In recent years, transcranial photoacoustic (TPA) imaging has become a popular modality for diagnosis of brain disorders. However, due to the presence of skull, TPA images are strongly degraded. Acoustically, this degradation is mainly categorized into the phase aberration, mode-converted shear waves, and multiple scattering. Previous studies numerically investigated the effects of mode-converted shear waves and multiple scattering on TPA images while the phase aberration caused by the skull was ignored and a conventional delay-and-sum method was employed for reconstructing TPA images. In this paper, we investigate these effects while a refraction-corrected image reconstruction approach is used to form TPA images. This approach enables separating the effects of phase aberration, mode-converted shear wave and multiple scattering. A realistic human temporal bone based on a MicroCT was used in the numerical model. In average for all the absorbers, the power of the artifacts caused by the mode-converted shear wave and multiple scattering are -13.7 dB and -20.1 dB when the refraction is corrected during image formation, respectively. These values were -7.9 and -18.8 if the conventional reconstruction is used. Accounting for phase aberration enables accurate quantification of the effects of the mode-converted shear waves and multiple scattering, which is necessary for evaluating the methods developed for degrading these effects.