The Einstein-Rosen-Podolsky (EPR) paradox gives an argument for the incompleteness of quantum mechanics based on the premise of local realism. The general viewpoint is that the argument is compromised, because local realism is falsifiable by Bell or Greenberger-Horne-Zeilinger (GHZ) experiments. In this paper, we challenge this conclusion, by presenting alternative versions of the EPR paradox based on premises not falsifiable by the GHZ and Bell predictions. First, we explain how the Bohm-EPR and GHZ paradoxes can be demonstrated using macroscopic spins Ŝθ formed from qubits realised as two macroscopically distinct states. This establishes an "all or nothing" incompatibility between quantum mechanics and macroscopic realism (MR). However, we note different definitions of MR. For a system in a superposition of two macroscopically distinct eigenstates of Ŝθ , MR posits the system to have a definite value for the outcome of Ŝθ . Deterministic macroscopic realism (dMR) posits MR regardless of whether the physical interaction U θ determining the measurement setting θ has actually occurred. In contrast, the weaker assumption, weak macroscopic realism (wMR), posits MR for the system prepared after U θ . We show that the GHZ paradox negates dMR but is consistent with wMR. Yet, we show that a Bohm-EPR paradox for the incompleteness of quantum mechanics arises, based on either form of MR. Since wMR is not falsified, this raises the question of how to interpret the EPR paradox. We revisit the original EPR paradox and find a similar result:The EPR argument can be based on a contextual version of local realism (wLR) alluded to by Bohr, which is not falsifiable by Bell or GHZ experiments. The premises wLR and wMR posit realism and no-disturbance for systems prepared with respect to a pointer basis (after U θ ), leading to further predictions giving consistency with quantum mechanics.