In the household recode (SPSS) for Angola MIS 2011, I weighted by hhmemwt calculated as hv012*hv005/1000000 which I think is the usual procedure for the wealth index. Normally, running frequencies on HV270 (wealth quintiles) then shows that 20% of the weighted dataset is in each quintile.

However, in this case, each quintile did not have 20% of the population in each. Rather the distribution (according to HV270) was as follows: Poorest - 11.9%, Poorer - 12.3%, Middle - 16.5%, Richer - 28%, Richest - 31.4%.

I think I may have missed or misunderstood something. Could you let us know why this might be?

Thanks for your help,

Kenzo]]>