Differential Susceptibility

People can share a household, school, workplace or neighbourhood and still be affected in different ways. One person may respond strongly to conflict, instability or support, while another appears less affected.

But sharing a setting does not mean receiving the same experience. People may encounter different treatment, notice different signals or interpret events through different histories. Their behaviour may also change how others respond to them.

Even after accounting for these differences, another possibility remains: people may differ in how strongly particular environments affect them.

Differential susceptibility is a model of this variation. It proposes that some characteristics can be associated with greater responsiveness to environmental conditions—not only for worse under adversity, but also for better under support.

The same responsiveness that increases difficulty in one environment may increase the capacity to benefit in another.

A relationship, not a type of person

Differential susceptibility does not divide humanity into sensitive and insensitive kinds of people.

It describes a conditional relationship among at least four elements:

• a characteristic of the person;
• a measured environmental condition;
• a particular outcome;
• a period of development.

A person might show greater responsiveness to caregiving quality during childhood but not to every school, workplace or social environment encountered later. Responsiveness might appear in emotional outcomes without appearing in cognitive or physical ones.

Changing the environment, outcome or developmental period can change the observed relationship.

This is why susceptibility is not a diagnosis or permanent identity. A group-level difference in average responsiveness cannot reliably predict how one individual will respond.

For worse—and for better

The defining feature of differential susceptibility is a for-better-and-for-worse pattern.

Imagine a characteristic such as high emotional reactivity. Under harsh, unpredictable or rejecting caregiving, children with that characteristic might show comparatively greater difficulty. Under warm, predictable and responsive caregiving, they might show comparatively more positive development.

If both sides of that pattern are supported, reactivity does not look like a simple weakness. It looks like greater responsiveness to environmental quality.

“Better” and “worse” refer here to the measured outcome—not to the value of the person. Someone can show greater responsiveness in one part of development without being generally advantaged or disadvantaged.

The pattern must also be demonstrated rather than assumed. Greater harm under adversity is not enough.

Susceptibility is not the same as vulnerability

Differential susceptibility is related to, but distinct from, two other environmental-response models.

In a diathesis–stress or vulnerability model, a characteristic increases the likelihood of difficulty when conditions are adverse. When the adversity is absent, outcomes become more similar. The characteristic is associated with greater risk, but not necessarily with greater benefit.

Vantage sensitivity describes the opposite side: some people benefit more from favourable conditions or interventions without showing corresponding additional harm under adversity.

Differential susceptibility includes both directions. The proposed susceptibility characteristic is associated with greater disadvantage in unfavourable environments and greater benefit in favourable ones.

These are models of conditional patterns, not permanent categories of people. Researchers must compare how well each model fits the observations.

Support is not merely the absence of harm

A study cannot establish for-better-and-for-worse responsiveness if it measures only different degrees of adversity.

Low conflict is not necessarily warm support. The absence of neglect is not the same as responsive care. Not being excluded does not necessarily create belonging.

If an environmental scale runs from severe harm to little harm, its favourable end may represent relative safety rather than a genuinely positive environment. That can still reveal vulnerability, but it cannot automatically demonstrate enhanced benefit under support.

A strong test of differential susceptibility must represent meaningful variation across both:

• adverse, restrictive or unsupportive conditions;
• supportive, responsive or enriching conditions.

What counts as support may also depend on age, culture, social expectations and outcome. Researchers cannot assume that a measure has the same meaning in every population without testing that assumption.

What might make responsiveness differ?

Researchers have investigated several possible susceptibility factors, including:

• temperamental and emotional reactivity;
• behavioural inhibition;
• physiological stress responses;
• neural responses to social or environmental signals.

None of these is a validated diagnostic test of general susceptibility.

Temperament is among the most frequently studied possibilities. Some children react more intensely to novelty, frustration, threat or changes in social conditions. This reactivity may make an unsupportive environment more disruptive. It may also make responsive caregiving or effective teaching more influential.

Temperament and differential susceptibility are not the same thing. Temperament describes broader differences in emotional and behavioural tendencies. Differential susceptibility asks whether a characteristic changes the relationship between an environment and an outcome.

A temperamental characteristic may predict greater responsiveness in one study but not another. Findings vary across ages, measures and developmental outcomes. Reactivity should therefore be treated as a probabilistic candidate—not as evidence that a child belongs to a fixed susceptibility group.

Physiological and neural studies offer possible mechanisms. A system that responds strongly to environmental signals might register both threat and support more intensely. But biological responses change across tasks, situations and developmental periods. They may also reflect earlier experience rather than an inborn sensitivity.

No current biological measure can reliably determine how responsive a person will be across environments.

What genetic evidence can—and cannot—show

Genetic differences may contribute to variation in environmental responsiveness. That does not mean there is a single “susceptibility gene.”

Earlier studies investigated particular variants connected with serotonin or dopamine systems. Some reported that a variant predicted poorer outcomes under adversity and greater benefit under support or intervention.

Many of these candidate-gene findings have not replicated consistently. Interaction effects are especially vulnerable to small samples, multiple analyses and selective publication. A plausible biological explanation does not compensate for weak replication.

More recent research uses polygenic scores that combine information across many genetic variants. This better reflects the distributed character of genetic influence. Even so, reported polygenic interactions with environments are heterogeneous, usually small and difficult to estimate reliably. They do not permit dependable individual classification.

Genetic analyses also have to distinguish interaction from gene–environment correlation.

Genetically influenced characteristics may be associated with the environments provided by biological relatives. A person’s behaviour may evoke particular responses from others. People may also select environments partly related to their dispositions.

An apparent genetic moderation effect may therefore coexist with—or partly reflect—non-random environmental exposure.

As Genetic Influence Is Not Genetic Destiny establishes, genetic influence is probabilistic and develops through experience. Genetic measures cannot show that susceptibility is fixed, universal or beyond change.

How researchers test the pattern

Differential susceptibility is commonly tested through a statistical interaction: does the association between an environment and an outcome vary according to a proposed susceptibility characteristic?

A statistically significant interaction is not sufficient.

Researchers must examine the shape of the relationship across the environmental conditions that were actually observed. The data need to contain meaningful adverse and supportive conditions, with enough observations to estimate outcomes credibly on both sides.

Estimated lines may appear to cross when plotted. But crossing lines do not validate the model by themselves.

A convincing analysis must consider:

• whether the crossover lies within the observed environmental range;
• how uncertain its location is;
• where the compared groups or gradients differ reliably;
• whether the data fit vulnerability or vantage-sensitivity models better;
• whether nonlinear relationships could produce the apparent interaction;
• whether floor or ceiling effects distort the outcome;
• whether the finding survives correction for multiple testing;
• whether it replicates in independent evidence.

An estimated crossover outside the observed data cannot demonstrate for-better-and-for-worse susceptibility. Extending lines into an unmeasured region is extrapolation, not observation.

The interaction also remains an association unless the study design supports a causal conclusion. Cross-sectional evidence cannot establish temporal order. Longitudinal studies can improve temporal interpretation but may still contain confounding.

Randomised interventions provide stronger evidence about an intentionally changed environment. Suppose a parenting intervention produces larger improvements among children with a proposed responsiveness characteristic. The trial supports differential benefit. It establishes full differential susceptibility only if evidence also supports comparatively poorer outcomes under unfavourable caregiving.

Otherwise, the finding may fit vantage sensitivity.

Susceptibility changes across development and context

Environmental responsiveness can vary across:

• childhood, adolescence and adulthood;
• family, peer, educational and institutional settings;
• emotional, behavioural, cognitive and physical outcomes;
• brief events and accumulated exposure;
• cultural and material conditions.

A characteristic that matters during one developmental period may become less influential, more influential or influential in a different way later. The meaning of an environment can also change as capacities, relationships and social demands change.

Developmental timing is therefore part of the model. Susceptibility cannot be separated from the question: susceptible to what, for which outcome and when?

Different environmental forms matter as well. The social response of a caregiver is not interchangeable with a school policy, an informational environment or an inaccessible institution. Physical, Social, Informational and Institutional Environments can create different demands and opportunities, even when they overlap in everyday life.

Transactions can amplify initial differences

Differential susceptibility fits within the broader processes described by Person–Environment Transactions.

A person’s characteristics can affect:

• what they notice in an environment;
• how they interpret it;
• the responses they evoke from others;
• which environments become accessible;
• how they act within those environments.

Environmental responses can then reinforce, redirect or reduce the importance of the original characteristic.

A child who repeatedly receives dependable support after expressing distress may develop different expectations from a child whose distress produces rejection or escalation. Over time, these transactions can generate continuity or change.

But a single interaction finding does not prove such a feedback loop. Transactional claims require evidence about sequence and repeated influence over time.

The process is also not inevitable. Later relationships, institutions and experiences can alter an established pathway.

Responsiveness does not mean responsibility

People do not freely choose all the conditions that shape them.

Children do not choose their caregivers, household resources or early institutions. Adults remain constrained by income, discrimination, disability, geography, law, employment conditions and unequal power.

A person may help select, evoke or modify part of an environment while having little control over its most consequential features. Reciprocal influence does not mean equal influence.

Differential susceptibility must therefore never be used to blame people for being affected by harmful conditions. It must not imply that individuals create the discrimination, deprivation, coercion or institutional restrictions imposed on them.

Nor does the model imply that adverse environments matter only for people showing greater measured responsiveness. Harmful conditions can affect everyone even when their average effects differ.

Environments Contain Behaviour because they distribute opportunities, constraints and available responses. Differential susceptibility operates within these structures; it does not replace them.

What the model changes

Differential susceptibility changes how certain characteristics can be interpreted.

A characteristic associated with difficulty under adversity may not be a pure deficit. Under supportive conditions, the same characteristic may be associated with greater benefit. This does not make it universally advantageous. It means its developmental significance depends partly on context.

The model also changes how average environmental effects are understood. An intervention can produce a modest average effect while participants respond differently. That variation may contain meaningful information, but exploratory differences need replication before they become explanatory conclusions.

Supportive environments therefore have an active role in developmental theory. Development is not only about reducing risk. Responsive relationships, effective teaching, predictable institutions and meaningful opportunities can generate positive change.

Some people may respond especially strongly to those conditions. But support does not guarantee superior outcomes, just as adversity does not make poor outcomes inevitable.

Present evidence does not justify screening people into susceptibility categories or restricting support to those predicted to respond most. Lower measured responsiveness is not invulnerability, and group-level findings cannot determine an individual’s need.

A conditional pattern, not a destiny

The most accurate description of differential susceptibility is deliberately specific:

A characteristic may be associated with greater responsiveness to a particular environmental dimension, for a particular outcome, during a particular period.

Changing any part of that statement can change the result.

The same person may respond strongly in one relationship but not another, or during one period of life rather than another. Life Transitions Destabilise Established Patterns because environments, roles and demands change—and the significance of earlier tendencies may change with them.

Differential susceptibility does not tell us that some people are destined to suffer more or flourish more. It explains why characteristics associated with difficulty in one environment can sometimes also create greater potential for benefit in another.

Responsiveness is not weakness, superiority or identity.

It is a relationship between a developing person and the conditions in which development occurs.