Many high-z galaxies have recently been detected with the James Webb Space Telescope (JWST), providing a better understanding of the evolution of the early universe. The physical properties of these galaxies are often inferred from spectral energy dis...
Many high-z galaxies have recently been detected with the James Webb Space Telescope (JWST), providing a better understanding of the evolution of the early universe. The physical properties of these galaxies are often inferred from spectral energy distribution (SED) fitting, but it is unclear how accurate this process is for such early galaxies. To address this issue, we use simulated SEDs at z=6 from the SPHINX20 cosmological simulation and employ BAGPIPES to quantitatively estimate the uncertainties in the recovered stellar masses, star formation rates, and stellar metallicities from mock observations with JWST NIRCam photometry. Even without dust and emission lines, we find that intrinsic stellar continuum fitting can overestimate stellar masses by 0.2 dex on average (up to 0.7 dex in low-mass galaxies), due to age and metallicity degeneracies. Dust and nebular emission further exacerbate these biases, with emission lines alone causing mass uncertainties by 0.1 dex on average, primarily by altering the inferred stellar population. Including bands that do not contain strong emission lines, such as F410M, reduces the overestimation of stellar mass to within 0.7 dex by separating the effects of emission lines from the contribution of old stars. We also find that adopting a best-fit model or likelihood-weighted estimates yields more accurate properties. Finally, while the stellar mass functions are reasonably recovered, the star formation main sequence varies significantly depending on the chosen fitting models. Our results underscore the importance of careful modeling choices when interpreting high-redshift photometric data, particularly for galaxies that are emission-line-dominated or low-mass with recent bursts of star formation.