This paper develops a framework to analyze partial population experiments, a generalization of the cluster experimental design where clusters are assigned to different treatment intensities. The framework allows for heterogeneity in cluster sizes and outcome distributions. The paper studies the...