Detection curves for bacterial detection are derived using Bayesian modelling. We show how to merge and use results from heterogeneous spiking experiments. Differences between B. anthracis and B. cereus and several food matrices were found. The interaction between matrix and bacteria type was non-ignorable.