Expert Witness Profiler

Category: Data Science Expert Witness

  • Statistics Expert Allowed to Opine on Over-Detention

    Statistics Expert Allowed to Opine on Over-Detention

    Plaintiffs Alanna Dunn, Reginald Haymon, Adam Day, Eric Zeider, Cameron Leonard, and Jason Wilson filed this class action alleging that the deliberate indifference of Cuyahoga County and the Cuyahoga County Sheriff’s Department caused them to be over-detained in the Cuyahoga County Jail after the legal basis for their detention ceased to exist.

    Plaintiffs retained Lacey Keller, a data scientist, as an expert witness to review and standardize files produced by the County relating to releases occurring between February 23, 2021 and December 31, 2023 while the County retained Dr. Sean Malone, a consultant who primarily works in statistics, finance, and economics, to rebut Keller’s analyses.

    The County sought to preclude Keller’s report while Plaintiffs sought to preclude Malone’s report.

    Data Science Expert Witness

    Lacey R. Keller is a seasoned data scientist with over 15 years of experience applying data to litigation, law enforcement, and investigations.

    She has been deposed nearly 30 times and has testified in over a half-dozen trials. She also joined the Washburn University faculty in 2025.

    Want to know more about the challenges Lacey Keller has faced? Get the full details with our Challenge Study report.

    Statistics Expert Witness

    Sean T. Malone is a consultant who primarily works in statistics, finance, and economics. He teaches finance and statistics at Trinity University.

    Discover more cases with Sean Malone as an expert witness by ordering his comprehensive Expert Witness Profile report.

    Discussion by the Court

    Lacey Keller

    Plaintiffs’ counsel asked Keller to conduct a variety of analyses, including calculating the time it took the County to release individuals after a triggering action, among other analyses on the timing of certain events relating to releases.

    For detainees with release-triggering event and no holds on their release, Keller found most were released in under 6 hours, and 95% were released in under 10 hours.

    In most cases, Keller found a booking hold added an additional 90 minutes to the release time. Once a booking hold was removed, the release usually occurred within the hour.

    Analysis

    The County argued that Keller’s report is unreliable and cannot be used to either (a) show the average time it took to release a detainee; or (b) ascertain whether a person is a member of the proposed class because the detainee’s time to release was greater than 12 hours.

    For the relevant time period, Keller identified 58,091 unique custody sessions. From there, Keller applied a series of exclusions to the data.

    The County argued that Keller’s methodology and opinions are unreliable because they solely rely on limited data not suited for determining release times.

    The County also faulted Keller’s report for not determining whether the data provided by Plaintiffs’ counsel could be used to determine actual time to release.

    First, information counsel provided to an expert does not render the opinions on which the expert relies insufficient. Second, there is a reasonable factual basis for Keller’s opinions. The report explained the sources from which the data came. It explained the exclusions applied to the custody sessions to obtain a sub-dataset that contains only custody sessions where there is an available release triggering event and subsequent release (among other information).

    Keller then explained how she conducted each calculation to arrive at her stated averages and conclusions. While the County vigorously challenges whether the underlying data can be used to calculate release times, those criticisms challenge the ultimate conclusions and not the reliability of Keller’s methodology or analysis.

    Whether the calculations are correct, accurate, or credible, is a separate question. As a result, the Court found that Keller’s methodology was reliable.

    Sean Malone

    Malone’s main conclusions are: Keller’s methodology for calculating time to release is unreliable because it is based on insufficient data and incomplete methods; Keller’s opinion is based on insufficient facts because it does not consider the individualized facts affecting time to release; Keller wrongfully groups together detainees leading to inaccurate time calculations; and Keller inappropriately excluded detainees which creates inaccurate time calculations.

    Analysis

    Plaintiffs argued that Malone is not qualified to opine on how to calculate over-detention of detainees because he is not an expert in release policies.

    Plaintiffs added that Malone’s expert opinions amount to “nothing more than legal conclusions about what facts are relevant and necessary” to an over-detention analysis.

    Lastly, Plaintiffs contended that Malone’s report is not the product of any reliable principles or methods.

    Malone’s experience in statistics plainly relates to his criticisms of Keller’s methodology for collecting data. It is appropriate for an expert in statistical analysis to criticize statistics prepared by another expert witness. And while Plaintiffs argued that Keller’s analysis is merely an “objective” calculation of the “time to release,” Malone’s report properly offered criticism of Keller’s underlying data, including her selection and purported manipulation of data. Malone’s criticism of the “limited” dataset Keller used to develop averages for “time to release” is appropriate rebuttal testimony.

    Because Malone is qualified to criticize the statistical analysis performed by Keller, and because his report properly challenges the data and methods used by Keller, the Court refused to exclude his report.

    Held

    • The Court denied the County’s motion in limine to preclude Lacey Keller’s expert report.
    • The Court denied the Plaintiffs’ motion to strike the expert report of Dr. Sean Malone.

    Key Takeaway

    Malone’s criticism of the dataset Keller used and the methods she used to calculate her “time to release” analysis is proper expert rebuttal.

    Case Details:

    Case Caption: Dunn V. Cuyahoga County
    Docket Number: 1:23cv364
    Court Name: United States District Court, Ohio Northern
    Order Date: March 31, 2026
  • Data Science Expert Was Not Allowed to Opine on Uncompensated Hours

    Data Science Expert Was Not Allowed to Opine on Uncompensated Hours

    Plaintiffs Keith Fischer, Michael O’Sullivan, John Moeser, Louis Pia, Thomas Barden, Constance Mangan, and Charise Jones, (collectively “Plaintiffs”), sought class action certification for their claims against Government Employees Insurance Company (“GEICO”) for failing to pay overtime wages in violation of the New York Labor Law (“NYLL”).

    Plaintiffs and the Class Members are current and former non-exempt employees of GEICO in the Special Investigations Unit (“SIU”). Plaintiffs relied exclusively on a damages model developed by their proposed expert, Dr. Catherine O’Neil, to show that damages are capable of class-wide determination.

    O’Neil described a model to measure lost wages by assigning an amount of time to each type of case-related activity (through a regression analysis of the Plaintiffs’ self-reported hours compared to case-related activities tracked in the GEICO SICM database), and then applying the assigned time-per-activity amounts to the case-related activity data for each Investigator to determine “actual” time worked by each.

    O’Neil contended that she would then subtract a putative class member’s reported hours from the hours derived from the regression model to determine the unreported and uncompensated hours for each month.

    Data Science Expert Witness

    Dr. Catherine H. O’Neil earned a Ph.D. in math from Harvard and previously taught Mathematics at the Massachusetts Institute of Technology and Columbia College. She is a data scientist who founded an algorithmic auditing company.

    In 2016 she wrote the book Weapons of Math Destruction: how big data increases inequality and threatens democracy. and in 2022 the book The Shame Machine: who profits in the new age of humiliation. 

    Fortify your strategy by reviewing a Challenge Study detailing grounds for excluding Catherine O’Neil’s expert testimony.

    Discussion by the Court

    Here, O’Neil contended that she can create a linear regression model to estimate the time investigators spent performing certain activities. But this model is nothing more, as she concedes, than a “thought experiment.” She has neither built the model nor applied it to the data available.

    As GEICO pointed out, O’Neil had access to seven months’ worth of SICM data for 34 Plaintiffs and putative class members and Plaintiffs’ testimony as to their estimated hours.

    By limiting her model to a “thought experiment” and failing to show the applicability of this model to even a subset of available data, the very data O’Neil contended she would use to build her model, O’Neil’s report has failed to show that her opinion is rooted in actual facts or data to properly assess class-wide damages; it amounts to a kind of “trust the expert” methodology. But such “ipse dixit” cannot satisfy Daubert.

    Contrary to Plaintiffs’ presentation, this is not a kind of “plug and play” expert analysis, where simple math—here a basic linear regression model—is the core of the model. The inputs into that model, including the allocation of time for a task, are based on a series of assumptions, which O’Neil has not tested or explored in any meaningful way, even if it was not necessary to build the final complete version of her model.

    For example, O’Neil assumed that certain investigator tasks will take a standard amount of time. However, the Court held that O’Neil has no known expertise or experience in doing the kind of investigations conducted by these kinds of employees, making merely accepting Plaintiffs’ accounts or her own uncredentialed assumptions problematic.

    Held

    The Court found Dr. Catherine O’Neil’s opinion unreliable for determining whether Plaintiffs have satisfied the requirements of Rule 23. As a result, the motion to certify a class was denied.

    Key Takeaway

    Plaintiffs have failed to show that O’Neil’s opinion “is the product of reliable principles and methods.” Rule 702‘s focus is “the scientific validity and thus the evidentiary relevance and reliability—of the principles that underlie a proposed submission. The focus, of course, must be solely on principles and methodology, not on the conclusions that they generate.” 

    Unsupported assertions “made without explanation or elaboration that would allow a fact finder to follow his reasoning and come to the same conclusion” are inadmissible expert opinions.

    Case Details:

    Case Caption: Fischer V. Government Employees Insurance Company
    Docket Number: 2:23cv2848
    Court Name: United States District Court, New York Eastern
    Order Date: February 20, 2026
  • Data Science Expert Witness’ Testimony About MLR Repricing Excluded

    Data Science Expert Witness’ Testimony About MLR Repricing Excluded

    In 1955, Congress created the Indian Health Service (IHS) to govern tribal healthcare. The IHS, now a sub-agency within the U.S. Department of Health and Human Services (HHS), continues to govern tribal healthcare today. Indeed, the IHS is the “principal federal health care provider and health advocate for Indian people, and its goal is to raise their health status to the highest possible level.” First, IHS funds and operates healthcare facilities—such as hospitals and clinics—which provide direct care to American Indians. Second, IHS separately funds Contract Health Service (CHS) Programs, which operate as a referral safety net such that, if an American Indian seeks a healthcare service that is unavailable at their direct-care IHS tribal facility, CHS Programs may refer that American Indian to a non-IHS healthcare facility. 

    Central to this case, in 2003, Congress passed the Medicare Prescription Drug, Improvement, and Modernization Act which authorized HHS to demand no more than Medicarelike rates (MLRs) from hospitals that provide services to tribes under a CHS Program, including those CHS Programs which tribes themselves orchestrate. 

     The Tribe retained ClaimInformatics—a healthcare “payment integrity firm” based in Connecticut to identify which of the Tribe’s claims were eligible for MLRs, and to “reprice” those eligible claims to determine if Blue Cross Blue Shield of Michigan (BCBSM) applied MLRs and, if not, how much the Tribe overpaid.

    On September 16, 2023, the Tribe produced the “Preliminary Expert Report of ClaimInformatics/Dawn Cornelis.” The Report noted that ClaimInformatics “repriced a total of 6,6641 claims” out of the 93,104 claims produced by BCBSM at that time. On November 6, 2023, BCBSM filed a Daubert motion seeking to exclude Cornelis’ proffered expert testimony about MLR and ClaimInformatics’ MLR repricing in this case.

    Data Science Expert Witness

    Dawn Cornelis created her own healthcare claim audit and recovery business—Claim Recovery Services. Claim Recovery Services closed nearly twenty years later in 2010, and Cornelis worked the next seven years in various payment integrity roles. In 2017, Cornelis cofounded ClaimInformatics and helped develop a payment integrity software known as “Claim Intelligence” which was used by ClaimInformatics to process and audit the Tribes claims with BCBSM in this case.

    Get the full story on challenges to Dawn Cornelis’ expert opinions and testimony with an in-depth Challenge Study. 

    Discussion by the Court

    BCBSM argued that (1) Cornelis lacked the training, education, and experience to testify as a MLR and re-pricing expert at trial under Federal Rule of Evidence 702, and (2) even if qualified, Cornelis’ expert testimony would be unreliable because the testimony merely “parrots” or “bootstraps” the calculations and conclusions of Linda Myrick.

    Qualifications

    BCBSM argued that Cornelis has no formal education in healthcare, insurance, tribal welfare, or any other study which would enable her to testify as an expert in MLR or MLR repricing.

    This Court, however, does not discount Cornelis’ substantial knowledge and experience related to healthcare insurance claim processing and auditing, generally. Yet, it held that despite Cornelis’ knowledge, training, and experiencing in processing and auditing healthcare insurance claims, generally, she does not have comparable knowledge, training, and experience in auditing tribal healthcare claims involving MLR, specifically. Also, Cornelis had no experience with MLR repricing in this specific case.

    As further proof of her lack of experience, Cornelis herself expressly denied her MLR repricing expertise on three separate occasions, noting that she only knows enough “to be dangerous.”

    To sum it up, Cornelis lacked the education, training, and experience to offer expert opinions on MLR repricing, both generally and as specifically applied to the Tribe’s claims with BCBSM.

    Reliability

    Even if Cornelis was qualified based on her training, education, and experience, her proffered opinion about BCBSM’s rates and the Tribe’s MLR repricing would be unreliable because this testimony does not concern her opinion. It would instead only concern Linda Myrick’s opinion.

    However thin the line may be between permissible reliance and impermissible “parroting,” Cornelis crossed it here. According to the Court, Cornelis conceded that all MLR re-pricing was conducted solely by Linda Myrick. Although Cornelis testified that she independently assured the accuracy of some CMS pricing tables which Myrick may have relied on when comparing the rates BCBSM charged the Tribe to the applicable MLRs, nothing suggested that Cornelis independently evaluated Myrick’s re-pricing calculations.

    The Court held that this testimony is just as unreliable under Daubert as it is unfairly prejudicial under Rule 403.  If Cornelis was permitted to testify about Myrick’s re-pricing calculations and conclusions, BCBSM would have no meaningful opportunity for cross-examination. The Tribe has indicated it will not call Myrick as an expert witness.

    In response to BCBSM’s meritorious motion to exclude Cornelis’s testimony, the Tribe attempted to bolster Cornelis’s qualifications and reliability by producing her “Declaration,” dated November 15, 2023—notably executed after the Parties’ Daubert deadlines and nearly one month after Cornelis’s deposition. To the extent Cornelis’s declaration contradicts her prior deposition testimony, the Court held that her declaration will be stricken and will not be considered when analyzing the propriety of her expert testimony at trial.

    Held

    The Court granted the Defendant’s motion to exclude Dawn Cornelis’s proffered expert testimony on Medicare-like Rate repricing.

    Key Takeaway:

    The law governing expert testimony distinguishes between permissible reliance and impermissible “parroting.” True, an expert may rely on the opinions and conclusions of other experts when forming their own independent conclusions throughout their own independent investigation. But an expert may not simply “parrot,” “echo,” “regurgitate,” or “bootstrap” the opinion or conclusion of another expert without any independent evaluation or analysis.

    Case Details:

    Case Caption: Saginaw Chippewa Indian Tribe Of Michigan Et Al V. Blue Cross Blue Shield Of Michigan
    Docket Number: 1:16cv10317
    Court: United States District Court for the Eastern District of Michigan, Northern Division
    Order Date: August 14, 2024
  • Class Size Calculations of Economics Expert Witness Held to be Reliable

    Class Size Calculations of Economics Expert Witness Held to be Reliable

    Spokeo owns and operates the website spokeo.com. It collects consumer and public data from various public sources and private vendors, associates that data with particular names, and publishes it online.

    Using proprietary algorithms and systems, Spokeo “attempts to collect and aggregate and merge all that data into persons, person objects, which are then designated with a unique [personal identifier or ‘PID’] for that person object.” That aggregated data, associated with a particular PID, can then be searched for by users of the website. It is also used to populate “teaser profiles,” which can be viewed by the public without a subscription to the website. The teaser profiles advertise additional personal information about the subject of the profile, including about their family, court records, sex offender registration status, marital status, and more. 

    The Plaintiffs, Aviva Kellman, Jason Fry, Nicholas Newell, Susan Gledhill Stephens, and William Williams V, found a teaser profile associated with their personal information.

    The Plaintiffs said that they did not consent to Spokeo’s use of their information on its website. They asserted that Spokeo’s publication of their personal information violated their statutory rights of publicity and common law rights regarding misappropriation of name and likeness. They sought class certification against Spokeo for four classes of people in California and Ohio.

    Spokeo filed motions to exclude the declaration and testimony of Plaintiffs’ experts, Michael Naaman and Steven Weisbrot. In response, the Plaintiffs also filed a motion to exclude Spokeo’s expert, David Alfaro.

    Class Size Calculations of Economics Expert Witness Held to be Reliable

    Economics Expert Witness

    Michael Naaman, Ph.D., is a senior consultant specializing in antitrust, econometrics, and machine learning. He has provided economic and econometric analysis in issues relating to patent infringement and intellectual property, false advertising, and antitrust disputes.

    Naaman has a decade of experience in the economic consulting industry. He received his Ph.D. in economics from Rice University, and he received a M.S. in statistics and B.S. in economics, math, and physics from Tulane University.

    Get in-depth insights into Michael Naaman’s expert witness experience by requesting his Expert Witness Profile today.

    Law Expert Witness

    Steven Weisbrot, Esq. has been responsible for the design and implementation of hundreds of court-approved notice and administration programs. He is President and Chief Executive Officer of Angeion Group, a leading provider of comprehensive settlement management services. Weisbrot is a licensed attorney in Pennsylvania and New Jersey.

    Gain a comprehensive understanding of Steven Weisbrot’s qualifications and casework history with his Expert Witness Profile report.

    Data Science Expert Witness

    David Alfaro is a Senior Managing Director and co-leads the Data & Analytics West Coast practice and is based in San Francisco. Over his 30-year career, Alfaro has led more than 200 engagements as an expert witness and expert consultant, nearly all of which have been in the investigations and disputes space.

    Moreover, he is an expert in the areas of collecting enterprise-wide information with extensive experience in complex, data-intensive analyses in response to government or internal investigations and litigation. In this capacity, Alfaro has provided formal and informal testimony to federal and state courts, the Federal Trade Commission (“FTC”), the Consumer Financial Protection Bureau (“CFPB”), the Federal Bureau of Investigation (“FBI”), the Securities and Exchange Commission (“SEC”), the Department of Justice (“DOJ”), the U.S. Attorney’s Office, the Financial Industry Regulatory Authority (“FINRA”) and other federal agencies.

    Get in-depth insights into David Alfaro’s expert witness experience by requesting his Expert Witness Profile today.

    Discussion by the Court

    Spokeo’s Motion to Exclude Declarations and Testimony of Michael Naaman and Spokeo’s Motion to Strike Naaman’s New Reply Declaration

    Spokeo filed a motion to exclude Naaman’s declaration and testimony, arguing that his class size calculations are unreliable and that his method to calculate damages is fundamentally flawed.

    Spokeo argued that Naaman failed to exclude Spokeo users, dead people, people who assigned their rights, duplicated profiles, profiles with inaccurate home addresses, and profiles that refer to people who are not real from his class size calculations. The Court considered the argument superfluous as Spokeo did not challenge numerosity. Even if Naaman should have excluded each of these, the classes clearly would still meet the numerosity requirements given the data upon which he relied and the evidence of Spokeo’s business model, which is apparently designed to have a teaser profile for every American adult.

    Also, Spokeo sought to exclude Naaman’s declaration and testimony about damages calculations, asserting that Naaman did not provide a method for calculating damages and instead simply multiplied the number of class members by the minimum statutory penalty sought by the Plaintiffs for the Viewed Prior to Purchase classes. 

    The Court held that, pursuant to Rule 702, it was a logical way to calculate damages in a case where the Plaintiffs seek the statutory minimum for damages, and this would help the trier of fact.

    Spokeo’s Motion to Exclude Declarations and Testimony of Steven Weisbrot

     Class Notification:

    Spokeo also moved to exclude the declaration and testimony from Weisbrot about class notification, asserting that it was irrelevant and that its methodology was unreliable.

    Weisbrot’s proposed notification method involved sending emails to potential class members using the email addresses posted on Spokeo’s teaser profiles, by publication in the media, and by website. Though his proposed method provided notice to the nationwide classes, he explained how he would and could use the same method on a narrower target audience if smaller classes were certified, such as statewide classes.

    Spokeo contended that Weisbrot’s notice plan will provide notice to all potential class members in the Purchase classes, and that this is overbroad because it is not directed only to members of the Viewed Prior to Purchase classes, for which notice is mandatory under Rule 23(b)(3). But the Federal Rules permitted notice to 23(b)(2) classes like the Purchase classes, and Spokeo offered no reason why notice should not be provided to them. 

    The Court rejected Spokeo’s argument that Weisbrot not offering a way to target solely Viewed Prior to Purchase members was merely another way to argue that the Plaintiffs were unable to identify their own class members from common evidence. Moreover, the Court found Spokeo’s argument about Plaintiffs self-identifying bizarre because claimants do not have to know pre-filing which class they are in.

    Notice Methodology

    Finally, Spokeo asserted that Weisbrot should not rely on the email addresses that Spokeo has in its possession and includes for teaser profiles because they might not be real or correspond to a real person. Despite Spokeo’s asserting that its email address data was inherently unreliable, and given the evidence about its use of data gathering and associating via personal identifier labels to connect names with addresses and other information, the Court found it highly likely that many of the email addresses were correct for many of the teaser profiles.

    Plaintiffs’ Motion to Exclude Declaration and Testimony of David Alfaro

    Plaintiff argued that Alfaro misrepresented evidence about Spokeo data vendors, was unqualified to opine on class size or Naaman’s methodology, was not an expert on class notice and could not opine on Weisbrot’s methodology, and provided improper legal conclusions. 

    The Court held that it did not rely on Alfaro’s declarations to assess Naaman’s declaration or the validity of Weisbrot’s methodology. In other words, the Court denied Plaintiff’s motion to exclude Alfaro’s testimony as moot because that was the driving basis for which the Plaintiffs challenged Alfaro’s declaration.

    Held

    The Court granted Plaintiffs’ motion for class certification for the California and Ohio classes, with certain amendments to the class definitions. The motion for the nationwide classes was withdrawn.

    To sum it up, the Court denied Spokeo’s motion to exclude the testimony of Michael Naaman and Steven Weisbrot. The Plaintiffs’ motion to exclude David Alfaro’s declaration and testimony was denied as moot.

    Key Takeaway:

    • Naaman simply multiplied the number of class members by the minimum statutory penalty sought by the Plaintiffs for the Viewed Prior to Purchase classes. The Court held it was a logical way to calculate damages in a case where the Plaintiffs seek the statutory minimum for damages.
    • Despite Spokeo’s assertion that its email address data was inherently unreliable, and given the evidence about its use of data gathering and associating via personal identifier labels to connect names with addresses and other information, the Court held that it was highly likely that many of the email addresses are correct for many of the teaser profiles. Hence, Weisbrot could rely on the email addresses that Spokeo had in its possession and remove junk, fake, or inaccurate emails, to the extent possible.