Journal of Education, 2016 Issue 66, http://joe.ukzn.ac.za Proxies and perplexities: What is the current state of adult (il)literacy in South Africa? John Aitchison (Received 3August 2015; accepted 26 November 2015) Abstract This article provides a detailed analysis of the data from a range of official sources that have been used to enumerate the number of people who can be described as totally or functionally illiterate and estimates whether illiteracy in South Africa can be reduced in the foreseeable future. The study examines the use of years of schooling (conventionally now set at Grade 7) as the proxy indicator of a person being functionally literate by the main sources, the General Population Censuses of 1996, 2001, and 2011 and the annual General Household Surveys and shows that these sources give somewhat contradictory and discordant estimates of the rate at which there is gradual decline of illiteracy in South Africa. Other indicators based on self reporting, also used in the Census and General Housing Surveys, show that a large number of adult South Africans have difficulty in reading, writing and calculating with numbers. The study also shows that the data presented by these surveys about participation in literacy and adult basic education and training classes is inaccurate. Note is made that currently South Africa does not make use of any means of direct testing of adult literacy. The article concludes with an exploration on whether South Africa is able to reach the goal of halving illiteracy by 2015. The target of such a reduction is necessarily based on what the baseline number of illiterates is as well as decision on whether full function literacy must be obtained or a merely a basic level of alphabetisation. Through a detailed estimation of the results of the Kha Ri Gude adult literacy campaign since 2008 a finding is made that the halving of illiteracy will be made, but only at the most basic level, and that attaining full functional literacy for all South Africans remains a major task. 112 Journal of Education, No. 66, 2016 What is the current state of adult (il)literacy in South Africa? The answers to this question: ! require an examination of a variety of data (much of it somewhat suspect) ! imply judgements about the successes and failures of the provision of adult education in the twenty years since South Africa became a democratic state ! provide the baseline data for the planning of the final eradication of illiteracy ! inform prognostications about how long that process will take. This article, which tries, inadequately, to answer the question, is broken down into two parts, the first looks at the data on illiteracy and the second at the whether the efforts to significantly reduce illiteracy have succeeded (or will succeed) in the near future. The data on adult (il)literacy levels The data on literacy levels in South Africa is derived from three types of data: ! proxy measures of literacy based on education (schooling) levels (recorded through census and other surveys) ! self-reports of literacy competency (usually collected through surveys) ! direct measures of literacy skills (usually done through smaller surveys or studies). Estimating literacy using proxy indicators Conventionally and until recently, UNESCO has recorded national statistics on illiteracy using proxy measures such as census data on education levels. Such proxy measures usually recorded “No schooling” as meaning totally Aitchison: Proxies and perplexities . . . 113 illiterate and “Below a certain grade level” as indicating functional illiteracy. In 1995 Harley et al. argued that in South Africa an adult with less than grade 7 was likely to be functionally illiterate (Harley et al., 1996, pp. 23-25), a position accepted by Statistics South Africa in its General Housing Survey reports (see Statistics South Africa, 2014, p. 24) and by the Department of Basic Education (see Department of Basic Education, 2014, p. 38). The estimates based on census data Table 1 summarises figures for the literacy and basic education levels of adult South Africans aged 15 and over, using data from the 1996, 2001 and 2011 General Population Censuses. (The age of 15 is chosen as the lower age limit because this is UNESCO usage in recording adult literacy levels.) Table 1 Literacy and basic education levels of South Africans aged 15 and over Level of education G eneral Population Census 1996 2001 2011 Full general education (grade 9 and more) 13.1 million (50% ) 15.8 million (52% ) 24.3 million (68% ) Less than full general education (less than grade 9) 13.2 million (50% ) 14.6 million (48% ) 11.5 million (32% ) None to less than grade 7 8.5 million (32% ) 9.6 million (32% ) 6.9 million (19% ) No schooling 4.2 million (16% ) 4.7 million (16% ) 2.7 million (8% ) These figures show that by 2001 there had been no decrease since 1995/1996 in the percentage of functionally illiterate adults (less than grade 7) and they had actually increased in raw numbers. Some 32% of the adult population of about 30½ million could therefore be regarded as functionally illiterate and the functional literacy rate (taking grade 7 education as the criterion of functional literacy) amongst the adult population had accordingly remained at 68% since 1996. There had been a very slight drop in the in the proportion of no 114 Journal of Education, No. 66, 2016 General Household Surveys are conducted annually by Statistics South Africa and collect 1 data from a sample of private households and residents in worker hostels in the whole country. The aim of the survey is to provide government departments and organisations with information on the progress of development in South Africa for monitoring and policy purposes. The survey covers six broad areas, namely education, health and social development, housing, household access to services and facilities, food security, and agriculture (Statistics South Africa, 2015, p. 10). A Community Survey, by comparison, is a mini census, undertaken in 2007 in South Africa, because of the large ten year gap between the 2001 census and the 2011 census. They also suggest lower overall population figures for South Africa than extrapolations of 2 the 2001 census data would predict. schooling illiterates but their raw number had increased. All provinces, except the Northern Cape had increasing numbers of adults with no schooling. These estimates assume the accuracy of the 2001 census figures, which also correspond to similar estimates in October Household Surveys from the late 1990s. The estimates informed the planning of the Kha Ri Gude literacy campaign (Aitchison, 2006). The corollary of these statistics is that the state system of adult basic education and training and its parallels in the business sector and non-governmental organisations had by 2001 failed to reduce the number or percentage of functionally illiterate people in South Africa. At best ABET provision was keeping the percentage of functional illiterates the same, though their raw numbers continued to grow. However, the situation shown in the 2011 census suggests a startling reversal of the growth in raw numbers of the poorly educated and a rapid decline in the percentage of those with less than grade 9 and grade 7. In addition, a new set of proxy based estimates now also challenge the 1996 and 2001 census based figures (and also, though to a lesser extent, the 2011 census data itself). The estimates based on post-2001 General household Surveys Subsequent to the General Population Census of 2001, a number of annual general household surveys present a suite of results of proxy-based estimates1 which generally suggest a lower and declining number of both total illiteracy2 (No schooling) and functional illiteracy (Less than Grade 7 amongst those Aitchison: Proxies and perplexities . . . 115 It needs to be noted that every year these figures are retrospectively revised downwards 3 (presumably by reweighting). The original General Household Survey figures in the report on each year from 2002 to 2013 have these No schooling figures: 11.8% , 11.2% , 10.8% , 10.4% , 10.5% , 9.3% , 8.7% , 7.4% , 7.0% , 6.6% , 5.8% , 5.6% . Only 2012 and 2013 remain unaltered. with only some primary education) as seen in Table 2 based on Figure 9 from the General Household Survey 2013 (Statistics South Africa, 2014, p. 23):3 Table 2 Literacy and basic education levels of South Africans aged 20 and over G eneral Household Surveys 2002 to 2013 Date of survey No schooling Some primary Both 2002 10.6% 17.0% 27.6% 2003 9.9% 15.8% 25.7% 2004 9.6% 15.5% 25.1% 2005 9.4% 15.0% 24.4% 2006 9.5% 14.3% 23.8% 2007 8.6% 14.3% 22.9% 2008 8.7% 13.9% 22.6% 2009 7.2% 12.3% 19.5% 2010 6.8% 12.1% 18.9% 2011 6.4% 11.5% 17.9% 2012 5.8% 10.9% 16.7% 2013 5.6% 10.7% 16.3% The General Household Survey report starts with an estimate of the No schooling group as having been 10.6% (3 016 000 people) in 2002 (indicating considerable divergence from the Census 2001 estimate of 17.9% for those aged 20 and over with No schooling) dropping to 5.6% (1 788 000 people) in 2013, both a percentage decline and a real number decline. The decline in numbers each year (except for 2006 and 2008 when there were small increases in the No schooling percentages) is erratic, ranging from 1% to 116 Journal of Education, No. 66, 2016 13%. Given that the explanation for the decline would presumably focus mainly on the fact that older people with no education would gradually be dying out and younger people who had benefited from the more general availability of schooling would be coming in, the erratic nature of the declines suggests some kind of error, whether of sampling or weighting (and retrospective reweighting – as it is clear that with each successive annual household survey the previous survey figures are retrospectively revised downwards (as shown in the example of the 2009 survey in the Table 3 below)). The 2013 general household survey claims a decline in total illiteracy (No schooling) and functional illiteracy (schooling of less than Grade 7) among adults aged 20 years and older from 27.3% in 2002 to 16.2% in 2013. Table 3 Percentage of adults (aged 20+) with less than Grade 7 education Year GHS of the year GHS 2009 GHS 2013 2002 28.9 27.9 27.3 2003 27.4 26.2 25.5 2004 26.5 25 2005 25.6 24.8 24.2 2006 25.1 23.7 2007 23.7 23.2 22.7 2008 22.4 22.4 2009 19.7 19.7 19.3 2010 19.2 18.8 2011 18.1 17.7 2012 16.5 16.5 2013 16.2 16.2 Aitchison: Proxies and perplexities . . . 117 Table 3 shows that the General Housing Surveys from 2002 to 2013 indicated a decline every year in the percentage of undereducated adults. The General Housing Survey of 2013 show how Statistics South Africa has also retrospectively re-weighted the figures for most years. The Department of Basic Education has made use of these General Household Survey reports and has issued reports based on their own analysis and calculations of the education statistics in the General Household Surveys from 2009 to 2013 (Department of Education, 2011, 2012, 2013, 2014) as well as a summary one on the 2002 to 2008 surveys (Department of Basic Education, 2010). Their latest summary (Department of Basic Education, 2014, p. 40) of literacy progress using the Grade 7 and higher proxy measure of functional literacy is shown in Table 4: Table 4 Percentage of adults aged 20 and above who have completed Grade 7 and higher, 2009 - 2012 Group 2009 2010 2011 2012 African 75.0 76.0 78.9 79.1 Coloured 83.9 85.5 72.3 86.4 Indian/Asian 95.5 92.1 92.6 92.0 W hite 99.8 98.8 98.0 97.5 All 79.4 80.0 80.7 82.2 [It might be noted that it is impossible for any country or population group in the world to actually have a 99.8% Grade 7 schooling achievement – because there is a percentage (usually at least 3% to 5% ) of every population which is largely uneducable because of major physical or mental handicap. Current estimates suggest that as many as half a million disabled South African children may not be in school (Human Rights W atch, 2015, pp. 74-75).] The Department of Basic Education’s General Housing Survey (GHS) 2012 Report – Focus on schooling (Department of Basic Education, 2014, p. 40) states: 118 Journal of Education, No. 66, 2016 As the Department of Basic Education (2014, p. 38) notes: “an adult who has completed 4 Grade 7 and above is regarded as literate. The completion of primary education is used as a proxy for measuring literacy; that is, it is assumed that the person is capable of reading, writing and doing some basic numeracy. This calculation is in line with the UNESCO Institute of Statistics calculations.” Overall the percentage of adults who are literate across all population groups has increased from 79% in 2009 to approximately 82% in 2012. This may be attributed to the introduction of Adult Basic Education and Training (ABET) now known as Adult Education and Training (AET), the Kha Ri Gude literacy campaign and other initiatives introduced to improve the literacy rates of adults by Government. This statement has a puzzling set of claims about the causes of the reduction in illiteracy (and ignores the most obvious one, the deaths of older illiterates). Firstly, the number of ABET learners had remained more or less static since the mid-1990s and the output of learners with a full ABET grade 9 equivalence was derisory (only 8 221 in the six years from 2001 to 2006). So- called AET is not the same as ABET as it includes both ABET (grade 1 to 9 equivalence) and adult Further Education and Training (grade 10 to 12 equivalence). Further, the Kha Ri Gude literacy campaign, whatever it achievements in reducing total illiteracy, has little impact on functional literacy statistics as its output level is grade 3 equivalence only. 4 Anomalies in the post-census 2001 surveys The anomalies in the post-census 2001 surveys are a cause for perplexity. Take the following example in Table 5 of a comparison between the Com m unity Survey 2007 and the General Housing Survey 2007 with the Census 2001 figure. The variations in the differences in the separate provincial declines in the number of illiterates are inexplicable (and they certainly cannot be explained away by assuming either systematic sampling, weighting or coding errors in the whole datasets for Census 2001, or the two 2007 surveys, or a selective genocide of illiterate adults over the age of 50). Aitchison: Proxies and perplexities . . . 119 Table 5 Decline in number of unschooled persons between 2001 and 2007 Census 2001 Community Survey 2007 Decline as % General Household Survey 2007 Decline as % KwaZulu-Natal 1 539 299 645 471 58% 617 000 60% Limpopo 858 681 511 714 40% 480 000 44% Eastern Cape 778 204 366 590 53% 399 000 49% Gauteng 515 747 277 285 46% 238 000 54% M pumalanga 468 747 307 740 34% 296 000 37% North W est 437 791 260 381 41% 230 000 47% Free State 257 140 132 110 49% 126 000 51% W estern Cape 167 619 94 724 43% 84 000 50% Northern Cape 91 304 76 358 16% 71 000 22% South Africa 5 114 532 2 672 373 48% 2 541 000 50% To be noted are the large declines in provinces with the highest illiteracy levels (KwaZulu-Natal, Limpopo and the Eastern Cape) and the tiny decline in the Northern Cape. Even more alarming are some other results in the General Household Survey 2009 (Statistics South Africa, 2010). For example, as shown in Table 6, it reports (p. 9) that there were only 18,000 person attending literacy classes at a time when the Kha Ri Gude adult literacy campaign had 613 643 learners registered that year (and of whom some 545 666 submitted their final 120 Journal of Education, No. 66, 2016 The Kha Ri Gude adult literacy campaign was launched in April 2008 after a two year 5 process of investigation and development. Some 357,195 learners were enrolled in 2008, 613 643 in 2009, 609 199 in 2010, 660 924 in 2011, 676 323 in 2012, and 562 926 in 2013. The South African Qualifications Authority has conducted an annual verification exercise involving a moderation of a sample of Learner Assessment Portfolios that are submitted from every learner and conveyed to the head office for moderation and verification. By the end of 2012 some 2 305 492 learners had passed this assessment and been recorded on the National Learner Record Database. The Auditor-General also conducted a verification exercise in 2014. Even allowing for some margin of error in the statistics it is incontestable that huge numbers of people participated in Kha Ri Gude classes. assessment portfolio at the end of the six month programme) (South African Qualifications Authority, 2010, 2013). 5 Table 6 Comparison between Kha Ri Gude registrations in 2009 and the General Household Survey 2009 tally of people in literacy classes Province Kha Ri Gude GHS 2009 GHS as % of Kha Ri Gude Eastern Cape 142 671 6 000 4.2% KwaZulu- Natal 133 486 2 000 1.5% Limpopo 103 828 4 000 3.9% Gauteng 75 678 1 000 1.3% Mpumalanga 55 971 < 1000 <1.3% Free State 50 984 4 000 7.8% North West 32 193 1 000 3.1% Western Cape 11 173 1 000 9.0% Northern Cape 7 654 < 1000 <13.1% 613 643 18 000 2.9% Even if one assumes that some literacy learners have been recorded as being in Adult Basic Education and Training Classes (the survey records 120 000 Aitchison: Proxies and perplexities . . . 121 The General Household Survey seems to have been somewhat confused (as many people 6 are) on the distinction between literacy classes and ABET classes. Buy even such confusion does not explain the overall miscount. attendees), this is still far short of the number of Kha Ri Gude registrations. It6 is a catastrophic miscount. So, on questions of accurate sampling, the General Household Survey of 2009 hardly inspires confidence. Table 7 shows the situation with the 2012 General Household Survey (Statistics South Africa, 2013, p. 9) and exposes an equally dire miscounting problem: Table 7 Comparison between Kha Ri Gude registrations in 2012 and the General Household Survey tally of people (aged 5+) in literacy classes Province Kha Ri Gude GHS 2012 GHS as % of Kha Ri Gude Eastern Cape 161 155 Less than 5 726 Less than 3.6% KwaZulu- Natal 148 687 Less than 1 688 Less than 3.9% Limpopo 109 035 Less than 1 002 Less than 0.9% Gauteng 88 821 Less than 1 543 Less than 1.7% Mpumalanga 56 907 Less than 659 Less than 1.2% Free State 52 892 Less than 397 Less than 0.8% North West 35 917 Less than 1 018 Less than 2.8% Western Cape 15 315 Less than 3 684 Less than 24% Northern Cape 7 589 Less than 162 Less than 2.1% 676 323 15 692 2.3% 122 Journal of Education, No. 66, 2016 So the general question arises: “If the General Household Survey can get this data so wrong, how can one trust any of these post-census 2001 surveys’ illiteracy statistics?” The problems with the two suites of proxy estimates Estimating South Africa’s illiteracy levels using proxy measures is complicated by what are now clearly two suites of statistics – the earlier Census and pre-2000 Household Surveys and the new set of Community and Household surveys. These latter surveys have come up with figures that suggest three options: ! the Census figures were wrong ! there has been a dramatic decline in illiteracy ! the recent survey figures themselves are inaccurate. Thus, for example, the 2007 Community Survey, found that the people with No schooling (the comparable cohort of people enumerated in the Census 2001 would now be aged 20 and over) totalled only 2,672,373 (an enormous difference of 2,048,253 from the census figure of 4,720,626). The October Household Survey of 2008 estimated that the Number of people adults aged 20 with No schooling was now about 2,451,856. This would suggest that the number of No schooling=Total illiterates would be about 9% rather than the 16% of the 2001 Census. If these new surveys, using proxy measures, are accurate, then the literacy problem is nearly halved. This may be considered good news, though the raw number of illiterate people is still a huge challenge, even if one now assumes that the actual number of functional illiterates may be less than the Census 2001 estimate of 9.6 million, possibly about 8.25 million (25% of adults) (Gustafsson et al., 2010, p. 21, p. 14). However, the continuing ambiguity of the statistics affirms the need for a well conducted literacy survey such as those run in Kenya and Botswana (which both showed the actual literacy levels to be lower than previously estimated) using direct measuring of literacy skills. Aitchison: Proxies and perplexities . . . 123 One needs to note the caution by Posel (2011, p. 43-44) that respondents my over-estimate 7 their literacy competence when faced with Yes/No options. Self-report measures of literacy Some recent South African community and household surveys have made use of self-report measures of literacy, though one should note the caveat of Gustaffson et al. (2009, pp. 3) that literacy data from a number of countries indicates that self-reported literacy rates tend to be higher than rates based on proxy measures and may lead to over-estimates of adult literacy. The General Household Survey of 2008 The General Household Survey of 2008 asked respondents whether they and their household members could read and write and found that 10.5% of adults aged 15 or over could not read or write, slightly more than the same survey’s proxy measure estimate of total illiteracy at about 8%.7 Table 8 Cannot read or write. General Household Survey 2008 Age group M ale Fem ale Both As % 15-19 86000 48000 133 000 3.8% 20-24 70000 44000 114 000 3.3% 25-29 92000 67000 158 000 4.5% 30-34 96000 95000 191 000 5.5% 35-39 84000 125000 209 000 6.0% 40-44 92000 165000 256 000 7.3% 45-49 138000 212000 350 000 10.0% 50-54 138000 257000 396 000 11.3% 55-59 153000 243000 396 000 11.3% 60-64 125000 227000 353 000 10.0% 65-70 127000 223000 349 000 10.0% 70-74 94000 177000 270 000 7.7% 75+ 100000 226000 326 000 9.3% Totals 1 394 000 2 108 000 3 501 000 100.0% 124 Journal of Education, No. 66, 2016 Of interest is that the raw number of illiterates peaks amongst people in their fifties and that adult illiteracy will, in the future, become predominantly a male problem, both trends evident from the Figure 1. Figure 1 These findings are replicated in the General Household Surveys in subsequent years. The General Household Surveys from 2009 to 2013 The General Household Surveys from 2009 to 2013 asked respondents a more nuanced set of questions about whether those aged 15 or more and with less than Grade 7 education could write their own name, read, fill in a form, write a letter, calculate the change they should receive, and read road signs on a scale of No difficulty/ Some difficulty/ A lot of difficulty/ Unable to do (Statistics South Africa, 2011, 2012, 2013). Aitchison: Proxies and perplexities . . . 125 Table 9 Literacy and numeracy skills of adults aged 15 + with no or less than Grade 7 level of education [General Housing Surveys 2009 to 2013] 2009 2010 2011 2012 2013 R eading Reading N o difficulty 50% 52% 52% 51% 50% Some difficulty 14% 11% 10% 10% 10% A lot of difficulty 10% 9% 9% 9% 10% U nable to do 27% 28% 29% 30% 30% Reading road signs N o difficulty 49% 48% 51% 54% 58% Some difficulty 15% 13% 12% 12% 11% A lot of difficulty 9% 11% 9% 9% 8% U nable to do 27% 28% 28% 26% 25% W riting W riting name N o difficulty 72% 71% 70% 70% 70% Some difficulty 5% 4% 5% 4% 4% A lot of difficulty 4% 4% 4% 4% 4% U nable to do 19% 20% 21% 21% 22% Filling in a form N o difficulty 32% 28% 32% 29% 29% Some difficulty 16% 15% 13% 14% 13% A lot of difficulty 14% 16% 15% 15% 14% U nable to do 38% 41% 40% 42% 44% W riting a letter N o difficulty 47% 47% 47% 48% 47% Some difficulty 13% 10% 11% 10% 9% A lot of difficulty 10% 10% 10% 10% 9% U nable to do 30% 32% 33% 32% 34% N umeracy Calculating change should receive N o difficulty 67% 71% 73% 74% 76% Some difficulty 11% 9% 8% 8% 7% A lot of difficulty 6% 6% 5% 5% 4% U nable to do 16% 14% 13% 13% 12% 126 Journal of Education, No. 66, 2016 However one needs to note the caveat from Gustafsson (2012) who states: 8 ... Very importantly, the way people judge their own level of literacy in these kinds of questions where you tell a fieldworker what you can do (as opposed to show what you can do, for instance in a test) seems to be unstable over time. I say this because people with the same level of education, in terms of highest grade attained, provide changing judgements of their level of literacy. Specifically, it seems South Africans have become more demanding of themselves, so over time they become less and less inclined to say they are literate, when controlling for grade. Past standardised tests have indicated that literacy levels by grade have remained more or less unchanged, so people with the same level of education are not becoming less literate in reality, it seems. I suspect that greater social demands to be literate are making people more aware of their own limitations. W hat does this mean for interpreting the data? Firstly, one needs to be extra careful when making comparisons over time. No change could in fact mean there’s an improvement, but people are not acknowledging this because they are becoming stricter on themselves. Secondly, I think we should present literacy statistics by level of schooling attained to make it easier to identify the anomalies I’m referring to. The Kha Ri Gude literacy campaign CEO, M cKay (2012b) commenting on these two 9 indices noted that: when we designed the questions I had hoped for a more nuanced interpretation. because we used the competences of the B otswana, Kenyan and LAM P surveys as indicators of different levels of competence so that we could see literacy/illiteracy as a continuum. T he approximate levels that could be denoted - based on the other surveys (which were direct testing) were: Level 0: the learner cannot write his/her name Level 1: the individual can write his/her name, demonstrates some form of emergent literacy and can read shop and road signs or labels on a package. Level 2: considered a suitable minimum for coping with demands of daily life. Can read a newspaper or book or other extended text. Level 3: Can do the above and can also produce text in the form of a letter or form. Respondent demonstrates higher-order information processing skills. Calculating change would say a lot about a degree of numeracy. Alarmingly these figures suggest a possibly worsening illiteracy problem amongst people with less than a grade 7 level of education – in 2008 some 27% were unable to read at all, in 2013 it was 30%; in 2008 some 38% could not fill in a form, in 2013 it was 44%. This may correlate with the known8 prevalence of functional illiteracy amongst children at school and people who have been through primary school (Smith-Greenaway, 2015). Statistics South Africa developed two possible indices of subjective literacy based on these questions in the 2009 to 2011 surveys. The first index took the No difficulty and Some difficulty answers to the questions on whether respondents could read (a newspaper or book) and write a letter. The second index added the question on filling in a form (Roux, 2012).9 Aitchison: Proxies and perplexities . . . 127 In terms of raw numbers, taking the components used in these two indices (reading, writing a letter and filling in a form), those who cannot do these or only with great difficulty are as follows: Table 10 Functionally illiterates among those aged 15+ with less than Grade 7 [General Housing Surveys 2009 to 2013] 2009 2010 2011 2012 2013 Read U nable to do 1536000 1 530 000 1 582 000 1 498 000 1 542 000 A lot of difficulty 538000 517 000 491 000 467 000 485 000 T otals 2 074 000 2 047 000 2 073 000 1 965 000 2 027 000 W rite a letter U nable to do 1 662 000 1 760 000 1 786 000 1 641 000 1 737 000 A lot of difficulty 577 000 571 000 533 000 510 000 473 000 T otals 2 239 000 2 331 000 2 319 000 2 151 000 2 210 000 Fill in a form U nable to do 2 087 000 2 249 000 2 198 000 2 122 000 2 261 000 A lot of difficulty 800 000 877 000 807 000 747 000 700 000 T otals 2 887 000 3 126 000 3 005 000 2 869 000 2 961 000 These numbers are more or less consistent with these surveys’ estimates of the numbers of number of people with no education or very low levels of education. Direct measures of literacy UNESCO’s Institute for Statistics is now demanding more direct measures of literacy from member countries (a good African example of such being the Kenya National Literacy Survey of 2007) (Aitchison and Alidou, 2009, p. 26) that will lead to better aligned policies. Gustafsson et al. (2010, p. 3) argue that South Africa should periodically test adult literacy levels and would be relatively inexpensive relative to the cost of adult education programmes. 128 Journal of Education, No. 66, 2016 The Kha Ri Gude literacy campaign had engaged with Statistics South Africa about testing 10 General Housing Survey 2009 respondents through reading a paragraph and writing a paragraph but the cost implications proved difficult and the suggestion to have a separate module during the fieldwork to focus on literacy did not materialise (M cK ay, 2012a). Gustafsson et al. (2010, p.4) estimate that “ if the quality of schooling in South Africa were 11 where it should be (at a level befitting a middle income country), GDP would be R550 billion higher than it currently is, or 23% above the current level. .. poor quality schooling at the primary level, which increases adult illiteracy in future decades, is undoubtedly a large, and arguably the largest, inhibitor of South Africa’s growth and development.” In South Africa the aforementioned 2008 October Household Survey and the General Household Survey 2009 did ask questions about whether household members could read or write (some 3,501,000 adults could not in 2008) and perform certain literacy skills (1,536,000 could not read in 2009) but these were not direct qualitative test-based measures. 10 However a number of direct measures of reading, writing and calculation skills directed at South African schoolchildren have indicated that, when tested with widely used international instruments, the South African schooling system is underachieving in literacy and numeracy (even when compared with other far poorer countries in the rest of Africa) (see the summary of these studies in the EFA Country Report of 2009 (Department of Basic Education, 2009, pp. 29-32)). Though the testing of reading and mathematical competencies amongst schoolchildren does not directly tell us about adult illiteracy, it does show that “there is a low correspondence between grade level and literacy level” (Posel , 2011, p. 41) and explain why many schoolchildren graduate into functionally illiterate adults. This would suggest11 that South Africa’s rate of functional illiteracy amongst adults will indeed be very high (indeed higher than the General Household Survey estimates since 2009 using proxy measures). Gustafsson et al. (2010, p. 39) note that one interesting measure is how may hours an adult spends reading each week and suggest that less than a hour could be a useful measure of functional illiteracy (and in which case about 25% of South African adults are functionally illiterate). Aitchison: Proxies and perplexities . . . 129 The Dakar Framework for Action has two of its six goals address the themes of adult 12 learning. These are Goal 3 – ensuring that the learning needs of all young are met through equitable access to appropriate learning and life-skills programme, and Goal 4 – achieving an improvement in levels of adult literacy by 2015 especially for women and equitable access to basic and continuing education for all. Reducing illiteracy – have the plans to halve illiteracy by 2015 worked? In 2000 South Africa committed itself to the revised (Dakar) Education for All goal of a 50% reduction in illiteracy by 2015. Has this commitment been12 adhered to? Setting targets In planning for the growth of a fully literate society one first has to make some further estimates of what reduction in illiteracy is already being made by existing programmes (and of course the death of elderly illiterates) and then, on the basis of the data gained from the various measures of literacy (described above), define the target numbers before planning programmes to achieve these literacy goals. As already indicated, the process of working out how many people in South Africa are illiterate is not an exact science. Numerous caveats and quibbles must be made about the data from proxy and self-reported measures. In addition all countries have a percentage of the population of sub-normal intelligence or severe mental or physical handicap and it is unrealistic to think that any but those countries with very high budgetary commitment to education can address their needs (in passing it is this problem that renders suspicious any country’s claim to have a literacy level that rises much above the 95% mark). But the bottom line is that, however one massages these estimates, South Africa definitely has several million illiterate adults who could benefit from literacy and/or adult basic education provision. Apart from contending with the divergencies between the census 2001 and 2011 and more recent household survey estimates, another, more pedagogical issue intrudes – what intervention genuinely guarantees the lifelong acquisition of literacy skills? Whilst people with no schooling can safely be 130 Journal of Education, No. 66, 2016 An Auditor-General of South Africa report (2014) notes both a 52% decline in Public Adult 13 Learning Centre enrolments between 2000 and 2013 (p. 11) and a failure to have Kha Ri Gude learners progress to ABET level 2 (grade 5 equivalency) “and continue to achieve functional literacy” (p.12). assumed to be illiterate, how many years of schooling is likely to result in the lifelong retention of literacy competency? Harley et al. (1996) in their major survey of South African literacy and adult basic education used Grade 7 as the retention indicator (a position that has been accepted by Statistics South Africa), but possibly many of those who dropped out of school before reaching grade 7 may well be functionally literate (Posel, 2011, p. 40; Smith- Greenaway, 2015). The Department of Higher Education used a more ambiguous indicator for functional literacy of “both read and write with understanding a short simple statement on their everyday life” (Auditor General of South Africa, 2014, p. 18). The major South African intervention to reduce illiteracy amongst adults, the Kha Ri Gude literacy campaign, only claimed to reach a grade 3 equivalence (ABET level 1) and the campaign drew attention to the need for follow up via the state system of Public Adult Learning Centres (which did not happen). 13 This leaves one in something of a quandary. If the grade 7 indicator is used then there is no prospect of South Africa reducing illiteracy by 50% by 2015. However, if the Kha Ri Gude outcome of grade 3 equivalence is accepted, the target may well be met, and evidence for this is analysed below. What is a realistic target? Realistic targets have to factor in a number of complex variables, that will reduce the number of people to far less than the raw statistics suggests. To start with, how many of the potential candidates are in effect uneducable – because of sub-normal intelligence or severe handicap? Probably about 1.7 million adults. About 2.4 million adults are over the age of 64. Should one invest in people this old? It may not be directly of economic benefit but many of the elderly play a crucial childcare role in South Africa and more literate grandparents may have positive long term effects. An obvious first target is to make literate the over 2 million unschooled adults even though they are a hard to reach group and many of the truly uneducable Aitchison: Proxies and perplexities . . . 131 1.7 million adults are in this group. The other obvious targets are those who dropped out of school after a year or so. It is likely that the bulk of the reachable target will be people who, for whatever reasons, dropped out or were extruded from schooling very early on. (The Kha Ri Gude literacy campaign statistics for 2008 show that unschooled people made up 36% of the learners and the largest group (48%) had only a year of schooling.) Making a very rough calculation on the basis of there being probably at least 8.25 million functionally illiterate people, one can deduct the 1.7 million uneducable and be left with a overall figure of 6.55 million. Half of this is 3.27 million. That would then be the initial Dakar goal related target. However, that 3.27 million would have to be the actually achieved number. Because of some inevitable drop-out and attrition from intervention programmes this requires a larger expanded target than 3.27 million. Thus, for example, the Ministerial Committee on Literacy report of 2006, which argued for a target of 3.8 million, developed a plan to enrol 4.7 million illiterate adults by the end of 2012 (Ministerial Committee on Literacy, 2006, pp. 17-18). What progress has there been towards such a target? Has the current provision of literacy and adult basic education successfully reached over 4 million people? The Department of Basic Education (2014, p. 40) makes the, already noted, claim: Overall the percentage of adults who are literate across all population groups has increased from 79% in 2009 to approximately 82% in 2012. This may be attributed to the introduction of Adult Basic Education and Training (ABET) now known as Adult Education and Training (AET), the Kha Ri Gude literacy campaign and other initiatives introduced to improve the literacy rates of adults by Government. Let us start with the Kha Ri Gude literacy campaign as it operated from 2008 to date rather than originally planned model (which aimed to enrol 4.7 million by 2012 with double the funding that was eventually granted to the campaign). A number of technical points need to be made. The majority of Kha Ri Gude participants have had no, or virtually no, schooling – Kha Ri Gude is targeting the genuinely functionally illiterate. In the first year of operation (2008) participants who completed the programme engaged in a formal externally 132 Journal of Education, No. 66, 2016 validated assessment process (overseen by the South African Qualifications Authority) and of these a negligible number failed. The results from 2009 to 2013 were also externally validated and were equally positive. The graph below shows the enrolments, completions and certifications of success for the campaign for the years 2008 to 2014 (the figures for 2014 are estimates). Figure 2 The cumulative totals are 4 111 080 enrolled, 3 326 216 completed, and most of those who completed were certificated as successful – 3 200 000. This is close to the campaign plan’s target of 3.8 million. Aitchison: Proxies and perplexities . . . 133 The Department of Basic Education (2014, p. 40) statement that Adult Basic Education and 14 Training (ABET) is now known as Adult Education and Training (AET) betrays a major conceptual confusion (in both the Department of Basic Education and the Department of Higher Education and Training). ABET refers to an adult equivalent of compulsory general school education and in South Africa that is still currently school grades 1 to 9. AET refers to all adult education at any level, both formal and non-formal, and therefore cannot be used as a synonym for formal ABET (or even ABET and Further Education and Training (FET)). Sadly, Kha Ri Gude now appears to be facing serious mismanagement problems and the 15 enormous potential it had to continue as a vehicle to deliver post-literacy education and training largely compromised. We also have to take into account the learners who gain basic literacy through other literacy and Adult Basic Education and Training (ABET) programmes.14 So far their impact on reducing illiteracy levels has been pretty insignificant (Gustafsson et al., 2010, p. 15). Table 13 in the EFA Country Report for 2009 records the long standing inability of the Department of Education to disaggregate figures for genuine ABET (up to NQF level 1) and those for students studying for the Senior Certificate (NQF level 4). One cannot therefore use this table to make reliable estimates on the ABET contribution to reaching the 50% reduction in illiteracy. Gustafsson et al. (2010, p.15) note a survey finding of about 50,000 per annum in ABET classes. The Independent Examinations Board had an average of about 40 000 examination entries over the years 2007 to 2009 (but these come from individuals writing several course examinations) in industry and SETA sponsored programmes. The General Household Survey of 2009 estimated 120,000 learners in “ABET” classes. Probably one can be generous and estimate 100 000 learners a year in ABET levels 1 to 4 programmes – and also assume, also overgenerously, a 50% success rate. Whatever the limitations of ABET provision, this output would certainly take the overall Kha Ri Gude plus ABET class cumulative total past the 3.8 million target. South Africa can make, indeed has made, the 50% reduction in total illiteracy. This achievement, albeit much of it only to the most basic level of alphabetisation, could never have been done without the Kha Ri Gude campaign’s first six years of work which were a model of effective service delivery (Aitchison and McKay, 2014) and women will have been the main beneficiaries of the progress made. 15 A caveat to this success is that basic literacy acquisition can be rapidly undermined by lack of use and further development of reading, numeracy and 134 Journal of Education, No. 66, 2016 associated skills. Unfortunately post literacy provision shows few signs of being geared up adequately. There is little evidence that the national Department of Higher Education and Training has geared up or is capable of further servicing the graduates of the Kha Ri Gude campaign (in spite of the Ministerial Committee on Literacy having warned about this back in 2006) (Ministerial Committee on Literacy, 2006, pp. 52-53). The EFA Country Report of 2009 very frankly acknowledged the failure of the ABET system to deliver on scale (only 8 152 learners had exited fully qualified from the system since 2001) and states that “As a mechanism for addressing the learning needs of adults with no or little basic education the apparatus of public adult education governance, provision, curriculum and support has evidently proved unequal to the task.” (Department of Basic Education, 2009, pp. 22–23). The current policy development and plans for a new system of community colleges and community learning centres is still too insubstantial to be evaluated as to its potential for making South Africa a more literate society. Can any conclusions be made? This article has attempted to analyse the available sources on adult literacy statistics in South Africa and found a fair degree of inconsistency among them. Some discrepancies are to be expected, particularly with reading and writing and their use in the daily functioning of people in a very complex society with high disparities in education and wealth and general living conditions. In addition we have an evolution of what kind of literacy or so- called literacies are needed to function in modern society. However, one can safely say that a more thorough form of direct testing of literacy capabilities is needed for South Africa to have a more reliable set of baselines from which to improve. Although some of the data sources (particularly the General Housing Surveys) have indicated that illiteracy (using Grade 7 schooling or equivalent as the proxy for functional literacy) is less of a problem, percentage wise, in raw numbers South Africa still has an embarrassingly large number of illiterate people. I would personally estimate that we have a practical target of about 5 million people still to be reached. It is clear that without the efforts of the Kha Ri Gude literacy campaign the number of totally or near totally illiterate people would be about three million people larger. But the follow up to this campaign has been more or less non- Aitchison: Proxies and perplexities . . . 135 existent. One can safely predict that unless something radical happens to the support and funding of literacy and adult basic education provision for undereducated South Africans, this will take a long time. References Aitchison, J.J.W. and Harley, A. (2006). South African illiteracy statistics and the case of the magically growing number of literacy and ABET learners. Journal of Education, 39, 89–112. Aitchison, J.J.W and McKay, V.I. (2014). Greening the data desert – a case study of South Africa’s Kha Ri Gude mass literacy campaign. Unpublished paper. Auditor-General of South Africa. (2014). Discussion extract of the management report of the Auditor-General of South Africa on the performance audit of the Department of Higher Education and Training. Pretoria: Auditor-General of South Africa. Department of Basic Education. (2009). Education for all (EFA) country report, 2009: South Africa. Pretoria: Department of Basic Education. Department of Basic Education. (2010). Report on the general household survey (GHS): Education focus. Pretoria: Department of Basic Education. Department of Basic Education. (2011). Report on the 2009 general household report: Focus on schooling. Pretoria: Department of Basic Education. Department of Basic Education. (2012). General household survey (GHS) 2010: Focus on schooling. Pretoria: Department of Basic Education. Department of Basic Education. (2013). General household survey (GHS) 2011: Focus on schooling. Pretoria: Department of Basic Education. Department of Basic Education. (2014). General household survey (GHS) 2012 report: Focus on schooling. Pretoria: Department of Basic Education. 136 Journal of Education, No. 66, 2016 Education for All. (2000). Dakar framework for action. Paris: UNESCO. Gustafsson, M., Van der Berg, S., Shepherd, D., and Burger, C. (2010). The costs of illiteracy in South Africa. Stellenbosch: University of Stellenbosch. Gustafsson, M. 2012. Personal communication (email) to McKay, V., 23 April 2012. Harley, A., Aitchison, J.J.W., Lyster, E. and Land, S. with Fotheringham, R., Hallowes, D., Hargreaves, L., McCubbin, C., Tothill, S., and Tuchten, G. (1996). A survey of adult basic education in South Africa in the 90s. Johannesburg: SACHED Books. Human Rights Watch. (2015). “Complicit in exclusion” South Africa’s failure to guarantee an inclusive education for children with disabilities. Johannesburg: Human Rights Watch. McKay, V.I. (2012a). Personal communication (email) to Libago, J., 18 April 2012. McKay, V.I. (2012b). Personal communication (email) to Libago, J., 22 April 2012. Ministerial Committee on Literacy. (2006). Ministerial committee on literacy: Final report. 19 June 2006. Pretoria: Department of Education. Posel, D. (2011). Adult literacy rates in South Africa: A comparison of different measures. Language Matters 42(1), 39–49. Roux, N. (2012). Personal communication (email) to Narsee, H. and Libago. J., 17 April 2012. Smith-Greenaway, E. (2015). Educational attainment and adult literacy: A descriptive account of 31 Sub-Saharan Africa countries. Demographic Research 33, article 35, 1015–1034. South African Qualifications Authority. (2009). SAQA’s role in the pilot implementation of the Kha ri gude adult literacy campaign. Pretoria: South African Qualifications Authority. Aitchison: Proxies and perplexities . . . 137 South African Qualifications Authority. (2010). Ascertaining the integrity of the marking of the Kha Ri Gude literacy campaign Learner Assessment Portfolios Year Two: 2009. Pretoria: South African Qualifications Authority South African Qualifications Authority. (2011). Ascertaining the integrity of the marking of the Kha Ri Gude literacy campaign Learner Assessment Portfolios Year Three: 2010. Pretoria: South African Qualifications Authority. South African Qualifications Authority. (2012). Ascertaining the integrity of the marking of the Kha Ri Gude literacy campaign Learner Assessment Portfolios Year Four: 2011. Pretoria: South African Qualifications Authority South African Qualifications Authority. (2012). Ascertaining the integrity of the marking of the Kha Ri Gude literacy campaign Learner Assessment Portfolios Year Five: 2012. Pretoria: South African Qualifications Authority. South African Qualifications Authority. (2012). Ascertaining the integrity of the marking of the Kha Ri Gude literacy campaign Learner Assessment Portfolios Year Six: 2013. Pretoria: South African Qualifications Authority. South African Qualifications Authority. (2012). Ascertaining the integrity of the Kha Ri Gude literacy campaign Learner Assessment Portfolios Year Seven: 2014. Pretoria: South African Qualifications Authority. Statistics South Africa. (2003). General household survey July 2002 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2004). General household survey July 2003 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2005). General household survey July 2004 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2006). General household survey July 2005 (Statistical release P0318). Pretoria: Statistics South Africa. 138 Journal of Education, No. 66, 2016 Statistics South Africa. (2007). General household survey July 2006 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2007). (Statistical release P0301). Pretoria: Statistics South Africa. Statistics South Africa. (2008). General household survey 2007 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2009). General household survey 2008 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2010). General household survey 2009 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2011). General household survey 2009 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2011). General household survey 2010 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2012). General household survey 2011. Revised. (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2013). General household survey 2012 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2014). General household survey 2013 (Statistical release P0318). Pretoria: Statistics South Africa. Statistics South Africa. (2015). General household survey 2014 (Statistical release. John Aitchison Professor Emeritus of Adult Education at University of KwaZulu-Natal aitchisonjjw@gmail.com mailto:aitchisonjjw@gmail.com