There is remarkably little 'hard' data on how living
standards in Africa have changed durïng adjustment.
In particular, there are few countries with nationally
representative household consumption surveys
conducted at two or more points of time during the
period.2 This helps explain the cautious language
used in a recent high profile World Bank report: 'the
poor are probably better off and almost certainly no
worse off' as a result of economic reforms (World
Bank 1994). Here we use the example of Uganda
to show the problems of comparability that may arise
even where two such surveys do exist. Where surveys
have very different designs, they are unlikely
to be comparable without adjustment and, in the
case of Uganda, we question whether they can be
reliably compared at all. lt is not surprising that
survey design matters: what is striking about the
Ugandan example is just how sensitive the results
seem to be.