Shale anisotropy must be quantified to obtain reliable information about reservoir fluids, lithology, and pore pressure from seismic data. Failure to account for anisotropy in seismic processing can lead to errors in normal moveout (NMO), dip moveout (DMO), migration, time-to-depth conversion, and amplitude variation with offset (AVO) analysis. Kriged estimates of density, vertical compressional-wave, and shear-wave velocities were derived from vertical wells in an area of interest in the Horn River resource play in northeastern British Columbia. The kriged predictions were compared with measured logs along the trajectory of a deviated well. Whereas the density comparison provided a blind test of kriging accuracy because density is scalar and independent of well deviation, the same comparison for sonic velocities revealed that they are systematically higher than the kriged vertical velocities. This difference was used to estimate anisotropy parameters at the location of the deviated well. The fact that the higher velocities observed are caused by anisotropy was confirmed subsequently by using the derived anisotropy parameters to apply a nonhyperbolic moveout correction η to flatten gathers from a seismic survey 10 km north of the area of interest, within which the anisotropy parameters were estimated.