Design, implementation and evaluation of a national campaign to distribute nine million free LLINs to children under five years of age in Tanzania

Table 2 Components of the discrepancy between 2008 National Bureau of Statistics (NBS) estimates for children under five years and the actual LLIN order in Southern and Lake Zones

Source of Variance	Range (%)	Average (%)	Remaining variance after exclusion of each factor (%)
Buffer estimated by VEOs*	1.7-11.9	5	36.8
Additional 5% village buffer†	-	5	31.8
DMO 5% buffer‡	-	5	26.8
Additional nets due to bale rounding**	3-7	5	21.8
Over-registration of children aged five years††	0-20	7	14.8
Estimate of NBS projection error***	-	14.8	N/A
Total difference between order size and NBS projections	12-68	42	42

*Village Executive Officers (VEOs) estimate of the number of children under 5 years missed on the registration reports. This is capped at 5% of the total village order
†This automatic 5% buffer at the village level was removed following an analysis of the over-registration
‡This buffer was shifted to a 5% ward level buffer to better facilitate reallocation to village level
** This buffer was removed following an analysis of the over-registration
†† This estimation is derived from a comparison of age distribution between 4,000 register book entries and the NBS projections
*** This figure comes from the subtraction of all other factors

ISSN: 1475-2875