There is a specific moment in high-stakes professional assessment when a practitioner feels they have the read. The interview is building. The subject has shifted. Something in the room has changed. And the practitioner, drawing on years of operational experience, arrives at a judgment they feel certain about. That feeling of certainty is not a neutral sensation. It is information. The problem is that under pressure, it is often the wrong kind.
What the numbers say
The research on human credibility assessment is not ambiguous on this point. A meta-analysis by Bond and DePaulo (2006), drawing on over 24,000 judgments across 206 studies, found that the average human accuracy rate at distinguishing truthful from deceptive behaviour sits at 54 percent. Chance is 50. The gap between trained professionals and untrained lay people is, on average, less than two percentage points. Experience does not close it. Seniority does not close it. And confidence, the research shows, predicts almost nothing.
That last finding is the one that matters most for practitioners operating under load. A separate meta-analysis by DePaulo and colleagues found that the correlation between confidence in a credibility judgment and the accuracy of that judgment is 0.04. Effectively zero. The practitioner who feels certain they are reading a situation correctly is, statistically, no more likely to be right than the one who feels uncertain - and under conditions of operational stress, the certainty tends to increase while the accuracy does not follow.
What happens under cognitive load
This is not a training failure or an individual weakness. It is a documented feature of how human judgment operates under cognitive load. When the stakes are high and the pressure is real, the brain's threat-response systems draw resources away from the slower, more deliberate processing that accurate behavioural assessment requires. What replaces it is faster, more pattern-driven - and more prone to confirmation. The practitioner narrows onto the cues that fit the emerging picture and becomes less responsive to the ones that complicate it. The judgment feels sharper. The evidence base for it has quietly contracted.
Research by Reyes and colleagues (2015) adds a further layer. High biological reactivity to stress was shown to significantly reduce what the researchers termed metacognitive accuracy - the ability to distinguish when your own judgments are correct from when they are not. In plain terms: under stress, practitioners lose calibration. They become less accurate about their own accuracy. The internal signal that is supposed to flag uncertainty stops functioning reliably.
The practical consequence
The practical consequence of this is specific and serious. It is not that practitioners under pressure make more mistakes. They make mistakes with higher conviction, pursue those conclusions with more behavioural commitment, and are less receptive to evidence that would revise the read. In a credibility assessment context, this is the condition under which false conclusions about a person's honesty, intent, or reliability are most likely to be reached - and acted on.
What the literature does not offer is a simple corrective. Awareness of the problem does not resolve it. Knowing that stress degrades metacognitive accuracy does not restore that accuracy during a live assessment. The practitioner who has read this article will not be meaningfully better protected against this failure mode when they walk into a high-pressure interview tomorrow - not unless something structural has changed in how they approach the assessment process itself.
That structural question - what actually changes practitioner accuracy under load, and what the evidence says about how to build it - is the subject of the work Nuance Intel does with organisations whose practitioners operate in exactly these conditions. It is not addressed by confidence training or accumulated experience. The research is clear on that point. The starting place is a different kind of assessment altogether.
