We simulated diffusion of behaviour change over fifteen real-world networks with seven network interventions under both simple and complex contagion. We found that structural network properties affect both the diffusion outcome and the relative effectiveness of the different interventions, with confounding effects that were inconsistent with results expected from mathematical analysis. These results suggest that comprehensive studies are needed to identify the effects of structural properties on diffusion in real-world networks. Further, researchers attempting to identify the effect of individual properties must measure a range of properties to avoid incorrect attribution.

我们模拟了行为变化在十五个现实世界网络中的传播情况，并在简单和复杂的传染下采用了七种网络干预措施。我们发现结构网络的性质会影响扩散结果和不同干预措施的相对有效性，其混杂效应与数学分析预期的结果不一致。这些结果表明，需要进行综合研究来确定结构特性对现实网络中扩散的影响。此外，试图确定单个属性影响的研究人员必须测量一系列属性，以避免错误的归因。