Bayesian Framework for Moderated Mediation Using Covid-19-caused Natural Experiment: Modeling Home Advantage in Soccer

Observational studies are being used more and more in psychology and medicine since they provide a wealth of data for real-world issues. Their biggest drawback is the lack of falsification due to the control mechanisms of control conditions being unavailable. However, the Covid-19 pandemic and the isolation policies related to it have provided an environment in which researchers can use natural experimental design to establish causal pathways in phenomena. Here we demonstrate how Covid-19-related changes can be used to investigate causal effects behind Home Advantage (HA), a robust phenomenon in which sport teams are more successful when they play in front of their fans. HA theories assume that the crowd support spurs home players to better performance and biases referees, and that these two factors in turn influence the result. Covid-19 has provided the perfect control condition for disentangling the causal links of the HA as sport teams have played at home but without the presence of fans. Using our newly developed Home Advantage Mediated (HAM) model, which considers all individual factors and their interrelations simultaneously instead of in isolation as was previously the case, we demonstrate how Covid-19 enables us to disentangle the processes behind the HA phenomenon. Besides throwing new (modeling) light on one of the most robust phenomena in sport, our paper also provides information about the practical implementation of mediation and moderated mediation mixed-effects models in the Bayesian framework. Similar implementations can be adapted in other medical and social science fields.