Investigating the myriad features of molecular transport in materials yields fundamental information for understanding processes such as ion conduction, chemical reactions, and phase transitions. Molecular transport especially impacts the performance of ion-containing liquids and polymeric materials when used as electrolytes and separation media, with applications encompassing battery electrolytes, reverse-osmosis membranes, mechanical transducers, and fuel cells. Nuclear magnetic resonance (NMR) provides a unique probe of molecular translations by allowing measurement of all mobile species via spectral selectivity, access to a broad range of transport coefficients, probing of any material direction, and investigation of variable lengthscales in a material, thus, tying morphology to transport. Here, we present new concepts to test for and guarantee robust diffusion measurements. We first employ a standard pulsed-field-gradient (PFG) calibration protocol using (2)H(2)O and obtain expected results, but we observe crippling artifacts when measuring (1)H-glycerol diffusion with the same experimental parameters. A mathematical analysis of (2)H(2)O and glycerol signals in the presence of PFG transients show tight agreement with experimental observations. These analyses lead to our principal findings that (1) negligible artifacts observed with low gyromagnetic ratio (γ) nuclei may become dominant when observing high γ nuclei, and (2) reducing the sample dimension along the gradient direction predictably reduces non-ideal behaviors of NMR signals. We further provide a useful quantitative strategy for error minimization when measuring diffusing species slower than the one used for gradient calibration.