Can you apply machine learning to create change for children in South Africa? We have teamed up with Africa's biggest community of data scientists Zindi for this exciting competition. Total prize money is US$3000.
The challenge is to use machine learning techniques to identify which early learning programme factors contribute to better learning outcomes in children, by predicting a child’s Early Learning Measurement - ELOM - score. ELOM categorises children's development as “on track”, “falling behind” or “falling far behind". The best entries from this competition will support better interventions that make optimal use of limited resources to ensure South Africa’s children are thriving. Deadline is 30 April 2023. More details are here: https://zindi.africa/competitions/datadrive2030-early-learning-predictors-challenge
Did you know that only 3 out of 10 of the poorest children in South Africa attending Early Learning Programmes start school on track, compared to 8 out of 10 children from the richest households?
We will never achieve our national development plan goals of reducing inequality and eliminating income poverty unless we give ALL children quality early learning opportunities so that they may start school ready to learn.
This is what powers DataDrive2030 - the commitment to closing the opportunity gap in early childhood.
Did you know that only 3 out of 10 of the poorest children in South Africa attending Early Learning Programmes start school on track, compared to 8 out of 10 children from the richest households?
We will never achieve our national development plan goals of reducing inequality and eliminating income poverty unless we give ALL children quality early learning opportunities so that they may start school ready to learn.
This is what powers DataDrive2030 - the commitment to closing the opportunity gap in early childhood.
Ensuring that the ELOM tools are fair was crucial to our development process.
The fairness of a test is seen when individuals of the same ability but from different groups (e.g. cultures or languages) have the same chance of succeeding on the test.
They are not disadvantaged by the test procedure or the items due to their background. An example would be that the images used to test a child’s vocabulary may be unfamiliar to children from some backgrounds resulting in a poor score that does not represent their knowledge of words.
As we come to the end of a busy year, we wish our partners and peers a happy holiday season and a joyful and peaceful new year. We look forward to continued collaboration in 2023!
Papama Mateza is a Monitoring and Evaluation Officer at early childhood organisation Sikhula Sonke and one of the first people in South Africa to train as an ELOM assessor.
He is a dedicated advocate for equal early learning opportunities for all children and has seen first hand how data-driven decisions can improve programmes.
"We used the tools in 2017 to assess Sikhula Sonke’s programmes, found gaps, and used these insights to strengthen the programme.
There was a measurable improvement when we did the follow-up assessment," he says. Read Papama's story https://bit.ly/3HgWvBt
We have 7 short years left to achieve our 2030 goals of ensuring every single child in South Africa has access to quality early childhood education so that they can start their formal schooling with the right learning foundations in place.
The clock is ticking! Right now, less than half of children who attend an early learning programme in SA start school on track. And for those children who are not able to access an early learning programme, the percentage who are developmentally on track is likely to be considerably lower.
The decision to include ‘2030’ in our name speaks to the urgency of our work. We want to mobilise data to drive change by 2030. One of the ways we do this is by providing tools and support to ECD operators to enable them to assess the strengths and weaknesses of their programmes so that they can take the necessary action.
By working with hundreds of organisations across the country, we aim to achieve 'micro change at scale'.
Read testimonials from our clients and partners on our website
At DataDrive2030, we are committed to making early childhood data accessible, understandable and actionable by all.
One of the ways in which we do this is to place large ELOM datasets on an open access repository and to invite researchers from across the world to engage with these data.
Our datasets are used by students for masters and PhD theses, they support secondary research, inform policy development and planning, and are used to enhance the design of early learning programmes.
We are proud to report that as of November 2022, we have had 24,000 page views and almost 3500 downloads!
Find our datasets in the link in bio.
Children begin learning in their homes from day one. The ELOM Home Learning Environment Tool measures characteristics of young children's homes that are associated with the development of early language, basic numeracy and cognitive functioning.
The tool looks at things like how many learning resources are in the home, how many different learning activities caregivers do with their children, and how much quality time caregivers and children spend together.
If you have a pre-schooler at home for the holiday and are looking for ways to keep them busy, we recommend a wide variety of games that promote play for learning and development.
Read more about the ELOM Home Learning Environment Tool in the link in bio.
We developed our Early Learning Measurement (ELOM) tools with respect for the diverse socio-economic and linguistic backgrounds of South Africa's children and families.
ELOM tools are available in all 11 official SA languages. Caregiver interviews and child assessments are done in the language the participant is most comfortable using, with assessors who are fluent in that language.
And because we love data, we thought we would share some recent stats on language of assessment. This graph shows the language spread for more than 5000 assessments completed so far in 2022, using our ELOM tools.
Read more about our ELOM tools in the linked in bio.
We are committed to making early childhood development data accessible to all because everyone can play a role in closing the opportunity gaps in early childhood.
Have a look at some recent statistics about how and where our open access datasets are being used.
ELOM datasets are available in the link in bio.
Less than half of 4 year old children attending an early learning programme in South Africa are able to do the cognitive tasks expected of a child their age. Children who start school without the right learning foundations in place will find it difficult to cope with the Grade R curriculum, and helping these children to catch up places a massive burden on Grade R teachers.
Improving the quality of our early learning programmes must be our overall goal if we are to increase the proportion of children who are developmentally on track when they enter the Foundation Phase of schooling.
While a lot of emphasis is often placed on improving things like infrastructure, the data tell us that relationships between children and their practitioners and between children themselves is key to promoting children’s well-being and improving early learning outcomes. This policy brief puts forward recommendations for improving the quality of teaching and learning in South African early learning programmes, at scale.
Find link in bio
