We determined the dimensions of standard polystyrene latex spheres suspended in water at different sample concentrations by dynamic light scattering and turbidimetry. Applying two different schemes for the decorrelation of multiple scattering, we show that the samples exhibit distortions of the auto-correlation function due to multiple scattering even at moderate volume fractions, which prohibits the correct determination of the particle radius. The cross-correlation functions, however, are free from these distortions. The recently deployed three-dimensional crosscorrelation setup is superior to the commercially available twocolor machine, as more turbid samples are accessible. In order to verify the results obtained by dynamic light scattering, we performed turbidity measurements with the same samples. This method is inherently free from multiple scattering contributions. We observe a systematically smaller radius in turbidimetry than in dynamic light scattering. The deviation, however, is only slightly outside the accuracy range of the measurements. We discuss possible origins for this deviation and show that our measurements are compatible with a hairy layer present on the particle's surface.