<p>This study examined beliefs about deception cues among police officers and lay people, focusing on three questions. It aimed to replicate the no-difference hypothesis between these groups, explore the relationship between confidence in cue validity and cue frequency expectations, and test attribution bias by comparing expected cue frequency in lies told by oneself, a close other, and a stranger.&#xa0;Participants included 124 police officers and 182 lay people. Bayesian analysis was used to test the no-difference hypothesis about deception cue beliefs. The study also analyzed the link between metacognitive confidence and expected cue frequency and assessed attribution bias across relational contexts (self, close others, strangers).&#xa0;Police officers and lay people showed similarly inaccurate beliefs about deception cues, confirming prior findings. Higher metacognitive confidence was linked to expectations of more frequent deception cues. Attribution bias emerged, with participants expecting cues like gaze aversion and posture changes to be more frequent in strangers than in themselves or close others.&#xa0;The findings highlight the need for improved deception detection training for police officers. Training should correct false beliefs, enhance confidence calibration, and reduce outgroup attribution biases. It could eventually replace beliefs with valid knowledge and prevent the occurrence of erroneous deceptiveness judgments.</p>