The recent development of the Early Learning Outcome Measure (ELOM), a culturally appropriate test for children aged 50 to 69 months for all eleven South Africa’s official languages, offers benchmarks against which children’s development can be tested. This paper compares the programme gains (value added) of participants in five early childhood development (ECD) programmes against the age-performance gradient of the ELOM test in two large scale cross-sectional studies to assess the possible influence of such programmes on early learning and learning outcomes. This goes beyond what was attempted in the programme evaluation, which did not have a control group against which to compare score gains. The cross-sectional relationship between age and learning gains is thus used as a counterfactual that can be regarded as a control group against which to evaluate programme gains. ‘Effect sizes’ estimated in this way range between 9% and 72% of a standard deviation under conservative assumptions. These are surprisingly large, considering that these five programmes served mainly children who would typically attend no-fee schools, and that three of the programmes are playgroups with only 2½ to 8 hours of contact time per week.