4&5 YEARS ASSESSMENT TOOL

The tool contains 23 items in a game-style format that children find fun and engaging. The children are assessed one at a time, in a quiet, appropriate space at their ECD centre or elsewhere. Data capture on the ELOM 4&5 is fully digitised using Survey CTOa digital platform that allows you to collect high-quality data using mobile phones, tablets, or computers, even when you’re offline (on an appropriate tablet). Instructions for assessors are colour-coded for ease of use, and stop rules are hard coded into the tool in such a way that if the child does not succeed on a specific item, the system automatically progresses to the next relevant item.

The assessment takes approximately 45 minutes per child.

The domains chosen for ELOM 4&5 underpin the South African early learning curriculum: Gross Motor Development; Fine Motor Coordination and Visual Motor Integration; Emergent Numeracy and Mathematics; Cognition and Executive Functioning; and Emergent Literacy and Language.

Once the overall domains had been decided, a scan of available tools used in similar exercises to the ELOM was undertaken. A particular focus was on measures with established validity and reliability which had been developed for use in South Africa and the region or in similar socio-cultural and socio-economic contexts. In designing the items used in the tool, the team also drew extensively on research literature, the South African National Curriculum Framework for Children from Birth to Four, and interviews with experts, including ECD practitioners and Grade R teachers. An initial long list of items were tested using accepted psychometric practices (see technical manual for detail) before selecting the final 23 items used in the tool.

A child’s performance on each item in each of the five ELOM 4&5 domains is awarded a raw score, which is then transformed into a scaled score. In each domain, item scaled scores are summed to provide a domain Total Score out of 20. The five domain scores are then summed to derive the ELOM 4&5 total score out of 100. Scores for each domain and for the total fall within one of 3 performance bands: On Track, Falling Behind and Falling Far Behind (in earlier versions this last category was referred to as At Risk). The cut off points for On Track, Falling Behind and Falling Far Behind for children aged 50 to 59 months and 60-69 months were set empirically and in consultation with key stakeholders (in 2016).

ELOM 4&5 Years Direct Assessment was validated on a sample representative of a range of socio-economic backgrounds of South African children, with children from five of the main language groups (isiXhosa, isiZulu, Afrikaans, Setswana, English). 1,331 randomly selected children from 173 schools in three provinces were assessed.

Yes, internationally accepted psychometric practice was followed to check for fairness. Analyses showed that the ELOM did not discriminate between children from different socio-economic or language backgrounds.

Yes they do. In the development process, we looked at three types of reliability:

1. Internal reliability (also known as consistency) is a measure of the consistency of a test-taker’s performance on the test questions. If all the questions on the test measure related skills, then its internal reliability will be high.
2. Test-retest reliability refers to the consistency (stability) of the test over time. This form of reliability is assessed by administering a test to the same participants on two occasions (for young children normally within a week), and correlating the two scores. A significant test-retest correlation of .75 or above is considered acceptable
3. Inter-rater (scorer) reliability is evident when the scores awarded by different assessors of the same individual are correlated. In a study where more than one assessor is testing a number of children or adults it is important that all have the same approach to scoring the test. This form of reliability is established during assessor training. It can be established by calculating Cohen’s Kappa (among other methods). Values of 0.4 to 0.75 are considered moderate to good and a kappa of >0.75 represents excellent agreement.

The team followed accepted practices in setting standards. This requires the use of empirical dataEmpirical data is acquired by observation or experimentation using accepted methods. Child test scores and observations of caregiver-child interactions would be two examples and the judgments of qualified people. The ELOM team used data from the standardisation sample to explore how children from the five economic quintiles performed. The task was to set standards for each domain and for the ELOM Total that was neither too high nor too low. They had to be aspirational but realistic. The preliminary standard was set at the score achieved by the top 40% of the children in the age validation sample (the 60th percentile). Children from all quintiles were represented in that band (fewer in quintile 1 and more in quintiles 4 and 5). A consultation with expert stakeholders in the ECD sector (both in government and civil society) resulted in the finalisation of this standard. It will be re-examined and changed if necessary as experience with the ELOM is accumulated.

No, ELOM is not a tool for assessing the school readiness of individual children. It is not designed to assess intelligence or diagnose developmental delay. However, it can be used to identify children who are significantly behind the standard expected for their age. They could then be referred for specialist assessment.

No, the ELOM 4&5 Years Targeting Tool was designed to quickly identify individual children within a programme that need additional early learning support.

Children’s performance on the five domains is more important than the total score, as these give programme managers a clear indication of how their programmes are performing, and where attention is needed to bring the children up to standard.

Costs depend on the amount of support required from the DataDrive2030 team and a cost estimate is provided once a request is received. We aim to keep the costs as low as possible and to subsidise costs for smaller non-profit organisations to enable their use of the tools.

The tool is designed to be used by an accredited ELOM assessor or trained professional psychometrist e.g. occupational therapist.

This tool is available in all of South Africa’s 11 official languages.