Individual Submission Summary
Share...

Direct link:

Teacher quality and student learning metrics: Evidence from northern Nigeria (RANA)

Wed, March 28, 1:15 to 2:45pm, Hilton Reforma, Floor: 1st Floor, Business Center Room 1

Proposal

By using more sophisticated program monitoring methodologies, it is possible to embed high-quality applied research and formative evaluation into programming, at very little cost, and at scale. We showcase data collected from the Reading and Numeracy Activity (RANA) project implemented by FHI 360 in Northern Nigeria to improve Hausa-language learning outcomes in Grades 1-3 in government and Integrated Qur'anic schools. The program includes a robust monitoring framework that uses program data to identify the links between teacher effectiveness, program fidelity, and student learning growth. As such, we identify/quantify teacher effectiveness using a teacher value-added model (VAM) where teachers’ performance during the school year is evaluated based on their students’ learning growth from the beginning through the end of the year using EGRA and EGMA assessments.
The empirical strategy relies on a longitudinal (within-year) sample of students from each school under the RANA umbrella. This method enables us to identify each teacher’s contribution to the gains made by each student over the course of the school year. In other words, we identify a teacher’s value-added to their students’ learning growth by estimating the difference between students’ predicted performance had they been taught by an average teacher and their performance under the current teacher. The empirical approach necessitates a robust regression analytic framework that accounts for measurement error in the assessments used to measure student learning, a complete longitudinal record for most students in the sample, teacher fixed effects estimation to identify teacher value-added, and lastly, application of the Empirical Bayes shrinkage estimator to account for possible small sample bias in teacher effects. The final analytic stage consists of correlating teacher VAM scores with measures of program fidelity of implementation, teacher characteristics, and classroom composition variables. This approach will inform project research as well as key stakeholders regarding specific classroom strategies and pedagogies associated with successful teachers, as well as identifying teachers who require additional resources and support from the project.
The within-year longitudinal design enables accurate measurement of learning growth, where traditional cross-sectional designs cannot. This approach enables us to link student learning gains, over time, to individual teachers to identify teacher value-added and ultimately aggregate teacher performance metrics to the school level to highlight schools requiring further assistance. Further, we assess the equity implications by treating teachers as inputs into the educational production function and examining the distribution of effective and ineffective teachers across different equity dimensions both at the student and teacher levels. For instance, we examine whether effective teachers are clustered in specific areas or are accessible only to students of a certain socioeconomic standing. Lastly, we couple teacher performance metrics with our Fidelity of Implementation tool results to distinguish associations between teacher/classroom practices and teacher performance, enabling the program to provide specific guidance and support to teachers. As such, we are able to use this information to construct statistical profiles of the schools and teacher attributes stratified by teacher performance classification.

Authors