[en] We consider the Biswas–Chatterjee–Sen model on Barabasi–Albert networks. This system undergoes a continuous phase transition from a consensus state to a disordered state by increasing a noise parameter q over a critical threshold q _c. The noise parameter is defined as the probability of the affinity between two neighbors being negative, modeling Galam contrarians. We obtained the critical exponent ratios , , and by finite-size scaling data collapses, as well as the critical noises. Our numerical data is consistent with the critical thresholds q _c being a linear function of the inverse of network connectivity z, and with an asymptotic value of q _c  =  0.3418, close to the value of the critical noise for the complete graph. (paper: classical statistical mechanics, equilibrium and non-equilibrium)