Is it true that people believe in divination more when times are hard?
Short answer: there is fairly solid correlational evidence that people turn to divination and ritual more when anxious, but the proposed mechanism, that inducing loss of control produces superstition, has not held up in experiments. These two tiers of evidence are often merged into one, and that merging is where the controversy comes from. The observational tier is about what happens in life; the experimental tier is about what happens inside people's heads. The first is fairly certain; the second remains open.
One thing about tone needs saying right away: turning to ritual in hard times is not a sign of weakness. Much research in psychology and anthropology treats it as a coping strategy, a functional behaviour under uncertainty, not a simple cognitive defect.
Malinowski and Trobriand fishermen: magic when the open sea is dangerous
Bronislaw Malinowski, in the essay Magic, Science and Religion in the collection Science, Religion and Reality edited by Joseph Needham, Macmillan, New York, 1925, recorded an observation that has become classic: fishermen in the Trobriand Islands performed rites and magic when fishing in dangerous open waters, but did not when fishing in the calm lagoon. He placed that observation within a broader argument: magic appears where technique and knowledge are insufficient to control outcomes.
It must be clear what this is and is not. It is an ethnographic observation in a specific community, recorded in a specific period, not a psychological law applying to everyone everywhere. Malinowski did not measure superstition levels, did not compare control groups, and did not conclude about individuals. He described a behavioural contrast within one community, between two kinds of trips with different risk levels. That is precisely why it is a starting point for a hypothesis, not evidence closing it.
Kind of knowledge: School: MalinowskiSchool's interpretation. How one school or one book reads it; other schools may read it differently. The five labelsA notable conceptual point: Malinowski distinguished magic from religion by purpose. Magic, in his usage, always aims at a specific outcome, while primitive religion expresses states of mind such as sadness, hope, despair without pursuing practical ends. The distinction is not absolute, since worship of spirits also has practical ends, but it helps clarify which kind of behaviour we are discussing. When this article says "turning to divination", it means the magic side, the side aiming at outcomes.
Keinan 1994 and 2002: from missile-hit areas to a knocking-on-wood experiment
Giora Keinan, in a 1994 article in the Journal of Personality and Social Psychology, compared people living in areas hit by missiles with people living in less threatened areas, and found the first group showed higher magical thinking. This is a context-based correlational study: high-anxiety context accompanies higher magical thinking. It does not tell us the direction of causation, nor whether anxiety increases superstition or whether people who are already superstitious tolerate anxiety better.
Keinan himself, in a 2002 article in Personality and Social Psychology Bulletin, investigated knocking-on-wood behaviour and found that stress increased this behaviour in people with a high need for control. This is an important methodological shift: from comparing two groups in two contexts to manipulating stress in the laboratory. But the result must be read within its scope: the effect was reported as stronger in the high need-for-control group, meaning it came with a condition, not as a flat effect for everyone.
Keinan's two studies sit on different tiers. The 1994 study is observational; the 2002 study is experimental. When someone cites "Keinan proved stress increases superstition", they are merging the two. A more accurate statement: one study found a correlation between a threatened context and magical thinking; another found that laboratory stress increased a specific superstitious behaviour in people with a high need for control.
Newspaper archive data: Germany 1918 to 1940 and the US in crisis periods
Vernon R. Padgett and Dale O. Jorgenson, in a 1982 article in Personality and Social Psychology Bulletin, analysed German newspaper content from 1918 to 1940 and argued that the level of economic threat predicted the level of superstition appearing in newspapers. This is a very different approach: not asking living people, not manipulating them, but counting what was written over a long period. Its advantage is covering many decades; its disadvantage is that what is counted is newspaper content, not people's beliefs.
A similar approach has been used for the United States during crisis periods, with the general idea that superstition in newspapers varies with economic cycles. It must be said clearly that this is tier B evidence: someone proposed it and there are results, but it is not a body of results replicated often enough for the research community to consider it closed. Moreover, the data scope must be stated in the sentence itself: Germany 1918 to 1940, the US in specific crisis periods, all historical data from abroad, not data about Vietnamese people.
An economic study on the value of superstition follows this direction too. Research on licence plate auction prices in Hong Kong found that the value attached to numbers considered lucky and numbers considered unlucky varied with the macroeconomic environment, and varied asymmetrically between the lucky and unlucky sides. This result is about the licence plate market in Hong Kong, not about divination behaviour in Vietnam. It is useful as a demonstration that superstitious beliefs can leave traces in economic data, not only in stories.
Why did the "induced loss of control" experiment fail to replicate?
Kind of knowledge: UnprovenNot enough evidence. Someone has claimed it, but no source or study is strong enough to confirm it. The five labelsHere the story turns. Jennifer A. Whitson and Adam D. Galinsky, in a 2008 article in Science, reported that when made to feel a loss of control, participants perceived more illusory patterns in random sequences. This is tier B: someone proposed it and there is a report, but it has not been robustly replicated.
Michiel van Elk and Paul Lodder, in a 2018 article in Collabra: Psychology, ran seven experiments to replicate that effect and did not: a manipulation reducing personal sense of control did not increase perception of illusory patterns. This is tier A in a different sense: a publicly reported replication failure, with multiple independent experiments, and results going against the original expectation. When an effect fails to replicate across many experiments, the correct statement is that the effect has not held up, not that it has been proven false under all conditions.
This does not overturn the observational tier. That people turn to ritual more when anxious still has evidence at the ethnographic and correlational levels. What has been toppled is a specific proposed mechanism to explain that observational tier: that loss of control produces perception of illusory patterns, and that perception of illusory patterns produces superstition. That causal chain may be partly right, may need other conditions, or may not be the main mechanism. Currently, there is no basis to assert it is the main mechanism.
Kind of knowledge: UnprovenNot enough evidence. Someone has claimed it, but no source or study is strong enough to confirm it. The five labelsOne other explanatory branch is worth noting but lacks sufficient evidence to conclude: some authors suggest that a tendency toward superstitious behaviour may have evolutionary roots, because in environments where falsely attributing causation is sometimes cheaper than missing a real danger. This is a theoretical argument and model, not empirical results on humans, so it should be read only as a hypothetical direction.
Comparison table of main studies
| Study | Data scope | Method | Main result | Evidence tier |
|---|---|---|---|---|
| Malinowski 1925 | Trobriand fishermen, recorded early 20th century | Ethnographic observation | Magic when going to dangerous open sea, none when fishing in lagoon | A (classic observation, not experiment) |
| Keinan 1994 | People in missile-hit areas | Group comparison by context | Threatened group had higher magical thinking | A (correlational) |
| Keinan 2002 | Experimental participants | Stress manipulation | Stress increased knocking-on-wood in high need-for-control people | A (experiment, with condition) |
| Padgett and Jorgenson 1982 | German newspapers 1918 to 1940 | Archive content analysis | Economic threat predicted superstition in newspapers | B (not robustly replicated) |
| Whitson and Galinsky 2008 | Experimental participants | Induced loss of control | Reported loss of control increased illusory pattern perception | B (failed to replicate) |
| van Elk and Lodder 2018 | Participants, seven experiments | Replication of loss-of-control manipulation | Effect did not replicate | A (replication failure) |
| Salemink 2010 | Contemporary Vietnam | Anthropological analysis | "Spiritual security" explains turning to mediumship under anxiety | A (specialist research) |
What about Vietnam?
Kind of knowledge: School: SaleminkSchool's interpretation. How one school or one book reads it; other schools may read it differently. The five labelsOscar Salemink, in the chapter Ritual Efficacy, Spiritual Security and Human Security: Spirit Mediumship in Contemporary Vietnam in A World of Insecurity edited by T. H. Eriksen, E. Bal and O. Salemink, Pluto Press, 2010, analysed the turn to mediumship in contemporary Vietnam through the concept of "spiritual security", and argued that the search for spiritual security through religious and magical practice often occurs when anxieties cannot be resolved by other means. He placed this concept within a broader framework called existential security, extending from the concept of human security usually defined in economic, political, physical or ecological terms.
Salemink identified four types of anxiety requiring spiritual adjustment: physical and mental health problems needing healing; economic anxiety and market risk, that is, seeking luck; existential anxiety related to death; and the deliberate embracing and managing of risk. He also viewed religious behaviour as a way of establishing personal identity in Vietnam.
This is anthropological research, not quantitative research. It gives a conceptual framework to understand why people turn to ritual, based on observation and interviews, but does not provide numbers on the proportion of people consulting diviners across economic periods. At present, there is no Vietnamese quantitative data on how the number of people consulting diviners varies with economic cycles. This is a real gap, and it should be stated as a gap rather than filled with speculation from foreign data.
Another study on Vietnam takes a narrower direction: seeking spiritual security in traffic, that is, religious and magical behaviours to control risk when travelling under complex traffic conditions. This approach is not about divination in general, but about a specific domain where physical risk and the need for control meet. It fits Malinowski's argument: ritual appears where technique cannot sufficiently control outcomes.
What is an effect that is real but arises from belief called?
Kind of knowledge: UnprovenNot enough evidence. Someone has claimed it, but no source or study is strong enough to confirm it. The five labelsThere is a type of effect that needs to be named correctly. When a person believes a behaviour will bring a result, that belief can change their behaviour, and if behaviour changes, the result may change accordingly. For example, someone who believes a ritual helps them calm down may actually become calmer, and that calmness changes how they handle situations. This effect is real, and it is not evidence that the ritual predicts anything about the future.
The correct statement is: this is an effect arising from belief and behaviour, not from predictive ability. It should not be called "the system predicts correctly", because that attributes to ritual a power the data do not support. Nor should the effect be denied, because that ignores part of the reality of spiritual life. Keeping these two things separate is the condition for discussing this topic without falling into either extreme.
Why do the observational and experimental tiers diverge?
There are several reasonable explanations for this gap, and they are not mutually exclusive.
First, experiments measure something much narrower than life. Perceiving illusory patterns in a random sequence on a screen is a very different task from deciding to attend a ceremony or consult a diviner after a major event. The laboratory manipulation lasts minutes; life anxiety lasts months or years.
Second, the observational tier may be confounded by things other than sense of control. A worsening economic context may increase community need, free time, permission to express beliefs, or simply make the topic more discussed in newspapers. Each of these could produce a correlation without the mechanism "loss of control produces illusory patterns".
Kind of knowledge: UnprovenNot enough evidence. Someone has claimed it, but no source or study is strong enough to confirm it. The five labelsThird, the true mechanism may not be loss of control in general, but a specific kind of loss of control, such as loss of control over a threat that could harm oneself or loved ones. This is a hypothesis, with no closing results.
What this article does not say
This article does not say divination is true or false, does not say believers are ignorant, and does not say all ritual is superstition. It also does not say Vietnamese people believe in divination more or less than people elsewhere, because there is no comparative data in the sources used here.
What it does say is: looking at the data, we see two different tiers with different degrees of certainty. The observational tier, including ethnography and context-based correlations, is fairly solid and shows anxiety accompanying more turning to ritual. The experimental tier, the hypothesis that inducing loss of control directly increases superstition through illusory pattern perception, has not held up. Keeping these two separate is the most honest reading currently available.
As for young people and divination, the parts on pressure, uncertainty and the Barnum effect have been discussed in another article on this site, see Why young people go to diviners: pressure, uncertainty and what psychology calls the Barnum effect. The article you are reading focuses on evidence and rebuttal, and does not repeat that article's content.
