Evaluation of the 2007 Ethiopian Census Data with an Emphasis on Coverage in Amhara Region

Abstract

Ethiopia’s 2007 census results are compared with the Central Statistical Agency’s
(CSA) forward projections from the 1984 and 1994 censuses for 2008. Given the
undercounts in all regions in both the 1984 and the 1994 censuses, the 2007 census
was expected to produce a higher population count for all regions than CSA’s
forward projections for 2008. Instead, the 2007 census gave a national total that
was lower than the projected number by five million people. Three-fifths of the
undercount was in Amhara Administrative Region. Evaluations using Whipple and
Myers methods and consistency checks based on sex ratio and proportions of
children under five also point to deficiencies in the 2007 census. The significant
shortfall in Amhara cannot be explained fully based on available data. The
possibility remains, however, that universally acknowledged problems with census
taking - errors in data collection, processing, and reporting - may have affected
Amhara far worse than the other regions. Our review of literature and data analysis
found no evidence favoring the argument that a higher level of demographic
transition caused the shortfall in Amhara Region. Census data need to be
scrutinized critically even when there are reasons to suppose that they are accurate.