Succeeding against the odds: A quantitative assessment of the effectiveness of IkamvaYouth

Author(s): Servaas van der Berg , Lewis Sidney McLean

Stellenbosch Working Paper Series No. WP05/2017
Publication date: June 2017
Abstract:
The work presented here is essentially an update to the previous evaluation conducted by RESEP (Research on Socio-Economics Policy), located in Stellenbosch University’s Department of Economics: Against the Odds: An Evaluation of the IkamvaYouth Programme. This update to Against the Odds provides a more rigorous quantitative assessment of the contribution of the IkamvaYouth programme to learners’ performance at school. IkamvaYouth’s contribution to the performance of its learners – the ‘Ikamvanites’ – is estimated in five matric subjects: English as a First Additional Language, Mathematics, Mathematical Literacy, Physical Sciences and Life Sciences. Controlling for socioeconomic and demographic factors, estimates for the 2014 sample of matriculating Ikamvanites suggest that the IkamvaYouth programme is associated with dramatic improvements ranging from a low as high as 0.293 standard deviations (about six percentage points) for Physical Sciences up to a high of 0.502 standard deviations (about ten percentage points) for Life Sciences. We check the robustness of these results with a subsample of Western Cape learners who could be matched with their 2011 grade 9 Systemic Test results. Matching makes it possible to control for learners’ pre-programme performance and thus for factors such as students’ pre-programme motivation and ability. However, the matching process also restricts the sample to a subset of high-performing learners (learners who reached matric without failing a grade between grade 9 and matric) that, it is argued, are not a good counterfactual for how Ikamvanites would have performed had they not participated in the programme. As a result, estimates of the effectiveness of the programme obtained from the matched sample would under-represent the effectiveness of the programme, and should thus be treated as lower-bound estimates of the effectiveness of IkamvaYouth. These lower-bound estimates of the impact of the programme range from a high of 0.30 standard deviations (about 5.5 percentage points) for Life Sciences, still approximately a year of learning, to a low of 0.036 (about one percentage point) for Physical Sciences. Thus, in sum, the findings suggest that IkamvaYouth does indeed add substantially to the performance of its learners, helping them to succeed against the odds.

JEL Classification:
I21, I24

Keywords:
IkamvaYouth, programme evaluation, matric results

PDF (1 MB)