Understand the analysis before it becomes the argument.
Independent expert and consulting support for matters involving statistical evidence, complex datasets, predictive-model validation, machine-learning model evaluation, and algorithmic decision systems.
Consulting expert support
Not every technical problem requires a designated testifying expert. Counsel may first need to understand what the data show, whether an analysis is reproducible, or where a methodology is vulnerable.
Consulting engagements can include early technical assessment, review of data and analytical methods, reproduction of results, sensitivity analysis, review of opposing expert work, identification of methodological issues, and support in understanding complex quantitative evidence.
Potential engagement questions
- Can the reported result be independently reproduced?
- Do the source data support the analysis?
- Are inclusion, exclusion, or cohort rules outcome-sensitive?
- Are key assumptions explicit and defensible?
- Does the selected statistical method fit the question, data, and claimed conclusion?
- Has a predictive model been adequately validated?
- Do subgroup or fairness-related analyses support the claim being made?
- Would reasonable alternative analytical choices materially change the result?
Testifying expert services
Where the subject matter fits the appropriate scope of expertise, Stuart Jones Consulting may accept engagements requiring independent expert opinions, reports, deposition, or testimony.
Any opinion is developed from the evidence and methods appropriate to the question presented. Retention does not imply a predetermined conclusion, and compensation is not contingent on the substance of an opinion or the outcome of a matter.
Possible areas of testimony
- Statistical methods, quantitative evidence, and interpretation
- Regression, forecasting, uncertainty, and anomaly analysis
- Data quality, lineage, and reproducibility
- Predictive-model specification, performance, stability, validation, and limitations
- Machine-learning model evaluation and algorithmic decision systems
- Subgroup performance and fairness-related statistical analysis
Typical stages of support
Early assessment
Determine whether a quantitative issue is material, what evidence would be needed, and whether the available data can answer the question posed.
Discovery and technical record
Help identify the analytical materials necessary to reproduce or evaluate the work, including data, documentation, code, assumptions, model artifacts, and methodological descriptions where relevant.
Analysis and rebuttal
Reproduce reported results, test sensitivity to assumptions, evaluate model performance, and distinguish substantive methodological problems from differences in reasonable analytical judgment.
Communication
Translate technical findings into clear explanations suitable for counsel, reports, deposition preparation, and fact-finders when testimony is part of the engagement.
Independent by design.
Engagements are accepted only after conflict review and an initial assessment that the requested work falls within the proper scope of expertise. Opinions are developed from the evidence and methods appropriate to the question presented, not from a desired case outcome.
Have a data-intensive matter?
Start with the general subject, parties for conflict review, and a non-confidential description of the technical issue.