheet until May 19, 2005, with the notation that it was filed in court on October 15, 2004. The date stamps, also, on the affidavit of plaintiffs' counsel in support of their motion for class certification, the appendix of foreign authority, and the affidavit of Shapiro, indicate that they were filed with the court in October 2004, but they do not appear on the docket sheet until May 19, 2005, with the notation that they were filed in court on October 12, 2004.
In opposing the plaintiffs' motion by asserting that the plaintiffs could not satisfy the requirements for a class action, Wal-Mart addressed not only the plaintiffs' G.L.c. 149, § 148, claim, but also the plaintiffs' other claims. With respect to predominance of the plaintiffs' G.L.c. 149, § 148 claim, Wal-Mart argued that the plaintiffs could not rely on Wal-Mart's records to prove this claim because "that proposed analysis will raise far more questions than it answers because whether an associate was actually deprived of pay and, if so, how much, will require an individualized analysis as to each alleged instance of time shaving." Defendant Wal-Mart Stores, Inc.'s Opposition to Plaintiffs' Motion to Recertify the Case as a Class Action, at 43. Further, the plaintiffs' theory as to its analysis of the record "fails to comport with reality because it requires them to assume that the edited punch does not reflect the time that the associate actually worked. Whether that is true cannot be ascertained from the records themselves and, instead, would require speaking to the associate or the person who made the edit in the time records." Id. Similarly, in its claim that the plaintiffs could not establish predominance as to its other claims, Wal-Mart asserted that "[p]laintiffs' theory for pursuing their meal and rest break claims on a class-wide basis rests on their assumption that missing punches on Wal-Mart's time records mean that an associate actually missed a meal period or rest break. That assumption fails to take into account that an [associate] may simply have forgotten to punch in and out for a meal period or rest break." Id. at 27. The plaintiffs' proposed expert testimony, according to Wal-Mart, could not overcome these individualized issues.
Wal-Mart quoted Shapiro as testifying in his deposition that he would not review any Time Adjustment Requests ("TARs") because they were "outside of [the] scope of what I was doing." Defendant Wal-Mart Stores, Inc.'s Opposition to Plaintiffs' Motion to Recertify the Case as a Class Action, at 44 n. 37 (alteration in original), quoting Shapiro Deposition, pp. 234-35. As discussed further below, in his preliminary March 2006 report filed with this court, and in his testimony at the July 2006 Daubert-Lanigan hearing, Shapiro indicated that he intended to consider TARs. His more recent report, however, appears to be at least somewhat inconsistent with that stated intent.
This court (Murphy, J.) certified the plaintiffs as a class action in a decision dated December 30, 2004 ("December 2004 Decision"). The court declined "to reconsider its initial finding that [the named plaintiffs'] issues of fact or law predominate over individualized inquiries . . . [because] it is not prepared at this stage to make such a determination." December 2004 Decision, at 5. Instead, the court focused on the plaintiffs' G.L.c. 149, § 148, claim. This claim satisfied the numerosity requirement because it would be impractical to require joinder of the more than 50,000 Wal-Mart associates who were allegedly subjected to Wal-Mart's alleged time-shaving practices. The claim of one of the named plaintiffs "that Wal-Mart illegally clocked her out of work one minute after she had returned from a break [is] typical of the claims of the other class members who were allegedly subject to a practice of altering their payroll records to deprive them of time earned." Id. at 7. The adequacy requirement was also satisfied as the court found no potential conflict between the named plaintiffs and the class members.
With respect to commonality, the court held that if the named plaintiffs' "allegations that Wal-Mart engaged in a systematic practice of time shaving are true, then the interest of each class member is identical. Each class member has an interest in determining Wal-Mart's practices regarding [associate] payroll records and if these practices were legal." Id. at 6 (footnote omitted). The court held that the plaintiffs' G.L.c. 149, § 148, claim also satisfied the predominance requirement, finding that the "plaintiffs have presented sufficient evidence to enable the Court to form a reasonable judgment they have met their burden in establishing predominance. Whether Wal-Mart had a systematic practice of time shaving is an issue that will predominate." Id. at 8. Finally, the court held that certification was a superior method for a fair and efficient adjudication of the controversy because "[i]t is extremely unlikely that the individual class members would have a realistic opportunity to litigate such a complex case." Id.
Wal-Mart appealed the December 2004 Decision to the Appeals Court, and the Appeals Court (Lenk, J.) affirmed that decision in an order dated May 18, 2005 ("May 2005 Decision"). The Appeals Court noted the procedural history of the case and ultimately held that "because of Judge Fahey's ruling [allowing the plaintiffs to amend their complaint to add the G.L.c. 149, § 148 claim], the Wage Act claim was no longer waived but instead fully resuscitated and any concerns about adequacy of representation attendant upon such waiver evaporated." May 2005 Decision, at 5. Therefore, "the concern that prompted the McHugh, J. remand [in the June 2004 Decision] no longer" existed. Id. at 5.
In a separate petition, Wal-Mart appealed this court's (Fahey, J.) December 2004 decision clarifying and reconsidering its order allowing the plaintiffs to amend their complaint. The Appeals Court (Lenk, J.) addressed this separate appeal in the May 2005 Decision and reported it to a panel of justices of the Appeals Court. May 2005 Decision, at 6. The Appeals Court issued its 1:28 decision on that issue on June 30, 2006, affirming this court's decision allowing the plaintiffs to amend their complaint.
After Wal-Mart filed this motion for decertification of the class and exclusion of the plaintiffs' expert Shapiro, but before the court ruled on that motion, Wal-Mart filed a motion for partial summary judgment. This court (Murtagh, J.) allowed in part and denied in part that motion, leaving open Counts I, VII, and VIII; Count II or IV, depending under which theory the plaintiffs choose to proceed; and Count IX insofar as the plaintiffs are limited to the time period of August 21, 1998, through December 31, 2005. This court also allowed Wal-Mart's motion for summary judgment on all of the plaintiffs' claims seeking compensation for missed, interrupted, and shortened meal periods. An additional effect of this partial allowance of Wal-Mart's motion for summary judgment is that Counts III, V, and VI are dismissed.
The court was awaiting Shapiro's final report before ruling on Wal-Mart's motion.
In December 2004, the plaintiffs filed a motion for partial summary judgment which this court (Murphy, J.) denied in a December 24, 2004, decision.
BACKGROUND I. Wal-Mart Timeclock Procedures
Massachusetts law and Wal-Mart's own policy provide that an associate who works for six hours is entitled to an unpaid thirty-minute meal break. See G.L.c. 149, § 100 ("No person shall be required to work for more than six hours during a calendar day without an interval of at least thirty minutes for a meal"). A Wal-Mart associate who works three to six hours is entitled to one paid fifteen-minute rest break; if the associate works over six hours, he is entitled to two paid fifteen-minute rest breaks. No policy of Wal-Mart allows associates to waive meal or rest breaks. Wal-Mart requires its associates to clock in upon arriving at work, and to clock out at the end of their shift. Associates must also clock in and out for meal breaks. Prior to February 9, 2001, associates were required to clock in and out for rest breaks. If an associate clocked in more than fifteen minutes after clocking out for a rest break, any time over fifteen minutes would be electronically deducted from the associate's record of hours worked.
It is undisputed that after February 9, 2001, and "effective February 10th, [Wal-Mart] . . . no longer require[d] Hourly Associates to clock in or out for their break periods." Memorandum titled "Changes in Payroll Processes, from Don Harris to "All Facility Managers," dated January 4, 2001, Exhibit 16 to Affidavit of Carolyn Beasley Burton in Support of Plaintiffs' Opposition to Wal-Mart's Motion to Exclude Testimony of Dr. Martin Shapiro.
These clocking in and clocking out policies are communicated to the associates during their orientation. If associates fail to follow these policies, they are formally disciplined through a process called "coaching." Coaching is a graduated process that begins with a verbal coaching, then, if the conduct continues, progresses to a written coaching. Often, a single coaching addresses more than one incident of alleged associate misconduct.
Wal-Mart's Time Clock Exception Report ("Exception Report") lists all associates who are missing punches. Exceptions that show up on the Exception Report include, inter alia, short meal breaks, long meal breaks, too few meal breaks, too many meal breaks, failure to clock out at the end of the day, and, prior to February 2001, short or long rest breaks or too few or too many rest breaks. Each exception on the daily Exception Report was to be investigated by the store manager or personnel manager in order to determine what corrections, if any, were needed to be added to the associate's time clock entries of the previous day. In the case of missing swipes for a rest break earned by the associate, the store managers were expected to inquire of the associate whether he or she had taken the break. If the break had been taken but not recorded, a Time Adjustment Request ("TAR") was expected to be completed and signed off on by the manager and associate. The time record was then to be edited to reflect the correction, and the final payroll record was now to include the insertion of the rest break. This process was to be followed even though there was no pay consequence for the associate.
Another Wal-Mart record is the Time Clock Archive Report which shows the times each associate worked for the entire pay period, and indicates the times each associate clocked in and out for his shifts, his meal breaks, and, prior to February 9, 2001, his rest breaks. Managers must investigate each incident on this report as well. Yet another Wal-Mart record is the Time Clock Punch Error Report which, on a daily basis, shows any clocking in or out errors an associate makes. Managers must correct the errors appearing on this report before they are able to finalize the daily payroll.
II. The Experts' Reports
The plaintiffs have presented the report and testimony of Shapiro, a professor in Emory University's Department of Psychology. He received a bachelor's of arts degree in psychology from Yale University, a Ph.D. in psychology and mathematics from Indiana University, and a juris doctor degree from Emory University. In this case, Shapiro applies his knowledge and experience as a statistician in formulating expert liability and damage opinions based on Wal-Mart's records.
Wal-Mart challenges Shapiro's analyses and methodologies with the report and testimony of Dr. Denise Neumann Martin ("Martin"). Martin has a bachelor's of arts degree in economics and French from Wellesley College, and a master's of arts degree and a Ph.D. in economics from Harvard University. She is currently the senior vice president of NERA Economic Consulting. The expertise she applies is also based on her knowledge and experience concerning the analysis of large databases.
The Revised Scheduling Order signed by this court (Murtagh, J.) on January 23, 2006, directed that expert witness disclosure ("including opinions and bases thereof") was due by March 15, 2006, for both the plaintiffs and Wal-Mart. In a report dated March 15, 2006 ("Shapiro Initial Report"), Shapiro proposed a methodology by which he would analyze Wal-Mart's timekeeping records. He wrote that, "[u]sing the finalized Class List [which, he asserted, he had yet to receive from Wal-Mart at that time] and the other electronic and written materials produced by Wal-Mart, [he] shall be able to identify the amount of time worked by the class members without compensation and breaks, as well as the corresponding wages owed to such class members." Shapiro Initial Report, ¶ 11. Wal-Mart's expert, Martin responded to Shapiro's report in her own report, dated April 17, 2006 ("Martin Initial Report"), in which she concluded that Shapiro's methodology was flawed for a number of reasons, including that Shapiro provided no assurances that he could accurately identify violations or improprieties, conduct that Martin maintained required individualized investigations.
As discussed further below, this methodology differs from the understanding this court (Murphy, J.) had previously with respect to Shapiro's role. See January 2004 Decision, at 7 (acknowledging plaintiffs' assertion that they intended "to produce expert statistical analysis [e.g., Shapiro Report] which will be probative and admissible" in order to "discount individualized non-actionable reasons for Wal-Mart's alleged failures to compensate"). The plaintiffs explain that the records themselves are sufficient to disclose specific information required to determine violations.
The cover sheet to the Martin Initial Report actually states, "April 17, 2005" but presumably that date is a typographical error.
Both Shapiro and Martin testified before this court at a hearing that took place over three days, from July 26 through July 28, 2006; their testimony was based on their Initial Reports. The hearing proceeded on the understanding that the methodology employed by Shapiro in his forthcoming final report would not differ from that in the Shapiro Initial Report, and, to the extent such methodology was not apparent in the Shapiro Initial Report, as explained in the testimony he gave at the hearings. Shapiro filed his final report ("Shapiro Report") with this court on October 6, 2006, and, as discussed further below particularly with respect to the one-minute clock-out, Shapiro altered his methodology without explanation.
At the time, the court was concerned that going forward with the Daubert-Lanigan hearing would not be meaningful if Shapiro made changes in his final report. The court believes that it received assurances from the plaintiffs that there would be no changes in Shapiro's methodology and that Shapiro's final report would only differ from the Shapiro Initial Report by providing a quantification of the violations attributed to Wal-Mart and a computation of damages. However, Shapiro's final report, filed in October 2006, included unexpected changes in his methodology in certain areas, and some of those adjustments appear to have been made in response to criticisms raised at the Daubert-Lanigan hearing.
In the Shapiro Report, Shapiro writes that he believes that he has "used appropriate methods for demonstrating the amount of time worked by class members without compensation and class-wide damages . . . [and that he] used appropriate methods for calculating class-wide damages and individual class member damages." Shapiro Report, ¶ 5. Martin responded to the Shapiro Report in her own report ("Martin Report"), dated October 9, 2006, in which she concludes that the Shapiro Report is speculative, unreliable, and inaccurate as he fails to take into consideration factors beyond the unreliable records themselves. See Martin Report, at 2-5.
Shapiro restricted his data analyses to the 67,506 class members appearing on the final electronic list with which Wal-Mart provided the plaintiffs on August 18, 2006. Shapiro Report, ¶ 10. To conduct his analyses, Shapiro "reviewed a voluminous amount of written and electronic materials produced by Wal-Mart in this and other like cases." Id. ¶ 6. Among the electronic materials he reviewed are "the following reports that are derived from the timeclock data: `Week-to-Date Hours and Expense Summary,' `FT/PT Associate Exceptions Report,' `Timeclock Activity Log,' `Timeclock Archive Report,' `Timeclock Punch Error Report,' `Timeclock Punch Exceptions Report,' and `Week-to-Date Supervisor Expense Summary.'" Id. ¶ 7. Shapiro also reviewed "Wal-Mart's policies and procedures concerning time-keeping, including but not limited to corporate policies relating to clocking in and out for work as well as the procedures for editing and/or modifying payroll." Id. ¶ 8. Finally, Shapiro "reviewed Wal-Mart's corporate policies and sworn testimony from hourly associates, store managers and corporate representatives concerning the policies and procedures utilized at Wal-Mart to ensure that hourly associates record their work and break time." Id. ¶ 9.
In the Shapiro Report, Shapiro quantifies occurrences of missed and interrupted rest periods occasioned by Wal-Mart, time shaving in the form of Wal-Mart-inserted one-minute clock-outs and meal breaks, and off-the-clock work performed by cashiers. Shapiro then calculates damages for such occurrences. Shapiro's opinions on these issues, and Martin's responses, are analyzed in the Discussion section, below.
III. The Plaintiffs' Liability Evidence
The plaintiffs' class liability evidence falls into a number of categories. Without attempting to be exhaustive in cataloging all of the evidence the plaintiffs will rely upon, these categories, beyond Wal-Mart's timekeeping databases, include the following:
a) Wal-Mart's corporate internal communications, before and during the class period, which acknowledge the widespread existence of time shaving and off-the-clock work caused by payroll pressure.
b) Grass roots complaints, some from Massachusetts associates, and cashier survey results for several years indicating ongoing problems with large numbers of associates, all pressured to cut short or miss their meal or rest breaks.
c) Audit Reports of store internal audits from September 1999 through July 2000, also indicating that a significant number of associates might not have been or in fact were not obtaining the meal and rest breaks to which they were entitled. None of the results were specific to Massachusetts stores.
d) A 1999 Store Audit Program which instructs auditors to determine missed breaks by using Wal-Mart's payroll records in the same way the plaintiffs and Shapiro are now using them to identify missed breaks. This document is supplemented by a series of audits in 2000 which implemented these instructions.
e) An audit report of July 17, 2000 ("Shipley Audit"), supplemented by Wal-Mart corporate follow-up treatment, which also indicated, based on a large number of exceptions appearing on the Exception Report, that Wal-Mart was not providing to associates in significant numbers the meal and rest breaks to which they were entitled. The Shipley Audit involved 128 stores across the United States, two of which were from Massachusetts. No specific results from the two Massachusetts stores are available.
f) Depositions and affidavits of the class representatives and of other Wal-Mart associates, a small number from Massachusetts and others from outside Massachusetts, who state they missed meal or rest breaks because of understaffing at their stores, and, in fewer numbers, who state they were subject to unauthorized meal insertions and one-minute clock-outs; and
g) Testimony of a relatively small number of current and former associates of Wal-Mart at corporate and store levels, acknowledging the practice of cost-cutting policies and their adverse impact on the ability of associates to obtain meal and rest breaks, acknowledging in fact the unauthorized existence of store managers using the one-minute clock-out and inserting meals, and acknowledging that Wal-Mart's policies were designed to encourage and condone such conduct.
The role intended for Shapiro in establishing the plaintiffs' liability case is not very clear based on the presentation made to the court. The plaintiffs have argued at times that he will testify as to both liability and damages, and at other times that his testimony will be focused on damages only. Neither the Shapiro Report nor his testimony clarifies the situation. It appears that in addition to quantifying damages, the plaintiffs will seek to have Shapiro pronounce with the imprimatur of an expert, with minor, non-quantifiable exceptions, that each instance of missing rest break swipes means that an associate missed his rest break because of Wal-Mart's staffing deficiencies; that every shortened rest break also results from the same deficiencies; that every one-minute clock-out not amended within seven days and every meal insertion recorded on the finalized payroll records represents time shaving by Wal-Mart; and that every instance of a cashier's presence at a cash register as reflected on the POS Database at a time when the cashier did not clock himself in on the Wal-Mart time clock system represents an associate working for Wal-Mart off the clock. In other words, the plaintiffs will have Shapiro imbue Wal-Mart's records with the ability to speak for class members or qualifying representatives in identifying specific instances when they have been injured or harmed by Wal-Mart actions or policies.
DISCUSSION I. Motion to Decertify the Class
The class as presently certified consists of approximately 67,500 past and current Massachusetts hourly associates. In the ten-and-a-third years of the class period, these associates worked at forty-seven different Wal-Mart stores, each with its own local management which certainly did not remain constant at all stores during the entire class period. Not all of the forty-seven stores were open for business during the class period, and some of them did not open until late in the class period.
Not surprisingly, given these circumstances, the somewhat conditional nature of the early certification orders, and the skeptical comments of Judge McHugh of the Appeals Court, this court, now that discovery is complete, is confronted with a request from Wal-Mart to decertify the class. Wal-Mart asserts that the class as certified is overbroad and is likely to consist of a not insubstantial number of associates who suffered no injury as a result of any breach of contract or violation of labor laws attributed to Wal-Mart. In pressing this point, Wal-Mart, once again, claims that the plaintiffs failed to satisfy the predominance requirement of Mass.R.Civ.P. 23(b), maintaining that Wal-Mart's records of associate's time cannot be read or interpreted in the relevant periods as an accurate description of what an associate was actually doing or why he was doing it. According to Wal-Mart, only individualized inquiry of the associates will reveal what a particular associate was doing (e.g., taking a rest break or not) or why he was doing it (e.g., logged in early from a break to converse with a co-worker). A need for such individualized inquiry to eliminate uninjured associates would of course not allow the plaintiffs to meet the predominance requirement of Rule 23(b) and decertification of the class would follow.
The predominance requirement as set forth in Mass.R.Civ.P. 23(b) requires that "the questions of law or fact common to the members of the class [as established by satisfying the commonality requirement of Mass.RCiv.P. 23(a)] predominate over any questions affecting only individual members . . ."
Even at this stage of the proceedings, the plaintiffs still carry the burden of providing sufficient information for the court to make a reasonable judgment that they meet all of the requirements for class certification under Rule 23(b). See Weld, 434 Mass. at 87 ("The plaintiffs bear the burden of providing information sufficient to enable the motion judge to form a reasonable judgment that the class meets the requirements of rule 23; they do not bear the burden of producing evidence sufficient to prove that the requirements have been met" (emphasis added)); see also Smilow v. Southwestern Bell Mobile Sys., 323 F.3d 32, 40 (1st Cir. 2003) ("Where . . . common questions predominate regarding liability, then courts generally find the predominance requirement to be satisfied even if individual damages issues remain"). The plaintiffs dismiss Wal-Mart's challenge to certification claiming that the United States Supreme Court's decision in Anderson v. Mt. Clemens Pottery Co., 328 U.S. 680 (1996) ("Mt. Clemens"), and a subsequent decision of the Ninth Circuit of the United States Court of Appeals, Rural Fire Prot. Co. v. Hipp, 366 F.2d 355 (9th Cir. 1966) (" Rural Fire"), compel this court to confirm the existing class certification of the pending matter. This court disagrees.
Neither Mt. Clemens nor Rural Fire involved a class action. Rather, they both involved actions brought under the Fair Labor Standards Act ("FLSA"). See 29 U.S.C. § 201 et seq. Under the FLSA there is specific authorization for an employee on behalf of similarly interested employees or the Secretary of Labor to initiate a case in either federal or state court for violations of the FLSA. See 29 U.S.C. § 216(b). No claim for violation of the FLSA has been asserted in the pending case. Under the FLSA, to become beneficiaries of litigation against the employer, the employees must expressly authorize the case to be brought on their behalf, or, in other words, "opt in" to the litigation as a member. See id. ("No employee shall be a party plaintiff to any such action unless he gives his consent in writing to become such a party and such consent is filed in the court in which such action is brought"). Under Mass.R.Civ.P. 23, class members have no right to "opt in" or "opt out" of the proposed class.
Mt. Clemens addressed claims of FLSA violations for some 1,200 piece-work employees who worked at a single pottery plant. 328 U.S. at 682, 684. These employees checked in and out on a time clock twice a day. Id. at 683. They were paid, however, only for the time from the succeeding even quarter hour after employees punched in to the quarter hour immediately preceding the time when they punched out. Id. at 683-84. The question became whether during those quarter-hour periods the employees were entitled to be paid for walking time from the timeclock to their work stations, for preparation activities necessary before productive work could begin, and for waiting time before or after a shift period. Id. at 684.
A master appointed by the United States District Court found that walking time and waiting time were not compensable. Id. at 685. He also found that the preparation work was not compensable because the employees had produced insufficient evidence from which the amount of such work could be definitely determined. Id. The District Court agreed with the Master with the significant exception that he believed that employees began productive work after an allowance for clocking in and walking time; the District Court established its own measurement for the allowance. Id. The District Court calculated an average time for the clocking-in process and walking time and subtracted this time from the period determined by the time punched in the morning to the scheduled quitting time. Id. at 685-86.
The Sixth Circuit Court of Appeals reversed and upheld the Master, concluding that the District Court's measurement was based on speculation. Id. at 686. The Court of Appeals further held that the employees had the burden of establishing not only that they did not receive wages for which they were entitled under the FLSA, but also the extent of the time worked for which they did not receive payment as required by the FLSA. Id. The employees could not satisfy their burden through an estimated average of the uncompensated work. Id.
The Supreme Court reversed the Court of Appeals, holding that the Court of Appeals and the Master had imposed too stringent a burden of proof on the employees. Id. The Supreme Court held that where an employer's records are inaccurate or incomplete in allowing a precise determination of the amount or extent of uncompensated work proven to have been performed by the employee,
an employee has carried out his burden if he proves that he has in fact performed work for which he was improperly compensated and if he produces sufficient evidence to show the amount and extent of that work as a matter of just and reasonable inference. The burden then shifts to the employer to come forward with evidence of the precise amount of work performed or with evidence to negative the reasonableness of the inference to be drawn from the employee's evidence. If the employer fails to produce such, the Court may then award damages to the employee, even though the result be only approximate.
Id. at 687-88.
In Rural Fire, the Ninth Circuit Court of Appeals affirmed the decision of the District Court which found that a company that provided private fire department services to a range of customers had violated the FLSA by failing to pay minimum hourly regular and overtime rates to its employee-firemen. 366 F.2d at 357. The employer argued on appeal that the employees' evidence of the number of hours worked was insufficient to support the District Court's findings of how much they had been underpaid. Id. at 359. The District Court had noted that the employer's own business records were an adequate source of information upon which to decide the hours worked. Id. The Court of Appeals, relying on Mt. Clemens, affirmed that use of the employees' time records was appropriate. Id. In so doing, the Court of Appeals stated:
That case [ Mt. Clemens] also establishes that the use of the employer's records to establish the amount of time worked is proper, and that if the employer has kept proper and accurate records the employees' burden of proof is discharged by merely producing those records. That is what the appellees did in this case, and we agree with the district court that since the records were kept in accordance with the practices of the employer and not those of the employees, and since the employer had found the records adequate for years for purposes such as income tax, social security, and insurance, it can hardly now claim that its records are improper. But even if we were to agree that the records which were kept by appellant were not accurate, the appellant would not be helped, for the Anderson v. Mt. Clemens Pottery Co. opinion goes on to state that where the employer has kept poor records the employees need only establish the amount of work performed as a matter of reasonable inference. Here the records, kept by management . . ., provide an adequate basis upon which to determine the hours worked and in light of its long continued reliance on these records for a variety of purposes, appellant's efforts to impeach them were a failure.
Id. at 359-60.
The plaintiffs argue that the principles of Mt. Clemens and Rural Fire should be applied in this case. The plaintiffs assert that if those principles are applied, the court must conclude that the missing swipe on the finalized Time Clock Archive Report for an earned rest break is adequate class-wide proof that the break was in fact not taken by the associate and that payment is due the associate. The plaintiffs also argue under Mt. Clemens and Rural Fire that their liability evidence coupled with their widespread evidence of missing and shortened rest breaks, time shaving, and off-the-clock work during the class period plus Wal-Mart's time records showing one-minute clock-outs and inserted meal breaks are proof enough of unpaid work performed by associates. Likewise, the plaintiffs claim that the results of the merged Wal-Mart POS and timekeeping databases show that work was performed by associates off the clock.
The court believes that the plaintiffs are misconstruing and misapplying the principles of Mt. Clemens and Rural Fire. The time records of Wal-Mart are not adequate of themselves to establish entitlement to compensation in a case like this, and Mt. Clemens and Rural Fire hold nothing to the contrary.
The decision in Secretary of Labor v. DeSisto, 929 F.2d 789 (1st Cir. 1991) (" DeSisto"), confirms this conclusion. In DeSisto, also an FLSA case, a key issue concerned generally the adequacy of proof for establishing a violation of the FLSA and specifically the requirements for representative testimony where the employees' payroll records were inadequate in demonstrating the extent of unpaid work. See 929 F.2d at 792-94. The First Circuit, relying in Mt. Clemens, pointed out that, in the absence of adequate records, the burden on the Secretary of Labor (substituting for the employees) was minimal. Id. at 792, 793. The court went on to state, however, that the burden is not nonexistent. Id. at 793. The court further recognized that not all employees needed to testify to prove a violation or to recoup back wages. Id. Testimony and evidence from representative employees was acceptable to meet the Secretary's initial burden. Id. at 792. Applying that reasoning to the facts before it, the court held in DeSisto that one employee could not represent 244 employees who held a variety of positions at varying locations. Id. at 793. First-hand knowledge of the requirements of each job position at each location would be required to prove a violation of the FLSA. Id.
The court interprets this triumvirate of FLSA cases as follows:
a) for every employee there must be proof of a violation, i.e., that the employee performed work for which he was not properly compensated;
b) a representative's testimony will suffice for proof of violation to other employees so long as the representative has first-hand knowledge that all employees similarly situated to him were subjected to the same treatment; and
c) that the records, with use of reasonable approximation techniques where the records are imperfect, will be acceptable proof of the extent of work performed by the employee, leaving the burden of proof with the employer to establish the employee's actual hours of work.
In this case, at a minimum, the plaintiffs would need to establish that every associate class member was actually working during the period for which there are missing swipes for an earned rest break, and that the associate did not take the rest break and/or that the associate did not voluntarily come back early from a break and was in fact performing work during the minutes the break was short of fifteen minutes. Moreover, the plaintiffs would need to establish that every associate class member was the actual associate working on the operator-accountable register which he was logged on to during the alleged "off-the-clock" period; that every associate class member performed work during the allegedly unauthorized management-inserted meal periods; and that every associate class member performed work for the shifts in which allegedly unauthorized one-minute clock-outs appear. However, the plaintiffs' trial plan does not include testimony of witnesses representative of the different positions held in each of the forty-seven Massachusetts stores covering the entire class period who could make the case of violations experienced by all class members. The plaintiffs' liability evidence is mostly non-specific to Massachusetts and will not satisfy DeSisto. The few depositions and affidavits of Massachusetts associates are clearly not representative in the sense demanded by DeSisto. If Mt. Clemens' principles, as explained by Rural Fire and DeSisto, govern this case as urged by the plaintiffs, then the court should decertify the existing class.
In the class action against Wal-Mart before the State of Minnesota District Court, Dakota County, the court (King, J.) agreed with the plaintiffs' argument in that case that "the procedural scheme in Anderson applies in this case," as well as with Wal-Mart's argument "that the Plaintiffs are misapplying the procedure in their arguments." Braun v. Wal-Mart Stores, Inc., Civil No. C0-01-9790, at 2 of Memorandum (Dakota Dist.Ct. April 19, 2006) (King, J.). Anderson, the court held, "initially places the burden on an employee to at least make a sustainable claim that he or she performed work which was not compensated." Id. Out of the class of approximately 57,000 people, however, only "a few hundred [associates] . . . directly claim that they were not paid for their work . . . This is hardly the `substantial' number that is required for a class action to go forward." Id. at 2-3 of Memorandum. The court held that Anderson requires that any class member "who wants to be compensated for paid work has to first come forward and claim that they did work for which they were not compensated." Id. at 3 of Memorandum (bold in original). To that end, and to avoid decertification, the court ordered that "a random sample survey . . . be taken of class members." Id. The plaintiffs' claims in Braun included alleged violations of Minnesota's FLSA ("MFLSA"). Id.; see Wal-Mart Stores, Inc. v. Lopez, 93 S.W.3d 548, 559-60 (Tex.App. 2002) (rejecting plaintiffs' reliance on FLSA and Mt. Clemens to grant class certification where plaintiffs had not previously raised FLSA).
The court proceeds further, however, to consider whether different types of proof may satisfy the requirement for Rule 23(b) class certification. The plaintiffs suggest, with dependence largely on Shapiro, that the Wal-Mart records are so error-free that they can be relied upon to identify specific associates who suffered missed breaks, time shaving, and off-the-clock work. The records are error-free, they claim, because the timekeeping system was designed to be error-free. Each exception, which included missed rest break swipes, was to be investigated daily by store managers and corrections were to be made for any inaccuracies. The finalized payroll record would incorporate any corrections, whether or not there was any pay consequence to the associate. For manager-generated one-minute clock-outs and meal insertions, again the managers were to talk with the associate and obtain an authorization from the associate or make an agreed adjustment. The off-the-clock work revealed by combining the POS Database with the timekeeping database is self-authenticating, according to Shapiro. He urges that overlapping the two different databases uncovers time when an associate was logged on to the POS Database but not logged in on the timekeeping database. According to Shapiro, these occurrences signal time that the associate was necessarily working off-the-clock.
Dealing first and most importantly with the subject of missed rest breaks, the court concludes that there is no class-wide evidence presented by the plaintiffs, including anything offered by Shapiro, that negates the need for individualized inquiry of associates or of representative testimony as defined by DeSisto. Under Massachusetts law, there is no requirement that paid rest breaks are to be afforded employees. The plaintiffs' claim for missed rest breaks is based upon a theory of implied contract derived from the Associate Handbook, orientation communications, and written policies. There is nothing in the law which precluded associates from waiving or taking advantage of the rest break opportunities provided to them. The court is aware that there exist no Wal-Mart policies allowing a waiver of rest breaks and, in fact, that there were policies in place, including disciplinary, which were intended to discourage associates from not taking their earned rest breaks. Such policies, however, were no guarantee that associates would always choose to take advantage of their rest break opportunities.
The plaintiffs argue and have presented evidence that skeletal staffing policies driven by cost-cutting and other bottom-line concerns have coerced some Massachusetts associates to miss or shorten their rest breaks so that associates could complete their work and attend to customer needs. The plaintiffs supplement this argument with other evidence non-specific to individual Massachusetts stores that missed breaks by significant numbers of associates was a common occurrence. Wal-Mart, on the other hand, has presented evidence from certain Massachusetts associates and managers that associates voluntarily chose not to take their breaks. A compensable violation for breach of implied contract exists for missed rest breaks only if Wal-Mart directly or indirectly caused associates to miss or shorten their breaks. A compensable violation does not exist if an associate chose freely to miss or cut short a rest break opportunity available to him or her.
Individual questions regarding the amount of damages does not preclude class certification. Waters v. Earthlink, Inc., Civil No. 2001-00628, 2006 WL 1549685, *6 (Barnstable Super.Ct. May 10, 2006) (Kane, J.). A class action is not an inappropriate vehicle when individual questions of the amount of damages exist for each potential class member. Id., citing Weld, 434 Mass. at 92. There is a critical distinction, however, between the existence of any harm and the measurement of that harm. Id., citing Newton v. Merrill Lynch, Pierce, Fenner Smith, Inc., 259 F.3d 154, 189 (3rd Cir. 2001). "Where proof of fact of damage requires evidence concerning individual class members, the common questions of fact become subordinate to the individual issues, thereby rendering class certification problematic." Id. (emphasis added), quoting Newton, 259 F.3d at 188. Therefore, when the court must explore whether any harm actually resulted, individual questions predominate and class certification is not appropriate. Id. Aspinall v. Philip Morris Cos., 442 Mass. 381 (2004), provides, in short, that in a G.L.c. 93A certification, de minimis inclusion of uninjured persons in a class will be tolerated. The application of this principle to certification under Mass.R.Civ.P. 23 is uncertain, but in making this decision, the court has assumed that the de minimis inclusion is allowable.
In this case, to determine whether any associate was injured — e.g., whether an associate was coerced to miss a rest break or not — mandates individualized inquiry of the associates. In this court's view, such a determination cannot be made on a class-wide basis. Cf. Ungar v. Dunkin' Donuts of Am., Inc., 531 F.2d 1211, 1226 (3d Cir. 1976) (denying certification in antitrust action where each plaintiff "individually, must prove that his purchases were coerced as an element of establishing a prima facie case of illegal tying"); Hewitt v. Joyce Beverages of Wis., Inc., 97 F.R.D. 350, 354 n. 9 (N.D.Ill. 1982), aff'd, 721 F.2d 625 (7th Cir. 1983) ("There has been no allegation of a uniform agreement in which the distributors were forced to adhere to the defendants' resale price policy. Therefore, any overt action . . . would be individualized in nature and thus not appropriately dealt with in the class action format"); DiCostanzo v. Hertz Corp., 63 F.R.D. 150, 152 (D.Mass. 1974) ("[C]ommon questions are not the predominating questions where there must be proof of coercion of individual franchisees and proof of injury to the individuals"); Fletcher v. Cape Cod Gas Co., 394 Mass. 595, 603-04 (1985) (affirming Superior Court's denial of certification where "issue of liability could not be decided on a class wide basis" because "questions relating to proximate cause of specific injuries, . . . knowledge of defect or danger, adoption of representations, reliance upon misrepresentations, and time of discovery all require[d] individualized proof"). But see Lockwood Motors, Inc. v. General Motors Corp., 162 F.R.D. 569, 580 (D.Minn. 1995) ("[C]oercion may be implied on a class-wide basis when evidenced by a uniform agreement common to the class"). The class of approximately 67,500 associates is premised and justified on the existence of class-wide evidence of rest break violations. By Shapiro's calculations, the number of associates in each category impacted by the alleged incidents of one-minute clock-outs, meal insertions, and off-the-clock work represent only fractions of the 67,500 class now certified and could not support the broad class presently certified. Because the existence of coerced rest breaks cannot be properly treated on a class-wide basis, the class as certified is clearly overbroad and must be decertified.
In the California class action against Wal-Mart, the Alameda County Superior Court made a similar finding, denying "the motion to certify a class to pursue monetary relief for rest break violations. At the class certification stage, the Court recognized that individualized issues would predominate on rest break claims for monetary relief because `[t]he mere fact than an [associate] missed or took a short rest break is insufficient to state a claim — an [associate] must demonstrate that the employer coerced or encouraged the [associate] to forego their break or take a short break.'" Savalgio v. Wal-Mart Stores, Inc., Civil No. C-835687, at 52-53 (Alameda Super.Ct. Sept. 17, 2006) (Sabraw, J.), quoting Class Certification Order of November 6, 2003, at 23.
There is no expectation that the plaintiffs will provide the kind of comprehensive representative testimony allowed by DeSisto or a close substitute. Perhaps the need for a representative with first-hand knowledge to testify as expressed in DeSisto could be avoided if there were an alternative for assuring that all class members were harmed or deprived of pay by Wal-Mart. Although this court has made no determination that such an alternative exists, interviews or surveys of associates identified from the class lists and representatives of the stores and job positions at the stores covering the class period might have some merit. Such efforts, however, were not undertaken by Shapiro or anyone else on the plaintiffs' behalf. The absence of such representative testimony or any alternatives necessitates individualized inquiry of the approximately 67,500 past and present associates to determine, for example, if that individual associate actually missed a rest break and, if so, whether he did so of his own volition or because Wal-Mart's understaffing policies left him with no choice.
Even if the required coercive element for missed rest breaks did not bar class treatment for the 67,500 Wal-Mart associates, class certification is still improper in this case. As discussed in the next section, the unreliability of Shapiro's opinions and, indirectly, of Wal-Mart's records do not provide an acceptable basis for supporting a finding of harm, with de minimis exceptions, to 67,500 members of the associate class. The plaintiffs rely on Smilow, 323 F.3d 32 (1st Cir. 2003), as support for their argument that regular business records, such as timekeeping records in this case, can be used to establish the fact of injury and not just the amount of damages. Smilow is distinguishable from this case.
Each plaintiff in Smilow "signed a standard form contract for cellular phone services with" the defendant, 323 F.3d at 35, which the plaintiffs purported "guarantee[d] free incoming call service." Id. at 34. The court found that the terms of the contract, identical for all class members, constituted a common factual basis; the common question of law was whether the contract terms precluded the defendant from charging for incoming calls. Id. at 39. The defendant argued that certification was improper because, even if there were common questions of law and fact, "individual issues predominate[d] on damages." Id. at 40. The court held, however, that where "common questions predominate regarding liability, the courts generally find the predominance requirement satisfied even if individual damage issues remain." Id. (Emphases added.) Moreover, common issues predominate even where individual factual determinations regarding damages "can be accomplished using computer records, clerical assistance, and objective criteria — thus rendering unnecessary an evidentiary hearing on each claim."
Potential liability in Smilow therefore arose out of the contract all of the plaintiffs signed. See id. at 39, 40-41. The defendant's records were capable of identifying those parties harmed by their receipt of incoming calls, as demonstrated by the plaintiffs' expert who stated that he could extract from the defendant's records those customers who had received incoming calls, those customers who were billed and paid for the incoming calls, and actual damages for each class member. Id. at 40. The existence of those events, as shown on the records, was therefore beyond challenge, unlike in this case where the mere existence of, for example, missed rest break swipes, is subject to varying explanations, not all of which constitute improper conduct by Wal-Mart. Furthermore, the defendant's liability in Smilow rested on an interpretation of the contract all of the plaintiffs signed; the plaintiffs' expert would use the defendant's records only to calculate damages. See id. at 40-41. Here, however, the plaintiffs intend to rely on Wal-Mart's records not only to calculate damages, but also to establish liability. Smilow is accordingly inapposite.
Weld is similarly distinguishable from this case and does not provide support for the plaintiffs' argument that Wal-Mart's records, alone, are sufficiently accurate to establish class-wide liability. In Weld the plaintiffs seeking class certification were CVS pharmacy customers whose prescription drug profiles CVS had scanned, without the plaintiffs' consent or knowledge, pursuant to the selection criteria that pharmaceutical companies provided to CVS. Id. at 89. The plaintiffs then received letters "printed on CVS letterhead . . . [with] the name of the sponsoring pharmaceutical manufacturer appear[ing] at the bottom. Each letter recommended a different course of action . . ." Id. at 83. The defendants asserted "that the determination of liability and damages . . . turn[ed] on individualized questions regarding the content of each letter, the customer's medical condition, and the subjective emotional reaction of each customer to the letter received." Id. at 91.
The court disagreed, holding that liability could be established without reference to the content of each letter or to the reaction each recipient had to the letter. Id. at 92. Rather, liability arose out of the defendants' "single course of conduct[,]" and whether the defendants' program violated statutory and/or common law "turn[ed] largely on common questions of law and fact regarding the duty CVS owed to its customers and the defendants' conduct in implementing the program." Id. Therefore, as in Smilow, where the fact of signing the contract was sufficient to trigger potential liability, see 323 F.3d at 29, the fact that the plaintiffs received a letter in Weld established that the defendants used their information without the plaintiffs' consent or knowledge; it was up to the jury to determine whether the defendants' conduct was unlawful. See 434 Mass. at 92.
The facts in Weld, while similar to those in Smilow, are distinguishable from those before this court. In Weld, the plaintiffs all received a letter as a result of the defendants' program; there was no variation as to the reason. See id. Conversely, here, while Wal-Mart's records arguably show which associates missed rest breaks, or were subject to meal break insertions or one-minute clock-outs, or were not paid for off-the-clock work reflected in the POS Databases, each instance does not necessarily constitute a potential violation. Consequently, there is variation as to the reasons for the occurrences the plaintiffs point to as improper.
Thus, nothing in Smilow or Weld permits this court to extend the holdings of Mt. Clemens and its progeny to hold that, beyond establishing damages, Wal-Mart's records alone can establish Wal-Mart's alleged wrongdoings on a class-wide basis. Moreover, and as discussed further below, the plaintiffs have failed to demonstrate that Shapiro's analysis of the records supplies the records with any credibility as to their accuracy. Wal-Mart's motion to decertify the class is accordingly ALLOWED.
II. Motion to Exclude Shapiro's Testimony A. Standard of Review
"That a person qualifies as an expert does not endow his testimony with magic qualities." Canavan's Case, 432 Mass. 304, 313 (2000), quoting Boston Gas Co. v. Assessors of Boston, 334 Mass. 549, 579 (1956). The purpose of the Daubert-Lanigan test, therefore, "is to prevent an expert from offering testimony to a fact finder that is not based on reliable methodology." Id. at 315; see Commonwealth v. Patterson, 445 Mass. 626, 649 (2005) (defining trial judges' "traditional role as gatekeepers . . . [as] protecting juries from evidence that had little chance of being reliable"). "[T]he burden is on the proponent of expert testimony to demonstrate its reliability, not on the opposing party to refute it." Palandjian v. Foster, 446 Mass. 100, 112 n. 7 (2006).
In making this determination as to the admissibility of the expert testimony, the court's focus "must be solely on principles and methodology, not on the conclusions that they generate." Daubert, 509 U.S. at 595. An "expert's mere assertion that a methodology is reliable is [not] sufficient to pass the [Daubert-] Lanigan test absent any other evidence showing its reliability." Canavan's Case, 432 Mass. at 315. Moreover, "nothing in . . . Daubert. . . requires a . . . court to admit opinion evidence that is connected to existing data only by the ipse dixit of the expert. A court may conclude that the there is too great an analytical gap between the data and the opinion proffered." General Electric Co. v. Joiner, 522 U.S. 136, 146 (1997).
The Supreme Court in Daubert set forth five factors, not "a definitive checklist or test[,]" to assist the trial court in conducting its "preliminary assessment of whether the [expert] testimony is sufficiently valid and of whether that reasoning or methodology properly can be applied to the facts in issue." 509 U.S. at 592-93; see Kumho Tire Co. v. Carmichael, 526 U.S. 137, 141, 149 (1999) (extending Daubert's application to testimony of "experts who are not scientists[,] . . . [including] testimony based on `technical' and `other specialized knowledge'"); Canavan's Case, 432 Mass. at 313 (applying Daubert-Lanigan analysis to "expert testimony based on, inter alia, personal observations and clinical experience"). These factors consider whether the methodology the expert utilized to reach his conclusion "can be (and has been) tested[;] . . . whether the [methodology] . . . has been subjected to peer review and publication[;] . . . [what] the known or potential rate of error [is;] . . . [whether] standards controlling the [methodology's] operation [exist and are maintained;]" and whether the methodology is generally accepted in the relevant community. Daubert, 509 U.S. at 593-94.
The trial court not only has "latitude in deciding how to test an expert's reliability, and to decide whether or when special briefing or other proceedings are needed to investigate reliability, [but it also has latitude in] decid[ing] whether or not that expert's relevant testimony is reliable." Kumho Tire Co., 526 U.S. at 152 (emphasis in original). Moreover, "[n]ot all of the factors identified in Daubert will be applicable in every case; a trial judge accordingly has broad discretion to determine how to assess the reliability of expert testimony." Palandjian, 446 Mass. at 111. In fact, in Lanigan, the Supreme Judicial Court "accept[ed] the basic reasoning of the Daubert opinion[,] . . . [but held that] general acceptance in the relevant . . . community will continue to be the significant, and often the only issue." 419 Mass. at 26; see Palandjian, 446 Mass. at 112 (noting "broad discretion that Lanigan and its progeny provide trial judges in assessing the reliability of expert testimony).
B. Reliability of Shapiro's Methodology
Shapiro contends that he can identify from Wal-Mart's records those associates who experienced the missed or shortened breaks, meal insertions, and one-minute clock-outs and will use that information to determine entitlements to damages, amounts of damages, allocation of any damages. A fundamental error in the methodology employed by Shapiro in his formulation of his liability and damage opinions regarding rest breaks, inserted meal breaks, one-minute clock-outs, and off-the-clock work is his failure to take steps to verify or test his assumption that the records of Wal-Mart, with de minimis exceptions, are accurate in revealing violations of Wal-Mart's contractual commitments to its associates and/or of Massachusetts labor laws. In other words, he failed to explain whether the reality of how the system was employed matched the way the system was designed to work. See Daubert, 509 U.S. at 593 ("[A] key question to be answered in determining whether a theory or technique is . . . knowledge that will assist the trier of fact will be whether it can be (and has been) tested"). He concedes no possibility of a meaningful margin of error or deviation in the time records or in his created database concerning off-the-clock work despite the fact that the class was originally certified by this court (Murphy, J.) based in part on the plaintiffs' assurance that Shapiro, as their expert statistician, would "discount individualized non-actionable reasons for Wal-Mart's [alleged] failures to compensate." January 2004 Decision, at 7; see id. (noting that there "demonstrably are . . . numerous innocuous reasons why a particular [associate] did not receive compensation for labor").
In Shapiro's opinion, nothing could be more accurate than the contemporaneous business records of Wal-Mart and no surveying or interviewing of associates, regardless of what might be revealed, should even be considered as having sufficient value to alter or discount in any way the conclusion to be drawn from those business records. See Hearing Transcript (July 27, 2006), at 263 ("There is no reason to look beyond those contemporaneous records. They are, by any argument, the best record of what happened"). Shapiro reasons that the design of the timekeeping system, coupled with the close monitoring and enforcement of the system requirements reflected by the TARs, assured him that the Wal-Mart records were extremely reliable with only de minimis errors possible. See id. ("The system was designed to make them the best records"); see also id. at 231-32 ("[I]f you look at the electronic data, you can see the enforcement or carrying out of the company policies which are all designed to make the data reliable and trustworthy for the purposes we're putting it to"). Further, he suggests that the records, in effect, are self-verifying, thereby rendering redundant and unnecessary any after-the-fact verification testimony.
Wal-Mart's finalized payroll records, according to Shapiro, are therefore the most accurate assessment of events which in some instances occurred many years ago. In reaching this conclusion, Shapiro has made a judgment that any witness testimony or statements that have been furnished by Wal-Mart and which challenge the accuracy of the Wal-Mart records must be disregarded as inconsequential without conducting an investigation into whether the records correspond to the reality of the system Wal-Mart had in place. Shapiro has the same view about the value of interviewing any group of associates or other store associates after the events were recorded in Wal-Mart records.
Shapiro dismisses the thought that his talking with associates sometimes years after a payroll was finalized would yield any information reliable enough to cast doubt on the accuracy of Wal-Mart's contemporaneous records. See id. at 270-71. He questions whether anyone can seriously believe that an associate will remember if or why he or she, for example, missed or took a shortened rest break on a given date years ago. See id. at 269. This valid observation, however, does not excuse his total acceptance of the records as the chronicle of reality. While there would certainly be no expectation that an associate would have such a memory of a given date, specific memories are not necessarily required to legitimize Shapiro's methodology.
From the list of those he considers eligible for relief for violations, i.e., the class list, Shapiro might make representative inquiries such as whether the associate always took the breaks to which the associate was entitled; whether the associate would at times not clock in or out for breaks because of forgetfulness, inconvenience, or other reasons; if the associate failed to clock in or out for a break, whether he always executed a TAR; whether the associate was careful to review his time records or otherwise verified that he was receiving appropriate pay for hours worked; whether the associate failed to log out on an operator-controlled register and someone else used the register under his identity; whether the associate always clocked in on the timekeeper system before going on a register; whether the associate ever authorized a meal break insertion by management without completing a TAR.
Moreover, Shapiro opines that, even if what an associate or store manager might later say to contradict the record is true, the number of such "innocent" incidents must be de minimis in view of Wal-Mart's regular investigation of exceptions and corresponding adjustments made to payroll records. In other words, Shapiro alleges that the Wal-Mart timekeeping system and enforcement practices would not allow room for inaccuracies in the time record beyond a de minimis level. Neither the Shapiro Report nor Shapiro's testimony, however, reveals what he considers to be de minimis reflections on the accuracy of Wal-Mart's records.
Finally, Shapiro points to Wal-Mart's internal audit reports and the Shipley Audit as demonstrating that Wal-Mart reported missing breaks to senior corporate management using only finalized payroll records as he has done, i.e., a missing break swipe equates to a missing break in fact. Wal-Mart has provided evidence that the audit procedures were based on erroneous assumptions, do not reveal anything specific about what was occurring in Massachusetts stores, and do not reveal anything about what was happening after mid-2000. Regardless, the bigger point is that Shapiro ignored this evidence in total without making an effort to determine whether there was any truth to Wal-Mart's claims that the records did not tell the complete story.
The court believes that it is a statistician's job to understand the limitations in the data he is collecting and organizing before drawing any conclusions from it. See SMS Sys. Maint. Servs. v. Digital Equip. Corp., 188 F.3d 11, 25 (1st Cir. 1999) ("[A]n expert must vouchsafe the reliability of the data on which he relies . . ."); see also Santos v. Chrysler Corp., 430 Mass. 198, 205 (1999) (excluding expert's testimony where "factual foundation for the expert's opinion was insufficient"). Accordingly, the court concludes that Shapiro's rigid refusal to do anything to ascertain the accuracy of the records on which his opinions are based by determining whether reality comported with the policies Wal-Mart had in place makes his methodology unreliable. See Canavan's Case, 432 Mass. at 315 ("The purpose of the Lanigan test is to prevent an expert from offering testimony to a fact finder that is not based on reliable methodology . . . [The court] cannot conclude that the expert's mere assertion that a methodology is reliable is sufficient to pass the Lanigan test absent any other evidence showing its reliability"); see also Ruiz-Troche v. Pepsi Cola of Puerto Rico, 161 F.3d 77, 81 (1st Cir. 1998) ("[W]hile the methodology remains the central focus of the Daubert inquiry, this focus need not completely [disregard] judicial consideration of an expert's conclusions. Rather, trial judges may evaluate the data offered to support an expert's bottom-line opinions to determine that data provides adequate support to mark the expert's testimony as reliable").
What follows is a discussion of each area the Shapiro Report addresses, as well as Martin's analyses of Shapiro's methodology, this court's specific conclusions, and, finally, this court's ultimate conclusions with respect to Shapiro's methodology as a whole and his ability to testify as an expert.
1. One-Minute Clock-Outs
The plaintiffs allege that Wal-Mart managers clocked out associates one minute after they clocked in to work in order to minimize payroll expenses. An associate subject to a one-minute clock-out would only be paid for one minute of work even though he worked his entire shift. Wal-Mart disputes this characterization and claims that the one-minute clock-out was merely used as a placeholder when a manager saw that an associate had an odd number of swipes on a given day because, for example, he clocked in at the beginning of his shift, but he did not clock out at the end of his shift.
Therefore, the dispute notwithstanding, a one-minute clock-out entered by store managers was not a reflection of the time an associate worked, and the very act of inserting the "placeholder" rendered the associate's record inaccurate. The store managers had other options, such as inserting the clock-out at the end of the associate's scheduled shift, so that the associate would receive payment even if the manager failed to follow up on the insertion. The selection of the one-minute clock-out was an option that disfavored the associate. It put the associate at risk that if the one-minute clock-out was not corrected, the associate would lose credit for hours he actually worked. But see Robin Miller Affidavit, ¶ 13 ("Sometimes . . . the associate would not get back to [her] until after [she] had finalized payroll for the week. In that case, to ensure that they were paid in full, [she] would calculate the number of hours that needed to be added . . . Then [she] would add those hours to a day in the new pay week. This addition would not appear as an edit to the associate's time for the day with the one-minute swipe. Thus while the one-minute swipe would still appear, the associate would be paid in full").
Shapiro "tabulated the number of computer-edited clock-outs one minute after the recorded beginning of the day's work shift or one minute after the recorded return from a break." Shapiro Report, ¶ 20. If "corrective editing" had been performed "in the timekeeper data during the contemporaneous pay period[,]" the one-minute clock-out would have been eliminated and "thus, . . . not included in [his] analysis." Id. Therefore, "a one-minute clock-out was counted only if the one-minute clock out remained uncorrected at the close of the current week's payroll." Id.
Shapiro also reviewed Wal-Mart's payroll records in order "to determine if an adjustment to regular pay had been made before the end of the next payroll period following the one-minute clock-out." Id. ¶ 21. If such an adjustment had been made, he would not have included the corresponding one-minute clock-out in his calculations. He further "analyzed Wal-Mart's cash payment records within the timekeeper database to determine if a cash payment had been made within seven days of the end of the payroll period in which the one-minute clock-out occurred." Id. Similarly, he would not have counted a one-minute clock-out for which an associate had already received cash payment. Shapiro noted that because "Wal-Mart retained 86.38% of the biweekly Timeclock Archive Reports . . . [and because] an additional 7.8% of the data were unusable due to the illegibility of the paper records[,]" he had to "calculate the total number of observed and extrapolated instances of uncompensated one-minute clock-outs and the expected total monetary loss." Id. ¶ 24.
At the July 2006 hearing, however, Wal-Mart store manager Gerard Roy testified that cash payments were not necessarily recorded.
For the time period of August 21, 1995, through December 30, 2005, Shapiro counted 29,176 uncompensated one-minute clock-outs. He calculated that the lost wages were $879,238. For the time period of August 22, 1998, through December 30, 2005, Shapiro counted 21,383 uncompensated one-minute clock-outs. He calculated that the lost wages were $740,365.
Shapiro presumably made these calculations using two different time periods because of the two different statutes of limitations involved. See G.L.c. 149, § 150 (providing, "[a]ny employee claiming to be aggrieved by a violation of section . . . 148 . . . may, . . . within three years of such violation, institute and prosecute in his own name and on his own behalf, or for himself and for others similarly situated, a civil action"); G.L.c. 260, § 2 (setting forth six-year statute of limitations for contract actions).
In response, Martin points out that Shapiro changed his methodology from the Shapiro Initial Report, dated March 2006. In the more recent Shapiro Report, Shapiro writes that he has determined "whether subsequent compensation was made in a 7-day period following the pay period, rather than a 5-week period following the" one-minute clock-out, which was his intention, as stated in the Shapiro Initial Report. Martin Report, at 6 (emphases added); see Shapiro Initial Report, ¶ 28. Originally, in the Shapiro Initial Report, Shapiro described his methodology with respect to the one-minute clock-out as follows:
I then shall analyze Wal-Mart's payroll records for an additional five-week period, to see if the aggrieved [associate] had received any one-time payments for the amount deleted from his or her timekeeping record by virtue of the one-minute clock-out. (The Wal-Mart Payroll/Scheduling Guide provides for a five-week window during which week-earlier . . . payments may be made through the Associate Database system. After five weeks, any payroll adjustment must be telephoned in to [Wal-Mart headquarters in] Bentonville).
Shapiro Initial Report, ¶ 28. Martin illustrates the problem with Shapiro's "new approach" with an example using an actual associate and his corresponding records:
five days before the end of the pay period ending May 30, 2002, an [associate] with associate ID 920 in store 1906 had a one-minute punch inserted by a manager. Under Dr. Shapiro's previous methodology [set forth in the Shapiro Initial Report, ¶ 28,] this one-minute punch would have been excluded from Dr. Shapiro's estimate of alleged damages as the associate received a flat pay adjustment 24 days later. However, under Dr. Shapiro's new approach, this flat pay adjustment will not be considered as it occurred more than 7 days after the end of the pay period.
Martin Report, at 6-7. By failing to consider the records five weeks after the alleged one-minute clock-outs, therefore, Shapiro overstates his estimate of alleged damages. Id. at 2, 6-7.
In July 2006, Shapiro told this court under oath that the most accurate and non-arbitrary way to calculate one-minute clock-outs was by looking at Wal-Mart's payroll records for five weeks after the one-minute clock-out occurred. See Hearing Testimony (July 27, 2006), at 221. Shapiro, however, altered that methodology in the final Shapiro Report without explanation.
At the October 11, 2006, final pre-trial conference at which the Shapiro Report was discussed, plaintiffs' counsel stated that they instructed Shapiro not to look beyond seven days after the close of the pay period in which the one-minute clock-out occurred because of the terms of G.L.c. 149, § 148, on which they base their one-minute clock-out claim (Count IX). The effect of following that instruction makes Shapiro's calculation of damages flowing from those instances of one-minute clock-outs inaccurate. Section 148 of G.L.c. 149, concerning the late payment of wages, provides that,
[e]very person having employees in his service shall pay weekly or bi-weekly each such employee the wages earned by him to within six days of the termination of the pay period during which the wages were earned if employed for five or six days in a calendar week or to within seven days of the termination of the pay period during which the wages were earned if such employee is employed seven days in a calendar week, or in the case of an employee who has worked for a period of less than five days shall, within seven days after the termination of such period, pay the wages earned by such . . . employee-during such period . . .
"The purpose of the weekly wage law is clear: to prevent the unreasonable detention of wages." Boston Police Patrolmen's Ass'n, Inc. v. City of Boston, 435 Mass. 718, 720 (2002). "As such, it bestows upon individual workers the right to recover not only their lost wages, but also treble damages, reasonable attorneys fees, and costs." Newton v. Commissioner of the Dep't of Youth Servs., 62 Mass.App.Ct. 343, 345 (2004) (emphasis added), citing G.L.c. 149, § 150.
Assuming for purposes of this discussion that every instance of a one-minute clock-out is attributable to improper conduct on the part of Wal-Mart, the seven-day period after the end of the pay period in which the one-minute clock-out occurred would establish a violation of the statute where Wal-Mart made no payment corresponding to the one-minute clock-out during that time. By looking five weeks beyond the end of the pay period in which the one-minute clock-out occurred, however, Shapiro would be able to determine if Wal-Mart ever paid the associate for the one-minute clock-out, a factor relevant to calculating damages. If Wal-Mart did pay the associate at some point, the associate would not be entitled to damages in the form of lost wages, as the associate was already paid, but the associate would be entitled to interest earned on the unpaid wages from the time the wages should have been paid to the time when they were actually paid. See Parow v. Howard, Civil No. 02-1403, 2003 WL 23163114, *4 (Middlesex Super.Ct. Nov. 12, 2003) (Graham, J.) [17 Mass. L. Rptr. 149] (determining "that damages [were] to be awarded to the Plaintiffs based on the interest (calculated at a rate of twelve percent per annum) earned on the overtime wages for the first time period beginning when the wages were unlawfully withheld until payment was received"); see also Munson v. Valente, Civil No. 04-3626, 2005 WL 2461883, *3 (Middlesex Super.Ct. Aug. 4, 2005) (Murphy, J.) [19 Mass. L. Rptr. 672] (holding that, where plaintiffs had been paid wages later than statute permits and where treble damages were appropriate, treble damages were limited "to the `time value' of the period of illegal deprivation of the wages, and sets . . . interest factor of three . . . percent. . . to be applied[;]" and noting that "trebling of the actual wages withheld would be both a `double recovery' windfall for the . . . plaintiffs and draconian against defendant").
Neither the plaintiffs nor Shapiro, not to mention the court, can make this assumption. For this reason, too, Shapiro's methodology is unreliable.
Damages and liability are intertwined in this case, and Shapiro's methodology purports to establish violations and calculate the damages flowing from those violations. As a consequence of employing this methodology, however, the calculations for the class are wrong in the first instance, rendering certification pointless. Shapiro's methodology with respect to one-minute clock-outs is therefore unreliable.
2. Missed and Interrupted Rest Breaks
Wal-Mart associates receive paid fifteen-minute rest breaks for every three hours of work. According to Shapiro, a rest break violation "occurs if an [associate] has worked for three or more hours and: (1) there is no rest break; or (2) the rest break is interrupted before the completion of the required fifteen minutes. A work period is defined as the time between the first and last time-clock entry minus the length of all unpaid meal periods." Shapiro Report, ¶ 28. Using the same database "as those used in the analyses of one-minute clock-outs[,]" Shapiro calculated the number of earned rest breaks and compared that number "to the number of rest breaks actually recorded in the shift." Id. (emphasis added). Wal-Mart changed its policy requiring associates to clock in and out for rest breaks after February 8, 2001, and Shapiro accounts for this change by extrapolating data from pre-February 8, 2001 to post-February 8, 2001. In the Martin Report, Martin writes that Shapiro's "analysis of rest break data is speculative and unsupported." Martin Report, at 2.
Given the absence of a statement to the contrary and given the logic in, and necessity of, ascertaining whether associates were already compensated for missed and/or interrupted rest breaks, it is likely that by this statement, Shapiro is indicating that he considered the payroll records to determine if there had been an adjustment in pay for each allegedly affected associate as well as the records of cash payments.
a. Missed Rest Breaks Prior to February 8, 2001
With respect to missed rest breaks, Shapiro "calculated the monetary value. . . as one-quarter of an hour at each affected [associate's] hourly pay-rate on the date of the missing rest break." Shapiro Report, ¶ 29. For the time period of August 21, 1995, through February 8, 2001, Shapiro observed 3,859,643 missed rest breaks on the 86.38% of the records that Wal-Mart retained. After "correct[ing] for the missing and illegible Time Clock Archive Reports, . . . the total number of observed and interpolated missed rest breaks increased to 4,911,737." Id. He calculated the monetary value as being $9,935,594. Shapiro performed the same calculations for the time period of August 22, 1998, to February 8, 2001. He observed 1,969,051 missed rest breaks; that number increased to 2,074,250 after Shapiro "corrected for the missing and illegible Time Clock Archive Reports . . ." Id. ¶ 30. He calculated the monetary value as being $4,728,922.
This 1998 date is likely in response to this court's September 2006 summary judgment decision limiting the time period for plaintiffs' Count IX to August 21, 1998, through December 31, 2005.
Martin maintains that Shapiro's "analysis is based on flawed and incorrect assumptions." Martin Report, at 8. First, associate deposition and affidavit testimony establishes that associates regularly took breaks without clocking in or out, and that there was no incentive for managers to enforce the swiping requirement because rest breaks were already paid. The records therefore do not "provide an accurate depiction of actual rest break behavior." Id. Second, associate deposition and affidavit testimony indicates that associates also regularly skipped their rest breaks. Again, as a consequence, "use of the timekeeping data to determine whether Wal-Mart denied associates their rest breaks will necessarily be inaccurate, unreliable and speculative." Id. The aspect of Shapiro's methodology with respect to rest breaks that this court finds most troublesome is Shapiro's failure to interview any associates or to conduct any sort of survey in order to determine the accuracy of Wal-Mart's records, despite his awareness that Wal-Mart associates and managers in depositions and affidavits swore that the time keeping regimens of Wal-Mart were not always followed when it came to rest breaks. Some associates claimed that they ignored the requirement to check in and out for breaks because the process would take too much time and interfere with their enjoyment of the break. See e.g., Kimberly Clark Affidavit, ¶ 12; Kerry Pasquarosa Affidavit, ¶ 10. Others claimed they would simply forget to clock in and out for breaks. See, e.g., Kristine McKim Affidavit, ¶ 10; Leeann Newman Affidavit, ¶ 12. Some managers asserted that not all of these incidents would be investigated because there was no payroll consequence. See, e.g., Susan Stewart Affidavit, ¶ 11 (stating that, as personnel manager, she "was never particularly strict in the way in which [she] dealt with missed 15 minute rest breaks punches . . . Associates were getting paid for the 15 minute rest break anyway so these punches didn't matter"). Adjustments to time records were also not made for the same reason, that the associate was paid regardless of whether she or he took the break.
Shapiro has further ignored the evidence Wal-Mart has provided supporting its position that Wal-Mart did not force associates to skip their earned rest breaks and that associates often neither clocked out for rest breaks nor filled out TARs to remedy those missed swipes. See, e.g., Alain Desjardin Deposition, at 32 (stating that no one at Wal-Mart ever "denied [her] the opportunity to take a rest break"); Diane Raczka Deposition, at 48-49 (stating that no one "at Wal-Mart ever refused to let [her] take a rest break" and that, over her seven years at Wal-Mart, she never "hear[d] an associate . . . complain that they [sic] weren't allowed to take their rest breaks"); Robert Rolandelli Deposition, at 129 (stating that, as store manager, he placed responsibility for taking earned rest breaks in associates themselves); Kimberly Clark Affidavit, ¶ 12 (stating that "[t]here were times" when she did not clock in and out for her rest breaks because she "did not want to waste [her] time walking backwards and forwards from the clock"); Kerry Pasquarosa Affidavit, ¶ 10 (stating that she rarely clocked in and out for rest break because she "found it very annoying and a waste of time having to walk back and forth to punch for a break that [she] was getting paid for anyway"[;] and that she "never filled in a [TAR] to show" her rest breaks); see also Kristine McKim Affidavit, ¶ 10 (stating she occasionally forgot to clock in and out for rest breaks, but that she "did not fill out a [TAR] when [she] forgot to punch for . . . [her] rest breaks, because [she] was getting paid anyway"); Leeann Newman Affidavit, ¶ 12 (stating that he often forgot to clock in and out for rest breaks, but that he "did not fill out a [TAR] when [he] forgot to punch out for rest breaks").
The plaintiffs have provided some evidence supporting their position that Wal-Mart forced associates to skip their earned rest breaks. See, e.g., Coaching for Improvement Form for Donna Prouty (November 4, 1998) (noting in "Associate's Comments" section that she "need[s] [her] breaks, a[n] ongoing problem in . . . Grill [where she worked]"); Telephone Complaint from Monica (December 8, 1998) ("Caller says it is difficult to get breaks . . . due to lack of help in the store [in Avon, Massachusetts]"); Michael Warrington Letter to "Mr. Stringer" (April 24, 2000) ("Nobody seems able to cover our breaks and meal times. This passed [sic] Saturday and Sunday . . . I [was]. . . stuck in the Garden Center because they had nobody [scheduled] to work the whole weekend"); Marcia Perry Letter to "Cynthia" (May 3, 2002) ("I have gone many shifts without a break because there is no coverage for my position. I am working longer hours, taking shorter, or no breaks . . ."); Richard Aubut Deposition, at 45-46 (stating that he and co-workers would not have been able to complete truck unloading tasks without working through breaks; "[t]hat's the only way [they] could get . . . [unloading tasks] all done"); Kristy M. Mitchel Deposition, at 22-23 (stating that she complained about not getting earned rest breaks while working as cashier); Joanne Perreault Deposition, at 5, 30 (stating that when she worked at Wal-Mart she did not receive earned rest breaks because store was understaffed, and that she quit her job at Wal-Mart because she was not "getting enough breaks"); Elaine Polion Deposition, at 71-72, 81 (testifying that, more than once per week, she did not receive both earned rest breaks, although she always received one earned rest break; and testifying that she never failed to punch out for rest breaks); Crystal Salvas Deposition, at 162-63 ("[O]ne to two times a week [she] would get [her] two 15-minute breaks . . . On the one or two times a week when [she] did not get [her] two 15-minute breaks, [she] . . . got at least one 15-minute break . . . So . . . one to two times a week [she] would not get her full allotment of 15-minute breaks[.]").
It appears that the associate received this verbal coaching because when she took a rest break, she left the snack bar counter unattended and failed to "respond to a page over the intercom for several minutes." Coaching for Improvement Form for Donna Prouty.
The date is not clear, but it appears to be November 4, 1998.
Presumably "Cynthia" is a store manager or personnel manager as this letter is in the context of requesting a raise. Marcia Perry Letter ("I don't mind doing the work I am doing, I am only asking to be compensated for my efforts. I feel that a dollar an hour raise would not be too much to ask due to the increase of work I am expected to do").
Presumably, Shapiro believes that the evidence the plaintiffs provide are characteristic of the class as a whole, with Wal-Mart's evidence constituting mere deviations on a de minimis scale. Neither Shapiro nor the plaintiffs, however, have demonstrated that these associates are in any sense representative of all Wal-Mart associates. Of course, Wal-Mart's witnesses on these subjects can also not be considered representative of all Wal-Mart associates. Given that the size of the class as currently certified is over 67,000, neither of the parties have provided an adequate sampling to demonstrate the extent to which associates, either voluntarily or involuntarily, did not take earned rest breaks.
Another troubling aspect of Shapiro's methodology is his testimony that the pattern of missed breaks assured him that the missed swipes in Wal-Mart's time records were in fact missed breaks as opposed to indications of forgetfulness or convenience by the associate. He observed that there was a greater incidence of missing swipes in the second or later part of a shift for which the associate was entitled to two breaks. See Hearing Testimony (July 27, 2006) at 238-39. Because the associates demonstrated by clocking in and out in the early shift period that they were aware of the requirement for doing so, Shapiro reasoned that the missing swipes in the later part of the shift meant that the associate missed the break. Shapiro's logic is not compelling, and, again, reliance on the records alone is insufficient. In the later part of their shifts the associates might have felt less energized to take the trouble to walk to the time clock. Shapiro's logic does not address the incidents described in testimony and affidavits of associates taking their breaks but forgetting or choosing not to clock in or out for breaks, or of associates voluntarily choosing not to take their breaks.
Shapiro's logic similarly does not address the incidents for which the associates received coaching, the very fact of which indicates that people were not necessarily clocking in or out for breaks. Wal-Mart's coaching practices suggest that before the coaching stage that resulted in a written record, there might very well be oral coachings with respect to an associate's objectionable behavior such as taking a break but failing to check in and out for breaks. Shapiro did not compare the written coachings given to associates because of their failing to clock in and out for breaks with TARs to determine whether in all or most instances the finalized payroll records on which Shapiro relies were a result of edits reflecting that the associate took the rest break but failed to check in and out for it. In fact, Shapiro did nothing to account for the instances that an associate might have been coached orally for failing to check in and out for a break which the associate took. See, e.g., Verbal Coaching Agreement of Dina MacBarron (November 30, 1998) (describing verbal coaching of associate who "fail[ed] . . . to punch out for break on" five occasions). As he spoke with no Massachusetts Wal-Mart associates or manager, Shapiro could not know whether any oral coachings resulted in a recognition of the taken break on the finalized payroll record.
At the pre-trial hearing, counsel for plaintiffs conferred to the court that Shapiro had not performed this task.
Presumably this coaching form concerns an associate working in a Massachusetts Wal-Mart.
Shapiro's dismissal of Wal-Mart's evidence of breaks taken but not recorded as not affecting the accuracy of the records, and his failure even to mention them in the Shapiro Report contribute to a finding of unreliability of Shapiro's methodology. He does not — and, this court believes, cannot without speaking to a cross-section of associates currently contained within the class — allow for any margin of error with respect to his analysis of the records. Instead, he equates each instance of a missed swipe with improper conduct on Wal-Mart's part. Shapiro's methodology of relying on the records without any sort of investigation as to their accuracy is accordingly unreliable for identifying actual missed breaks.
b. Missed Rest Breaks After February 8, 2001
After February 8, 2001, Wal-Mart no longer required its associates to swipe in and out for rest breaks. To calculate the post-February 8, 2001 missed rest breaks when associates no longer had to clock in and out for their rest breaks, Shapiro "calculated the percentage of shifts which had no missing rest breaks for each calendar year from 1995 to 2000." Shapiro Report, ¶ 33 (emphasis added). He "then performed a regression analysis, predicting percentage of shifts without missing rest breaks as a function of calendar year." Id. He used that analysis "to predict the percentage of shifts without missing rest breaks" in the five-year period of 2001 through 2005. Id. (emphasis added). Thereafter, he multiplied the predicted percentages by the actual number of shifts in the years 2001 through 2005. He "calculated that 3,945,965 shifts of the total 8,470,137 shifts would have no missed rest breaks from 2001 to 2005." Id.
As an initial matter, the plaintiffs take the position that this change in policy reflects Wal-Mart's reaction to the wage and hour actions being filed against it throughout the country and that this elimination of the swiping requirement was Wal-Mart's attempt to complicate associates' abilities to prove violations. Wal-Mart disputes this implication, asserting that the change in rest break policy resulted from internal complaints from associates about having to take the time to clock in and out for rest breaks. According to Wal-Mart, the change in policy is therefore an example of Wal-Mart's receptiveness to associate suggestions and its willingness to structure an associate-friendly workplace. See "Payroll Processes Question and Answers," Exhibit 16 to Affidavit of Carolyn Beasley Burton in Support of Plaintiffs' Opposition to Wal-Mart's Motion to Exclude Testimony of Dr. Martin Shapiro ("As [Wal-Mart's] facilities have gotten larger, it takes Associates quite a bit of time to walk back and forth to the time clock to clock in and out for breaks. [Wal-Mart] ha[s] also recognized that the personnel manager spends a lot of time on the administrative process of the payroll process. [Wal-Mart] feels] that [its] Associates are responsible to take their breaks as they are intended and [Wal-Mart] do[es] not need to monitor that process.").
Shapiro then performed a regression analysis "on the number of shifts with one or two missed rest breaks during the time period from 1995-2000." Id. ¶ 34. He determined "that among the shifts with missed meal breaks, less than 60% of the shifts had only one missed rest break and that more than 40% of the shifts had two or more missed rest breaks." Id. (footnote added). He "assign[ed] one missed rest break to 60% of the shifts with missing rest breaks and assigned two missed rest breaks to 40% of the shifts with missing rest breaks." Id. He concluded that "there were 6,297,913 missed rest breaks in the period 20012005." Id. ¶ 35, After multiplying that number by one-quarter of the average hourly pay-rate during that time period, he determined the monetary value to be $15,288,184.
It is unclear if "meal" is a typographical error; if not, Shapiro does not explain why he focuses his attention on shifts where there are missed meal periods in order to calculate missed rest breaks.
Martin writes that Shapiro's extrapolation method is "speculative, [and] unfounded . . ." Martin Report, at 9. As "his calculations and conclusions are not reliable even for the period prior to February 9, 2001 [,]" he cannot legitimately rely on those records to calculate post-2001 damages. Id. Moreover, Shapiro "compounds the unreliability of his extrapolation by failing to account for events at Wal-Mart during his five year extrapolation period that would have changed rest break behavior[,] . . . [such as the] various initiatives regarding meal and rest break compliance" that Wal-Mart undertook. Id. As a result of his ignoring "all possible effects that these initiatives had on rest break behavior, [and] instead assuming that the patterns he observes in missed rest break swipes, for the period 1995 through 2001, remain true for his 2001 through 2005 extrapolation period[,] . . . [Shapiro's] already speculative and unreliable estimates likely are increasingly unreliable over time." Id.
Given the unreliability of Shapiro's methodology as applied to the records from pre-February 8, 2001, this court credits Martin's arguments. As a consequence, this court reaches the same conclusion it reached with respect to his methodology as applied to the pre-February 8, 2001, records. Additionally, as Martin points out, Shapiro ignores Wal-Mart's post-2001 initiatives to encourage rest break compliance and sets forth no basis for his assumption that the extrapolation from one time period to the other is proper. Cf. Muñoz v. Orr, 200 F.3d 291, 301-02 (5th Cir. 2000) (excluding expert statistical testimony where expert "began his analysis with the assumption that [the challenged] . . . promotion system discriminated against Hispanic males[;] . . . fail[ed] to consider other variables such as education and experience as explanations for any observed discrepancy between promotion rates[;] relied on the plaintiffs' compilations of data, which gives rise to a `common-sense skepticism' regarding the expert's evaluation[;] . . . did not seek to verify the information presented to him").
The above discussion notwithstanding, another striking issue arises out of Shapiro's extrapolations, both in the context of the POS Database and in the context of rest breaks. The plaintiffs have asserted throughout Shapiro's involvement in this litigation that Shapiro's analysis of Wal-Mart's records will enable them to attribute damages to specific class members, as opposed to calculating an aggregate amount. Given that Shapiro extrapolates where the records are incomplete and/or illegible, Shapiro's claim that he can identify who was injured is, to a certain extent, an exaggeration. It is unclear how Shapiro would attribute damages to particular class members when he is extrapolating data. Neither Shapiro nor the plaintiffs addressed this matter.
The level of speculation involved in this extrapolation is worsened by the fact that the forty-seven stores at which the class members worked over the course of the ten-year class period were not all open during the 1995-2000 time period from which Shapiro extrapolates data. For example, the Shipley Report was distributed in July 2000, and the plaintiffs commenced this action in October 2001. Shapiro has made no mention of accounting for, or at the very least considering, whether stores that opened after those dates were more aggressive about enforcing the rest break policy than before the Shipley Report was performed. See Bazemore v. Friday, 478 U.S. 385, 400 n. 10 (1986) (Brennan, J., concurring in part, joined by all Justices) (holding that while "omission of variables [in regression analysis] may render the analysis less probative than it otherwise might be," such omission affects "probativeness, not its admissibility" as long as analysis "accounts for the major factors"[;] and noting that "[t]here may, of course, be some regressions so incomplete as to be inadmissible as irrelevant").
This criticism is based on data set out in the POS section of the Shapiro Report and is also discussed below in the context of the POS Database. Specifically, Shapiro writes that "[i]n 1995, there were data for 22 stores; in 1996, there were data for 25 stores; in 1997, there were data for 26 stores; in 1998, there were data for 30 stores; in 1999, there were data for 32 stores; in 2000, there were data for 35 stores; in 2001, there were data for 39 stores; in 2002, there were data for 42 stores; in 2003, there were data for 43 stores; in 2004, there were data for 45 stores; and, in 2005, there were some data for 47 stores." Shapiro Report, ¶ 48.
In the class action against Wal-Mart before the Superior Court of California, Alameda County, Shapiro testified at trial that he "generated individual rest break `violation' estimates based on swiping records for rest breaks in a six-week period in 2001." Savaglio v. Wal-Mart Stores, Inc., Civil No. C-835687, at 24, ¶ 65 (Alameda Super.Ct. Sept. 27, 2006) (Sabraw, J.). The court found his extrapolation analysis "unpersuasive" because the time span from which he extrapolated was "too small and narrow in scope to support Dr. Shapiro's conclusions regarding rest breaks." Id. at 24, ¶ 66. Moreover, "[t]he data alone cannot distinguish those instances in which [associates] voluntarily elected to skip their breaks or forget to swipe in and out for their breaks or were forced to miss their breaks. The data was not vetted for accuracy through [associate] interviews or other surveys nor was there any effort to consider the impact of Wal-Mart's rest break policies on the accuracy of the swipe data." Id. "There is simply no evidence one way or the other whether any of the . . . exceptions identified in . . . [the records] resulted from Wal-Mart coercion or pressure . . . [W]hen associates took, but did not swipe for, rest breaks, those rest breaks are not violations and are not reflected in the electronic data." Id. at 26, ¶¶ 71-71. The California court accordingly held that "Dr. Shapiro's rest break `violation' extrapolations from the pure rest break swipe record alone should be given little weight." Id. at 50, ¶ 150; see id. at 50, ¶ 151 (relying on Pacific Gas Elec. Co. v. Zuckerman, 189 Cal.App.3d 1113, 1135-36 (1987), for proposition that "[w]here an expert bases his conclusion upon assumptions which are not supported by the record, upon matters which are not reasonably relied upon by other experts, or upon factors which are speculative, remote or conjectural, then his conclusion has no evidentiary value"); see also Toubiana v. Priestly, 402 Mass. 84, 91 (1988) ("A mere guess or conjecture by an expert witness in the form of a conclusion from basic facts that do not tend toward that conclusion any more than toward a contrary one has no evidential value." (quoting Kennedy v. U-Haul Co., 360 Mass. 71, 73-74 (1971))).
c. Interrupted Rest Breaks
Shapiro "calculated the monetary value of these interrupted rest breaks as the corresponding fraction of an hour at each affected [associate's] hourly pay-rate on the date of the interrupted break." Shapiro Report, ¶ 31. Shapiro limited his calculations of interrupted rest breaks to the time period prior to February 8, 2001, the date on which Wal-Mart no longer required its associates to swipe in and out for rest breaks. Id. ¶ 33 n. 4. For the time period of August 21, 1998, to February 8, 2001, he observed 1,605,696 interrupted rest breaks. After "correct[ing] for the missing and illegible Time Clock Archive Reports[,] . . . the total number of observed and interpolated interrupted rest breaks increased to 2,033,597." Id. ¶ 31. He calculated the monetary value as being $763,209. He conducted the same analysis for the time period of August 22, 1998, to February 8, 2001, in which he observed 957,224 interrupted rest breaks. After "correct[ing] for the missing and illegible Time Clock Archive Reports[,] . . . the total number of observed and interpolated interrupted rest breaks increased to 1,024 337" Id. ¶ 32. He calculated the monetary value as being $402,540.
Presumably, Shapiro so limited his calculations because he understood he would not be able to extrapolate with any degree of reliability the incidents of interrupted rest breaks prior to February 8, 2001, to the time period after that date.
Martin has two criticisms of Shapiro's methodology with respect to interrupted rest breaks. First, Shapiro "assumes that any short rest break was involuntarily short, even though there is no basis in the data for making this assumption." Martin Report, at 8. Second, Shapiro identifies as violations those rest breaks "that were only slightly short." Id. As "it is natural that associate swipes are not precisely on time, causing their rest breaks to be slightly more or less than 15 minutes, . . . [and as,] under Wal-Mart's policy, rest breaks lasting at least 12 minutes would not have been flagged as exceptions [on the Exception Report], so would not have been investigated by managers[,]" the records on which Shapiro rely are inaccurate. Id. at 9. Martin contends that Shapiro did characterize as an interrupted rest break, and, consequently, a violation, those rest breaks lasting between twelve and fifteen minutes. When those rest breaks between twelve and fifteen minutes are excluded from Shapiro's analysis, "the number of rest break exceptions calculated using Dr. Shapiro's approach drops by 67.9 percent . . ." Id.
As with the missed rest breaks, Shapiro's methodology assumes that the records on which he relies for calculating interrupted rest breaks establish violations with complete accuracy without setting forth any basis to support this assumption. Shapiro dismisses as de minimis any evidence indicating that associates voluntarily cut their rest breaks short even though common sense indicates that associates would come back for any number of reasons, not just that they were compelled to do work.
For example, an associate could clock in early from his break, then chat with another associate before returning to work; an associate's own watch could be a few minutes faster than the Wal-Mart clock; if it is cold outside, an associate could decide to clock back in early and, stand around for a few minutes rather than stand outside for the full fifteen minutes.
Wal-Mart has provided evidence of associates voluntarily cutting their rest breaks short. See, e.g., Kimberly Clark Affidavit, ¶ 13 (stating that "[s]ometimes, [she] might not take [her] 15-minute rest breaks in full . . . No one has told [her] to do this . . . [and] [she] know[s] that if [she] wanted to take [her] 15-minute breaks in full [she] could do so"); Kenneth Fleet Affidavit, ¶ 13 (stating that "[t]here are . . . times when [he] make[s] the choice not to take a full 15-minute break because [he] would rather spend [his] time on the job than sitting around. [He] consider[s] this to be a completely voluntary decision . . . [and] [he] ha[s] never been ordered or felt pressured to return early from either a meal or a rest break"); Ann Pierni Affidavit, ¶ 12 (stating that "usually" she does "not stay on break for the full 15 minutes because [she] get[s] bored and do[es] not need to be on a break that long. This is [her] own choice"); see also Carol Kozlowski Affidavit, ¶ 13 (stating that occasionally she takes a slightly shorter or longer rest break, "but [she] find[s] that the time generally evens out . . . No one has ever suggested that [she] should miss a . . . rest break . . . [She] ha[s] never found that [her] breaks are interrupted"). The plaintiffs have provided considerably less evidence supporting their position that Wal-Mart forced associates to cut their rest breaks short in order to work. See, e.g., Tina DaSilva Deposition, at 100 (testifying that the statement she made in her affidavit that her rest breaks were never interrupted was not true); Crystal Salvas Deposition, at 163 (stating that "there were other times when [she] got her second break, but [she] was called back [s]o [she] always got at least one full 15-minute break . . . [o]n [her] six-hour shift").
In his deposition, however, Kenneth Fleet testified that he felt pressured to come back early from his rest breaks after he signed the affidavit in October 2003. Kenneth Fleet Deposition, at 85-86.
Common sense further indicates that associates cannot possibly clock in at exactly the fifteen minute mark, thereby making Shapiro's failure to subtract from his calculations those early clock-ins that do not even show up on the Exception Report an error that contributes to a finding of unreliability. Shapiro's methodology cannot bolster the records to establish liability and it is equally unreliable for calculating damages.
3. Point of Sale Database
The plaintiffs allege that Wal-Mart failed to compensate the plaintiffs for time worked on the POS Database and rely on Shapiro's analysis of the records to calculate the amount of time, and the damages resulting therefrom, that an associate was logged off of the timekeeping records but logged onto the POS Database. In order to log onto a register, the associate must enter his "private four-digit password." Shapiro Report, ¶ 39. "Cashiers are instructed to log off of cash registers before leaving the registers." Id. ¶ 41. As discussed further below, any off-the-clock work that the associate performed while logged on to the POS Database and simultaneously logged out of the timekeeping database is arguably the fault of the associate himself.
The POS Database "contains records of the time each individual cashier logged on or logged off a cash register, cashier identifying information, and the time of each sale transaction." Id. The POS Database's records contain this information for "every cash register in every store in Massachusetts for the entire period August 21, 1995 through December 31, 2005[,] . . . [and there are separate files] for each store for each day of the week." Id. ¶ 37. The POS Database "contains two types of relevant records" both of which contain the cash register identification number and the operator number of the cashier. Id. ¶ 39. One type of record also contains the name of the cashier and the "logon and logoff times[,]" while the other type of record also contains the time of the register transaction. Id.
Shapiro combined the POS Database data "from a representative, unbiased sample of 25 of the 47 stores for which data was produced" because the volume of data from the POS Databases was "very large[.]" Id. ¶ 37. Shapiro lists the twenty-five stores he used as representatives by their store numbers, but he does not explain why he chose the stores he did. "Of the 25 store sample there were 21 stores open in 2001, 24 stores open in 2002 and 2003, and 25 stores open in 2004 and 2005." Id. ¶ 48.
In order to identify off-the-clock work, Shapiro "merged the Timekeeper tables in the Associate database with the [POS] Database and re-sorted the records by operator identification number." Id. ¶ 42 (footnote added). Shapiro used "the Operator table . . . [to] link . . . [the] cash-register operator identification number [used in the POS Database] and associate identification number [used in the timekeeper tables]." Id. ¶¶ 40, 42. He further confirmed he was merging the timekeeper tables and POS Database correctly by comparing "the cash-register operator's name in the [POS] [D]atabase . . . to the associate's name in the Associate database . . ." Id. Shapiro admits that "nine percent of the [POS] records contained Operator identification numbers which were not linked in the Operator table to an [associate's] Associate identification number. Those records were not included in [his] subsequent analyses." Id. ¶ 43. He further limited his analysis by date because archived electronic timekeeper data were available only for 2001 through 2005 and "there are no missing time clock data during this time period." Id. ¶ 38.
Shapiro limited the timekeeping data to the same twenty-five stores he used for the POS Database data. Id. ¶ 37.
Once Shapiro merged the two databases, "[t]he resulting file contained the interleaved cash register actions and Timekeeper records of each cashier in chronological order for the twenty-five Massachusetts stores in the time period 2001-2005." Id. ¶ 44. In order to "ensure proper matching of Associate and Operator and to minimize the effect of any mismatching upon the results if any mismatches were to occur[,]" id. ¶ 45, Shapiro placed the following additional restrictions on his analysis:
the registers had always been operator-accountable registers and never register-accountable registers (¶ 44);
the associate identification number must be linked to the operator identification number in the Operator table (¶ 45);
"if the operator is already operating the cash register when he/she goes off-the-clock, the measurement of the cash register off-the-clock work time excludes any continuing transactions and does not begin until the operator logs onto the cash register again with his/her private password" (¶ 45);
the off-the-clock work does not exceed eight hours (¶ 45); and
the off-the-clock work begins "within 18 hours of the previous on-the-clock work or must end 18 hours before the beginning of the next on-the-clock work" (¶ 45).
Shapiro applied this methodology to the records and determined that between 2001 and 2005, "there were 298,054 cash register operations off-the-clock in the twenty-five stores [which is measured as] 695,551 minutes (11,592.5 hours). The off-the-clock cashier activity occurred in every store every year that each store was open." Id. ¶ 46. Shapiro acknowledges that as a result of Wal-Mart's "introduction of a software alteration in 2003 which prevented [associates], without an over-ride by management [associates], from logging on to a cash register unless they were clocked in for work" the number of off-the-clock hours decreased. Id. For this reason, when extrapolating back to the period before 2001, Shapiro used the average result from 2001 to 2002 only.
After conducting this extrapolation, Shapiro concluded that the total number of off-the-clock hours in 1995 was 1,464.3, resulting in $9,767 in unpaid wages; 4,662.6 hours equaling $32,029 in 1996; 4,745.0 hours equaling $34,306 in 1997; 5,476.0 hours equaling $40,796 in 1998; 5,840.0 hours equaling $46,311 in 1999; and 6,387.5 hours equaling $51,611 in 2000. Id. ¶ 49. The total unpaid wages for 1995 through 2000 is $214,820. Id. The total amount of unpaid wages for the class period, 1995 through 2005, is $389,105. Id. ¶ 51. "The total reflects the decrease in off-the-clock work beginning in 2003 when Wal-Mart began electronically preventing cashiers from logging onto cash registers unless clocked into the time clock or accompanied by a manager override." Id. ¶ 51.
Martin disputes that Shapiro's methodology is able to "generate a reliable estimate of off-the-clock work. Indeed, no reliable estimate of off-the-clock work can be generated solely from the data Dr. Shapiro uses." Martin Report, at 10. First, Martin states that associates "may (either intentionally or inadvertently) operate even an operator-accountable register using another associate's operator ID . . . [such as] when, for example, one associate leaves the register but fails to log off." Id. Martin conducted her own examination of the POS Database for a particular store using operator-accountable registers during the time period of January 1, 2001, through October 31, 2005. "In the data [Martin] observed shifts that exceed 21 hours in duration. Such findings clearly prove that associates must have left registers without logging off." Id. Martin acknowledges that Shapiro is able to exclude "absurd situations" from his calculations by restricting the off-the-clock work to shifts that "last no more than eight hours, and . . . to shifts that begin within 18 hours of the previous on-the-clock work or end within 18 hours of the next on-the-clock work." Id. She also points out, however, that his restriction does not remedy the unreliability of his methodology because Shapiro still classifies the remaining instances of apparent off-the-clock work as violations. "Nothing in the data Dr. Shapiro is using can distinguish such situations from any situations that are actually improper." Id.
Second, Martin notes "that the time clock governing the POS system was not always synchronized to the time clock governing the timekeeping system." Id. at 11. Moreover, "the POS system tracks time in hours, minutes and seconds, whereas timekeeping data is tracked in hours and minutes. If the timekeeping clock differed from the POS clock by even a few minutes, it could lead to numerous instances which Dr. Shapiro considers off-the-clock work that were not truly time off-the-clock." Id.
Third, Martin raises several issues with respect to TARs. Shapiro does not appear to have reviewed "the TAR forms produced in this case so has made no adjustment for these payments." Id. at p. 12. That notwithstanding, often associates made time adjustment requests verbally, and some TARs were lost or destroyed. Even where TARs were submitted properly, they were not always accurate.
For example, some associates failed to swipe out for meal breaks even though they took a meal. Some of those associates filled out TARs so that their timekeeping data was adjusted to reflect actual hours worked . . . [A] cashier on an operator-accountable register who took a meal break from 11:45-12:15 might simply indicate that she took a half hour meal break `around noon,' . . . [and] her manager would enter a meal break from 12:00-12:30 in the timekeeping data. In this case, Dr. Shapiro would find that the cashier was using the POS system from 12:15 until the end of her meal break . . . [and] wrongly attribute 15 minutes of off-the-clock work to this associate.
Id. at p. 11.
Finally, Shapiro gives no "reported basis" for his decision to use twenty-five stores to calculate off-the-clock work for all of the Massachusetts stores, and he has "not done any evaluation of the fundamental assumption underlying this vast extrapolation — namely, that the amount of alleged off-the-clock work would be the same across all stores in Massachusetts. Id. at p. 12. Shapiro's unsupported assumption "that the alleged off-the-clock hours that he estimates at 25 stores with at least some operator-accountable registers would be a reliable basis for assuming off-the-clock work at the remaining 22 stores" further renders his methodology unreliable. Id.
Interestingly, with this methodology, Shapiro suggests that Wal-Mart's timekeeping records accurately register when associates log in and out except when the POS Database contradicts that record. Shapiro has therefore placed a preeminence on the POS Database over the associates' own logging out and logging in practices. This methodology appears to run counter to the methodology he applies elsewhere, i.e., using the timekeeping records to establish liability and calculate damages with complete accuracy.
The situation for which the plaintiffs seek compensation was consequently caused by the plaintiffs themselves; there is no claim that Wal-Mart forced associates not to clock in or out, and, in fact, the plaintiffs rely on Wal-Mart's policy that associates must clock in and out as support for their assertion that Wal-Mart's records are accurate. Moreover, Shapiro has not inquired of any group of cashiers whether they believed they ever worked off the clock. Before he can reliably testify that his conclusions are accurate with respect to liability and damages, Shapiro must attempt to verify whether the POS Database is a better reflection than the timekeeping record the associate created through clocking in and out as to when the associate was actually at work and not at work. See Ruiz-Troche, 161 F.3d at 81 (holding that data on which expert relies must provide "adequate support to mark the expert's testimony as reliable").
That notwithstanding, this court accepts Martin's criticisms of Shapiro's methodology, including her discussion of the TARs. Shapiro's strict adherence to the times listed on the timekeeping records does not take into consideration that the times might not be exact given that the associate who filled out the TAR did not write the exact time he took his meal or rest break. Assuming Shapiro could accurately match up the timekeeper records with the POS Database, any discrepancies he found are not necessarily discrepancies, and it is unlikely that Shapiro could correct for such occurrences even if he did consider them. Therefore, Shapiro's methodology of presuming that the times are accurate and of matching those times to the corresponding times on the POS Database to determine liability and calculate damages is unreliable.
With respect to Wal-Mart's argument that Shapiro has done nothing to investigate, let alone account for, synchronization issues that might exist from combining the POS Database and the time clock database, this court points to and credits the July 2006 testimony of Chris Ferguson ("Ferguson"), an employee of Wal-Mart's Information Systems Division at its Arkansas headquarters. Ferguson testified that, as of August 2001, Wal-Mart implemented a process that synched the systems seven times per day. See Hearing Transcript (July 27, 2006), at 45. Accordingly, given this testimony, any argument as to the synchronization of the databases, or the lack thereof, concerns the weight of Shapiro's testimony rather than the reliability of his methodology.
Compounded with these flaws in his methodology is Shapiro's discounting entirely the testimony of Wal-Mart associates that the associate at an operator-accountable register might check out for a break without logging off the register and another associate might step into her spot to cover and operate the register still logged on with the now-absent associate's identification number. For example, Alicia Chandonnet, a former Wal-Mart cashier, stated in her affidavit that although she "was good about signing on and off the cash register[,] [she] observed . . . other cashiers who would not always remember to sign off so when someone took their place on the cash registers, they were not using their own code." Alicia Chandonnet Affidavit, ¶ 8; see also Kenneth Fleet Affidavit ¶ 17 (stating that when he "go[es] to use a [non-operator-accountable] register and it is still open, meaning that someone has recently logged on and used the register[,] . . . [he] do[es] not always log the other person off and log [him]self in"). Consequently, in those instances, the associate Shapiro identifies as having worked off-the-clock was not the actual associate using the register at that time.
Here, too, this court credits Martin's own examination of the POS Database for operator-accountable registers in one store for a particular time period that revealed shifts that exceeded twenty-one hours. Even though Shapiro expressly excludes off-the-clock work in excess of eight hours, Shapiro has no evidence on which to base his assumption that the explanation for the record showing an associate worked twenty-one hours off-the-clock is different from the explanation for the record showing an associate worked eight hours off-the-clock, or even one hour off-the-clock. Thus, Shapiro's methodology of using only these records renders his analysis deficient as he has not done the necessary investigation to reach any conclusions regarding the accuracy of the records.
Furthermore, Shapiro's methodology employs a double extrapolation which, when considered together, also contributes to a finding of unreliability. First, the amount of POS Database data he received was very large, covering forty-seven stores and spanning from 1995 to 2005. This large volume of data was presumably unmanageable, therefore Shapiro considered only the POS Database data from twenty-five of the forty-seven stores. Shapiro acknowledges that, "[o]f the 25 store sample there were 21 stores open in 2001, 24 stores open in 2002 and 2003, and 25 stores open in 2004 and 2005." Shapiro Report, ¶ 48.
Second, because electronic timekeeping records do not exist for years prior to 2001 and because of the 2003 software upgrade theoretically preventing associates from logging onto the POS Database while logged out of the timekeeping database, Shapiro used the average off-the-clock hours from 2001 to 2002 and extrapolated backwards to 1995 to 2000. Therefore, the data from 2001 to 2002 undisputably does not concern the entire class, and the results from 1995 to 2000 are artificially constructed through the use of that incomplete data. These two extrapolations thereby contribute to a finding of unreliability. Moreover, in 1995, only twenty-two stores were open. Shapiro's methodology that extrapolates data from twenty-five stores to a time period when fewer stores were open is unreliable, not only for the methodology itself, but also because Shapiro has failed to explain whether and how he took this discrepancy into account.
Although this issue arguably does not concern reliability, it is unclear how Shapiro calculated results for the years 2003, 2004, and 2005 when Wal-Mart had implemented a computer system to prevent off-the-clock work on the POS Database. Shapiro does not explain this phenomenon, nor does it appear that he conducted any sort of investigation.
4. Meal Period Insertions
Finally, the plaintiffs allege that, "in response to corporate instructions, . . . [Wal-Mart] store[s] made concerted efforts to start inserting meals into the records where it looked as if none had been taken . . ." Karen Smith Deposition, at 118; see id. at 125 ("Don was instructed about all the meal exception reports, there needs to be nothing on them, payroll needs to be lower, cut hours . . . [T]here had to be no one coming up on the meal exception report and no one was coming up on over the six hours." (footnote added)). "Wal-Mart states that its policy was that time clock records should be edited to accurately reflect meal periods actually taken by associates where the associates neglected to punch out for them." Wal-Mart's Answers to Plaintiffs' Seventh Set of Interrogatories, at 6-7. Wal-Mart managers therefore had the authority to insert meal periods into the timekeeping records.
The deposition excerpt the plaintiffs provided from which these sentences are taken does not identify "Don," but Smith's testimony suggests that Don was her superior at the store in which he worked and that the "home office" communicated these directives to him rather than to her. See id. Therefore, Smith's understanding as to meal period insertions "was based on what Don told [her] directly." See id.
Shapiro "counted the number of meal periods of exactly 30 minutes or exactly 60 minutes which were added by management personnel to [associate] timekeeper records." Shapiro Record, ¶ 52. Shapiro only included in his calculations "inserted meals of 30 or 60 minute duration which were added [by management] to shifts otherwise created by [associate] swipes." Id. From January 6, 2001, through December 16, 2005, Shapiro counted 7,622 inserted meal periods of exactly thirty minutes, and 5,950 inserted meal periods of exactly sixty minutes. Using hourly pay rates, Shapiro calculated that the monetary value of the meal period insertions was $94,779. Shapiro anticipates that this total will "decrease after the [TARs] are considered and deductions are made based on valid" TARs. Id. ¶ 54.
Martin notes Shapiro's failure to consider TARs in his analysis, and adds that even if he did consider TARs, "his approach would still remain unreliable. Testimony exists from managers and associates indicating that TARs were not always used or may not have been saved and such changes could be requested verbally." Martin Report, at 6. According to Martin, therefore, the absence of a corresponding TAR does not necessarily render the edit unauthorized. Moreover, "[p]rior to May 2001, Wal-Mart policy did not require the use of TARs and associates could authorize a change by initialing the exception report. However, Dr. Shapiro has done nothing to incorporate this additional source of data into his analysis." Id. Accordingly, Martin maintains that Shapiro's analysis on this point is unreliable.
In the Martin Report, Martin has two sections titled "Dr. Shapiro's Analysis of Inserted Meals is Unreliable," one of which is section IV and the other which is section VI. The two sections are not identical, but neither are they contradictory.
It is particularly concerning to this court that Shapiro's methodology with respect to meal break insertions does not take TARs into consideration at all. Even if he had, however, the TARs themselves, when required, would not make the records entirely accurate for the purpose for which Shapiro seeks to use them because there is evidence that associates did not always follow the requirement, but rather made adjustment requests verbally. Shapiro ignores the evidence that meal insertions were made on the spot without completion of a TAR after discussion between manager and associate in which it might have been acknowledged that the associate simply forgot to clock in or out for a meal that was taken. See, e.g., Paul S. Antonelli Deposition, at 158-59 (stating, as assistant manager, that he would sometimes correct associate's failure to clock in and out for meal break by inserting meal break without having associate filling out TAR); Robert Rolandelli Deposition, at 121 (stating that, as store manager, if associate told him she forgot to clock out for lunch, he would not "force the associate to go get a [TAR] and sign off on it. [He] would just go ahead and enter it into the system right then and there. There's no issue or no authorization"). But see Gerard Roy Deposition, at 49 (stating that, as store manager, he would occasionally add time to associate's record without TAR, but that he "would never subtract time from an associate's record without a paper TAR"). Shapiro similarly ignores testimony that a manager would insert a meal without a TAR for an associate who had a regular shift and always took a meal break; the manager would orally confirm with the associate the propriety of the insertion at a later time. For example, Paula Branchand, a Wal-Mart personnel manager, stated in her affidavit that "[i]f [she] know[s] that someone has a very regular shift pattern and it is so obvious to [her] that they [sic] have simply missed one of their punches, [she] will just fill this in for them . . . and check with the associate when [she] next see[s] them." Paula Branchand Affidavit, ¶ 12.
Shapiro has not established that he has a basis on which to conclude that every instance of a manager-inserted meal break is an instance of improper conduct by Wal-Mart. In fact, Shapiro's application of his methodology to the records reveals that the incidence of meal period insertions over the class period are quite rare and would not suggest a systematic attempt to take earned compensation from associates. For an hourly associate, failure to obtain payment for a 30- or 60-minute period would not be insignificant and would be noticed at least by some of the associates who experienced a meal insertion for time they actually worked; it is illogical for Shapiro to make the blanket assumption that no associate noticed the missing compensation. Under such circumstances, Shapiro was not free to assume that an inserted meal represented shaving of an associate's work time without talking with associates and assuring himself that such insertions were unauthorized and unjustified.
Shapiro applied his methodology to the time period of January 6, 2001, to December 16, 2005, which this court will construe as a full five-year period for purposes of this discussion. Shapiro counted 7,622 meal period insertions of thirty minutes in duration, which breaks down to 1,524.4 meal period insertions per year, or approximately 127 per month, or approximately 29 per week. Further, Shapiro counted 5,959 meal period insertions of sixty minutes in duration, which breaks down to 1,190 meal period insertions per year, or approximately 99 meal period insertions per month, or approximately 23 meal period insertions per week. This court also notes that, as Shapiro has not noted otherwise, these calculations refer to all of the stores open at the relevant time. Therefore, in 2001, when there were thirty-nine stores open, Shapiro Report, ¶ 48, and approximately twenty-three meal period insertions per week, each store did not engage in meal period insertions.
Consequently, Shapiro's methodology on this point, as well, is unreliable.
C. Conclusion
In the June 2004 Decision, the Appeals Court acknowledged that in the January 2004 Decision, this court (Murphy, J.) "effectively left for another day a question that determines whether the action can truly proceed on a class basis." June 2004 Decision, at 5. "[C]ertification is of doubtful durability unless the plaintiffs successfully meet their burden of showing that the statistical evidence is reliable enough to be admissible and that its admission would not violate Wal-Mart's due process rights." Id. at 6. This court finds that the plaintiffs have not met this burden.
Daubert and Lanigan instruct that the court has the gatekeeper function of assuring itself that before an expert is allowed to testify that the methodology employed by the expert is reliable or sound. See 509 U.S. at 597 ("[A] gatekeeping role for the judge, no matter how flexible, inevitably on occasions will prevent the jury from learning of authentic insights and innovations"); 419 Mass. at 25-26 ("The trial judge has a significant function to carry out in deciding on the admissibility of . . . [an] expert's opinion. If the process or theory underlying . . . [an] expert's opinion lacks reliability, that opinion should not reach the trier of fact"). An opinion of an expert with impressive credentials has inherent persuasiveness and the ability to mislead. There is a danger that a jury could be influenced to accept an opinion lacking appropriate foundation simply because the expert is willing to express it and the court is permitting him to do so.
The court accepts that in all likelihood, Wal-Mart's timekeeping records reflect instances of Wal-Mart misconduct. The court's concern with respect to Shapiro's methodology is his refusal to acknowledge the possibility that the timekeeping records contain inconsistencies beyond a de minimis level. Shapiro has pre-determined that nothing reported in the affidavits and deposition transcripts — which indicate that breaks were taken but not reported, that TARs were not completed, that one-minute clock-outs were occasionally utilized, and that associates used operator-accountable registers while not logged in properly — or the coaching reports — which indicate that at least some associates in fact continued not to clock it or out for breaks they took — could ever result in more than de minimis errors in the time records. Shapiro's decision to disregard, without at least limited investigation, the Wal-Mart produced affidavits and deposition testimony and other documentation which indicate that timekeeping procedures were not always followed contributes to the finding that his underlying methodology is unreliable.
As discussed further above in the context of Wal-Mart's motion to decertify the class, a class is not to be certified if such a class contains persons who were not harmed by the alleged wrongful conduct. In this case, the affidavit and deposition testimony reveal that within the class as it is currently certified, there are associates included who were not harmed by Wal-Mart's conduct. Coaching reports disclose associates who violated rest break rules for checking in and out for breaks they took and these associates may very well be within the certified class but not injured by any Wal-Mart conduct. How many more uninjured class members might exist is unknown. Only individual inquiry of the associates or perhaps a representative sampling of associates can provide answers as to how many and whether that number is de minimis. A de minimis answer cannot be assumed as it was by Shapiro. If some inaccuracies in the records are recognized, then the magnitude or range of inaccuracy must be accounted for in some way. Shapiro's methodology is deficient in that it fails to look beyond the records and to assess the degree of inaccuracy in them.
Other courts have found Shapiro's methodology, or its equivalent, wanting. See Basco v. Wal-Mart Stores, Inc., 216 F.Sup.2d 592, 603 (E.D.La. 2002) (finding that plaintiffs' statistical analysis intended "to determine the global number of rest hours and meal breaks taken by [associates] and the number of hours worked off-the-clock for the entire workforce in Louisiana during a period of time . . . ignores the highly individualized issues involved in this case including[ ] . . . the myriad of reasons why any [associate] may have missed a meal or work break or worked off-the-clock"); Hermanson v. Wal-Mart Stores, Inc., 290 Wis.2d 225, 230 (Wis.Ct.App. 2006) (pointing out that "the issue is not `why' but whether the [associates'] self-generating data accurately reflect that a break was missed at all[,]" and holding that Wal-Mart would be entitled to test through discovery and examination at trial plaintiffs' statistical analysis compiled "by a professor in Emory University's Department of Psychology [presumably Shapiro,]" something that would require not only the examination of each and every member of the proposed class, but, also, their co-workers and supervisors" (emphasis in original)); Wal-Mart Stores, Inc. v. Lopez, 93 S.W.3d 548, 560-61 (Tex.App. 2002) (denying class certification because plaintiffs' proposed statistical analysis of Wal-Mart's records "will preclude any individual inquiry in the presence of the jury regarding the formation of each alleged contract or the varied circumstances surrounding each [associate's] missed breaks or off-the-clock work when determining Wal-Mart's liability: . . . [n]or will Wal-Mart be able to question each [associate] as to individual issues such as the number of breaks missed, amount of time worked off the clock, or amount of wages in front of the jury"); Osuna v. Wal-Mart Stores, Inc., Civil No. C200143 19, 2004 WL 3255430, *6 (Pima Super.Ct. Dec. 32, 2004) (Cornelio, J.) (finding that Shapiro's "`formulaic methodologies' . . . [could not] supplant the individualized inquiry necessary to establish Wal-Mart's liability . . . [because] [i]t is easy to conceive of any number of reasons why an [associate] might have chosen to skip a rest break . . . or even provide gratuitous work off-the-clock"); Kuhlmann v. Wal-Mart Stores, Inc., Civil No. 01-CV-008080, at 17 (Milwaukee Circuit Ct. Jan. 9. 2004) (Wells, J.) (denying class certification where "[r]eliance solely on the records of Wal-Mart, as . . . plaintiffs [sic] experts [including Shapiro] propose, will not tell the whole story"); Jackson v. Wal-Mart Stores, Inc., Civil No. 01-40751-NZ-5, at 9 (Saginaw Circuit Ct. March 13, 2003) (Borrello, J.) (rejecting plaintiffs' "allegation that statistical evidence [as analyzed by Shapiro] can be introduced to meet the predominance requirement. . . instead of conducting numerous individual inquiries . . . [and holding that] only way to determine how the [p]laintiffs' alleged damages occurred would be to require all 96,000 [p]laintiff[s] to testify about their individual harms and how those harms occurred. Any other means of determining liability and damages would deprive [Wal-Mart] [its] right to have a jury determine these issues"). But see Braun v. Wal-Mart Stores, Inc., Civil No. 3217, at 10 (Philadelphia Ct. of Common Pleas Dec. 27, 2005) (Bernstein, J.) (granting class certification and relying "primarily on the expert analysis [by Shapiro and another] of computer records to conclude that the systematic loss of contractual break and meal time in Pennsylvania stores has been prima facie demonstrated. It thus becomes a factual determination [for the jury as to why these statistically significant demonstrated discrepancies between the recorded time records and unalterable company policy exists").
Accordingly, for the foregoing reasons, Shapiro's methodology is unreliable and Wal-Mart's motion to exclude his testimony is ALLOWED. Shapiro will not be allowed to give testimony to the effect that missed rest break swipes equal missed breaks; that meal insertions mean time was shaved from the associate's record; that rest breaks shorter than fifteen minutes represent time for which Wal-Mart owes compensation; or that Wal-Mart records show off-the-clock work. He will also not be permitted to testify regarding damages for one-minute clock-outs or for the other claim categories just discussed because, other than Shapiro's own testimony of class-wide violations, there is no basis for his conclusions.
ORDER
For the foregoing reasons, Wal-Mart's motion to decertify the class is ALLOWED and Wal-Mart's motion to exclude the expert testimony of Dr. Martin Shapiro is ALLOWED.