<p>This study conducted two experiments to investigate how punishment probability affects dishonest behavior and how this effect interacts with individual differences in sensitivity to punishment. Experiment 1 employed a public goods game paradigm to elicit participants’ experiences of potential penalties, while Experiment 2 assessed participants’ dishonest decisions, which could potentially lead to punishment. Both experiments systematically manipulated punishment probability (i.e., low, medium, and high) and measured participants’ sensitivity to punishment. Dishonest behavior was classified into two categories based on intensity: low-degree and high-degree. Results showed that high punishment probability decreased high-degree dishonest behavior compared to low probability, while medium probability showed a similar but non-significant trend. For low-degree dishonest behavior, medium punishment probability unexpectedly increased the incidence of dishonesty, and this effect was modulated by individual differences in sensitivity to punishment. Specifically, individuals with high punishment sensitivity exhibited greater behavioral restraint under medium punishment probability compared to those with low sensitivity. The findings underscore the differential impact of punishment probability on reducing dishonest behavior of varying degrees and highlight the significance of sensitivity to punishment in moderating the effectiveness of deterrence strategies.</p>