TRIP (transformation induced plasticity) in steels is made responsible for some remarkable features such as a good compromise between ductility and strength which is highly relevant in lightweight design. For the purpose of studying the TRIP effect itself an appropriate testing material, in this case a Cr-Ni-Mo-Al-Ti maraging steel transforming completely from austenite to martensite slightly above room temperature has been chosen, which has extensively been tested in a series of previous experiments by Nagayama et al. In order to reproduce these experimental findings numerically, a unit cell approach based on a regular cubic array of finite elements (RVE, representative volume element), each one accounting for an initially austenitic domain that is subject to martensitic transformation is presented in this thesis. During transformation triggered by cooling below the martensite start temperature, randomly selected single elements of the unit cell are toggled from austenite to martensite following a given kinetic law (element toggling technique). At the same time the unit cell is subjected to different external loading paths, including constant hold-stresses as well as unloading cases. At integration point level the total strain can be decomposed into an elastic, a plastic, a volumetric and a deviatoric orientation related strain contribution and their values can be worked out. The finite element results at integration point level, i.e. the micro level allow to calculate the averaged quantities on the RVE level, i.e. the meso level. This way the magnitude of the TRIP strain can be determined. Moreover, each contribution to the TRIP strain, i.e. the plastic strain due to the Greenwood-Johnson effect as well as the orientation strain due to the Magee effect can be identified. The numerical procedure also allows to work out the average phase stresses. This information serves as valuable input for mean-field models of martensitically transforming materials. For constant load levels the unit cell approach works well. When unloading cases are considered, the expected backflow of TRIP strain, as seen in the experiments by Nagayama et.al, is not adequately reproduced by the model. Nevertheless, the final values of the TRIP strain after complete transformation satisfactorily agree with the experimental findings, even for unloading cases.