There is an unpleasant paradox faced by almost everyone who has conducted interviews: the most confident candidates are not always the most competent. Sometimes it is quite the opposite.
This is not a coincidence or a quirk of specific individuals. This is the Dunning-Kruger effect, described in 1999 by psychologists David Dunning and Justin Kruger: people with low competence in a certain area systematically overestimate their skills—simply because they also lack the very skills needed to recognize their own incompetence.
In an interview, it looks something like this. A candidate with genuinely strong expertise sees all the nuances and limits of their knowledge—they are aware of how much they still do not know, and speak more cautiously, with caveats, sometimes doubting themselves out loud. A poorly prepared candidate simply does not see these limits—they think they answered well because they lack the experience to notice the mistake.
A systemic bias occurs: if HR or a hiring manager primarily evaluates the confidence of delivery—which is convenient, fast, and subjectively pleasant—there is a risk of systematically preferring those who sound more confident over those who know better.
What is done about this in practice: structured interviews with clear criteria instead of a general impression, practical tasks instead of conversations about skills, and calibration sessions among interviewers to cross-check the feeling of "they sounded confident" against the actual quality of answers.
Confidence is not an indicator of competence. Sometimes it is its opposite.