Adolescents experiencing physical abuse

The 2017 Optimus study on child sexual victimisation was commissioned by the UBS Optimus Foundation and carried out by researchers from the Centre for Justice and Crime Prevention and the Department of Psychology and Gender Health and Justice Research Unit at the University of Cape Town.

The study consists of a population-based survey and an accompanying school-based survey, as well as a qualitative agency component. The household survey was based on a multi-stage stratified sample that was designed to produce a nationally representative sample and which used province, geographic area (urban/rural) and race group as explicit stratification variables (see p23 of the technical study report for more detail). A power allocation of 0.4 was used to increase the sample size in the smaller strata. The schools, on the other hand, were clustered according to the Enumerator Areas (EAs) identified in the household survey and so this data is not representative of the school population.

Between 5 and 10 interviews were conducted in each EA for the household study, and one adolescent aged 15 – 17 years was interviewed per selected household. Where a given household included more than one child in the required age group, one was randomly selected using either a Kish Grid or interviewing the young person whose birth date occurred the earliest in the year. Active informed consent was obtained from parents and informed assent was obtained from the adolescent. At the schools, a total of 30 interviews were completed; 10 learners each were randomly selected from grades 10 to 12. Passive parental consent was sought at schools, so that parents were requested to complete signed consent forms if they did not want their child to participate in the study.

The household sample consisted of 5 631 participants, while the school-based survey comprised 4 086 learners. The refusal rate for households was 5.2%, and 3.9% for learners at schools.
The study included both interviewer administered questionnaires and a self-administered component. In reporting the findings, we have drawn on the self-administered data where possible as we consider them to be more reliable given the anonymity involved. There does appear to have been a greater willingness to disclose in the self-administered component, with rates reported in this component generally being higher than rates reported in the interviewer-administered component.

A strength of the prevalence survey is that it captures incidents that are not reported to the authorities and therefore do not appear in administrative data. The Optimus survey drew on different sites (home and school) and different methods of collecting the data (interviewer and self-administered questionnaires), and provides a sense of how the approach taken can influence the levels of reporting.

A potential limitation was the ethical requirement of obtaining parental consent in the household survey. Consent and informed assent is essential to research ethics, but there is a concern that parents who may themselves be abusers may have refused to consent and therefore biased the results.

A further limitation is that the data cannot be disaggregated beyond provincial level, and in some cases the small numbers reporting various forms of abuse makes disaggregation of the weighted data even to this level difficult. Obtaining data at a lower level (e.g. district) is important for understanding where violence against children is most prevalent, and which groups of children are most at risk.

Lastly, this survey fills an important data gap in understanding the prevalence of sexual violence against children (in the context of other forms of violence), but it is not a regular survey and does not provide routine surveillance data that can be used to monitor trends over time.