We present halo-independent methods to analyze the results of dark matter direct detection experiments assuming inelastic scattering. We focus on the annual modulation signal reported by DAMA/LIBRA and present three different halo-independent tests. First, we compare it to the upper limit on the unmodulated rate from XENON100 using (a) the trivial requirement that the amplitude of the annual modulation has to be smaller than the bound on the unmodulated rate, and (b) a bound on the annual modulation amplitude based on an expansion in the Earth's velocity. The third test uses the special predictions of the signal shape for inelastic scattering and allows for an internal consistency check of the data without referring to any astrophysics. We conclude that a strong conflict between DAMA/LIBRA and XENON100 in the framework of spin-independent inelastic scattering can be established independently of the local properties of the dark matter halo.