To evaluate the difference in quality of hatchery-reared fingerlings released into the same field at the same time, we propose simultaneous estimation of mortality rates taking the correlation between two series of recovery data into account. We consider the joint likelihood function as an extension of the bivariate normal distribution. For a test of mortality rates of two groups released, the likelihood ratio test is used. As an example, we applied this procedure to the case of red seabream Pagrus major in the Seto Inland Sea. We obtained maximum likelihood estimates for four models. The correlation coefficient was estimated to be over 0.9, and two models which assumed no correlation were rejected. For the two models of the correlation, the likelihood ratio test could not reject the hypothesis of no difference in the two groups. We also found that the X2 test of recovery rates under a contingency table is invalid for the overdispersion and correlation of recovery data.