From Learning Data to Policy Evidence

This upcoming one-week residential course, designed around the African Fellows in Education Policy (AFEP) and open to other African education researchers through generous funding from the Gates Foundation.

Overview of the course

The one-week residential course will be held at the University of Cape Town, from 22-26 February 2027. This five-day course develops participants’ ability to judge whether their data are adequate for their research question, and which research design lets them answer the policy question they care about. Rather than focusing on estimation techniques, the course builds a conceptual understanding of the main research designs used in education policy research: what policy question each answers, what it requires of the data, and its limitations for internal and external validity. Participants work on their own research throughout the week, with dedicated writing time every day.

Who this course is for

This course is ideal for AFEP Fellows and other African education researchers and postgraduate students who are working on a paper or research project using education data. The course will benefit those who wish to:

Sharpen a policy question into a specific, answerable research question.
Assess whether their data are fit for purpose before “pressing regress”.
Understand the strengths and limitations of experimental and quasi-experimental designs, and choose one that fits their question and data.
Make concrete progress on writing their own paper, with feedback from facilitators and peers.

What you will learn

This five-day course combines lectures, guided discussions of published African studies, and workshops on participants’ own research. Key focus areas include:

Learning data in Africa: The range of national, regional, international, household and administrative learning data available across the continent, what each measures, and whom it covers.
Data fit for purpose: How outcome and treatment measurement, coverage, selection, variation and timing determine whether data can answer a research question.
The regression toolkit: Ceteris paribus reasoning, omitted variable bias, dummy variables and interactions, bad controls, and what selection on observables can and cannot deliver.
Research designs and the questions they answer: Randomised controlled trials, fixed effects, difference-in-differences, regression discontinuity and instrumental variables: the counterfactual each relies on, its key assumption, what it needs from the data and its limitations.
Internal and external validity: Weighing credibility against relevance, and recognising when careful descriptive research is the right contribution.
Writing your working paper: Structuring a paper, writing a compelling introduction, motivating your data choice, and writing clear data, methods and limitations sections.

Indicative course programme

Day 1: Learning data, research questions and fit. Participants pitch their research question, followed by an overview of learning data across Africa and a framework for judging whether data are adequate. Participants audit their own data and begin writing their introduction.
Day 2: Descriptive and causal questions. Sessions on descriptive versus causal policy questions and an intuitive approach to regression, with a hands-on example on private schooling and learning. Participants draft the data section of their paper.
Day 3: Experiments and fixed effects. Guided discussions of a randomised evaluation in Kenya and a within-household study of private schooling in East Africa, with short replication exercises. Participants test whether these designs fit their own research.
Day 4: Difference-in-differences, regression discontinuity and instrumental variables. Guided discussion of a curriculum reform evaluation in Tanzania and other African examples, followed by a design clinic on participants’ own questions.
Day 5: Validity, limitations and writing. A session on internal and external validity and the value of descriptive research, a writing workshop on methods and limitations, and presentations of each participant’s research design.

How the course will be structured

The course will feature a combination of lectures, guided discussions of published papers, short replication exercises, small group design clinics and dedicated writing time. Participants receive a reading guide for three core papers before the course. Exercises use R, and equivalent Stata materials will be provided for participants who prefer Stata.

The hub will offer follow-up support via online sessions with facilitators in the months following the course.

A key goal of the course is to build and strengthen a community of African education policy researchers. To foster networking, social events will be held throughout the week.

Prerequisites

The course is designed for participants with research experience who are working on a paper or research project using education data. Basic skills in data analysis and regression in R or Stata will be assumed. Before the course, participants will be asked to read three papers with the reading guide provided and to submit a one-page note on their research question and data. Participants must bring a laptop with R (or Stata) installed. The course will be conducted in English.

Two courses, one application. This course is one of two back-to-back, one-week courses on African learning data. Course 1, Foundational Learning Data in MICS: FLS and FLM 2.0 15-19 February 2027, is co-taught with UNICEF and focuses on analysing and communicating MICS foundational learning data. Course 2, From Learning Data to Policy Evidence 22-26 February 2027, focuses on research design and on whether data are adequate to answer a research question. Participants may take either course or both.

How to apply

Please complete the course application form.

We encourage you to complete the form as soon as possible. The deadline is 8 November. Spaces are limited, and you will be notified if your application is successful. For inquiries or further assistance, please email us.