IS THE PROPORTION OF MATRICULANTS WITH POST-MATRIC QUALIFICATIONS DECLINING?

He [the Statistical General] said parents were better skilled compared with their children, arguably owing to changes in the education system after 1994. “What we know is that the nursing schools, the teacher training schools and where people used to do trades, [those] schools were closed and they were converted into part of the university system. The net result has been they, those who are 15 to 34, do not have the skills of their parents, so hence the skills crisis, in part, in the country.” – Mail and Guardian, 18 June 2014

1. The importance of avoiding faulty analysis

At first sight, it might appear that post matric training is in relative decline. The 2016 Community Survey enables one to calculate, for each year of age, the proportion of people with Grade 12 who have post-matric education. Figure 1 presents the results, with age converted into date of birth.

The decline seems to be long-standing and substantial. But Figure 1 clashes with other information. For example, the number of graduates from public universities between 1990 and 2014 increased by 3.9% per annum on average, comfortably ahead of the 1.7% growth in Grade 12 enrolments. And there is a substantial contribution by private sector universities and colleges, not measured by the Department of Higher Education and Training.

It turns out that Figure 1 is misleading, for two reasons.

The first is that the acquisition of post-matric qualifications takes time. Based on the increase in first post-matric qualifications between the censuses of 2001 and 2011, the cumulative proportion of qualifications received at successive ages can be calculated. The results are presented in Figure 2.

Figure 2 shows that the cumulative percentage does not reach 95% until age 35. Younger age groups are still in the process of acquiring post-matric qualifications. So one would expect that the curve in Figure 1 would slope downwards after the 1980 birth cohort. At a future date, the curve would be higher in this range, the more so for later cohorts.

However, this observation does not explain the decline between the 1960 and 1980 birth cohorts. There is a second reason why Figure 1 is defective. To qualify for post-matric education, one needs not only to have reached Grade 12, but also be in a possession of a (National) Senior Certificate. Had the NSC pass rate been constant, the curve in Figure 1 would have been the same shape, but at a somewhat higher level. But the pass rate has not been constant, as was shown in Youth Brief 1. If the denominator is adjusted to reflect the (N)SC pass rate twenty years after birth, a modified estimate of the proportion of successful NSC candidates who attained post matric education can be calculated. Figure 3 shows the result.

Figure 3 changes the conclusion dramatically. About 40% of people born before 1957 with (National) Senior Certificate obtained a post-matric qualification. The percentage then increased rapidly over a couple of years, and for those born between 1960 and 1980, it fluctuated between 50% and 60%.

2. A more detailed account of post-school education since 1970

It is possible to gain a more detailed understanding of what has happened since 1970, by comparing the results from successive censuses and community surveys. A tabulation by five year age groups is set out in the Appendix.

Table 1 sets out the population with post-school education in each year, along with the growth rates in the interval between the years.

Table 1 - Population reporting post-school education (PSET), 1970 to 2016

	Population	Growth rate
1970	323 512
1980	697 730	8,0%
1985	966 930	6,7%
1991	1 141 478	2,8%
1996	1 434 508	4,7%
2001	2 139 785	8,3%
2006	3 493 508	8,5%
2011	4 139 876	4,3%
2016	4 368 497	1,1%

Estimates based on self-reporting, especially when adjusted, are open to question. So, are the estimates in Table 1 plausible?

One way of testing them is to compare:

1. The growth of National Senior Certificate passes year by year

2. The growth of the population age 35-39 between census and community survey years seventeen years later, on the assumption that the average age when qualifying for the NSC is 20.

3. The growth in births twenty years earlier.

Figure 4 graphs the three superimposed series with the NSC pass date as the reference date.

Figure 4 shows that the gap between the growth of the birth cohort size and the growth of both NSC passes and people reporting post-school education was very large in the second half of the twentieth century. Moreover, there is considerable alignment between the growth of NSC passes and the growth of people reporting post-school education. The reason why the growth in post-school education was higher than NSC passes in the early years was that, given the low level of NSC passes then, many people with Grades 8 to 10 received post-school certificates and diplomas in those years. Similarly, remembering that the PSET graphed against 1968-1974 refers to the position seventeen years later, the consequent low growth in PSET reflects weak economic growth in the late 1980s. Thereafter, the NSC and PSET curves are more closely aligned.

Accordingly, the low growth of PSET between 2011 and 2016 reflects the low growth in NSC passes in the late 1990s, rather than a falling off the proportion of people with NSC going on to post-school education. Figure 3 suggests that the growth rate in PSET can be expected to pick up again in the next fifteen years, though never again to the high levels experienced between 1950 and 2000, because cohort sizes have stabilized and the proportion of learners reaching Grade 12 and passing the NSC is approaching its upper limit.

Charles Simkins

Head of Research

charles@hsf.org.za

Appendix

Estimates of the population with post-school education by five year age groups

The raw numbers in the sources have been adjusted in two ways:

1. The old homelands of Transkei, Bophuthatswana, Venda and Ciskei were excluded from the 1980, 1985 and 1991 censuses. Separate censuses were taken in Bophuthatswana, Venda and Ciskei in 1991. For other years, estimates have been made by interpolation.

2. Compared with the 2015 revision of the United Nation’s World Population Projections, it appears that the population in older age groups was under-enumerated in the 2011 census, and more strongly in the 2016 community survey. Statistics South Africa weights the community survey population in accordance with its unpublished demographic model. The 2016 average weights per age group fall off sharply at older ages (they should be more or less flat), suggesting that the demographic model is based on mortality estimates which are too high. Accordingly, the age groups are re-weighted to accord with the United Nations estimates.

Table 1 - Post-school qualifications

	1970	1980	1985	1991	1996
15-19	4575	5618	11190	8545	13435
20-24	47752	95684	123148	132847	161503
25-29	62068	140575	190278	218456	271510
30-34	50290	124038	165976	211317	262371
35-39	38435	100317	140427	169739	230337
40-44	33442	73597	108795	132981	172313
45-49	29263	53315	81438	103811	122704
50-54	22713	43697	59807	75370	88651
55-59	19065	35729	48088	52091	64360
60-64	15909	25158	37783	36321	47322

Total	323512	697730	966930	1141478	1434508
Annual growth: 15-64		8,0%	6,7%	2,8%	4,7%
Annual growth: 35-39		10,1%	3,4%	1,9%	3,1%

Table 1 continued	2001	2007	2011	2016

15-19	39296	89035	66255	33721
20-24	277998	495170	465118	413685
25-29	401348	565205	736853	720232
30-34	379077	575053	677062	660858
35-39	314535	526490	586155	595398
40-44	254346	407921	495620	527114
45-49	183647	322129	392939	492624
50-54	129249	231584	322103	386581
55-59	93745	163747	236651	317726
60-64	66545	117175	161120	220558

Total	2139785	3493508	4139876	4368497
Annual growth: 15-64	8,3%	8,5%	4,3%	1,1%
Annual growth: 35-39	3,2%	5,3%	1,1%	0,2%