||The lack of normalized protocols to perform dynamic material tests could lead to inconsistent results compared to normalized static tests ones, or between different dynamic test campaigns. Such inconsistencies may be problematic when an accurate identification of strain rate dependent material models is needed, from creep to fast dynamics loadings. For instance, changes in the dynamic specimen in-plane geometry compared to the normalized static ones proved to influence the apparent elastic modulus of CFRP materials. In the present work, a pre-normative study of the dynamic specimen geometry is presented. First, a geometrical criterion is proposed for the [+/-45]s specimens that are commonly used (Rosen tension tests) to study the shear behaviour of composite materials, in order to answer the above mentioned consistency issue between static and dynamic test results. Then, several biases are pointed out which appear when using the (various) normative static test exploitation rules to identify the material properties from dynamic test results, which end up to the proposal of general recommendations to calculate CFRP elastic apparent modulus the same and accurate way for static and dynamic tests, and identify advanced visco-elastic models.