A short, hands-on workshop: five two-hour meetings over two weeks, working with real surveys from raw files to a reproducible result.
The second link pulls the session notebooks into your hub account. Click it again after any update — your own edits are kept.
Most graduate training inverts the priority. Students learn fixed effects and GMM in some detail, but rarely learn how a survey is designed, where it breaks, and what its structure does and does not let them identify. Yet the design choices — multistage sampling, rotating and refreshment panels, attrition, the household that splits between rounds — are exactly what determine whether a clean-looking regression means anything.
A careful analyst needs both halves, and the second is the one usually missing. The workshop follows Deaton's progression from design to description to demand to dynamics, working throughout with data from the region.
What a careful analyst should know before touching the data. Cross-section, repeated cross-section, and panel, and what each lets us identify; weights, stratification, and clustering; attrition versus comparability across rounds.
Constructing consumption aggregates, then poverty, inequality, and the shape of the distribution, by region and across rounds. The discipline of describing the data carefully before modelling it.
Engel curves, budget shares, equivalence scales, and spatial price deflation, building to a first demand system. This is where household surveys earn their keep for policy.
Pseudo-panels built from repeated cross-sections, then genuine panel methods — fixed effects, two-way fixed effects, and clustered inference — on a true multi-wave panel.
Participants run a small reproducible analysis of a survey of their choosing. We close with where the methods break down: short panels and dynamic models, measurement error, and endogeneity.
The sessions run on the workshop hub, in your browser. You will get a username and set your own password at first login — it must be at least seven characters. Bring a laptop if you have one.
If you would like the software on your own machine as well — worth doing, since it is what you will use after the workshop — the installation instructions assume no prior Python and take you through it a step at a time. Do it at home on a good connection: it is roughly a 1.3 GB download, and twenty laptops installing at once in the classroom will simply stall.
You do not need a World Bank Microdata account. Registering is free and worth doing eventually for your own research, but nothing in the five sessions requires it.
The hands-on work uses LSMS_Library, an open-source Python package giving a uniform interface to the household surveys of roughly forty countries. We use Ghana's GLSS as the running local example and Uganda's National Panel Survey for the panel material.
World Bank microdata cannot be freely redistributed, so the sessions work from harmonised extracts prepared in advance and staged on the hub. This repository contains code and text only — no survey data.