Experimental studies regularly show that third-party punishment (TPP) substantially exists in various settings. This study further investigates the robustness of TPP under an environment where context effects are involved. In our experiment, we offer a third party an additional but unattractive risky investment option. We find that, when the dominated investment option irrelevant to prosocial behavior is available, the demand for punishment decreases, whereas the demand for investment increases. These findings support our hypothesis that the seemingly unrelated and dominated investment option may work as a compromise and suggest the fragility of TPP in this setting.