To better understand the results of the disclosure of social responsibility information and promote the related policies, this paper empirically analyses the relation of earning per share (EPS), social contribution value per share (SCVPS) and abnormal returns based on the data from China’s A-share market for 2009 to 2015. This paper compares the Random-effects GLS regression and the Fixed-effects (within) regression using a Hausman test and uses the Fixed-effects regression to test the relation between SCVPS and abnormal returns. This paper proposes that the SCVPS will provide incremental information beyond EPS for investors to predict abnormal returns. Previous research indicates that there is a positive relation between EPS and abnormal returns. This paper predicts that the data will indicate a positive relation between EPS and abnormal returns and that using the official formula to calculate the SCVPS for a large number of companies to run the regression, the results will show that SCVPS can provide incremental information to predict abnormal returns. This study provides new insights into calculating SCVPS that are not limited to the disclosures of the companies to run the regression in large data to find the SCVPS value.