This presentation summarizes qualitative data from user observations and semi-structured interviews with teachers and coaches, as well as Kenyan education officials on their use of Tangerine:Coach (previously Tangerine:Tutor) data for decision-making and teacher support. Tangerine:Coach is open-source software developed by RTI International that has been deployed in support of over 25,000 schools nationwide in Kenya since 2015, and is currently being introduced or scaled up also in Cambodia, Jordan, Liberia, Sierra Leone, West Bank, and Uganda. In Kenya, a typical coach visit to a school includes taking stock of classroom conditions, observing teacher practice, appraising student ability, and providing pedagogical support to teachers. Tangerine:Coach is designed with the intent of supporting this process - the software applies logic to combine data from surveys, classroom observations, and student assessments into a feedback report tailored for each coaching conversation. Coaches regularly upload their data to a central server which are then published to an online dashboard and sent by email to education officials. In theory, this data-sharing enables ongoing monitoring of coach activities and student progress and allows for timely targeted support (e.g., through additional school visits or professional development) to schools, teachers, or coaches as needed. Yet, how are these different actors actually using the data and dashboards? What changes are they making in their coaching or resource allocations? What data format is most accessible to actors across the different levels of the education system?   This prevention was developed by Lucy Wambari, Timothy Slade, and Carmen Strigel for the 2018 mEducation Alliance Symposium.

Do children with deficits in both rapid automatic naming (RAN) and phonological awareness in Chichewa experience more difficulty with reading growth than children with single or no deficits? To answer this question, we developed a RAN measure of common objects and a phonological awareness measure and administered them four times during a school year. Students who scored poorly on one of these measures as compared to their peers were classified as having a deficit. A compelling finding was Single Deficit and Double Deficit groups have similar average scores at Time 1; by Time 3 and Time 4, the Single Deficit group pulls away, more than doubling the average score of Double Deficit counterparts at Time 4. This study presents a tool to help to identify children at risk for reading difficulties. Children with a Double Deficit profile may need differentiated instruction.