Philosophy · 6 min read
What Cross-Cultural Experiments Show About Sacrificial Moral Judgments
Recurring and variable moral judgments across countries, and why survey patterns do not settle moral truth.

What Cross-Cultural Experiments Show About Sacrificial Moral Judgments
In a sacrificial moral dilemma, a participant is asked whether harming one person is acceptable if doing so would save several others. The best-known cases involve a runaway trolley: redirecting it onto a track where it will kill one person, or pushing a person from a bridge to stop it before it reaches five. Responses are often described as moral intuitions.
These experiments can reveal patterns in reported judgments. They do not reproduce emergency behavior, and they do not determine which action is morally right. A survey response is evidence about how a participant understood and evaluated a vignette under particular conditions. Philosophers disagree about how much weight such responses should carry in normative argument.
Cross-cultural work addresses an additional problem. Many early studies recruited university students in affluent Western countries. If a pattern recurs across countries and languages, a theory that treats it as a local convention becomes less plausible. If it varies, a theory built from one sample needs qualification. Neither outcome by itself identifies the psychological cause.
Three distinctions recur in many samples
A large study by Edmond Awad and colleagues recruited about 70,000 participants in 42 countries. Published in Proceedings of the National Academy of Sciences in 2020, it presented sacrificial dilemmas designed to vary three factors. Was harm produced by action or inaction? Did physical force from the decision-maker directly affect the victim? Was the victim's harm used as a means to save others or merely foreseen as a side effect?
Across the combined sample, participants were generally less willing to endorse intervention when it required action, personal force, or using the victim as a means. These broad tendencies resemble distinctions long discussed in moral philosophy, including the act-omission distinction and the doctrine of double effect. The study did not show that participants consciously applied those doctrines. Similar response patterns can arise from emotion, learned rules, causal understanding, social expectations, or the wording of the cases.
The interactions were less uniform than the main tendencies. The combined effect of personal force and instrumental harm appeared especially strong in Western countries. Country-level variation was systematic enough to resist a simple universal-versus-cultural binary. Some distinctions recurred widely, while their relative weights and combinations varied.
An earlier web-based study led by Marc Hauser collected thousands of responses from participants in more than one hundred countries, though the geographic distribution and self-selection limited national comparisons. It also reported broad similarities in judgments across demographic groups. That result helped motivate claims of a shared moral faculty, but later work has shown that a recurring average can coexist with meaningful variation in framing, sample composition, and interpretation.
Sampling across countries does not automatically represent cultures
Country labels are rough proxies. People within a country differ by language, region, religion, class, education, age, and political institutions. Online volunteers are more likely to be literate, connected, and interested in the task than the population around them. Translating a vignette can preserve literal content while changing connotation or perceived realism.
The 42-country study used a large sample and a common design, which helps comparison. It was not a probability survey of humanity. The authors reported country-level associations with measures including relational mobility and individualism, but ecological correlations are not individual causes. If countries with one social indicator differ in average judgments, it does not follow that the indicator explains a particular person's answer.
The category “Western” also compresses differences. Joseph Henrich, Steven Heine, and Ara Norenzayan coined the acronym WEIRD—Western, educated, industrialized, rich, and democratic—to criticize behavioral science's dependence on unusually narrow populations. The critique was not that results from such samples are false. It was that their generality must be tested rather than assumed. Cross-national trolley studies are one response, but they still draw heavily on participants reachable through common research platforms.
Vignettes measure interpretation as well as evaluation
Sacrificial dilemmas are deliberately sparse. That makes variables easier to manipulate, but participants may fill gaps differently. Is the person on the bridge large enough to stop the trolley? Does the agent have legal responsibility? Is there time to search for another option? Does refusing to act count as allowing harm or avoiding an uncertain intervention? A researcher can specify these facts, yet increasing detail can introduce new ambiguities.
Order matters in some experiments. Seeing an extreme case first can change how a later case is understood. Wording such as “kill,” “sacrifice,” or “save” can shift attention. Asking what is “acceptable,” “permissible,” or what the participant “would do” poses different questions. A response-time limit may favor a quick aversion, while deliberation may let participants calculate outcomes or infer the experimenter's purpose.
Real behavior brings consequences, fear, uncertainty, and accountability that text lacks. It would be unethical to create an actual lethal choice. Virtual reality and nonlethal simulations can add sensory immediacy, but they remain models. The appropriate conclusion is not that vignettes are useless. It is that they isolate judgment under description, not conduct under danger.
The philosophical disagreement concerns what follows
One view treats considered intuitions as evidence for or against moral principles. If people judge redirecting the trolley differently from pushing a person, a theory should explain the distinction or show why it is mistaken. Experimental evidence can then reveal whether the intuition is stable, widely shared, or sensitive to irrelevant factors.
A skeptical view argues that framing, order, emotion, and demographic variation undermine confidence in intuition as a reliable guide. Yet variability alone does not show that a judgment is wrong. Perceptual judgments also vary under poor conditions without making vision worthless. The normative task is to decide whether the source of variation is relevant to the moral question.
A third approach uses experiments diagnostically. Instead of trying to validate or discredit intuition as a category, it asks which psychological processes generate particular judgments. Fiery Cushman, Liane Young, and Joshua Greene, among others, have distinguished evaluation of outcomes from evaluation of actions and intentions. Such models can explain patterns, but moving from a causal explanation to a moral conclusion requires an additional argument. Knowing why a person judges an act wrong does not establish whether it is wrong.
The Stanford Encyclopedia of Philosophy's review of experimental moral philosophy emphasizes this division of labor. Experiments can test descriptive premises embedded in philosophical arguments. They cannot replace conceptual analysis or justify a moral principle by counting votes.
A bounded contribution to moral inquiry
The strongest cross-cultural result is neither uniformity nor radical difference. Large studies find recurring aversion to action, direct force, and using a person as a means, alongside variation in how those features combine and how strongly they matter. That finding constrains theories claiming that one response pattern is either plainly universal or merely peculiar to one classroom sample.
What remains contested is the source of the patterns and their normative relevance. Shared judgments might reflect common cognitive architecture, widespread social learning, common features of the vignettes, or several processes that converge on the same answer. Variation might expose bias, track morally relevant social experience, or result from translation and sampling.
Cross-cultural experiments make the empirical question better. They show who answered which case under stated conditions. Philosophy must still argue about what features should matter, how competing reasons are weighed, and when an intuition deserves revision. The data can discipline that argument without settling it.
Sources consulted
4 sources, every one with a link you can follow.
- Edmond Awad and colleagues, “Universals and Variations in Moral Decisions Made in 42 Countries by 70,000 Participants,” PNAS (2020).
- Joseph Henrich, Steven Heine, and Ara Norenzayan, “The Weirdest People in the World?,” Behavioral and Brain Sciences (2010).
- Stanford Encyclopedia of Philosophy, “Experimental Moral Philosophy,” substantive review.
- Marc Hauser and colleagues, “A Dissociation Between Moral Judgments and Justifications,” Mind & Language (2007).
