This article investigates scientists' commitment to underperforming research projects based on the concomitant consideration of their past success and social environments. Based on escalation of commitment and network theory, the model hypothesizes that past success triggers the commitment to underperforming projects but that the strength of this influence varies depending on the characteristics of decision makers' social networks. Results from the analysis of 3,072 scenario assessments nested within 96 scientists show that the positive relationship between past success and continued investment in underperforming projects is more positive when the network is larger, when the ties within the network are stronger, and when feedback from network partners is predominantly positive. Surprisingly and contrary to model predictions, results also show that the relationship between past success and scientists' tendency to commit to underperforming projects becomes stronger with lower communication frequency with network partners. This study extends current research by exploring the boundary conditions of the impact of decision makers' social environment on commitment to failing projects. Further, it adds to literature on the downside of success by emphasizing that decision makers, particularly those in some social environments, are driven to commit additional resources to underperforming -and potentially failingprojects. Decision makers acting in such environments should be aware that they are prone to overinvestment of resources, and the findings of this study can help them increase their awareness. Based on this study's results, decision makers (including scientists) can thus better reflect on and improve their research project evaluations. Finally, the findings of this study open up various opportunities for future research.