The ATLAS collaboration calibrates its b-tagging algorithms using a number of data-driven methods. The b-tagging algorithms, used to separate b-quark jets from light-quark jets, are widely used in top, Higgs, and Exotics analyses in ATLAS. The algorithm’s performance is measured for light-quark jets, charm jets, and bottom jets. In some cases multiple methods can be used to measure the performance: methods using di-jet data and a pure sample of top-quark pair production are be used to measure the bottom-quark jet performance. These methods have complementary strengths – the di-jet methods tend to have the smallest errors at low jet energy, and the ttbar methods tend to have smaller errors at high jet energy. We used a minimum likelihood method to fit the results, profiling common systematics, to combine the results. For results on common data we have also understood the statistical correlation of our methods. This poster will describe the details of the combination.