I am investigating variation in education attainment by birthmonth using the Malawi Census 2008. Tabulating birthmonth as a data consistency check reveals that the number of monthly births between January to June is more than double that for monthly births for July to December. This is strange and other sample surveys like the DHS 2010 do not show a similar pattern. Is this a coding error, a sampling error or is it even an error in the first place? Tabulation results are attached.
Malawi Census 2008_b-3.docx (12.5 KB)