Motivated Causal Reasoning and Responsibility for Civilian Casualties in Military Conflicts

Causal judgments are ubiquitous in politics and are crucial for assigning responsibility and blame. But causal judgments are also tricky, especially when complex causal structures are involved, such as joint production, omissive causation and double prevention. Cognitive science has demonstrated that people are more likely to pick factors as ‘causal’ when they make a difference for the outcome across a range of counterfactual scenarios, with the scenarios sampled based on their statistical and prescriptive normality. We propose that this makes causal judgements susceptible to motivated reasoning, and ingroup favouritism in particular. We hypothesize that people will be less likely to assign causal efficacy and responsibility for counternormative outcome to groups they support, and that the bias will be greater for more complex causal structures. We test these propositions in a pre-registered survey experiment run on quota-based representative sample in Poland. The context of the experimental vignettes is a military conflict between Russia and Ukraine. We find that in all scenarios respondents assign significantly higher causal power and responsibility to the attackers when the attackers are Russian rather than Ukrainian, consistent with our theory and the very high levels of public support for Ukraine in Poland. Contrary to our expectations, responsibility of the attackers is not significantly lower when they hit a public building as a result of defending combatants moving there rather than when unprovoked. Unexpectedly, in a scenario where the civilians get killed after moving to where the defenders are, the defenders get a higher share of the blame when they are identified as Ukrainian rather than Russian. This is consistent with people holding their favored party to higher moral standards for protecting civilians.