The risk of predation can drive trophic cascades by causing prey to engage in antipredator behavior (e.g. reduced feeding), but these behaviors can be energetically costly for prey. The effects of predation risk on prey (nonconsumptive effects, NCEs) and emergent indirect effects on basal resources should therefore depend on the ecological context (e.g. resource abundance, prey state) in which prey manage growth/predation risk tradeoffs. Despite an abundance of behavioral research and theory examining state-dependent responses to risk, there is a lack of empirical data on state-dependent NCEs and their impact on community-level processes. We used a rocky intertidal food chain to test model predictions for how resources levels and prey state (age/size) shape the magnitude of NCEs. Risk cues from predatory crabs (Carcinus maenas) caused juvenile and sub-adult snails (Nucella lapillus) to increase their use of refuge habitats and decrease their growth and per capita foraging rates on barnacles (Semibalanus balanoides). Increasing resource levels (high barnacle density) and prey state (sub-adults) enhanced the strength of NCEs. Our results support predictions that NCEs will be stronger in resource-rich systems that enhance prey state and suggest that the demographic composition of prey populations will influence the role of NCEs in trophic cascades. Contrary to theory, however, we found that resources and prey state had little to no effect on snails in the presence of predation risk. Rather, increases in NCE strength arose because of the strong positive effects of resources and prey state on prey foraging rates in the absence of risk. Hence, a common approach to estimating NCE strength – integrating measurements of prey traits with and without predation risk into a single metric – may mask the underlying mechanisms driving variation in the strength and relative importance of NCEs in ecological communities.