Sampling
AACG has experts with more than 40 years of experience performing and critiquing statistical sampling design, sampling, statistical analysis and extrapolation from statistical samples to the broader population from which they were drawn.
In some cases, scientific sampling is needed to obtain information about a large population in an efficient and cost-effective manner in order to estimate the magnitude of the alleged issue and/or calculate damages. However, statistical samples are often performed incorrectly. AACG economists are experts at identifying fatal flaws in sampling methods that prevent them from being representative of the broader population or used in court. We also have extensive experience critiquing statistical issues related to class certification.
AACG experts understand that while there are a core set of underlying principles driving scientific sampling and related statistics, the requirements for an individual statistical sample and analysis will depend on the questions and level of precision needed to answer the issues in each individual case. Our deep understanding of these statistical sampling principles allows us to develop innovative sampling approaches that fit the nature and magnitude of the disputed issues, while maintaining the scientific properties of statistical sampling.
Selected Experience
Intellectual Property
Copyright in Computer Code
AACG provided expert support for Oracle in its copyright infringement case against SAP. AACG experts provided statistical analysis and rebuttal testimony related to marginal costs of production and damage. The analysis involved developing statistical samples of copyrighted material to review and then providing an estimate of the total number of copyright infringements based on a statistical extrapolation from the sample to the population as a whole. At the core of the marginal cost analysis was AACG’s review of a regression analysis performed by another expert, and a demonstration that an over-reaching restriction in that regression analysis caused bias in the results so severe as to completely invalidate them.
In a second copyright case, AACG investigated the prevalence of copyright infringement involving computer code that was downloaded by the infringing company from the copyright owner’s website. AACG developed the statistical sampling process to randomly select specific instances of recorded downloads for review of copyright infringement. Based on the findings from the scientific sample, AACG provided estimates of the number of incidents of copyright infringement.
Copyright Infringement through Video Posting
In a dispute involving a claim of copyright infringement of video material posted to a website, AACG performed statistical analysis to determine the extent of the infringement, including the number of views. As part of the analysis AACG investigated whether there was any scientifically valid reason that the results of a sampled population by plaintiffs’ statistician could be extended beyond the population from which the sample was taken.
Labor and Employment
Meal and Rest Breaks
In an off-duty meal period litigation in California State court, AACG economists used scientific sampling to select and examine time records and claims about missed meal breaks or over time hours and extrapolated the results to the broader class.
Telecommunications
For more than a decade, AACG economists ran the performance metrics reporting for several major telecommunications companies. The monthly reporting requirements called for the development and ongoing use of sophisticated statistical models which were performed on a monthly basis under strict FCC reporting deadlines and requirements. This work called for estimation of confidence intervals for complex sampling models, including both large and small sample techniques. It also required development of sample size to achieve precision goals and confidence interval requirements.
Related to this work, we provided testimony to the FCC and State Commissions. Our experts have testified at hearings and provided affidavits to the FCC about statistical sampling methods and the calculation of confidence intervals in complex statistical sampling designs. Several of the methods discussed below, including ratio estimators, probability samples, stratified samples and bootstrap estimates of confidence intervals, were issues in this more than decade-long work.
Finance
Banking Client
In a matter involving banking services, an AACG economist was the statistical sampling expert, drawing the sample that would be analyzed and then extrapolating the results to the broader population.
Healthcare
Hospital Payments
For a series of hospitals, our experts developed the sampling plan and statistical analysis used to estimate the magnitude of over-payments and under-payments by Medicare and Medicaid. This work required development of samples from a large dataset, stratified samples, dollar weighted sample plans, and ratio estimators. These analyses have strict precision requirements for the confidence intervals and called for the determination of sample sizes that would have a high likelihood of achieving those precision goals.
Government
The DOJ required the collection of data and related economic modeling and extrapolation based on the sample. Responsibilities on the project included:
• Analysis of sampled results and extrapolation to the broader population
• Preparation of the report
• Discussed results under oath with the DOJ and parties interested in the analysis