Infrared (IR) and Raman spectroscopic features of silicate glasses are often interpreted based on the analogy with those of smaller molecules, molecular clusters, or crystalline counterparts; this study tests the accuracy and validity of these widely cited peak assignment schemes by comparing vibrational spectral features with bond parameters of the glass network created by molecular dynamics (MD) simulations. A series of sodium silicate glasses with compositions of [Na2O](x)[Al2O3](2)[SiO2](98-)(x) with x = 7, 12, 17, and 22 were synthesized and analyzed with IR and Raman. A silica glass substrate and a crystalline quartz were also analyzed for comparison. Glass structures with the same compositions were generated with MD simulations using three types of potentials: fixed partial charge pairwise (Teter), partial diffuse charge potential (MGFF), and bond order-based charge transfer potential (ReaxFF). The comparison of simulated and experimental IR spectra showed that, among these three potentials tested, ReaxFF reproduces the concentration dependence of spectral features closest to the experimentally observed trend. Thus, the bond length and angle distributions as well as Si-Q(n) species and ring size distributions of silica and sodium silicate glasses were obtained from ReaxFF-MD simulations and further compared with the peak assignment or deconvolution schemes-which have been widely used since 1970s and 1980s-(a) correlation between the IR peak position in the Si-O stretch region (1050-1120 cm(-1)) and the Si-O-Si bond angle; (b) deconvolution of the Raman bands in the Si-O stretch region with the Q(n) speciation; and (c) assignment of the Raman bands in the 420-600 cm(-1) region to the bending modes of (SiO)(n) rings with different sizes (typically, n = 3-6). The comparisons showed that none of these widely used methods is congruent with the bond parameters or structures of silicate glass networks produced via ReaxFF-MD simulations. This finding invokes that the adequacy of these spectral interpretation methods must be questioned. Alternative interpretations are proposed, which are to be tested independently in future studies.