Polarimetry is one of the most promising types of remote sensing for improved characterization of atmospheric aerosol. As an important polarization data source in the world currently, the Directional Polarimetric Camera (DPC) onboard Gaofen-5 satellite can provide measurements of the Stokes vector, the polarization radiance (L_p), and degree of linear polarization (DOLP). The calibration accuracy of these polarization quantities will affect the retrieval error of aerosol parameters, and the corresponding impacts need to be analyzed quantitatively. For this purpose, the degree of freedom for signal (DFS) and the posterior error for aerosol model parameters are calculated based on the optimal estimation inversion framework and information content analysis theory. The results show that the information contained in the Stokes vector is the highest, followed by DOLP and L_p. The corresponding total aerosol DFS are 7.5, 6.1, and 5.2, respectively. The imaginary part of the complex refractive index (m_i^f) and the effective variance (v_eff^f) reduce significantly when using L_p in the retrieval. This indicates that the two parameters are sensitive to the polarization direction and measurement error. The average DFS of the fine-mode columnar volume concentration (V₀^f), the real part of the complex refractive index (m_r^f), and the effective radius (r_eff^f) are larger than 0.85, which can be well retrieved from DPC multiangle polarization observation. The inversion of coarse-mode aerosol parameters has high uncertainties, which is related to the aerosol type. Compared with adopting the Stokes vector in the retrieval, the posterior error of aerosol parameters increase 67.6% and 65.5% on average for adopting L_p and DOLP, respectively. The fine-mode columnar volume concentration (V₀^f) and effective radius (r_eff^f) are most affected. Evidently, the polarization direction has great value to improve the retrieval of aerosol. Among all the aerosol parameters, the posterior error of the real part of complex refractive index (m_r^f) is the smallest, and the imaginary part (m_i^f) has the highest inversion uncertainty.