Version Notes
Changes from Version 1 to Version 1.1:
1. Two new projects were appended. They are Nal'ibali Story Powered Schools (SPS) in 2017, 2018, and 2019 from South Africa; and Literacy Achievement and Retention Activity (LARA) in 2017 and 2019 from Uganda.
2. Three projects are excluded because the original source microdata for these records are not available. Users of Version 1 release that included these projects should treat the current file as authoritative and expect those country-years to be absent. These are Early Grade Reading Improvement Activity (MERIT) in 2016, 2017 and 2018 from Malawi; Tunoze Gusoma in 2022 from Rwanda; and a national assessment in 2021 from Zambia.
No variables from the Version 1 were dropped.
Changes from Version 1.1 to Version 2:
1. The new version expands coverage by adding 82 259 pupil records from 20 new project-rounds across seven new projects, including the first data from Somalia. The harmonised data file now holds 217 127 records, 15 countries, 24 projects, and 50 project-rounds.
2. Adds 771 variables, 2 880 in total, mostly oral reading items, second-language and written-comprehension subtasks, rapid naming, working memory, and a representation level field for every record. One variable, age_dob, is removed.
3. Corrects values for 11 035 previously published records in Kenya 2015, DR Congo 2015, and the South Africa SPS endline. The corrections cover age, month, and item-level codes, so analyses that used these variables should be re-run.
4. Puts every *_pcnt variable on a 0 to 100 scale, rebuilds the administrative division fields, and reorders and recodes the mathematics strategy flags. These changes alter 132 125 previously published records, but no record is removed and no harm_student_id changes.
