Developments in single-molecule experimental techniques have led to the ability to indirectly monitor trajectories of a biological macromolecule fluctuating at equilibrium in the absence or presence of an external biasing potential. Often, the experimental observable (such as the energy transfer efficiency between two fluorophores or extension along a pulling coordinate) is chosen as a putative reaction coordinate along which the kinetic behavior of the macromolecule is presumed to be described by a diffusion process. The observable, however, is not guaranteed to provide a useful reaction coordinate --- poor choices will not provide a separation of timescales between motion along the coordinate and relaxation of degrees of freedom orthogonal to it, causing such a diffusion model to fail to accurately describe dynamics. Here, we invoke the concept of the splitting probability between two absorbing boundaries as a test of the suitability of the observable as a reaction coordinate for equilibrium single-molecule experiments. We apply this test to the extension coordinate of a DNA hairpin system in an optical trap under constant applied force and to the fluorescence resonant energy transfer (FRET) efficiency of a labeled thermostable RNA in the absence of force.