Question NW3026 to the Minister of Basic Education to ask the Minister of Basic Education

Reply:

Answer to Portfolio Committee Question on Dropout 

There are two main data sources for estimating learner dropout rates for South Africa: STATS SA’s General Household Survey and Education Management Information System (EMIS) data. Both involve lags in between when the data is collected, when it gets consolidated and when it can be analysed.

STATS SA’s General Household Survey, which is conducted annually, is a nationally representative survey of South African households and collects information about the highest level of education attained by each member of the household. If a person’s highest level of education is grade 11, for example, and they are old enough to be unlikely to still be in the process of completing a higher level of education, it can then be assumed that they exited the education system at that point. The sample sizes are not large enough to permit precise estimates of dropout for each grade and province separately. The table below shows the estimated percentages of children reaching each grade, and conversely, the estimated percentages of children dropping out after each grade. The average across each grade comes to 4.5%. So, if we had to come down to a single statistic for the annual dropout rate, we could estimate that in recent years about 4.5% of all children in grades 1-11 exit the school system each year, though of course most of this dropout is after grades 9 to 11. Based on this methodology, data collected in a particular year is reflective of dropping out in the years preceding that year, but not necessarily of dropping out in that year. As the table shows, grade survival rates (or completion rates) have been increasing over time in recent years. Using GHS data from 2019-2021, the estimated percentage of youths who completed grade 12 had reached a figure of 57,0%.

The second approach is to use EMIS data to estimate departure ratios after each grade. The term ‘departing’ is used here because several forms of departing do not represent an educational failure: learners can leave the system due to death, emigration, transfer to a college and of course because they have successfully completed Grade 12. When analysing EMIS data, it is impossible to know the reason for the departure. These statistics also have limitations, mainly because for around 17% of learners who stay from one year to the next, national ID numbers cannot link the learners across the two years. This non-linkage problem reduces substantially to just 3% if variables other than the ID number are used, but even this presents challenges for the calculation of the flow statistics. A key question is whether apparent drop-ins from one year to the next are actually the same learners as apparent drop-outs from the previous year. Despite these difficulties, the department has estimated departure ratios using EMIS data, and these are presented in Table 2, which shows estimates of the percentages departing the school system at the end of 2018, 2019 and 2020.

Table 1: Survival rates and drop-out rates for each grade based on GHS data

2015-2017	2017-2019	2019-2021
	Percentage reaching at least this grade	Grade-specific dropout rate	Percentage reaching at least this grade	Grade-specific dropout rate	Percentage reaching at least this grade	Grade-specific dropout rate
Total cohort	100%		100%		100%
No schooling		0.6%		0.6%		0.7%
Grade 1	99.4%	0.2%	99.4%	0.1%	99.3%	0.0%
Grade 2	99.3%	0.2%	99.3%	0.3%	99.3%	0.2%
Grade 3	99.1%	0.5%	99.0%	0.2%	99.1%	0.2%
Grade 4	98.6%	0.4%	98.8%	0.3%	98.9%	0.4%
Grade 5	98.2%	0.6%	98.5%	0.6%	98.4%	0.4%
Grade 6	97.6%	1.5%	97.8%	1.4%	98.0%	1.3%
Grade 7	96.2%	2.9%	96.5%	2.7%	96.7%	1.9%
Grade 8	93.4%	4.7%	93.9%	4.3%	94.9%	3.8%
Grade 9	89.0%	9.3%	89.9%	8.9%	91.3%	6.8%
Grade 10	80.7%	15.6%	81.9%	15.3%	85.1%	12.4%
Grade 11	68.2%	25.0%	69.3%	23.7%	74.6%	22.4%
Grade 12	51.1%		52.9%		57.9%

Notes: Own calculations using General Household Survey datasets, STATS SA.
Three years of GHS data are pooled together with the purpose of increasing the sample size and therefore constructing more stable estimates over time. For the GHS datasets of 2015-2017, only persons born between 1991 and 1993 were included; for 2017-2019, only persons born between 1993 and 1995 were included, and for 2019-2021, only persons born between 1995 and 1997 were included. These age ranges were decided on in order to ensure that persons included would have been old enough to have been unlikely to still be completing school but young enough so as to reflect recent trends in school completion and dropout.

Table 2: Departure ratios at the end of 2018, 2019 and 2020 based on EMIS data

Source: DBE Analysis of EMIS data for 2018-2021