We evaluated a quantitative amplitude analysis of 3D physical model reflection data acquired over an experimental phenolic layer that modeled a fractured medium with one set of vertical fractures. The phenolic layer was overlain by two isotropic layers, the uppermost being water, and the data acquisition was designed to avoid the interference of the primary and ghost events. The elastic stiffness coefficients and hence the anisotropy of the phenolic layer were known in advance from a previous traveltime analysis. The reflection amplitudes from the top of the phenolic layer required corrections to make them suitable for an amplitude study. In addition to the usual amplitude corrections applied to seismic field data, a directivity correction specific to the physical model transducers was applied. The corrected amplitudes along different azimuths showed a clear azimuthal variation caused by the phenolic layer and agreed with amplitudes predicted theoretically. An amplitude variation with angle and azimuth inversion was performed for horizontal transverse isotropy (HTI) parameters of the phenolic layer. We determined from the inversion results that from the azimuthally varying P-wave reflectivity response, it was possible to estimate HTI parameters that compared favorably to those obtained previously by a traveltime analysis. This result made it possible to compute the S-wave splitting parameter (historically determined from S-wave data and directly related to fracture density) from a quantitative analysis of the PP data.