Abstract
In human-autonomy teaming, calibrated trust is key to delivering assured performance. If trust is not well calibrated, interventions can be used to return the team's trust to a calibrated state. In this chapter, we propose a causal analysis approach to understanding the effects of teamwork interventions. We provide a model scenario for human-autonomy teaming involving an autonomy-supported navigation and threat identification task and introduce causal modeling and two of its common approaches: Bayesian networks and structural equation modeling. Following that, we construct a Bayesian network based on our model scenario, highlighting the ability of the causal model to reveal the importance of various factors in the human-autonomy teaming intervention. This chapter highlights the importance of more completely characterizing and analyzing teamwork and trust-related factors when designing and implementing interventions and presents the causal modeling approach as a useful way forward for human-autonomy teaming research.