In this work, a modified push-out test is proposed to study the load-slip relationship in steel-concrete-steel sandwich beams. This relation is one of the most important factors that are required in the analysis and design of steel-concrete-steel sandwich beams with partial interaction. The diameter of the connector is assumed to be variable while the other parameters are kept constants. The modification in this test in comparison with the standard test concentrated on the use of steel tube- concrete slab- steel tube instead of concrete slab- steel (I-Section)- concrete slab (in Standard Test) in order to be more compatible than the standard test in modelling the steel-concrete-steel sandwich or double-skin beams. The stud is suggested to be threaded along the whole length, connected to the tube by a nut (the separation between the layers is eliminated and only interlayer slip is assumed to exist) and the connector passes to the other tube through the concrete slab (thus the connector is subjected to double shear force). These cases maximize the function of this stud. Five dial gages are used; one at the base, and two at each side in order to measure the slip at each stud. An experimental relationship of load-slip is carried out to simulate the behaviour of this type of connection in steel-concrete-steel sandwich construction.