Valuing and revaluing education: what can we learn about measurement from the South African poor?

ABSTRACT

This paper reflects on the identification of relevant aspects of education for measurement purposes. It begins by reviewing some detailed lists of educational capabilities from disparate literatures. It then considers how ordinary South Africans perceive education by drawing on two open-ended surveys, and attempts to reconcile their views with different education lists. The main finding is that most abstract lists need to say more about the practical side of education (skills, information and knowledge for everyday living). They also need to embrace a more joined-up view of education that can incorporate linkages between different aspects of education and between education and other aspects of well-being (including mental states and material things). The final part of the paper makes the case for embracing the complexity and imprecision involved in measuring education and briefly sketches a methodological framework that can achieve this end.

KEYWORDS:

• Capability lists
• education
• practical skills
• complexity
• vagueness
• well-being
• South Africa

Acknowledgements

I am grateful to Elaine Unterhalter for encouragement and helpful comments, and to Shailaja Fennell, Jaqui Goldin, Peter Nolan, Mozaffar Qizilbash and the late Dudley Horner for many useful conversations over the years. I would also like to thank two anonymous referees for helpful comments.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes on contributor

David A. Clark is a Teaching Associate and Affiliated Lecturer in the Centre of Development Studies at the University of Cambridge. His research focuses on poverty and well-being from a human development perspective. His books include Visions of Development: A Study of Human Values (2002), The Elgar Companion to Development Studies (2006), Adaptation, Poverty and Development: The Dynamics of Subjective Well-Being (2012), and The Capability Approach, Empowerment and Participation (co-edited, forthcoming).

Notes

1. Sen’s approach is relevant because people have reason to value a broad range of capabilities rather than (just) income and happiness (see Clark 2002, 29–34).

2. My precis of Nussbaum’s (2006) capabilities differs slightly from Biggeri and Santi (2012).

3. In this respect, a systematic comparison between the ideal-theoretical list and the capabilities and values that emerge from the fieldwork would be insightful. This could have been complemented with an attempt to reconcile any differences in emphasis or substance, and to justify the capabilities that are included and excluded from the final list.

4. A strict comparison of the theoretical list with the final fieldwork influenced list reveals few substantive changes and identical wordings across domains (compare the final column of table 1 with table 3 in Wilson-Strydom 2016). The only apparent addition, is that the respect category has been augmented to include ‘receiving respect from others’ which echoes my own fieldwork findings (Clark 2002).

5. The adult list does include a reference to ‘lifelong learning’ that has no equivalent counterpart in the children's list.

6. A small number of interviews (n = 36) are excluded from the sample on the grounds that the wrong person was selected for interview.

7. The relationship between values, education and the formation of capabilities has been discussed by Vaughan and Walker (2012). In this paper, I refrain from entering this debate.

8. Similar approaches have been independently developed by Biggeri et al. (2006) and Biggeri, Ballet, and Comim (2011, chapter 4). Table 4 suggests that participants were able to draw meaningful distinctions between thresholds for ‘getting by’ and ‘living well’.

9. Initially attempts to apply this methodology to the ELS provide mixed results that are not easy to interpret. It is fairly clear, however, that people do not comprehensively adapt to straitened circumstances. Insofar as adaption takes place, it appears to vary across domains and is a partial process at best. For example, there is some evidence of adaptation in the housing and income domains, but not in the health domain. The results for education are harder to interpret, but overall ‘are not suggestive of adaptation’ (Clark and Qizilbash 2008, 539). More rigorous econometric analysis suggests a small positive correlation between actual and aspired education, but with several caveats (see Barr and Clark 2010). It is worth remembering that many adjustments to preferences (that happen to be consistent with this trend), do not reflect habituation or resignation to fate (see Teschl and Comim 2005, 233; Clark 2012, 72ff.).

10. Participants were asked to nominate up to five priorities of life. An ordinal ranking was then compiled based on a scoring system that gave greater weight to a person's first priority of life (mark of 5) than their least important priority (mark of 1).

11. Almost 42% of survey participants spontaneously mentioned education in Kwanonqaba compared to around 5% in Murraysuburg and Khubus. Younger and middle-aged people were also more likely to mention education (48% were under 35 years of age; and 84% were under 50 years of age).

12. In contrast to the responses to open-ended questions, the figure increases to 97.4% for Murraysburg and 100% for Khubus.

13. Interviewers were instructed not to prompt possible answers. In total, 77 distinct reasons for valuing education were identified. An ordinal ranking of these reasons was compiled based on a similar scoring system to before.

14. The PDS results are broken down by location but not race. The two main racial groups covered in the survey were divided between the township (black African) and the rural area (predominantly coloured). The same point applies to the ELS.

15. In developing countries, the criterion of ‘literacy’ is often used to help allocate scarce jobs.

16. Specifically, Figure 2 is based on the reasons for valuing jobs, housing, education, income, goods and services, food, clothing, family and friends, leisure, recreation, recreational activities, coca-cola, watching sports, television and cinema, and advertising. The focus is on ‘prominent linkages’ insofar as the figure is sensitive to the relative frequency particular reasons (or connections) were mentioned by survey participants.

17. Further examples can be found in Table 3 (see items 1b, 1f, 1g, 2b, 3c, 3d, 4c, 6a, 6b, 7f and 7h), which is discussed shortly.

18. As mentioned, participants only had the opportunity to cite up to four reasons for valuing education. The performance of the ‘collective list’ would improve further if specialist lists focusing on ‘higher education’ and ‘professional capabilities’' were removed (MW, WS and WM in Table 3).

19. Wilson and Ramphele (1989, 138) begin their chapter on the importance of education and literacy by recognising that

Any well-educated urban-dweller who has tried to milk a cow or to track game to shoot it for the pot can testify that the ability to read and write is not necessarily the most useful economic skill in some circumstances.

20. The boundary between means and ends is far more malleable than often thought. There is some evidence to suggest that even certain resources (such as food, water and electricity) might be valued for intrinsic as well as instrumental reasons (Clark 2002, 145). Frediani (2017) makes a similar point with respect to housing in Brazil.

21. The framework described below been applied to the conceptualisation and measurement of poverty in previous work. In Qizilbash and Clark's framework, education was one of the many dimensions of poverty. I am inclined to think that sub-dimensions of education (such as literacy, practical reason and life skills) should be treated as distinct dimensions of poverty in the broader framework (e.g. Clark and Qizilbash 2008, table 3). The approach set out in this paper may be useful for those who wish to focus on education. For convenience, sub-dimensions of education are referred to as ‘domains’ of education in this paper.

22. ‘Self-respect’ was endorsed by 96.8% of people in the PDS, and ‘self-esteem and respect’ was endorsed by 95.5% of people in the ELS.

23. The 5 percent rule (like the 95 percent rule) does not demand virtual unanimity; instead it allows for a margin of error in the interviewing process and reflects the fact that the vagueness framework is concerned with imprecision or ambiguity in articulating concepts such as ‘virtual unanimity’. Clark and Qizilbash (2008) explore the robustness of both rules with respect to the ELS.

24. There is no threshold data for one of the domains in Table 2 that qualifies as core (i.e. ability to plan life and make choices). In addition, a longer survey would have asked questions about additional education domains.

25. See also Alexander (2015, 252), who draws a corresponding distinction between measures and indicators.

Additional information

Funding

The PDS results this paper is based on was funded by the Cambridge Political Economy Society Trust, The Smuts Memorial Foundation, Girton College and The Thomas Carpenter Trust. The ELS results are based on joint work with colleagues funded by the UK’s Department for International Development.