We develop an empirical site‐amplification model for central and eastern North America (CENA) that is parameterized using two site variables: (1) the peak site frequency () and (2) the time‐averaged shear‐wave velocity in the upper 30 m (). The model is developed by first deriving empirical site‐amplification terms for seismographic sites in CENA. This is done by analyzing the residuals calculated from ground‐motion amplitude data with respect to a selected regional ground‐motion prediction equation that applies to hard‐rock reference site conditions. An amplification model is then developed that explains the residuals in terms of the two site variables ( and ). We consider two alternative cases. In the first case, we assume that is the main modeling parameter and then model any remaining residual effects in terms of . In this case, we obtain a frequency‐dependent scaling term that is similar in form to that obtained in previous studies for sites in western North America (WNA). However, the scaling term is less significant in amplitude for CENA in comparison to that for WNA, suggesting that is not as strongly indicative of site response. For the second alternative, assuming that is the main site‐effects parameter, a frequency‐independent scaling term is obtained for CENA, which is much smaller in amplitude compared to the scaling effect derived in the first approach. This shows that, using as the primary site‐effects modeling parameter, we remove most of the scaling effects that are implied by the data. Finally, we provide recommendations on the effective use of and to model site effects in CENA, differentiating between glaciated and nonglaciated sites. Glaciated sites show larger amplifications compared to nonglaciated sites, especially at intermediate‐to‐high frequencies, presumably due to the strong impedance contrast at the base of glaciated sites.