We investigate the approval mechanism (AM) for a common pool resource (CPR) game with three players, underlining the role of unanimity and majority rules. The game involves two stages. In stage 1, players simultaneously and privately choose a proposed level of extraction from the CPR. In the second stage, they simultaneously decide whether to approve or disapprove others' choices. If the group approves, players' first stage proposed extractions are implemented. Otherwise, a uniform extraction level, called disapproval benchmark (DB), is implemented onto each group member. We combine two approval rules, majority and unanimity, with two DBs, the minimum extraction level (MIN DB) and the Nash extraction level (NASH DB). These combinations offer four different treatments for testing the approval mechanism (AM). Our experimental findings show that the AM reduces signicantly over-extraction in each treatment, and that the unanimity rule is more effective than the majority rule to lower extractions. The MIN DB reduces more group extractions than the NASH DB. Finally, only the MIN DB with unanimity implements the Pareto-effcient extraction level.