Prey availability is one of the principal drivers of tiger distribution and abundance. Therefore, formulating effective conservation strategies requires a clear understanding of tiger diet. We used scat analysis in combination with data on the abundance of several prey species to estimate Amur tiger diet and preference at 3 sites in the Russian Far East. We also examined the effect of pseudoreplication on estimates of tiger diet. We collected 770 scats across the 3 sites. Similar to previous studies, we found that tigers primarily preyed on medium to large ungulates, with wild boar, roe, sika and red deer collectively comprising 86.7% of total biomass consumed on average. According to Jacobs' index, tigers preferred wild boar, and avoided sika deer. Variation in preference indices derived from these scat analyses compared to indices derived from kill data appear to be due to adjustments in biomass intake when sex-age of a killed individual is known: a component missing from scat data. Pseudoreplication (multiple samples collected from a single kill site) also skewed results derived from scat analyses. Scat analysis still appears useful in providing insight into the diets of carnivores when the full spectrum of prey species needs to be identified, or when sample sizes from kill data are not sufficient. When sample sizes of kill data are large (as is now possible with GPS-collared animals), kill data adjusted by sex-age categories probably provides the most accurate estimates of prey biomass composition. Our results provide further confirmation of the centrality of medium ungulates, in particular wild boar, to Amur tiger diet, and suggest that the protection of this group of species is critical to Amur tiger conservation.