Abstract
There is renewed global interest in ending violence in childhood: member states of the United Nations (UN) are now committed to the Sustainable Development Goals, which include the goal of ending all forms of violence – including violence in childhood, by 2030. A recognized tenet of policy reform, however, is that before a problem can be tackled, one has to be able to measure it. But the database on violence against women and children remains poor. There are too few indicators on interpersonal violence, and only a few countries report prevalence rates on all of them. This paper therefore collects values on indicators of violence against women and children from multiple data sources, and combines them to arrive at country-estimates of violence in childhood. We have used the imputation method, a statistical operation which generates one or more plausible replacements for each missing value in original data. The assigned “imputed” value for a missing entry is based on other information in the data, either the reported value of a similar country, or constructed from other responses of the country with the missing entry. Imputation does not claim to represent the missing value accurately. Rather, it provides a means of using existing data in the most efficient way. In particular, it allows information from countries that report incomplete data to be utilized appropriately in the estimation of summary measures.
In this case, we impute countries’ missing values for seven indicators of violence in childhood. Six of these directly reflect violence against children (VAC). They are child homicide, corporal punishment at home, peer violence in schools (two categories – bullying and fighting), and violence against adolescent girls (two categories – physical and sexual abuse). The final, indirect, indicator is violence against women (VAW, intimate physical and/or sexual violence experienced by women in the last 12 months), because when children witness domestic violence they too are adversely affected.