Due to their potential superiority in terms of speed of construction, performance, modularity and durability, steel-concrete structures are an increasingly preferred choice for critical constructions, such as high-rise buildings seeking accelerated construction. One of the technological (and regulatory) barriers to a more frequent use of SC structures arises from the challenge of conducting rapid and extensive non-destructive testing of the concrete matrix to ensure that it is free from critical concreting defects. The steel plates act as a permanent formwork that prevents access to the concrete. Cheap control methods, such as visual inspection, therefore cannot be applied. While the scientific community has largely addressed the development of new connectors to establish a robust bond between the steel plates and the concrete, limited attention has been awarded to exploring the impacts of concrete defects on the structural behavior. Given the dimensions and the inherent complexity associated with SC structures, the cost of experimental studies increases rapidly, especially when testing various types of defects at different locations. The numerical representation of a defect is more straightforward, considering that it can be defined as a parameter for the simulation. However, achieving accurate simulation of complete SC structures remains a challenging task due to the relatively large meshes involved, and finding a solution as cracks propagate through the structure is complex for the solver. In this contribution, the structural consequences of a defect are investigated through its impacts on the stud-concrete interaction at a local scale. Various experimental pushout tests from the literature [1] are simulated using the finite element method with damage mechanics. Defects are then introduced into the mesh or in the material parameters to respectively simulate holes in the concrete, or changes in the material properties resulting from poor concreting. A set of numerical experiments is then designed using the Box-Behnken method [2] to identify the most critical defect type and determine its severity through the response surface methodology.