How Statistics Expert Witnesses Get Excluded Under Daubert — and How to Survive the Cross

Daubert is not a medicine problem. Under Kumho Tire and the 2023 amendment to Federal Rule of Evidence 702, the trial court's gatekeeping applies to every form of specialized testimony — statistics included. A 20-year study of 2,842 challenges to non-medical experts found that roughly half of those opinions were excluded or partially excluded, and the single most-cited reason was “unreliable methodology.”

The exclusion rarely happens in a written motion alone. It is built, piece by piece, in the deposition cross-examination — where opposing counsel walks a statistics expert into conceding scope, methodology, or an assumption that unravels the whole opinion. Here are the three traps, and how a prepared expert answers each one.

The three ways statistics experts lose ground

Scope: testifying outside your lane

The cross-examiner's question sounds simple:

Your analysis showed a significant disparity — but you didn't control for the obvious confounding variables, did you?

Why it works: Omitted-variable / failure-to-control is the statistician's exclusion driver. Show your controls and sensitivity.

A stronger answer: “I specified the model with the relevant controls, performed sensitivity analyses, and disclosed the variables and the limits of the available data.”

Methodology: the reliability attack

The cross-examiner's question sounds simple:

Your sample wasn't representative of the relevant population, was it?

Why it works: Methodology / sampling. Document the frame.

A stronger answer: “I drew the sample per accepted methodology to represent the relevant population and documented the sampling frame and any limits.”

Assumptions: the one premise that sinks the opinion

The cross-examiner's question sounds simple:

You assume your model's functional form is correct — but other specifications give different results, don't they?

Why it works: The model-specification assumption. Test alternatives and disclose sensitivity.

A stronger answer: “I justified the specification, tested reasonable alternatives, and disclosed the sensitivity rather than presenting a single result as definitive.”

How to prepare for the cross before you're sworn in

Every one of those traps is defeatable — but not by reading your report one more time. The experts who survive the cross have done three things:

• Rehearsed the cross-examination out loud, repeatedly, against a realistic examiner — so the scope concession, the methodology defense, and the assumption hedge are second nature.
• Mastered the record, so that when counsel asks them to recall the one line buried in thousands of pages of the dataset, model, and methodology notes, they can produce it in seconds rather than fumble.
• Stress-tested the report against FRE 702 — finding the reliability gaps before opposing counsel does.

Questions? Contact us at [email protected].