Several building codes, such as the National Building Code of Canada, recommend that an effective stress ground response analysis be performed if a liquefiable stratum is identified within a soil profile. While constitutive models for total stress ground response analysis have been well verified against earthquake recordings, existing models for effective stress ground response analysis have yet to be thoroughly validated. This project compiles a dataset of seven sites that present a potential of liquefaction, that can be used to validate the predictions from simplified nonlinear 1D effective stress ground response analysis models. The dataset is composed of seven sites, five downhole seismic arrays and two centrifuge tests, and 98 ground motion recordings. The sites and ground motion records are selected to represent a broad range of soil properties, ground motion intensities, and excess-pore pressure generation levels. Five effective stress models are selected to model the sites and study the predictive capabilities of effective stress analyses using the software DEEPSOIL. The predicted and measured motions are compared in terms of spectral response and ratio of surface to downhole acceleration (i.e. amplification factor) spectra. The available models show the ability to capture liquefaction when it was triggered, but the results indicate that existing models are limited in their predictive capabilities of the surface spectra and highlight the need to address the current limitations of these models. In fact, total stress simulations were found to be at least as precise as the effective stress simulations.