On Angola, the sample results doesn't represent the nation as a whole. This is because samples cannot represent areas that are excluded from the frame (i.e., not covered by the survey operations), especially if those areas comprise 35% of the country as was the case in Angola.
You may of course combine the urban and rural components together for a "total" but it has to be labelled properly (see below).
In putting out the results a statement such as the following should accompany the tables: "The survey estimates do not represent Angola as a whole because the sample was restricted to territory that was regarded as secure by the Government. That territory comprised only about 65 percent of the population. An important point to note is that most of the secured territory in Angola is urban, while most of the unsecured territory is rural. For this reason the sample is disproportionately larger in urban communities.
However, since most of the urban community in Angola is secure, the urban part of the sample is a fairly accurate representation of all of urban Angola. By contrast, the rural sample only represents the rural secured territory; since a large proportion of rural Angola was not covered by the survey sample, it is not expected that the rural results would reflect rural Angola as well as the urban results reflect urban Angola.
The totals, that is, urban and rural combined, represent what can be described as Angola-Secured Territory but not the nation as a whole.
We advice users to keep these points in mind when analyzing the data.
