The main reasons for the low reliability of results from preclinical studies are the lack of prior sample size calculations and poor experimental design. Here, we demonstrate how the tools of meta-analysis can be implemented to tackle these issues. We conducted a systematic search to identify controlled studies testing established migraine treatments in the electrophysiological model of trigeminovascular nociception (EMTVN). Drug effects on the two outcomes, dural stimulation-evoked responses and ongoing neuronal activity were analyzed separately using a three-level model with robust variance estimation. According to the meta-analysis, which included 21 experiments in rats reported in 13 studies, these drugs significantly reduced trigeminovascular nociceptive traffic, affecting both outcomes. Based on the estimated effect sizes and outcome variance, we provide guidance on sample sizes allowing to detect such effects with sufficient power in future experiments. Considering the revealed methodological features that potentially influence the results and the main source of statistical bias of the included studies, we discuss the translational potential of the EMTVN and the steps needed to improve it. We believe that the presented approach can be used for design optimization in research with other animal models and as such deserves further validation.