Violence against women does not have a single cause. The strongest evidence points to an interaction between gender inequality, unequal power in relationships, attitudes that legitimise male control or violence, exposure to violence during childhood, harmful alcohol use, economic and social stressors, and weaknesses in institutions that are supposed to protect women. The World Health Organization identifies gender inequality as a root cause while emphasising that violence emerges through interacting individual, relationship, community and societal factors.
For South Africa, that distinction matters. Poverty, unemployment, alcohol misuse and crime can increase risk, but they should not be treated as automatic explanations for violence against women. Research has found that power, control and norms around gender and violence are important parts of the picture. At the same time, not every man exposed to economic hardship or harmful social norms becomes violent, and women are not responsible for the violence committed against them. The evidence supports a multi-level explanation rather than a single-cause theory.
Violence is not simply a product of poverty
Economic hardship is often cited when discussing South Africa’s GBV crisis, but the evidence is more nuanced than the claim that poverty causes violence.
The WHO identifies high levels of poverty and unemployment among community and societal factors associated with higher prevalence of violence against women. It also identifies gender discrimination, harmful gender norms, weak enforcement of laws and high levels of community violence as relevant factors.
That does not mean poverty mechanically produces violent behaviour.
Economic pressure can interact with other conditions. Financial dependence may make it harder for a woman to leave an abusive relationship. Unemployment can increase household stress. Lack of independent income can reduce a woman’s bargaining power. Weak access to housing and social services can also make separation from an abusive partner more difficult.
These are different mechanisms from saying that a poor household is inherently more violent.
A useful analysis therefore asks how economic conditions alter power, choices, exposure to violence and access to protection, rather than treating income as a simple cause.
Power and control are central to intimate-partner violence
One of the clearest themes in the evidence is the role of unequal power within relationships.
The WHO defines intimate-partner violence to include physical and sexual violence as well as psychological abuse and controlling behaviours. Its current evidence review identifies men’s controlling behaviour and high levels of inequality in relationship power among risk factors.
Control can take many forms.
It may involve deciding who a partner can see, restricting access to money, monitoring communications, controlling movement, dictating clothing or friendships, threatening to take children away, or using violence to enforce compliance.
Not every controlling behaviour results in physical violence, and not every abusive relationship follows the same pattern. But control becomes particularly important when it is used to restrict a woman’s autonomy or to punish her for exercising it.
That helps explain why violence cannot be understood only as a response to an argument.
In some abusive relationships, violence functions as a means of maintaining or restoring control.
Social attitudes can make violence easier to justify
Attitudes do not automatically cause violent behaviour, but evidence links acceptance of violence and norms that legitimise male control with increased risk.
The WHO identifies attitudes that justify violence as a risk factor and harmful gender norms that uphold male privilege, entitlement and restrictions on women’s autonomy as societal risk factors.
Earlier WHO prevention research similarly found strong associations between accepting violence and both experiencing and perpetrating intimate-partner and sexual violence.
The relevant issue is therefore not simply whether a society has “bad attitudes”.
It is whether certain beliefs make particular forms of control, coercion or violence appear legitimate, excusable or normal.
Examples can include beliefs that a man should have the final say in a relationship, that a woman owes a partner sex, that jealousy demonstrates love, or that physical punishment is an acceptable response to perceived disobedience.
These beliefs can exist alongside formal laws that prohibit violence.
South African research points to power and norms
South African research provides evidence that these issues cannot be reduced to imported theories about violence.
A large South African cross-sectional study published in 2002 examined risk factors for domestic violence among 1,306 women aged 18–49. The researchers found that domestic violence was most strongly related to women’s social status and the normative use of violence in conflict or as an exercise of power.
The study identified several associations, including childhood exposure to violence, alcohol-related conflict and relationship conflict.
Importantly, however, it did not find significant associations for every commonly proposed explanation. Variables including the partner’s employment, financial disparity, cohabitation, urbanisation and several other characteristics were not significantly associated with domestic violence in that particular analysis.
That is an important caution against turning individual risk factors into universal explanations.
Evidence from one study does not establish that a particular characteristic causes violence in every setting.
Gender norms operate through relationships and institutions
Gender norms are not simply individual opinions.
They can influence what families teach children, what communities tolerate, how institutions respond to complaints and how women and men understand their expected roles.
WHO describes gender inequality as operating across intimate relationships, households, communities, institutions and wider society. Unequal access to resources and decision-making can limit women’s ability to negotiate within relationships and increase vulnerability to violence.
This helps explain why legislation alone cannot eliminate the problem.
A law can establish that violence is illegal.
It cannot instantly change beliefs about entitlement, control or gender roles.
Nor can legislation by itself guarantee that police, courts, health services and communities will respond consistently when violence is reported.
Recent South African research also finds themes of male dominance
A qualitative South African study published in 2026 examined women’s perspectives on intimate-partner violence in rural Limpopo.
Researchers interviewed 31 women aged 18–40 and identified themes including traditional male dominance and beliefs about women’s submission. While most participants rejected the abusive behaviour presented in the study’s scenario, some considered it acceptable in circumstances where they believed a wife should submit.
Because this was a small qualitative study, it cannot be treated as representative of South African women as a whole.
Its value is different: it provides detailed evidence of how ideas about male authority and female submission can appear in people’s understanding of relationships and violence.
That complements larger quantitative studies rather than replacing them.
Economic dependence can affect a woman’s options
Economic inequality matters partly because it affects exit options and bargaining power.
A woman who depends on a partner for housing, food, school fees or other necessities may face greater practical barriers to leaving an abusive relationship.
This does not mean economic dependence causes the abuse.
The distinction is important.
The perpetrator remains responsible for the violence. Economic dependence can instead affect how much freedom a survivor has to resist, leave or seek protection.
WHO’s evidence framework identifies women’s economic empowerment and poverty reduction among the strategies for preventing violence against women. Its 2025 RESPECT women framework specifically includes economic empowerment and poverty reduction alongside relationship skills, services, safer environments, prevention of child and adolescent abuse, and transformation of harmful attitudes and norms.
Alcohol matters, but it is not an excuse
Alcohol frequently appears in research on violence against women.
WHO identifies harmful alcohol use as a risk factor for both perpetration and experience of intimate-partner and sexual violence.
South African research has also found associations between alcohol-related conflict and domestic violence.
But alcohol should not be treated as the underlying explanation for violence.
Most people who drink alcohol do not assault their partners.
The more useful interpretation is that harmful alcohol use can interact with existing attitudes, relationship conflict, impulsivity and social environments in ways that increase the likelihood or severity of violence.
That distinction matters for prevention because reducing alcohol-related harm can address one risk factor without pretending that alcohol is responsible for the broader system of gender inequality and control.
Childhood exposure can reproduce patterns of violence
Another recurring finding concerns exposure to violence during childhood.
WHO identifies a history of child abuse and witnessing violence in the family as individual-level risk factors associated with intimate-partner and sexual violence later in life.
South African research has also found an association between experiencing violence in childhood and later domestic violence.
This does not mean that a child exposed to violence will inevitably become either a perpetrator or a victim.
Rather, early exposure can form part of a pattern in which violence becomes familiar as a way of resolving conflict or exercising authority.
That is one reason prevention aimed at children and adolescents is relevant to the long-term response.
Community violence can shape the environment
Violence against women does not occur in isolation from other forms of violence.
WHO identifies high levels of violence and crime within communities as factors associated with higher prevalence of violence against women. It also identifies the availability of drugs, alcohol and weapons as relevant community or societal risk factors.
A community with widespread violence may provide fewer safe spaces, weaken informal protection networks and normalise aggressive responses to conflict.
But again, this is not a claim that general crime automatically produces intimate-partner violence.
Different forms of violence have different dynamics.
The evidence instead suggests that broader violent environments can interact with relationship-level risk factors.
Weapons can change the consequences of violence
The availability of weapons is particularly relevant to lethal violence.
An abusive relationship does not have to become more frequent to become more dangerous if access to a weapon changes the potential consequences of an assault or threat.
WHO includes weapon availability among community and societal factors associated with violence against women.
This is especially important when considering the transition from non-lethal abuse to homicide.
The presence of multiple risk factors does not allow investigators to predict that a particular relationship will end in femicide. It does, however, explain why prevention requires attention to circumstances that can increase the severity of violence.
Relationship conflict is a risk context, not a justification
Conflict occurs in healthy relationships as well as abusive ones.
Research therefore distinguishes ordinary disagreement from violence used to intimidate, punish, control or harm a partner.
South African research has found associations between frequent conflict and domestic violence.
But conflict should not be framed as something that “causes” violence in a way that distributes responsibility equally between victim and perpetrator.
Two people can disagree without one assaulting the other.
The relevant question is what happens when conflict occurs: whether one person uses threats, coercion, violence or control to impose their will.
Sexual entitlement is another important mechanism
Sexual violence is also connected to beliefs about entitlement and consent.
WHO’s evidence review identifies male sexual entitlement and beliefs that limit women’s ability to refuse sexual advances as important factors in sexual violence.
The principle is straightforward: consent requires the ability to choose freely.
A relationship, marriage, previous sexual activity or financial dependence does not turn another person’s body into property.
Where social norms treat women’s refusal as illegitimate or assume that men are entitled to sex, coercive behaviour can become easier to rationalise.
This is one reason sexual violence prevention cannot rely exclusively on criminal penalties. Understanding consent and challenging entitlement are also part of prevention.
Inequality can work through institutions
Power is not confined to the household.
WHO identifies gender discrimination within public and private institutions—including health, justice, police and education—as a societal factor associated with violence against women. It also identifies the absence or poor enforcement of laws addressing violence against women as a relevant factor.
This creates a second layer of risk.
A woman may experience violence from an individual perpetrator and then encounter institutional barriers when seeking help.
If complaints are dismissed, protection orders are poorly enforced, services are inaccessible or investigations are delayed, the formal existence of rights may not translate into practical protection.
For South Africa, this connects directly to the wider implementation problem examined in Why South Africa’s GBV Crisis Has Persisted Despite Laws and Government Programmes.
Why “toxic masculinity” is too simple an explanation
Public discussion often compresses these findings into a single phrase: toxic masculinity.
That phrase can point toward genuine issues, particularly rigid ideas of masculinity involving dominance, entitlement, aggression and control. WHO notes that harmful gender norms, including rigid notions of masculinity, can contribute to violence perpetration.
But it is not a complete causal model.
Men are not inherently violent, and masculinity is not itself a cause of violence.
The evidence concerns specific norms and behaviours: acceptance of violence, entitlement, control over women, sexual entitlement and rigid expectations about male authority.
That is a more precise and useful way to describe the evidence.
Why not every risk factor applies equally to every case
Violence against women is a heterogeneous problem.
Intimate-partner violence, sexual violence by strangers, family-related killings, trafficking, workplace harassment and femicide can involve different perpetrators, circumstances and mechanisms.
A factor associated with one form of violence may not explain another.
The WHO therefore uses an ecological approach that considers individual, relationship, community and societal factors together.
This approach is more useful than searching for a single variable that supposedly predicts all violence against women.
It also explains why prevention has to operate at multiple levels.
What the evidence does not support
Several popular explanations should be treated cautiously.
“Poverty causes GBV.”
The evidence supports an association between poverty, unemployment and higher prevalence in some settings, but poverty is not a sufficient explanation.
“Alcohol causes men to become violent.”
Harmful alcohol use is an established risk factor, but it does not explain violence by itself and does not remove responsibility from perpetrators.
“Gender norms alone cause violence.”
Norms are important, but violence results from interacting individual, relationship, community and societal factors.
“All men are potential perpetrators.”
The evidence does not support such a generalisation. Risk factors describe statistical associations and mechanisms; they do not determine an individual’s behaviour.
“Women stay because they accept abuse.”
This reverses the direction of responsibility and ignores economic dependence, safety concerns, children, housing, social support and the possibility that leaving can itself increase danger.
The evidence is strongest when it explains mechanisms without turning statistical risk into individual prediction.
What prevention needs to target
The evidence increasingly points toward prevention that operates across several levels rather than relying on one intervention.
WHO’s updated RESPECT women framework identifies seven broad strategies: strengthening relationship skills, empowering women, ensuring services, reducing poverty, making environments safer, preventing child and adolescent abuse, and transforming attitudes, beliefs and norms.
That framework reflects the complexity of the evidence.
If unequal power matters, prevention needs to strengthen women’s autonomy and relationship skills.
If economic dependence creates barriers to safety, economic empowerment matters.
If harmful norms increase risk, interventions need to change attitudes and expectations.
If childhood exposure to violence increases later risk, prevention needs to begin before adulthood.
If institutional failures allow violence to continue, laws and services must be implemented effectively.
Prevention has to be measured, not simply announced
The existence of a programme does not establish that it reduces violence.
WHO’s prevention guidance emphasises evaluation so that interventions can be assessed and improved.
That means programmes should ideally be evaluated against meaningful outcomes rather than only counting activities.
For example, the number of people attending an awareness session says less than whether attitudes changed, whether abusive behaviour declined or whether survivors became more able to access protection.
Likewise, providing economic-support services should be evaluated not merely by the number of beneficiaries but by whether those services improve economic independence and safety.
The evidence points to interacting causes, not a single culprit
The strongest conclusion from the research is that violence against women is multi-level and interconnected.
Power and control can operate within relationships. Gender norms can legitimise unequal authority. Economic inequality can restrict women’s options. Alcohol can increase risk in particular circumstances. Childhood exposure to violence can shape later behaviour. Community violence and weapon availability can increase danger. Weak institutions can allow abuse to continue or make protection harder to obtain.
None of these factors provides a universal explanation.
And none transfers responsibility for violence away from the person who commits it.
For South Africa, the practical implication is that prevention cannot depend solely on tougher laws, policing or public-awareness campaigns. The evidence points toward a combination of effective institutions, women’s economic and social autonomy, safer communities, prevention of violence in childhood, interventions addressing harmful relationship norms, accessible survivor services and sustained work to change attitudes that legitimise control and violence.
The most important question is therefore not which single factor “causes” GBV. It is which combination of conditions allows violence to emerge, continue or escalate—and which interventions can change those conditions before violence becomes lethal. That distinction is essential if South Africa is to move from describing the causes of GBV to preventing it.