¹ It is worth stressing, however, that the viability of a particular approach may depend both on the availability of specific data and on their suitability for manipulation. For example, the statistics most frequently used to draw causal inference require integral scales and thus varibales that are, or that can be treated as if they were, linear and continuous. The canons of contemporary social science, the expectations of manuscript reviewers, and the reward systems of universities often encourage researchers to make heroic leaps offaith in their analyses. Discarding an approach because of the unsuitability of available evidence is less well rewarded. For an insightful exploration of the tyranny of numbers, see Davis, Philip J. and Hersh, Reuben, Descartes' Dream: the world according to mathematics (Boston, 1986).Google Scholar

² That research is reported in Samoff, Joel, Tanzania:local politics and the structure of power (Madison, 1974), ch. 2,Google Scholar and ‘Education in Tanzania:class formatin and reproducation’, in The Journal of Modern African Studies (Cambridge), 17 1, 03 1979, Pp. 47–69.CrossRefGoogle Scholar

³ Like the rapid expansion of primary education two decades earlier, the explosive growth of private secondary schooling in Kilimanjaro in the late 1980s and early 1990s was significantly under-recorded. For both technical and political reasons, functioning but not-yet-registered private schools and their students were not included in government reports on secondary enrolment. Those students, however, compete with their officially recorded counterparts in the national Form 4 examination, and in seeking both further education and jobs. See Samoff, Joel, ‘School Expansion in Tanzania: private initiatives and public policy’, in Comparative Edzwation Review (Chicago), 31, 3, 08 1987, pp. 333–60.CrossRefGoogle Scholar

⁴ Rugumyamheto, Joseph A., ‘Manpower Forecasting in the United Republic of Tanzania’, in C.V., Youndi and Hinchliffe, Keith (eds.), Forecasting Skilled Manpower Needs:the experience of eleven countries (Brussels, Unesco/I.I.E.P., 1985), p. 242.Google Scholar

⁵ World Bank, Education in Sub-Saharan Africa: policies for adjustment, revitalization, and expansion (Washington, D.C., 1988), Table A-14, p. 138.Google Scholar Note here two of the problems that recur in these data: (i) different World Bank publications sometimes report different numbers for the same statistic – e.g. the 1970 figure of 16.0 per cent in 1988 was given as 19·7 per cent four years earlier in Wolff, Laurence, Controlling the Costs of Education in Eastern Africa: a review of data, issues, and policies (Washington, D.C., 1984); and (ii) there are inconsistencies even within a single publication – e.g. for some, but not all, of the years for which data are provided, the figures refer to ‘expenditures of the Ministry of Education only’, rather than to the total public spending on education.Google Scholar

⁶ Unesco, , United Republic of Tanzania: education in Tanzania, Vol. 1, Ovevrview (Paris,1989), p. 4.Google Scholar

⁷ Annual Joint Review of the Swedish Support of Education and Vocational Training 1988, Part 1, General Education (Dar es Salaam, Swedish International Development Authority, 1988), pp. 2–3.Google Scholar

⁸ From a World Bank assessment of education finance in Tanzania, May 1991, restricted distribution.

⁹ Note that on this, too, the statistics are inconsistent. Most sources agree that capital expenditures in recent years have averaged approximately 13 per cent of total education expenditures, as reported in Unesco, Education in Tanzania: sector review, Vol. 11, Technical Paper (Paris03 1989), p. 32.Google Scholar Some reports, however, indicate a much higher percentage. But according to the Ministry of Education's own reports (themselves the sources for national statistics, and thus for the data used by other agencies!), capital spending ranged from 26 to 46 per cent of total education spending from 1975/76 to 1980/81 – see The Ministry of Education Combined Annual Report for the Years 1976–1981 (Dar es Salaam03 1987), Appendix 23, p. 64, where it is noted that Tanzania did not spend a significant proportion of the budgeted funds, ‘largely attributed to the failure of contractors to complete their work as scheduled due to shortage or lack of building materials’Google Scholar

¹⁰ Ibid. p. 64.

¹¹ For example, in the World Bank's major policy study on African education, the Tanzanian figures for 1983 – but apparently not for the other three years with which they are compared – ‘refer to expenditures of the Ministry of Education only’. Education in Sub-Saharan Africa, technical note to Table A-4, p. 176.Google Scholar

¹² Rugumyamheto, , loc. cit. p. 242.Google Scholar