The goal of this study was to evaluate the performance of different splitting measurement techniques in the particularly complicated tectonic setting of subduction beneath Japan. We use data from the broadband Japanese F-net array and consider the methods of Silver and Chan (1991), Levin et al. (1999), and Chevrot (2000). We find that the results generally agree well, although discrepancies arise if the anisotropy beneath the station is more complex than the simple single-layer anisotropic model often assumed in splitting studies. A combination of multichannel and single-record methods may serve as a powerful tool for recognizing complexities and for characterizing upper-mantle anisotropy beneath a station.