Understanding the ecological assembly of parasite communities is critical to characterise how changing host and environmental landscapes will alter infection dynamics and outcomes. However, studies frequently assume that (a) closely related parasite species or those with identical life-history strategies are functionally equivalent, and (b) the same factors will drive infection dynamics for a single parasite across multiple host species, oversimplifying community assembly patterns. Here, we challenge these two assumptions using a naturally occurring host–parasite system, with the mussel Anodonta anatina infected by the digenean trematode Echinoparyphium recurvatum, and the snail Viviparus viviparus infected by both E. recurvatum and Echinostoma sp. By analysing the impact of temporal parasite dispersal, host species and size, and the impact of coinfection (moving from broader environmental factors to within-host dynamics), we show that neither assumption holds true, but at different ecological scales. The assumption that closely related parasites can be functionally grouped is challenged when considering dispersal to the host (i.e. larger scales), while the assumption that the same factors will drive infection dynamics for a single parasite across multiple host species is challenged when considering within-host interspecific competition (i.e. smaller scales). Our results demonstrate that host identity, parasite identity and ecological scale require simultaneous consideration in studies of parasite community composition and transmission.