We analyze theoretically a common experimental process used to obtain the magnetic contribution to the specific heat of a given magnetic material. In the procedure, the specific heat of a non-magnetic analog is measured and used to subtract the non-magnetic contributions, which are generally dominated by the lattice degrees of freedom in a wide range of temperatures. We calculate the lattice contribution to the specific heat for the magnetic compounds GdMIn₅ (M=Co, Rh) and for the non-magnetic YMIn₅ and LaMIn₅ (M=Co, Rh), using density functional theory based methods. We find that the best non-magnetic analog for the subtraction depends on the magnetic material and on the range of temperatures. While the phonon specific heat contribution of YRhIn₅ is an excellent approximation to the one of GdCoIn₅ in the full temperature range, for GdRhIn₅ we find a better agreement with LaCoIn₅, in both cases, as a result of an optimum compensation effect between masses and volumes. We present measurements of the specific heat of the compounds GdMIn₅ (M=Co, Rh) up to room temperature where it surpasses the value expected from the Dulong-Petit law. We obtain a good agreement between theory and experiment when we include anharmonic effects in the calculations.