Situational Appraisals and Criminal Violence: Lab-in-the-field Evidence from Kenya


How do people make decisions about their safety when they are imminently threatened by violence? In this poster, I design and field novel experimental manipulations in a large lab-in-the field omnibus study in Kenya to show how individual perceptions shape decision-making during simulated violence. Understanding decision-making during violence has implications for a wide range of political processes. Individual decisions form the micro-foundations of phenomena ranging from refugee crises to mass riots, and civilian resilience during political conflicts. Understanding individual decisions also sheds new light on questions in international security, like combat motivations in armies, and the dynamics of deterrence and escalation. Though important, these decisions are difficult to study. In the real world, judgments and decisions about physical safety are difficult to measure rigorously. In controlled settings, there are clear ethical (and practical) impediments to re-creating “realistic” life-or-death choices. To create realistic-within-reason decision scenarios, I field a large lab-in-the-field experiment (n = 1, 506) in Machakos County, Kenya that asks participants to make decisions about hypothetical violence after playing an economically-incentivized video game. Participants are randomly assigned to variants of the game that have higher or lower difficulty—to boost or diminish perceptions of control and agency—and clearer or more opaque information about the gameplay—to boost or diminish perceptions of uncertainty. To enhance emotional realism, all participants are further exposed to audio recordings validated to stimulate fear. After this novel experimental manipulation, participants choose strategies for the protagonists in two audio vignettes that describe instances of violence that are common in Machakos county, criminal mugging and domestic violence. To ensure that violence scenarios were realistic and relevant, vignettes were designed with input from focus groups drawn from the population of potential participants. Results show that manipulating incidental perceptions of control and uncertainty has substantial effects on strategy preferences. Participants primed to feel more “in control” are up to 7 percentage points more likely to respond to violence by approaching the source of the threat (either to attack it or bargain with it). Participants primed to feel that the future is more certain—through the game and through random exposure to more or less certain information earlier in the omnibus—are up to 32 percentage points more likely to respond to violence with conservative, less disruptive strategies (like hiding from danger, bargaining, or ignoring threats). Exploratory analyses of mechanisms suggest that the incidental treatments work by changing the assumptions participants make about relative power and uncertainty when interpreting the violence vignettes. Further analysis shows that treatment effects are moderated somewhat by personality traits, and by recent engagement in social decision-making processes, but not substantially by demographic or socioeconomic factors. This study contributes to research on decision-making about violence by showing that incidental appraisals of control and uncertainty shape decision-making, and also by showing that important appraisals can be manipulated; they are not simply derived from personality traits or socioeconomic millieu. The study also contributes to applied methods research on manipulating perceptions in experimental settings in the Global South, by providing a new, engaging, and easy-to-deploy way to manipulate incidental perceptions for study populations with varying levels of literacy.

PolMeth XL
Aidan Milliff
Aidan Milliff
Assistant Professor