Why Great Doctors Get Destroyed on Cross-Examination — and How Not To

By John Mahoney · June 2026 · 9 min read

A brilliant clinician walks into a deposition with thirty years of practice, board certification, and a wall of publications — and walks out having handed the other side three concessions, two absolutes, and an opinion that may not survive a motion to exclude. It happens constantly, and not because the doctor was wrong. It is that being an excellent physician and being an effective expert witness are two different skills, and the second is almost never taught in medical training.

That is the good news, oddly. The consistent message from the people who train experts for a living — from SEAK's deposition-skills work to the DOJ / National Institute of Justice "Law 101: Legal Guide for the Forensic Expert" — is that testifying is a learnable, teachable discipline. This article lays out what it consists of: the deposition rules, the credibility science, the certainty standard that keeps opinions admissible, the "hired gun" trap, and the practice method that moves the needle.

This article is for educational purposes only and is not legal advice. Admissibility standards, certainty-of-opinion phrasing, and the qualification of expert testimony vary by jurisdiction and are frequently reinterpreted by the courts. Retaining counsel should confirm the controlling requirements in every case.

The Skill Gap Nobody Warns You About

Clinical reasoning rewards thoroughness and saying everything that might be relevant — the differential, the caveats, the "on the other hand" — because in medicine, withheld context can hurt a patient. Cross-examination inverts those instincts: the clinician's reflex to be complete, helpful, and educational is exactly the reflex a good opposing attorney is counting on, because every volunteered sentence is a new door to walk through. That is why the smartest person in the room is so often hurt worst — intelligence is neither the problem nor the protection. The expert who improves recognizes that the deposition and witness stand are a separate performance environment, and learns its rules deliberately rather than discovering them mid-cross.

The Deposition Rules That Separate Survivors From Casualties

SEAK's catalog of common deposition mistakes and the NIJ's "Law 101" forensic-expert guidance converge on a short list of disciplines. None are about being clever — they are about restraint.

• Answer only the narrow question asked — then stop. The single highest-leverage habit. The question defines the scope of the answer; once it is answered, the job is done. Silence after a complete answer is not awkward, it is correct.
• Never volunteer. Unprompted elaboration is where depositions are lost — the helpful extra fact, the clarifying aside, the "well, technically." If the questioner wants more, they will ask.
• Do not assume every question deserves an answer. Some are vague, compound, or built on a false premise. "I can't answer that as phrased" is professional — not evasion.
• Do not lose your temper. A baiting questioner is often working to provoke the defensive answer that reads badly on a cold transcript and worse on video. Composure is a credibility asset; losing it is a gift to the other side.
• Come prepared to defend the foundation of every opinion — the factual basis, the methodology, the timing, and the consistency of each opinion with prior discovery (the same rigor a certificate-of-merit affidavit demands at the outset of a case). The cross-examiner's best material is a contradiction between today's answer and a prior report, an earlier deposition, or the record itself.

The "absolutely" trap

One discipline is so common and costly it deserves its own heading. The NIJ guidance is explicit: words like "absolutely" and "positively" are dangerous, because absolute means forever, without exception — a single counterexample, one atypical case in the literature, can detonate an answer stated too strongly. The same guidance warns against the mirror error: do not exaggerate, underestimate, or minimize. Calibrated language is not hedging; it is what makes the rest of the testimony trustworthy.

The Certainty Standard That Keeps Your Opinion Admissible

On the other side of the absolutes problem is a line just as fatal to cross the wrong way — this time to admissibility, not credibility. For an expert opinion to be admissible, it generally must be stated to at least a reasonable degree of medical certainty (or probability) — meaning more probable than not, greater than 50%. An opinion phrased as a mere possibility ("it could have," "it's possible that") risks being excluded as speculative. The two failure modes are mirror images.

Overstate the opinion — "absolutely," "without question" — and it shatters on cross at the first exception. Understate it below "more probable than not" and it may be excluded before a jury ever hears it. The skill is landing in the defensible middle — more likely than not, to a reasonable degree of medical probability — and being able to say why.

That defensible middle is what a later admissibility fight will test: the same methodology and basis questions that govern phrasing today drive a Daubert and FRE 702 reliability challenge tomorrow — so the expert who can articulate the basis of an opinion clearly at deposition is also the one hardest to exclude.

The Credibility Science: It Isn't About Being Right

Here is the part that surprises most physicians: with a jury, being correct is necessary but nowhere near sufficient. Research on how jurors evaluate experts — most prominently the Witness Credibility Scale (Brodsky, Griffin, and Cramer, 2010) — describes perceived credibility as a four-factor construct: confidence, likability, trustworthiness, and knowledge. Knowledge is the one physicians arrive with; the other three are performance variables, and two behave in counterintuitive ways.

Confidence is an inverted-U, not a ladder

The intuitive model is "more confident equals more believable." The data say otherwise. Research on expert confidence (Cramer and Brodsky) finds the relationship is curvilinear — an inverted U: medium confidence is rated the most credible, while both low confidence and maximum confidence score worse. Overconfidence reads as significantly less credible than a calibrated, moderate presentation — the expert who answers every question as if beyond all doubt reads not as authoritative but as a salesperson. The most believable witness sounds sure of what they know and measured about its limits, the same posture the "absolutely" rule and the certainty standard already demand.

Likability is dominant — but asymmetric

Of the four factors, likability does a surprising amount of the work, but not in the direction most assume. The research is asymmetric: being disliked carries a heavy penalty — jurors discount even strong evidence from an expert they find off-putting — while charm above a neutral baseline adds essentially nothing. So the lesson is not "be charismatic" but "do not be arrogant, condescending, or combative." The downside of being unlikeable is enormous; the upside of extra charm is roughly zero. An expert does not need to win the jury over with personality — only to avoid giving them a reason to dislike them, which again lines up with staying composed and not sparring with the questioner.

An honesty note on this science: most of the jury-credibility research above — the confidence curve, the likability asymmetry — was conducted in forensic psychology and psychiatry settings, often with mock jurors or undergraduate participants, not in medical-malpractice civil trials with physician experts before real juries. The mechanisms plausibly transfer, but the exact magnitudes for physician experts before actual juries have not been directly tested. Treat these as well-supported directional principles, not precise dials.

Adversarial Allegiance: The Empirical Core of the "Hired Gun" Problem

Trustworthiness, the third credibility factor, has a specific enemy backed by evidence. Cross-examiners love to imply that a retained expert is simply saying what the retaining side paid them to say — easy to dismiss as a cheap rhetorical move, except the research suggests something real underneath it. Work by Murrie and Boccaccini on adversarial allegiance found that experts unconsciously drift toward the side that retained them even when scoring objective, structured instruments. Given identical case materials, experts reached different conclusions depending on which side they believed had retained them, with effect sizes reported as large as Cohen's d = 0.85. The drift appeared to operate below awareness rather than as conscious dishonesty — which is exactly why the honest expert is not immune.

The implication is not despair but hygiene. An expert who wants to be — and to appear — genuinely independent can build that independence into how they practice:

• Keep a balanced caseload. An expert who works only one side of the "v." is the textbook hired gun; taking both plaintiff and defense engagements is far harder to caricature.
• Be consistent across cases. The same fact pattern should produce the same opinion regardless of who is paying. Inconsistency is the cross-examiner's favorite exhibit.
• Charge reasonable, transparent fees not contingent on outcome — removing an easy line of attack.
• Show your methodology. A transparent, reproducible method another qualified expert could follow to the same conclusion is the strongest answer to "you just reached the conclusion you were hired to reach."

The same balanced, transparent caseload that counters allegiance in perception also limits the unconscious drift in reality — the rare case where the credibility move and the right thing to do are the same move.

How Experts Actually Get Better: Deliberate Practice

Everything above is knowable from a weekend of reading, and none of it transfers to performance from reading alone — the whole reason great clinicians still get hurt. The lever that actually changes how someone testifies is deliberate practice: repeated reps under realistic, adversarial conditions, followed by specific critique. That is the design of the gold-standard programs — SEAK's flagship expert-witness course is built around live mock cross-examinations and depositions with individualized feedback, not lectures. You cannot internalize "answer only the question and stop" by agreeing it's good advice; you internalize it by getting burned for volunteering, in a safe setting, until the new habit overwrites the clinical one.

Two habits carry most of the gain:

1. Reps against a genuinely hostile questioner. A friendly colleague lobbing easy questions builds false confidence. The value comes from a questioner actively trying to make you volunteer, overstate, contradict yourself, or lose composure — because that is what the real cross will do.
2. Reviewing your own transcript. Reading back what you actually said — the buried "absolutely," the answer that ran three sentences long, the place you got defensive — is where abstract rules become concrete, personal fixes. The transcript does not flatter; it instructs.

The obstacle has always been access: live mock-cross courses are scarce, scheduled, and expensive. The constraint was never knowing what to do — it was getting enough reps to make the right habits automatic.

Bottom Line

Great doctors get destroyed on cross-examination not because they are wrong, but because testifying is a separate, unpracticed skill. Answer only what is asked and stop, refuse the absolutes, stay composed, and state opinions in the defensible middle — more probable than not, to a reasonable degree of medical probability — so they survive both impeachment and exclusion. Present with calibrated confidence, avoid arrogance and combativeness, and counter the hired-gun perception with a balanced caseload, consistent opinions, and a transparent method. Then practice it: the reading is the easy part, but the reps change the performance. The expert deposed a dozen times by a hostile questioner in a low-stakes setting — and who read every transcript back — is the one who stays calm, narrow, and credible when it counts.

Questions? Contact us at [email protected] or (856) 979-6525