Attitudes towards violence against women shape how people interpret behaviours, respond to disclosures, and support or resist policy and practice reforms (1, 2). When permissive or minimising attitudes are normalised, they can make it harder for people who have experienced violence to be believed, seek support and receive consistent institutional responses, while also weakening expectations of perpetrator accountability (1–6). Conversely, attitudes that recognise harm, affirm equality and support respectful relationships create conditions for prevention and early intervention to succeed (1, 2). Tracking attitudinal change over time is therefore central to understanding progress, identifying emerging risks, and strengthening primary prevention approaches (1, 2).
Australia’s National Research Organisation for Women’s Safety (ANROWS) conducts the National Community Attitudes towards Violence against Women Survey (NCAS) every four years. The NCAS benchmarks community attitudes and understanding related to violence against women, gender inequality and recognition of different forms of violence and control (7, 8). NCAS reporting typically summarises results using a set of validated scales that capture broader constructs, aligned with ‘Change the story’ and the evidence base on the drivers of violence against women (1, 7, 8).
The NCAS was designed to monitor population-level attitudes over time, and the scale scores are well suited to that purpose. By aggregating multiple items into validated measures, the NCAS provides a stable way to track broad progress and compare trends across jurisdictions and cohorts.
While useful for high-level tracking over time, scale averages can obscure heterogeneity in item responses and differences between population groups. Respondents with different attitudinal profiles may produce a similar aggregate score, and relatively stable scale averages can coexist with substantial variation in the beliefs underlying those scores. Targeted prevention practice requires insights at the level of specific attitudes rather than abstract scale scores. Programs and communications typically engage with concrete narratives such as myths about sexual violence, mistrust of people who have experienced violence, and assumptions about gender and leadership (1, 2, 9). Item-level analysis therefore complements scale reporting by identifying which attitudinal statements are most strongly associated with overall scale scores, and which beliefs are widely endorsed versus contested in the community. This identifies candidate areas for message testing, program refinement and more detailed monitoring.
This approach is timely because community understanding of ‘what counts as violence’ increasingly extends beyond physical assault to include coercive control, social isolation, economic abuse and technology-facilitated surveillance (10–12). Understanding which non-physical behaviours are consistently recognised as violence in Victoria, and where recognition is uneven, matters. This is because recognition influences help-seeking, bystander response and institutional decision-making (5, 6, 10–12). Similarly, community attitudes towards gender equality are shaped not only by overt endorsement of inequality, but also by more subtle patterns of minimisation, denial and backlash, which can undermine prevention and policy implementation even when explicit sexism is rejected (9).
Respect Victoria funded ANROWS to conduct an additional 1,500 interviews in Victoria as part of the 2021 NCAS, increasing the Victorian sample to enable item-level analysis of ‘high impact’ attitudes within the state (8). ‘High impact’ is an analytic shorthand for items that are either strongly aligned with overall scale scores or especially prevalent, contested or gender-patterned. It does not mean that an individual item has been shown to cause broader attitudes or violence, or that changing that item alone would produce population-level change.
This paper uses two approaches: (a) item–scale correlations to identify items most strongly associated with overall scale scores, and (b) descriptive patterns (agreement distributions, variability and gender differences) to identify which attitudes are widely shared, contested or gender-patterned in Victoria. Together, these complementary lenses support a practical prevention question: where are the beliefs that are both meaningful in the measurement framework and salient in the population, and are therefore candidates for prioritisation in messaging, education and further testing.
This exploratory secondary analysis addresses two research questions:
- Which specific attitude items most strongly differentiate overall scale scores in Victoria?
- How are responses to specific attitude items distributed in Victoria, and which items show the clearest gender differences?
This technical paper is intended primarily for researchers, evaluators and policy analysts seeking detailed information about NCAS item analysis in Victoria. Companion evidence and implications briefs translate its principal findings for policymakers, senior decision-makers and primary prevention practitioners.