The aerosol direct radiative effect (ADRE) is defined as the change in the solar radiation flux, <i>F</i>, due to aerosol scattering and absorption. The difficulty in determining ADRE stems mainly from the need to estimate <i>F</i> without aerosols, <i>F</i><sup>0</sup>, with either radiative transfer modeling and knowledge of the atmospheric state, or regression analysis of radiation data down to zero aerosol optical depth (AOD), if only <i>F</i> and AOD are observed. This paper examines the regression analysis method by using modeled surface data products provided by the Aerosol Robotic Network (AERONET). We extrapolated <i>F</i><sup>0</sup> by two functions: a straight linear line and an exponential nonlinear decay. The exponential decay regression is expected to give a better estimation of ADRE with a few percent larger extrapolated <i>F</i><sup>0</sup> than the linear regression. We found that, contrary to the expectation, in most cases the linear regression gives better results than the nonlinear. In such cases the extrapolated <i>F</i><sup>0</sup> represents an unrealistically low water vapor column (WVC), resulting in underestimation of attenuation caused by the water vapor, and hence too large <i>F</i><sup>0</sup> and overestimation of the magnitude of ADRE. The nonlinear ADRE is generally 40–50% larger in magnitude than the linear ADRE due to the extrapolated <i>F</i><sup>0</sup> difference. Since for a majority of locations, AOD and WVC have a positive correlation, the extrapolated <i>F</i><sup>0</sup> with the nonlinear regression fit represents an unrealistically low WVC, and hence too large <i>F</i><sup>0</sup>. The systematic underestimation of <i>F</i><sup>0</sup> with the linear regression is compensated by the positive correlation between AOD and water vapor, providing the better result.