We have found significant entropy-enthalpy compensation for the transfer of a diverse set of two-state folding proteins from water into water containing a diverse set of cosolutes, including osmolytes, denaturants, and crowders. In extracting thermodynamic parameters from experimental data, we show the potential importance of accounting for the cosolute concentration-dependence of the heat capacity change upon unfolding, as well as the potential importance of the temperature-dependence of the heat capacity change upon unfolding. We introduce a new Monte Carlo method to estimate the experimental uncertainty in the thermodynamic data and use this to show by bootstrapping methods that entropy-enthalpy compensation is statistically significant, in spite of large, correlated scatter in the data. We show that plotting the data at the transition midpoint provides the most accurate experimental values by avoiding extrapolation errors due to uncertainty in the heat capacity, and that this representation exhibits the strongest evidence of compensation. Entropy-enthalpy compensation is still significant at lab temperature however. We also find that compensation is still significant when considering variations due to heat capacity models, as well as typical measurement discrepancies lab-to-lab when such data is available. Extracting transfer entropy and enthalpy along with their uncertainties can provide a valuable consistency check between experimental data and simulation models, which may involve tests of simulated unfolded ensembles and/or models of the transfer free energy; we include specific applications to cold shock protein and protein L.