A further point made by Keith Lewin (personal communication) is that
very often the various studies treat factors as separate entities where
variance is unproblematic. It makes a great deal of difference how school
size, for example, varies from place to place. Treating things like the
existence of textbooks or science laboratories dichotomously often loses
important dimensions of school quality related to how they are distributed.
A further problem related to this at the policy level is that, whatever the
difficulties of disaggregation, it will generally be the case that selected
inputs on the margin have far more effect than flat rate increases.
None of these caveats mean that existing research evidence should be rejected; it is the
best evidence we have. However, it is important that it should not be applied blindly or
prescriptively, but rather used to assist decision making, with due attention to context.
On the other hand, it can be said with confidence that policy-makers who ignore the
research evidence take a considerable, and unjustifiable, risk.
[Previous Page] [Top of Page] [Next Page]
DFID Education Department, 94 Victoria Street, London, SW1E 5JL, UK, Tel: +44 (0)
20 7 917, [email protected]