Early numeracy skills are highly relevant for children’s mathematics learning at school, especially in the initial years when much mathematics learning relies on early numeracy competence. The aim of this study was to investigate the level of early numeracy skills in a sample of South African children in the first months of formal schooling. In this cross-sectional study, there were 443 first graders (206 girls and 237 boys) from Gauteng Province schools. The mean age of the children was 81.61 months (6 years 10 months) (SD 5.40 months). Their early numeracy skills were measured with the ThinkMath Scale. The main finding of this study was that there were statistically significant differences in early numeracy skills between the children when they started first grade. The differences were related to the home language of the first graders in the English medium schools, as well as the type of school (public vs. private). This article concludes that the numeracy competence of the children from the sample was notably varied in the beginning of their formal schooling, which has implications for teaching in the vastly different classroom populations that are all served by one national curriculum.
In the research about innate and core knowledge of numbers (Feigenson, Dehaene & Spelke
Knowledge about the development of early numeracy skills and their relevance for later mathematics learning has been increasing rapidly during last 10 years (Aunio & Räsänen
There are different known causes for early grades children’s low performance in mathematics, such as differences in cognitive functions or inadequate opportunities for learning essential mathematical skills (Geary
South African school children have demonstrated poor academic outcomes and levels of performance in mathematics (Spaull & Kotze
One of those potential components contributing to the development is the language development (Kleemans, Segers & Verhoeven
In the first grade, the South African curriculum includes numbers, patterns and relationships; patterns, functions and algebra; spaces and shapes; measurement; and data handling (Basic Education Republic of South Africa
Typically, the children in the private schools attend a year of school (kindergarten) before Grade 1, known as Grade R, whereas, in the public schools many children do not attend Grade R. Even though Grade R is the first year of compulsory schooling and forms part of the 4 years which comprise the ‘foundation phase’ (i.e. elementary school phase), because of affordability of transport and higher school fees (as Grade R is more expensive than day care pre-schools), many parents only place their children in formal schooling in Grade 1. Grade R has been compulsory since 2014, but many parents still adhere to the former notion that Grade 1 as the first year of school. Many parents do not see the benefit of Grade R, which is mostly focused on perceptual, early numeric and literacy development. There are currently many movements in the South African education system to try and change this perception and advocate for the necessity for children to attend Grade R.
The main aim of this study was to investigate the early numeracy performance of children starting primary school in a sample of South African Gauteng Province schools. We captured the aim into the following research questions:
What is the difference between boys and girls in early numeracy performance?
What are the differences in performance between the children with different home languages, all of whom learn in an English medium school and were tested in English?
What is the difference in performance between children in public and private schools?
What are the differences in performance in children between individual schools?
How do listening comprehension skills relate to numeracy skills?
How are low-performing children distributed according to individual schools and school type?
This study is part of the larger research project that investigates the early numeracy learning and evidence-based pedagogical support in South African schools. A total of 443 children (206 girls and 237 boys) participated in this cross-sectional study. The mean age of children in this study was 81.62 months (6 years and 10 months) (SD 5.40 months). The study was conducted in schools using English as a means of instruction in the greater Johannesburg area and Western Gauteng (
Demographic descriptives (age, gender, race, home language and low performance in numeracy skills) by school.
Variable | School |
||||||
---|---|---|---|---|---|---|---|
A | B | C | D | E | F | G | |
Type | Public | Public | Private | Public | Private | Public | Private |
Category | Mainstream | Mainstream | Remedial | Mainstream | Mainstream | Mainstream | Mainstream |
Location | Semi-rural | Urban | Urban | Semi-rural | Urban | Urban | Urban |
Socio-economic status (according to school fees) | Middle–low | Middle–low | High | Middle–low | Middle–high | Low | Middle–high |
2 | 2 | 2 | 3 | 2 | 3 | 3 | |
Class size ( |
23–25 | 35–38 | 7–8 | 35–40 | 18–20 | 35–40 | 15–18 |
48 | 76 | 16 | 106 | 38 | 104 | 55 | |
Boys (%) | 28 (41.7) | 47 (61.8) | 8 (50.0) | 56 (52.8) | 20 (52.6) | 46 (44.2) | 32 (58.2) |
Girls (%) | 20 (58.3) | 29 (38.2) | 8 (50.0) | 50 (47.2) | 18 (47.4) | 58 (55.8) | 23 (41.8) |
82.67 (4.46) | 80.68 (5.59) | 84.20 (7.04) | 82.34 (4.50) | 83.03 (4.06) | 78.73 (5.87) | 84.40 (4.16) | |
English (%) | 17 (35.4) | 50 (65.8) | 14 (87.5) | 1 (0.9) | 30 (78.9) | 11 (10.6) | 41 (74.5) |
Afrikaans (%) | 1 (2.1) | 4 (5.3) | – | 1 (0.9) | 1 (2.6) | 2 (1.9) | 4 (7.3) |
Setswana (%) | 25 (52.1) | 6 (7.9) | 1 (6.3) | 79 (74.5) | 1 (2.6) | 20 (19.2) | – |
isiXhosa (%) | 1 (2.1) | 1 (1.3) | – | 2 (1.9) | – | 8 (7.7) | 2 (3.6) |
isiZulu (%) | 1 (2.1) | 7 (9.2) | – | 10 (9.4) | 3 (7.9) | 35 (33.7) | 2 (3.6) |
Sesotho (%) | 3 (6.3) | 3 (3.9) | – | – | – | 14 (13.5) | – |
Other (%) | – | 5 (6.6) | 1 (6.3) | 13 (12.3) | 3 (7.9) | 14 (13.5) | 6 (10.9) |
9 (18.8) | 4 (5.3) | 0 | 25 (23.6) | 1 (2.6) | 45 (43.3) | 0 |
Low performance=performance at or under –1 SD of mean score of ThinkMath Scale.
The private schools in this sample have access to special educational support professionals, occupational therapists and speech therapists, who assist children who had been identified with difficulties in their learning and development. The public schools in this sample were supported by only few professionals in the local school district office, who have to serve more children than they can accommodate properly. Even though schools were supposed to have a school-based learning support team with suitable special education competence, they have difficulties in guaranteeing that all of these professionals have the necessary background to assist children in public schools sufficiently.
The children in this study were starting their primary school in Grade 1 at the time of the study. Children (
Numeracy skills were assessed using the ThinkMath Scale (Aunio & Mononen
Item examples of ThinkMath Scale.
There are dots in the white box. Tick the box that has fewer dots in it than the white box.
You see three numbers. Check through all three numbers carefully. Now tick the smallest number (e.g. 5, 8 and 4).
You see three numbers. Check through all numbers carefully. Now tick the largest/biggest number (e.g. 12, 15 and 11).
Here you see two numbers and a line. There’s a number missing where the line is. From the boxes next to the line, tick the number that belongs on the line (e.g. 4, 5, __ or 8, 7, __).
Count how many dots there are in the box altogether. Now tick/circle the number that gives the right answer (e.g. 9).
First look at the number. In the box next to it, colour or tick the same number of dots (e.g. 12).
Next to the letter A, there are numbers in the box. I will say one number. Tick that number (e.g. 8).
As the ThinkMath Scale was used for the first time in the South African context, we wanted to validate it in relation to another test, namely,
The children’s
A native speaker of both Finnish and English translated the ThinkMath Scale in English. The accuracy of translation was checked by the research team. Children’s skills were assessed in the beginning of their school year, in March 2016. As the
The reliability in terms of Cronbach’s alpha for the ThinkMath Scale was high,
Parametric tests were used in data analyses as all scales were normally distributed (skewness and kurtosis within -1 – 1).
Results are reported following the order of the research questions, and means, standard deviations and test values for each analysis are presented in
Means, standard deviations and test values in ThinkMath Scale by gender, race, home language, school type and school.
Variable | Post-hoc comparison |
||||
---|---|---|---|---|---|
All | 443 | 27.55 (9.50) | |||
0.171 | |||||
Boys | 237 | 26.97 (10.07) | |||
Girls | 206 | 28.21 (8.77) | |||
0.000 | |||||
English | 164 | 32.96 (8.27) | English > Setswana, isiZulu, Sesotho and other | ||
Afrikaans | 13 | 30.85 (9.97) | Afrikaans > isiZulu | ||
Setswana | 132 | 23.60 (7.83) | Setswana < English | ||
isiXhosa | 14 | 26.29 (10.91) | |||
isiZulu | 58 | 22.40 (8.41) | isiZulu < English and Afrikaans | ||
Sesotho | 20 | 23.05 (7.77) | Sesotho < English | ||
Other | 42 | 27.52 (9.67) | Other < English | ||
0.000 | |||||
Public | 334 | 24.58 (8.63) | |||
Private | 109 | 36.37 (5.95) | |||
0.000 | |||||
A | 48 | 24.58 (6.68) | A < B, C, E and G | ||
B | 76 | 31.17 (7.01) | B > A, D, E, F and G | ||
C | 16 | 31.56 (6.60) | C > A, D and F | ||
D | 106 | 23.91 (7.47) | D < B, C, E, F and G | ||
E | 38 | 35.66 (7.01) | E > A, B, D and F | ||
F | 104 | 20.74 (8.89) | F < B, C, D, E and G | ||
G | 55 | 38.25 (3.78) | G > A, B, C, D and F |
Post-hoc comparison with Bonferroni correction. Sign > indicated statistically significantly (
One-way ANOVA on ThinkMath total score × gender (two groups) showed that girls and boys were performing equally well,
One-way ANOVA on ThinkMath on total score × home language (seven groups) showed a statistically significant difference between the groups,
One-way ANOVA on ThinkMath Scale × school type (two groups) revealed a statistically significant difference between the groups,
One-way ANOVA on ThinkMath × school (seven schools) showed a statistically significant difference between the groups,
Pearson’s bivariate correlational analysis revealed that the Listening Comprehension Scale correlated statistically significantly with the ThinkMath Scale,
There were 84 (19%) low-performing children (performance at or under -1 SD of total mean score of the ThinkMath Scale) in these sampled Gauteng Province schools. The distribution of low-performing children in different schools is presented in
The aim of this study was to investigate the school beginners’ early numeracy knowledge in a small convenience sample of Gauteng Province schools. We were especially interested in the existence of low performance of early numeracy, as it is a risk for later (MLD) (Jordan
The main finding that emerged from this study was that children who enter Gauteng schools’ in the first grade have big heterogeneity in their early numeracy knowledge. This is a common finding also in the international context (Aunio & Niemivirta
The fact that most of the children were studying and tested in English that was not their home language can be important explaining factor. Taylor and Von Fintel (
The most striking result was that the low-performing children were almost all attending public schools that are not well-resourced to support the diversity of learners, especially with regard to the availability of professional support at school or from the school district officer. The children in school F, with most low-performing children, were found to be significantly younger than children in other schools, so young age can be one explaining factor (Jordan
This study gave an overview of the early numeracy skills of South African school beginners in the Gauteng Province in South Africa. To get a more clear and a reliable view of the numeracy skills of a larger and more representative sample of South African children, children from other provinces and in schools where the medium of instruction is not English should be included in the future studies. It is therefore important to have the test translated into the other 10 South African languages, or a selection of them, to find out how children perform on their home language. Furthermore, in this study, only numeracy and listening comprehension skills were measured. In order to explain children’s level of numeracy performance in more detail, other factors should be included, such as executive functions and family socio-economic status could be included in future studies.
Low performance in early numeracy is an important indicator for later learning difficulties in mathematics. It is possible to support these low-performing learners educationally so that later problems can be avoided. In the South African context we need to develop easy-to-use
In this study, the early numeracy skills of South African children were assessed. Children’s numeracy knowledge varied and differences were found related to home language of the first graders in the English medium schools, school type (public vs. private) and difference between individual schools. Those children who were attending a public school were performing lower than children in private schools. English-speaking children outperformed children who spoke some other language at home, except for Afrikaans. Adequate educational resources and targeted numeracy evidence-based pedagogical support should be aimed for public schools in low socio-economic areas specifically. It is also important to consider teacher development in collaboration with school-based support teams and district officials.
The authors acknowledge the limitations of the JSAIS. Although it is widely used in the South African context as a measurement of cognitive ability, the test was developed in the 1980s, and its content has never been updated. However, it was fast to administer and had scale for early math similar to ThinkMath Scale. Another possibility could have been the MARKO-D, which is an individual test developed from a German instrument standardised and normed for South Africa (Fritz
The authors declare that they have no financial or personal relationships which may have inappropriately influenced them in writing this article.
P.A. is the main person responsible for writing the article and making the analysis. R.M. has contributed to the writing of the article and making the analysis. L.R. has contributed to the writing of the article and was the main person responsible for data collection. M.T. participated in the process of writing.