High-quality data on violence against women and girls is crucial for understanding the types, patterns, causes and consequences of VAWG and designing effective programmes and policies. Using both quantitative data (numbers and statistics) and qualitative data (personal stories and experience) is important to understand the scale, trends and experiences of those affected by violence, as well as the impacts of different interventions. These sources generally use rigorous collection and analysis methods, are regularly updated, and often undergo peer review.

• Prevalence data: Several countries have systems for collecting national-level prevalence data, for example through Demographic and Health Surveys. Survey data can be found on the interactive data visualisation platform: WHO Global Database on the Prevalence of Violence against Women.
• Administrative data – such as records kept by health and social services, legal aid services and police – can also provide information on reported incidents and responses.
• Other good sources of data to understand VAWG can include government reports, academic research, and data published by recognised international organisations like the United Nations as well as practitioner-based knowledge from civil society organisations.

Qualitative data and practitioner-based evidence gathered from those working directly with women and girls affected by violence is crucial to complement and understand quantitative data. This provides on-the-ground insights, often filling gaps in official data and giving a more complete understanding of VAWG and the experiences of those affected.

There are multiple sources of VAWG data and it can sometimes be difficult to judge the accuracy and reliability of different data sets and their interpretation. It is important to analyse or seek advice on data collection methodologies used in studies, to be aware of potential biases and problems. Be aware that various sources and forms of data on VAWG might not capture the full scope of the issue, due to factors such as underreporting, cultural stigma, or lack of comprehensive data collection methods. Different sources might have biases or limitations depending on their methodology, perspective, or the context in which data was collected. It is important to approach the data critically and to use and triangulate data from a range of sources to get a more accurate and nuanced understanding of violence against women and girls.

Strengthening the capacity of local partners to use and collect data to inform the development of laws, policies and programmes, as well as track progress and advocate for change, is important for effectively addressing VAWG. Take the time to build data literacy by explaining how to interpret trends and any limitations. For example, an increase in reported incidents of violence against women and girls could indicate several things. Firstly, it might reflect an actual rise in the number of incidents occurring (prevalence). However, it could also suggest improved awareness of and/or confidence in reporting mechanisms, meaning that more victims are coming forward to report incidents that might previously have gone unreported (but the prevalence has not changed or may have even decreased). To improve understanding and use of data, it can also help to explain the importance of understanding context, particularly when sharing data with people who may not be familiar with data analysis but want to be able to make informed decisions about policy and programming.

Ethical guidelines for data collection, sharing and reporting must guide any VAWG data collection process. This is essential for all VAWG data, but particular principles and processes apply for especially sensitive data, including related to groups at high risk of discrimination, such as LGBTQI people, or data on adolescents under the age of 18. For example, the Violence Against Children and Youth Surveys (VACS) are nationally representative household surveys of adolescents and young people ages 13-24 and are considered good practice in terms of the way that data is collected and disseminated. Read more about how to conduct ethical data collection.

When using new data systems, such as artificial intelligence (AI), to analyse big data, it is important to consider any potential risks. For example, the Spotlight Initiative Latin America Regional Programme worked with Brazil’s National Justice Council to develop an innovative AI tool that strengthens the gender-responsive analysis of femicide by the justice sector. To support the ethical roll-out of the AI tool, Spotlight Initiative worked with partners to produce a guide for the justice sectors on how to use AI to support the analysis of femicide data.

Using data effectively for improved decision-making and advocacy is an ethical responsibility and involves creating diverse products tailored to various users. This requires thinking carefully about what types of information and presentation are suitable for different data users. For example, the list below sets out common types of information useful for different users.

• General Public: Key figures, trends, data visualisations, short stories
• Media: Summary tables, charts, trends, data visualisations, short stories
• Policymakers: Summary tables, charts, trends, data visualisations, short stories
• Development practitioners/CSOs/Gender specialists:  Summary and disaggregated data, analysis of results over time, metadata
• Data analyst, researcher: Complete, raw data set (without names)