Electromagnetic Acoustic Transducers (EMATs) are widely used in Nondestructive Evaluation (NDE) of metals, including the inspection of ferritic steels. Optimization of EMATs in ferromagnetic materials is often accomplished using computational simulations that account for the three main types of transduction mechanism, namely the effects of Lorentz forces, magnetization forces, and magnetostriction. The extensive literature on the Lorentz and magnetization mechanisms of EMATs has been historically scattered with alternative notations causing conflicting interpretations of the fundamental governing equations that often led to erroneous implementations of these equations in computational simulations. The purpose of this paper is to clarify this issue. The confusion is caused by the arbitrary nature of the separation of electromagnetic transduction force into Lorentz and magnetization components. This paper shows that apparently contradicting models are actually equivalent if the Lorentz and magnetization forces are chosen to correctly match each other, while mixing otherwise acceptable but mutually incompatible terms leads to unacceptable modelling errors. This paper also provides guidelines for the implementation of these transduction mechanisms in computational models for EMATs.