Gastrointestinal nematodes (GIN) have enormous global impacts in humans, wildlife and grazing livestock. Within grazing livestock, sheep are of particular global importance and the economics and sustainability of sheep production are greatly constrained by GIN infections. Natural infections are composed of co-infections with multiple species, and while some past work suggests species can interact negatively with one another within the same host, there is wide variation in reported patterns. Here, we undertook a systematic literature search and meta-analysis of experimental GIN co-infections of sheep to determine whether these experimental studies support the hypothesis of antagonistic interactions between different co-infecting GIN, and test whether aspects of parasite biology or experimental design influence the observed effects. A systematic search of the literature yielded 4848 studies, within which, we identified 19 experimental sheep studies comparing post-mortem worm counts across two co-infecting GIN species. Meta-analysis of 67 effects obtained from these studies provides strong evidence for interactions between GIN species. There was wide variation in the strength and direction of these interactions, but the global effect was significantly antagonistic. On average, there was a decrease in the number of worms of one species when a co-infecting species was also present, relative to a mono-infection with that species alone. This effect was dependent on the infectious dose and was rapidly lost after anthelmintic treatment, suggesting that live worms are required for the effect to occur. Individual parasite species varied in the extent to which they both exerted, and were subject to, these interspecies interactions, and these differences are more complex than simply co-localisation within the gastrointestinal tract. Antagonistic interactions between co-infecting GIN may feedback into their epidemiology as well as potentially affecting the clinical impacts of infection. Furthermore, the consequences of these interactions may be heightened when clinical interventions affect only one species within the co-infecting network. Whilst it was not possible to identify the causes of variation between GIN species in the impact of co-infection, these findings point to new avenues for epidemiological, clinical and mechanistic research on GIN co-infections.