Forced Rayleigh scattering (FRS) studies have been carried out on methyl yellow/alcohol solutions at two different probe wavelengths (633 and 543 nm). The signal shapes observed at the two wavelengths are quite different: using ethanol and 2-propanol as solvents, we observe at 633 nm a decay-grow-decay (DGD) shape similar to profiles frequently reported in the literature, while at 543 nm we observe a DGD shape in which the signal at the local minimum does not reach the baseline. In principle, the nonzero minimum can be accounted for by a difference (due to amplitude/phase-grating mixtures) in the phase shifts of light scattered from the photoproduct and ground-state molecules. To test this hypothesis, we first show in a straightforward manner that the signal can be approximated as the product of a polynomial and an exponential decay, which allows for data reduction of profiles with zero or nonzero phase-shift differences. Using this approach, the diffusion coefficients measured using the two probe wavelengths are found to be the same to within an uncertainty of 2%-3%. The results provide strong evidence that the difference in signal shapes is caused by amplitude/phase-grating mixtures within the methyl yellow/alcohol system, and imply that diffusion coefficients can be measured without difficulty for other FRS systems characterized by such mixtures.