Summary
excluding expert where "[a]ll of his publications [were] in different fields using different methodology from that required in th[e] case" and "[w]hile it is possible that [expert] ha[d] had sufficient 'on-the-job' training in his court related work, he ha[d] not produced sufficient information about that work to demonstrate his knowledge of statistical science in th[e] case"
Summary of this case from Cnty. of Maricopa v. Office Depot Inc.Opinion
Civil Action No. 96-7564-A
August 28, 2001 Order Denying Reconsideration October 2, 2001
Steven C. Thompson, Esq., David Abboud Thomas, Esq., More, Walters, Shoenfelt Thompson, Baton Rouge, Lousiana, for plaintiffs.
Shawn L. Holahan, Esq., Hoffman, Sutterfield, Ensenat Bankston, I. Harold Koretzky, Esq., Elvige, Cassard, Richards, Carver, Darden, Koretzky, Tessier, Finn, Blossman Areaux, New Orleans, Louisiana, for defendants.
RULING ON MOTION
This matter is before the court on the Motion of defendant, Albemarle Corporation, to Exclude the Testimony of Plaintiffs' Statistical Expert, Dr. Thomas Daymont (doc. 129). The magistrate judge, to whom the motion was referred, denied the motion (doc. 164) and the defendant has filed "objections" thereto which the court accepts as an appeal of the magistrate judge's ruling under 28 U.S.C. § 636 (b)(1)(A).
Needless to say, counsel for plaintiffs strenuously opposes the defendant's motion.
This case, along with a number of related actions, arises from a reduction in force conducted by the defendant in October 1993. All plaintiffs claim that their jobs were terminated because of their age in violation of the Age Discrimination in Employment Act.
The evidence of age discrimination upon which plaintiffs would rely is largely statistical. The issue, therefore, is significant.
Rule 702 of the Federal Rules of Evidence now formally incorporates the trial judge's gatekeeping function as articulated in Daubert v. Merrill-Dow Pharmaceuticals, Inc., 509 U.S. 592, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993):
Rule 702. If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion or otherwise, if (1) the testimony is based upon sufficient facts or data, (2) the testimony is the product of reliable principles and methods, and (3) the witness has applied the principles and methods reliably to the facts of the case.
The court has appointed its own expert to assist it in resolving the highly technical issues presented by this motion. The court selected William R. Schucany, Ph.D., whose doctorate is in statistics. Dr. Schucany is a Professor of Statistics at Southern Methodist University who has taught at the college and graduate level for more than thirty (30) years. The court's expert has submitted a written report to the court which has been shared with counsel for both sides.
The court requested Dr. Schucany to answer the following questions regarding plaintiffs' expert:
(1) Is Dr. Thomas Daymont qualified by knowledge, skill, experience, training and education to express opinions of the type set forth in his reports dated February 19, 1999 and November 12, 1999?
(2) Are the opinions of Dr. Daymont set forth in those reports:
(a) based upon sufficient facts or data;
(b) the product of reliable principles and methods; and
(c) has Dr. Daymont applied the principles and methods reliably to the facts of the case?
Dr. Daymont has a B.A. in Mathematics and Economics, an M.A. in Physical Education, a Ph.D. in Sociology, "and a minor in Statistics." Since July 1987, Dr. Daymont has been an associate professor in the Department of Resource Administration at Temple University in Philadelphia.
The court's expert reviewed Dr. Daymont's academic and professional qualifications, as well as all of his reports submitted to counsel for plaintiffs and answered the court's question number one as follows:
In my professional opinion Dr. Daymont does not deserve to be qualified in statistical science simply by virtue of his formal education and training. His Ph.D. in sociology is not adequate preparation for such statistical analyses. His C.V. does not provide convincing support in this regard. He is not a member of the American Statistical Association. Furthermore his published refereed journal articles on the specific topic of employment history comparisons are insufficient to raise his credentials to that level. The quantitative studies that he has published (four in Social Science Research, one each in Demography, Journal of Human Resources, and Social Science Quarterly) primarily pertain to linear-model evaluations of labor market effects on income using national longitudinal studies. As such these are fundamentally different from the methodology required here. He lists no peer-reviewed articles since 1992. Temple University has not promoted him to the rank of Full Professor, which would at least acknowledge his senior stature in his discipline.
On the other hand he very well may have gained enough expertise from his legal applications. Thus it is possible that he is adequately self-educated by his work on the 11 federal suits in which he testified and perhaps on some other cases about which he may have been consulted. However without easy access to that testimony I cannot judge the nature or quality of his statistical analyses. Consequently it is not clear to me that Dr. Daymont's credentials reach the threshold that would qualify him to express expert opinions on a statistical analysis of whether age was a factor in a reduction in force.
Thus, Dr. Daymont does not have sufficient academic credentials to qualify him to express expert opinions on a statistical analysis of whether age was a factor in a reduction in force. All of his publications are in different fields using different methodology from that required in this case. While it is possible that Dr. Daymont has had sufficient "on-the-job" training in his court related work, he has not produced sufficient information about that work to demonstrate his knowledge of statistical science in this case. Since the burden is upon the party claiming that a witness is qualified in a particular field, the court concludes that plaintiff has not met the first step of FRE 702 — that the "witness is qualified by knowledge, skill, experience, training, or education."
Accordingly, the court will not accept Dr. Daymont as a expert in statistical science principles as to whether age was a factor in a reduction in force in this case.
Assuming, however, that Dr. Daymont can and does produce sufficient information confirming his prior experience to show that he meets the minimum qualifications, there are problems with the opinions that he has submitted to counsel for plaintiffs.
The experts for the two sides in this case took different approaches to the issues. Dr. Daymont basically employed a "company-wide" analysis, assuming that all employees whose positions were eliminated were "similarly situated" and that the reduction in force was directed by the same individual. The defendants' expert disagrees with Dr. Daymont's approach and suggests that each department or unit of the company in which positions were eliminated, determined which positions were to be eliminated, independently of all other units.
Dr. Daymont submitted a report, dated August 5, 1999, in which he supported his original opinions and critiqued the opinions of defendants' expert, Dr. Siskin.
As to Dr. Daymont, the court's expert answered questions 2(a) as to facts and data, and 2(b), as to principles relied upon, in the affirmative. As to question 2(c), regarding application of principles and methods, however, the court's expert pointed out that:
With regard to 2(c) each party questions the reliability of some of the other's applications. My specific concerns follow:
i) To a great extent the dispute is a problem of interpretation of the phrase "similarly situated". At one extreme Dr. Daymont does a company-wide analysis, which can be a potentially misleading application of this methodology. In other words it is of questionable reliability to put all of the employees of the company into the same unrefined pool.
ii) There is merit to each party's general approach to the counts, depending upon which of the company's decision processes should be modeled. It may not be necessary for the court to decide whether Dr. Daymont's approach is confounding each individual's choice to decide on redeployment or whether the end result of both stages of selection and then ultimate termination should be expected to be age neutral. That is, if he provides genuinely parallel analyses of a) positions selected and b) RIF's, then the court may see whether his conclusions are the same either way.
iii) There may be some legitimate points to be made about the consistency of the pattern of disparities across twelve departments. However, the analysis of "median age [within each department]" does not directly address the legally protected class of employees 40 years of age and older. Moreover, even under the ideal model of no age effect, these 12 assumedly independent "above" and "below" outcomes may not have probability exactly equal to 1/2 in each department. Only if the total number of selections were such that no employee has precisely the median age, would this binomial distribution be technically correct. For these reasons, it is my opinion that Paragraphs 9. through 13. in Daymont's Affidavit dated August 5, 1999 are not reliable applications of appropriate methodology to the facts of this case.
These statements by Dr. Schucany, the court's expert, cast significant doubt as to the reliability of Dr. Daymont's conclusions in his affidavit of August 5, 1999, and question the reliability of the approach he used in formulating his own opinions.
In the 1999 affidavit, Dr. Daymont is critiquing the approach taken by the defendants' statistical expert, who considered each of twelve (12) departments of the company that eliminated employee positions.
Paragraphs 9 — 13 of that affidavit which the court's expert says "are not reliable applications of appropriate methodology to the facts of this case", read as follows:
(9) Specifically, I have been advised by counsel for the Plaintiffs that "In eleven out of twelve departments within the Siskin Subset, there were more `position eliminations' above the median age [within each department] than below the median age." Also, I have been advised that "In all twelve departments, there were more actual RIF's above the median age [within each department] than below the median age. Counsel for the Plaintiffs has asked me to conduct a statistical analysis to determine whether these patterns are consistent with an age-neutral "position elimination" process and an age-neutral process. The binomial distribution can be used to make this determination.
(10) In an age-neutral process, we would expect that, among the twelve departments identified by Dr. Siskin, about six would have more "position eliminations" above the median age and about six departments would have more "positions eliminations" below the median age. We would not expect that exactly six departments would have more "position eliminations" above the median age; we would expect some deviation from six due to chance or random factors. The question is: Is the actual finding of 11 or 12 departments a large enough deviation from the expectation of 6 of 12 departments to reject the assumption that the "position elimination" process was age neutral? The answer is: Yes.
11. Using binomial distribution, it turns out that if the process were age-neutral, there would be less than a one-in-a-hundred chance (probability = .003) that we would find 11 of 12 departments with more "position eliminations" above the median age than below. According to statistical theory and standard statistical practice, this is statistically significant. Thus, this is statistical evidence that the "position eliminations" process was not age-neutral in these departments.
12. An analogy can be made to the tossing of a coin. If we were to toss a "fair" coin twelve times, we would expect about six heads. The chance of obtaining 11 heads from 12 flips of a "fair" coin is less than one-in-a-hundred (probability .003). This would be statistical evidence that the coin is not fair.
13. Regarding actual RIF's, there is less than a one-in-a-thousand chance (probability = .00024) that we would find 12 of 12 departments with more RIF's above the median age if the decision making process were age-neutral. This is statistical evidence that the RIF process was not age neutral in these departments.
Dr. Daymont refers to "the binomial distribution" several times. The dictionary defines it as, "a frequency distribution of the probability that an attribute that occurs with a given probability among the members of a population will occur a certain number of times in a succession of samples of the population." Webster's Third New International Dictionary (1976).
In the above quoted paragraphs, Dr. Daymont, purportedly using the binomial distribution, concludes that because eleven of the twelve departments rather than about half that number showed more "position eliminations" above the median age than below, this is statistical evidence that the process was not age-neutral in these departments. The court's expert points out that analysis of median age within each department does not "directly address the legally protected class of employees 40 years of age and older", and that these paragraphs "are not reliable applications of appropriate methodology to the facts of this case.
One of the gatekeeper functions of the district court under FRE 702 is to shield the jury from expert opinion which is not a reliable application of principles and methods to the facts of the case. Since Dr. Daymont's opinions in his August 5, 1999 report do not meet that reliability standard, they will not be allowed to be presented to the jury.
Moreover, the court's expert, Dr. Schucany, notes that Dr. Daymont's use of a "company-wide" approach without further refinement, "can be a potentially misleading application of this methodology" . . . because "it is of questionable reliability to put all of the employees of the company into the same unrefined pool."
In paragraph 12 of his report of August 5, 1999, Dr. Daymont supports his conclusion that there is statistical evidence that the positions elimination process was not age-neutral in the company units by an analogy to the tossing of a coin. Dr. Daymont declares that in tossing a "fair" coin twelve times, "we would expect about six heads" and that a result of eleven out of twelve would be statistical evidence that the coin is not "fair", and by analogy statistical evidence that the position elimination process in the twelve departments was not age-neutral.
The court is independently aware that Dr. Daymont's expectations in his coin tossing analogy are not supported by scientific principle or published scientific data.
Dr. John Allen Paulos, whom all parties would certainly concede is truly outstanding in this field, discusses coin tossing in his book, Innumeracy, Mathematical Illiteracy And Its Consequences, Vantage Books, 1990.
Dr. Paulos explains that, while coin tossing over time does approach a half and half distribution, it may take a thousand or more flips to approach that distribution. At any given time during the coin tossing, there may be long runs of heads appearing many times in a row. Dr. Paulos sums the matter up as follows:
In terms of ratios, coins behave nicely: the ratio of heads to tails gets closer to 1 as the number of flips grows. In terms of absolute numbers, coins behave badly: the difference between the number of heads and the number of tails tends to get bigger as we continue to flip the coin, and the changes in lead from head to tail or vice versa tend to become increasingly rare. Id. at p. 58.
This leads Dr. Paulos to conclude that "Most people don't realize that random events generally can seem quite ordered." Id. at p. 59. It is troubling to the court that Dr. Daymont seems to be one of these people who does not realize that.
Other members of the statistical scientific community confirm the observation of Dr. Paulos:
How results that are not indicative of anything can be produced by pure chance — is something you can test for yourself at small cost. Just start tossing a penny. How often will it come up heads? Half the time, of course. Everyone knows that.
Well, let's check that and see. . . . I have just tried ten tosses and got heads eight times, which proves that pennies come up heads eighty percent of the time. . . . Now try it yourself. You may get a fifty-fifty result, but probably you won't; your result, like mine, stands a good chance of being quite a ways away from fifty-fifty. But if your patience holds out for a thousand tosses you are almost (though not quite) certain to come out with a result very close to half heads — a result, that is, which represents the real probability. Only when there is a substantial number of trials involved is the law of averages a useful description or prediction.
Darrell Huff, How To Lie With Statistics, W.W. Norton Co., Norton Paper Back Reissue 1993, p. 39. (A very helpful little book for laymen.)
The court in Daubert, supra, noted:
Faced with a proffer of expert scientific testimony, then, the trial judge must determine at the outset, pursuant to Rule 104(a), whether the expert is proposing to testify to (1) scientific knowledge that (2) will assist the trier of fact to understand or determine a fact in issue. This entails a preliminary assessment of whether the reasoning or methodology underlying the testimony is scientifically valid and of whether that reasoning or methodology properly can be applied to the facts in issue.113 S.Ct. 2786, at 2796.
This "gatekeeper" is presented with a proffered expert whose academic credentials do not qualify him to express expert opinions on a statistical analysis of whether age was a factor in a reduction in force, although it is possible that he may have gained enough expertise in practical work to be so qualified. That, however, has not been demonstrated by Dr. Daymont.
That same expert, although relying on "solid principles" in testing, does a company-wide analysis which is of "questionable reliability and potentially misleading" in support of his own opinions.
That expert's analysis of the opposing expert's opinions "are not reliable applications of appropriate methodology to the facts of this case." The expert's coin tossing analogy contradicts published data in the scientific field.
For all of these reasons, the court concludes that Dr. Daymont's opinions are not reliable applications of appropriate methodology to the facts of this case and he will not be accepted as an expert in statistical science as applied to a reduction in force in this case.
Accordingly, the ruling of the magistrate judge (doc. 164) is hereby REVERSED. The motion of defendant to exclude the testimony of Dr. Thomas Daymont (doc. 129) is hereby GRANTED and Dr. Daymont will not be accepted as an expert in statistics by the court.
RULING ON MOTIONS
This matter is before the court on three motions filed on behalf of plaintiffs which are related to a ruling of the court dated August 28, 2001 (doc. 229). They are: Motion to Reconsider, Alter or Amend Ruling Regarding Plaintiffs' Statistical Evidence (doc. 234); a Motion for Leave to Depose Court-Appointed Expert and For Related Relief (doc. 230); and a Motion for Leave to Retain New Statistical Expert (doc. 232).
On August 28, 2001, the court held (doc. 229) that plaintiffs' retained expert statistical witness was not qualified by knowledge, skill, experience, training or education to express expert opinions on a statistical analysis of whether age was a factor in the reduction in force which is at the core of this law suit, as required by FRE 702. The court further held that, assuming the witness could demonstrate minimum qualifications to express expert opinions, the opinions already expressed by the witness in his several reports are not reliable applications of principles and methods to the facts of this case, as required by FRE 702. The court declined to accept the witness as an expert in statistical science at the trial of this matter.
Background
The matter was presented to the court on a motion by the defendant to exclude the testimony of plaintiffs' statistical expert (doc. 129). On February 2, 2001 (doc. 181), the court informed counsel that it had decided to appoint an expert to assist the court in determining whether the opinion of the expert witness is "[a] based upon sufficient facts or data, [b] is a product of reliable principles and methods, and [c] whether the witness has applied the principles and methods reliably to the facts of the case." The court also informed the parties that the court does not anticipate that the court appointed expert will testify before the jury but "will submit a written report to the court and the parties."
The court sought suggestions from counsel as to the person to be appointed as the court's expert.
The parties submitted suggestions to the court which the court considered and rejected. The court selected its own expert and by a Notice to Counsel dated March 22, 2001 (doc. 188), the court advised the parties of its selection and sought suggestions as to the specific questions to be submitted to the court's expert.
The parties submitted volumes of information and suggestions, all of which the court rejected and by notice to counsel dated April 2, 2001 (doc. 190), the court informed the parties that it would submit all of the expert reports submitted by plaintiffs' expert witness to the court's expert and that the following specific questions would be submitted to the court's expert:
"`Is _________ qualified by knowledge, skill, experience, training and education to express opinions of the type set forth in his reports dated ___________ Are the opinions of ______ set forth in his reports dated ______ (a) based upon sufficient facts or data; (b) the product of reliable principles and methods; and (c) has _____ applied the principles and methods reliably to the facts of the case?'"
The court has reviewed the questions suggested by both sides and has concluded not to incorporate them."
The court sought to make it plain that the court's expert is not an expert witness but an expert in the science of statistics who would assist the court with the technical aspects of performing its role as "gatekeeper" under FRE 702 — a question of law.
Thereafter the court submitted the witness's reports and the court's questions to the court's expert and ultimately received a written report which was shared with the parties and, as noted above, on August 28, 2001, the court issued its ruling. (doc. 229).
Motion to Reconsider, Alter or Amend Ruling Regarding Plaintiffs' Statistical Evidence
Plaintiffs urge reconsideration of the court's ruling because plaintiffs take issue with some of the conclusions of the court's expert. After attacking the reasoning and conclusions of the court's expert, counsel for plaintiffs abandons some of their expert's opinions, thereby conceding the unreliability of those opinions.
Counsel for plaintiffs also objects that the report of the court's expert is "hearsay." The report is obviously not hearsay under FRE 801 (c) because it is not testimony before the jury.
Counsel for plaintiffs had ample time to comment upon the report of the court's expert prior to the court's ruling and the court sees no new information or new reason as to why the ruling of August 28, 2001, should be reconsidered.
Accordingly, that motion will not be granted.
Motion for Leave to Depose Court-Appointed Expert and for Related Relief
In support of the motion to take the deposition of the court appointed expert, counsel cites the provisions of FRE 706 which do indicate that a witness appointed by the court to testify before the jury may be deposed by any party and shall be subject to cross examination.
As noted above, in the notices sent to counsel prior to the appointment of the court's expert, it was made plain that this expert would not testify before the jury and that the questions submitted to the expert would all relate to the "gatekeeper" function to be performed by the court. As the specific questions propounded to the expert by the court demonstrate, this expert was appointed solely to assist the court in connection with the question of whether the plaintiffs' expert witness was qualified and whether his opinions were based upon scientific methodology properly applied. As counsel for the defendant points out in brief, such a role has been characterized as that of a "technical adviser" to the court and depositions are unnecessary. In re: Joint Eastern and Southern Districts Asbestos Litigation, 151 FRD 540, 544 (SDNY 1993), vacated the remanded on other grounds, 982 F.2d 721 (2nd Cir. 1992); Renaud v. Martin Marietta Corp., 972 F.2d 304, 308 n. 8 (10th Cir. 1992).
In short, the expert appointed by the court was a court expert but not a court expert witness.
Under these circumstances, the court will not grant the motion to depose the court's expert.
Motion for Leave to Retain New Statistical Expert
Finally counsel for plaintiffs requests the court to authorize him to search for and retain a "new" statistical expert to succeed the witness which the court has declined to accept. Counsel for the defendants points out that discovery cut-off and the deadline for designating experts have long since passed and that plaintiffs should not be allowed to "shop" for another expert witness.
The court agrees. These cases have been pending in this court since December 19, 1996, and all parties have had ample time to select, interview and depose all necessary witnesses, including expert witnesses. It is time that these matters proceed to ultimate resolution.
Accordingly, the motion on behalf of plaintiffs to Reconsider, Alter or Amend Ruling Regarding Plaintiffs' Statistical Evidence (doc. 234); Motion for Leave to Depose Court-Appointed Expert and For Related Relief (doc. 230); and Motion for Leave to Retain New Statistical Expert (doc. 232) are hereby DENIED.