The solubility parameter is frequently used to screen potential solvents for a variety of applications. Solubility parameters for pure components are well documented; however, solvents used in industrial settings are usually mixtures. In this work, the mixture solubility parameter is defined in terms of experimentally measured properties. Mixture solubility parameters calculated from experiments are compared to two different methodologies: calculations using an ideal mixture assumption and calculations using the Perturbed-Chain Statistical Associating Fluid Theory (PC-SAFT) equation of state. It is shown that for a wide range of mixtures, the ideal mixture assumption and PC-SAFT calculations both give good agreement with experimental mixture solubility parameters. The agreement of the ideal mixture calculations with experiment does not signify these mixtures are ideal. This occurs because Sigma(n)(i=1) x(i)delta(2)(i)(v(i,m)/v(t,m)) >> vertical bar Delta e(mix)/v(t,m)vertical bar and Sigma (n)(i=1)x(i)v(i,m) >> vertical bar Delta v(mix)vertical bar for these particular systems. Chloroform/ketone systems present a case for which the ideal mixture assumption cannot predict the experimental mixture solubility parameters, while PC-SAFT calculations yield good agreement. Finally, the natural ability of the PC-SAFT approach to calculate Hansen partial solubility parameters due to the appropriate partitioning of energetic degrees of freedom is briefly discussed.