Comparison of six high-redshift quasar spectra obtained with the Large Binocular Telescope with previous observations from the Sloan Digital Sky Survey shows that failure to correctly identify absorption and other problems with accurate characterization of the C IV λ1549 emission line profile in low signal-to-noise (S/N) data can severely limit the reliability of single-epoch mass estimates based on the C IV emission line. We combine the analysis of these new high-quality data with a reanalysis of three other samples based on high-S/N spectra of the C IV emission line region. We find that a large scatter between the Hβ- and C IV-based masses remains even for this high-S/N sample when using the FWHM to characterize the broad-line region velocity dispersion and the standard virial assumption to calculate the mass. However, we demonstrate that using high-quality data and the line dispersion to characterize the C IV line width leads to a high level of consistency between C IV- and Hβ-based masses, with <0.3 dex of observed scatter and an estimated ~0.2 dex intrinsic scatter, in the mass residuals.