In this work, the robustness of the sample preparation procedure for the determination of six tranquilizers (xylazine, azaperone, propionylpromazine, chlorpromazine, haloperidol, and azaperol) and a beta-blocker (carazolol) in animal muscle by LC/MS–MS was assessed through the experimental design methodology. A 2III7 − 4 fractional factorial design was performed to evaluate the influence of seven variables on the final concentration of the seven drugs in the samples, in accordance with what is laid down in Commission Decision No 2002/657/EC. The variation considered for each of those seven factors is likely to happen when preparing the samples, being the values chosen as level − 1, the nominal operating conditions. The results of the experimentation were evaluated from different statistical strategies, such as hypothesis testing using an external variance previously estimated, Lenth's method, and Bayesian analysis. Both Lenth's and Bayes' approaches enabled to determine the effect of every variable even though no degrees of freedom were left to estimate the residual error. The same conclusion about the robustness of the extraction step was reached from the three methodologies, namely, none of the seven factors examined influenced on the method performance significantly, so the sample preparation procedure was considered to be robust.
