This weekend plot is borrowed from a nice recent review on dark matter detection:
In all existing direct detection experiments, the target contains nuclei rather than single nucleons. Unlike in the spin-independent case, for spin-dependent scattering the cross section is not enhanced by coherent scattering over many nucleons. Instead, the interaction strength is proportional to the expectation values of the proton and neutron spin operators in the nucleus. One can, very roughly, think of this process as a scattering on an odd unpaired nucleon. For this reason, xenon target experiments such as Xenon100 or LUX are less sensitive to the spin-dependent scattering on protons because xenon nuclei have an even number of protons. In this case, experiments that contain fluorine in their target molecules have the best sensitivity. This is the case of the COUPP, Picasso, and SIMPLE experiments, who currently set the strongest limit on the spin-dependent scattering cross section of dark matter on protons. Still, in absolute numbers, the limits are many orders of magnitude weaker than in the spin-independent case, where LUX has crossed the 10^-45 cm^2 line. The IceCube experiment can set stronger limits in some cases by measuring the high-energy neutrino flux from the Sun. But these limits depend on what dark matter annihilates into, therefore they are much more model-dependent than the direct detection limits.