Specific volumes of high-density and low-density polyethylene liquids at several elevated temperatures and pressures were measured. The measured specific volumes were then used to estimate the thermal expansion coefficients (alpha = (1)/(upsilon) ((partial derivativeupsilon)/(partial derivativeT))(P)) and isothermal compressibility (beta = -(1)/(upsilon)((partial derivativeupsilon)/(partial derivativeP))(T)) of of the polymers. Two different approaches were used in which one was simply to fit the raw data by second order polynomials to obtain (partial derivativeupsilon/partial derivativeT)(P) and (partial derivativeupsilon/partial derivativeP)(T), while the other by the Sanchez-Lacombe (S-L) equation of state. It was found that the resultant alpha and beta obtained from the above methods differ significantly, indicating that the S-L equation of state may not be suitable for determining alpha and beta at elevated temperatures. When these two sets of alpha and beta were used to calculate the corresponding solubility parameters and then the Flory-Huggins interaction parameters (chi) of the polymers, the results also differ considerably. Nonetheless, chi obtained from the first method agrees well with the results obtained from small angle neutron scattering measurements while the S-L equation of state method does not. The current results suggest that solubility and interaction parameters obtained from pressure-volume-temperature experiments depend critically on the manner by which the data analysis is performed.