This study developed and tested the idea that training in simultaneous interpreting leads to the working memory advantage generally observed among interpreters. Answering the question of the working memory advantage possessed by interpreters, and its causal link with interpreting training, is of fundamental importance in the study of simultaneous interpreting. Very few studies in the field have been able to carry out a longitudinal study to test the evolution of working memory in interpreters. We conducted a longitudinal study consisting of two sessions of experiments designed to test working memory, in interpreting students at the University of Geneva and in a control group of translation students. Contrary to our expectations we did not observe significant changes in either test group, nor the expected working memory advantage. The results of our experiments, possible reasons for the unexpected outcome and areas to focus on in the future are discussed.