Who belongs inside our moral circle is one of ethics’ oldest questions. A new cross-cultural study suggests that the answer varies less from place to place than might be expected, at least when researchers examine the broad ordering of who and what people feel obliged to protect.
Across 6,561 university students from 41 cultural groups, researchers found substantial agreement in a previously observed hierarchy of moral concern. Family and friends consistently occupied the strongest position. People generally placed members of their own social groups above more distant human groups, while plants and people viewed as serious wrongdoers tended to sit much further down the hierarchy.
Yet the study also uncovered an important boundary to that apparent agreement. The most culturally variable comparisons involved human outgroups, non-human animals and the environment. In other words, cultures appeared to agree more readily about the special moral status of close relationships than about how humans, animals and nature should be ranked against one another.
The research, published in the European Journal of Social Psychology, offers a unusually broad empirical look at the structure of the moral circle. It does not tell us who should receive moral consideration. Instead, it asks a descriptive question: when people distribute moral concern across very different kinds of targets, does a common pattern emerge across cultures?
Testing whether a Western moral pattern travels
Previous research using the Moral Expansiveness Scale had suggested a relatively stable hierarchy. People tend to report the greatest moral obligation toward family and friends, followed by members of their own groups, then more distant people, animals and environmental targets, with serious wrongdoers often receiving the least concern.
The problem was that much of the evidence behind this structure came from the United States and other Anglosphere countries. A pattern can look universal when the populations used to discover it share substantial cultural history, institutions and values.
Işıl Ayça Akkuş and colleagues therefore re-analysed a large multi-lab dataset collected in 2018 and 2019. Their final sample contained 6,561 students recruited through 41 universities spanning countries and culturally distinct regions in Africa, Asia, Europe, Oceania and the Americas. South Africa was among the samples.
Participants were young overall, with a mean age of 21.60 years. About 63.37% were female and 29.72% male. The sample also leaned somewhat left politically, with a mean political-orientation score of 3.26 on a seven-point scale running from left or liberal to right or conservative.
That composition matters. This is a large international sample, but it is not a representative survey of 41 national populations. The researchers deliberately describe the units as cultural groups rather than treating every university sample as a miniature version of its country.
Thirty targets mapped onto a moral circle
Each participant completed the Moral Expansiveness Scale. The measure presents moral concern as concentric circles ranging from the highest level of moral obligation to being outside the moral boundary altogether.
Participants rated 30 targets on a four-point scale from 0 to 3. These targets were grouped into 10 broader categories. They included family and friends, ingroup members such as neighbours and co-workers, revered figures, stigmatized groups, human outgroups, high-sentience animals, environmental targets, lower-sentience animals, plants and villains.
The researchers did not simply compare every possible average with every other average. That would have produced tens of thousands of potential comparisons and would also have been difficult to justify because the scale did not show scalar measurement invariance across cultures. Instead, they focused on 13 relative rankings that had appeared consistently in earlier research.
For example, did a cultural group place family and friends above ingroup members? Did it place the environment above plants? Did it place human outgroups above high-sentience animals? Each of these comparisons could either reproduce the previously observed ordering or depart from it.
Eleven of 13 rankings typically survived the cultural test
The broad pattern was strikingly stable. Across the 41 cultural groups, the median group reproduced 11 of the 13 rankings from the earlier hierarchy. The mean was 11.32, with a standard deviation of 1.44.
Twelve cultural groups reproduced all 13 rankings. These included several Anglosphere samples, as might be expected given where much of the earlier research originated, but also Belgium, Japan, the Netherlands, Portugal, Singapore and Ukraine.
Even the most dissimilar group was not a wholesale exception. Malaysia reproduced eight of the 13 rankings, while French-speaking Canada, Chile, Germany and Poland each reproduced nine.
The endpoints of the moral hierarchy were particularly clear. Family and friends received an average moral-concern score of 2.89 out of 3 across cultural groups. Villains averaged 0.58. Villains were the lowest-rated group in all but three cultural groups: Hong Kong, Thailand and Uganda.
This does not mean people everywhere held identical moral values. A ranking only tells us which of two categories came first, not that cultures assigned them identical absolute levels of concern. The distinction is especially important because the measurement-invariance analysis did not support straightforward comparisons of mean scale levels across countries.
The strongest agreement appeared around close relationships
Three of the 13 relative rankings appeared in every one of the 41 cultural groups. Family and friends were prioritized over ingroup targets. The environment was prioritized over low-sentience animals. The environment was also prioritized over plants.
At the individual level, 91% of participants placed family and friends above ingroup members. This is one of the clearest findings in the study and fits a much broader literature showing that close relationships carry special perceived obligations.
But agreement weakened once the comparison moved away from close human relationships. Only 21 of the 41 cultural groups placed human outgroups above the environment. Human outgroups were prioritized over high-sentience animals in 26 groups.
The multilevel analysis sharpened this point. Of the variance in whether a particular ranking reproduced the earlier pattern, 40.73% was associated with which two target groups were being compared. By contrast, 13.36% was associated with which cultural group was being examined.
That result shifts the interpretation. The biggest source of disagreement was not simply that one country had one moral hierarchy and another country had a completely different one. Rather, some moral comparisons were inherently much more variable across cultures than others.
Animals and nature exposed the fault lines
The individual-level results make those fault lines even clearer. While aggregated cultural patterns looked relatively orderly, the average individual reproduced only 6.85 of the 13 typical rankings, with a median of seven.
Only 38% of participants prioritized human outgroups over the environment. Just 32% prioritized the study’s high-sentience animals over its low-sentience animals. The researchers caution that these labels may hide other considerations. Bees, for example, may receive concern because of their perceived ecological usefulness, while attitudes toward cows may be shaped by meat consumption.
Approximately half of participants prioritized high-sentience animals over human outgroup targets, and at the cultural-group level high-sentience animals outranked human outgroups in roughly one-third of samples.
This is where the study becomes especially relevant to contemporary ethical disputes. Arguments about animal welfare, environmental protection, refugees and distant populations often involve precisely these cross-category trade-offs. The research suggests that broad agreement about caring most for those closest to us can coexist with substantial disagreement over what happens further out in the moral circle.
Cultural distance from the United States mattered, but only partly
For 26 of the 41 cultural groups, the researchers could match their data to an existing multidimensional measure of cultural distance from the United States. Greater cultural distance was associated with a less similar moral hierarchy, with a correlation of r = -0.46 and p = 0.017.
That is meaningful evidence that culture contributes to the remaining differences, but it should not be read as proof that distance from the United States causes a different moral hierarchy. The cultural-distance index combines many dimensions, and some overlap conceptually with moral concern, including attitudes toward family and environmental protection.
The finding is better understood as a clue. The original hierarchy was developed largely from US and Anglosphere data, and the further a group was culturally from that reference point, the more likely some rankings were to differ.
A universal moral circle would be too strong a conclusion
The scale showed configural invariance for most target groups, but evidence for metric invariance was mixed and scalar invariance was not supported. This is a technical limitation with an important practical meaning: the same numerical score cannot automatically be assumed to carry exactly the same measurement properties in every cultural context.
The researchers therefore designed their main analyses around within-person and within-group rankings rather than claiming that, for example, one country’s mean moral concern was definitively higher than another’s.
The sample creates another important boundary. University students are younger, often more educated, and in this dataset likely more liberal and socioeconomically advantaged than their national populations. Shared features of university populations may themselves make the 41 groups look more similar. Representative national samples could reveal larger differences.
Individual-level measurement was also noisy. A multilevel model designed to explain individual variation accounted for only 15% of the variance. Each broad target group was represented by only three specific examples, meaning some apparent individual differences may reflect measurement noise or reactions to those particular examples rather than stable moral priorities.
Some labels may also travel poorly across cultures. Categories such as “stigmatized groups” were built from specific targets, including refugees and LGBT people, whose social position and cultural meaning can differ substantially between societies. The authors explicitly call for more work on cross-culturally meaningful target groupings.
What the study changes about the moral-circle debate
The strongest contribution of the study may be its refusal to force a choice between universality and cultural difference. Both appeared in the data.
There was a recognisable cross-cultural structure. Close relationships occupied a privileged moral position with remarkable consistency, and most cultural groups reproduced most of the rankings established in earlier work. At the same time, the outer portions of the moral circle were far less settled, especially when people weighed distant humans against animals or environmental entities.
That distinction matters for how moral disagreement is understood. Two societies may share the same broad intuition that family deserves exceptional concern while still disagreeing sharply about whether a coral reef, a cow, a foreign citizen or a socially condemned person deserves priority in a particular comparison.
The study is explicitly exploratory and hypothesis-generating rather than a preregistered test of a universal theory. Its value lies in mapping where moral hierarchies appear stable and where they fracture. Future research can now test why those specific fault lines emerge, ideally using representative samples and measures built to function comparably across cultures.
Source Information
Study: Exploring Cultural and Individual Differences in the Hierarchy of Moral Concern: Evidence From Student Samples Across 41 Cultural Groups
Authors: Işıl Ayça Akkuş, Charlie R. Crimston, Kelly Kirkland, Brock Bastian and Bastian Jaeger
Journal: European Journal of Social Psychology
First published: 14 September 2026
DOI: 10.1002/ejsp.70114
