Assessment of children in Grade R in four provinces

For this study, a sentinel approach to sampling was used. Two education districts were selected in each of the four participating provinces. These districts cover a wide range of school quintiles.

Schools were stratified by district and quintile group and ordered randomly. The first three or four schools in each randomly-ordered district-quintile stratum list were included in the sample. As the sample sizes per stratum were unequal (the researchers expected to sample 10 schools from each district, but there were three quintile groups) the rule was to have larger sample sizes in the lower-quintile groups (e.g. four schools in Q1, three schools in Q2-3 and Q4-5, respectively).

Learners were sampled on the day of the actual school visit, stratified by gender, with a target of 25 children in each school. Children were assessed using the ELOM 4&5 and the Social-Emotional Rating scales.

After cleaning the data, 2055 children were retained in the final sample, with roughly even numbers in Quintile 1, Quintiles 2 and 3 combined and Quintiles 4 and 5 combined.

Ideally data should have been collected in January/February 2019, at the start of the Grade R year and before children are exposed to the Grade R curriculum. However, final sign-off from the sponsor was only obtained at the end of February 2019. This delayed the start of the data collection process and children who participated in this study had already been exposed to one full term of Grade R. This is likely to have influenced child outcomes.

Download the data set here.