Abstract #65  -  Selection and adaptation of measures to evaluate an intervention to reduce intimate partner violence in South African adolescents.
Authors:   Presenting Author:   Dr. Marcia Russell - Pacific Institute for Research and Evaluation
Additional Authors:  Dr. Alan Flisher, Dr. Pamela Cupp, Dr. Rachel Jewkes, Dr. Cathy Mathews, Ms. Anik Gevers, Mrs. Chantel Lefleur-Bellerose,
Aim: This abstract is one of four that make up a symposium on “Developing a School Based Intimate Partner Violence Prevention Program for South Africa.” It describes procedures used to select and adapt measures for evaluation of the intervention, “Respect 4U.”
Method / Issue: Measurement selection was guided by the theoretical framework for our intervention. Work began with an inclusive list of measures used to assess model constructs in previous studies, many conducted by team members working with South African adolescents and young adults (AF, RJ, PC). This list was pruned to eliminate redundant items, and ethnographic work was conducted (AG) to ensure that the survey used terms that would be familiar to 8th grade learners in mainstream schools serving English-, Afrikaans, and Xhosa-speaking communities, and that these learners would be able to relate to experiences asked about. The survey was then translated from English into Afrikaans and Xhosa and back translated. Cognitive interviews were conducted (CLB, AG) with 8th grade English-speaking, Afrikaans-speaking, and Xhosa-speaking focus groups from mainstream schools to ensure that survey items were clear and comprehensible to learners from school settings targeted for the intervention. The questions were read to learners, and they were given an opportunity to suggest changes. It was notable that English-speaking and Afrikaans-speaking learners in focus groups made many suggestions for change, but those in Xhosa-speaking groups did not, claiming they understood the survey terms and questions. The facilitator suggested changes based on her work with Xhosa youth, and it was decided to further test the survey in focus groups of 10th grade Xhosa speakers from a magnet school. These learners had much more to say about changing specific words and terms in Xhosa and the way questions and response options were structured. Finally, paper versions of the survey were administered to classes in targeted schools. Students were spaced to provide privacy, and survey items were read to assist slow readers in completing and comprehending the survey.
Results / Comments: The ethnographic work provided assurance that learners understood the terms ‘boyfriend’ and ‘girlfriend’ and experienced these types of relationships. Some behaviours such as transactional sex were confirmed as happening often, whereas others, such as using substances to dry or tighten their vaginas, were not and were dropped. Findings from the cognitive interviews resulted in simplification of wording and sentence structures throughout the survey and deletion or rewording of ambiguous questions. Examination of findings from surveys administered in classrooms of targeted schools led to the elimination of redundant items and those endorsed by all learners or none.
Discussion: The survey will be administered using a personal digital assistant (PDA). Students will use earphones to hear a recording of the questions and key their answers into a handheld computer. This technology reduces problems related to low literacy, gives learners privacy in responding to sensitive questions, and facilitates data entry.
Go Back