The reliability of effective core potentials (ECPs) for estimating vibrational frequencies of transition metal (TM) complexes is assessed in relation to all-electron methods for main group compounds. Complexes with a multiple bond between a transition metal and chalcogen (O, S, or Se), chalcogenides, are investigated using the Stevens ECP/valence basis set scheme. Statistical treatment of the data indicates that ECPs, in addition to reliably modeling electronic structure, can be successful in estimating vibrational frequencies for TM complexes. As expected, theoretical prediction of vibrational data is not as accurate as the prediction of metric data for chalcogenides. However, agreement with experiment is still very good at the Hartree-Fock level of theory and is in even better accord upon the use of simple corrections to model well-known computational deficiencies (e.g., the neglect of anharmonic effects). Analysis of the data show interesting differences in predictive ability for first row transition metals versus second- and third-row analogues and oxo complexes versus their congeners with heavier chalcogens.