Approaches to understanding infectious disease dynamics often emphasise studies of multiple hosts or pathogens, but for many diseases, population structure may play a more important role. Changes in population structure may stem from heterogeneity of infection risk in the host population, pathogen polymorphism, spatial structure, or host population size structure. We review previous work demonstrating effects of population structure on baculoviruses of two insects in North America, focusing on heterogeneity in infection risk, which has consequences for both single epizootics and long-term host–pathogen population cycles. Baculoviruses’ simple biology and their insect hosts’ small size means that insect–baculovirus interactions provide experimentally tractable systems for testing models. Our research has combined mechanistic models with data using a combination of statistical model selection, Bayesian statistics, and time-series probes, allowing us to show how host variation affects disease dynamics and pathogen coexistence, and how variation is affected by host-plant resource quality, climate change, and host population size structure. We also discuss how Bayesian mixture models can make it possible to combine multiple sources of data collected across a range of scales.