We argue that for a single particle Bell's inequality is a consequence of noncontextuality and is incompatible with statistical predictions of quantum mechanics. Thus noncontextual models can be empirically falsified, independent of locality condition. For this an appropriate entanglement between disjoint Hilbert spaces pertaining to translational and spin degrees of freedom of a single spin-1/2 particle is invoked.