Abstract—In this research, traditional public goods experiment is modified to test the effectiveness of shared leadership on public goods provision. With the proposed shared leadership index, mean difference test, linear regression and local regression are employed to test the hypothesis. The results show significant positive relation between shared leadership and public goods contribution rate. It suggests management could employ shared leadership to improve group project performance where free riding might exist. Experimental method is used at the first time to measure shared leadership in the field of experimental economy and leadership.