Characterizing the physical properties of exoplanets , and understanding their formation and orbital evolution requires precise and accurate knowledge of their host stars .
Accurately measuring stellar masses is particularly important because they likely influence planet occurrence and the architectures of planetary systems .
Single main-sequence stars typically have masses estimated from evolutionary tracks , which generally provide accurate results due to their extensive empirical calibration .
However , the validity of this method for subgiants and giants has been called into question by recent studies , with suggestions that the masses of these evolved stars could have been overestimated .
We investigate these concerns using a sample of 59 benchmark evolved stars with model-independent masses ( from binary systems or asteroseismology ) obtained from the literature .
We find very good agreement between these benchmark masses and the ones estimated using evolutionary tracks .
The average fractional difference in the mass interval \sim 0.7 – 4.5 M _ { \odot } is consistent with zero ( -1.30 \pm 2.42 % ) , with no significant trends in the residuals relative to the input parameters .
A good agreement between model-dependent and -independent radii ( -4.81 \pm 1.32 % ) and surface gravities ( 0.71 \pm 0.51 % ) is also found .
The consistency between independently determined ages for members of binary systems adds further support for the accuracy of the method employed to derive the stellar masses .
Taken together , our results indicate that determination of masses of evolved stars using grids of evolutionary tracks is not significantly affected by systematic errors , and is thus valid for estimating the masses of isolated stars beyond the main sequence .