In this work, we employ trajectory-based simulations to predict the electronic coherences created by nonadiabatic dynamics near conical intersections. The mapping approach to surface hopping (MASH) is compared with standard fewest-switches surface hopping on three model systems, for which the full quantum-mechanical results are available. Electronic populations and coherences in the adiabatic representation, as well as nuclear densities, are computed to assess the robustness of the different methods. The results show that standard surface hopping can fail to describe the electronic coherences, whereas they are accurately captured by MASH at the same computational cost. In this way, MASH appears to be an excellent simulation approach for novel X-ray spectroscopies, such as the recently proposed transient redistribution of ultrafast electronic coherences in attosecond Raman signals (TRUECARS).