Social network graphs are often used to help informjudgments in a variety of domains, such as public health, lawenforcement, and political science. Across two studies, weexamined how graph features influenced probabilisticjudgments in graph-based social network analysis andidentified multiple heuristics that participants used to informthese judgments. Study 1 demonstrated that participants’judgments were influenced by information about directconnections, base rates, and layout proximity, andparticipants’ self-reported strategies also reflected use of thisinformation. Study 2 replicated findings from Study 1 andprovided additional insight into the hierarchical ordering ofthese strategies and the decision process underlyingjudgments from social network graphs.