Table 1.

Comparison between the two approaches used to estimate excess female under-5 mortality

	Hill and Upchurch7	Alkema et al2
Type	Prescriptive	Descriptive
Data	Historical series from four Northwestern European countries (England and Wales, France, the Netherlands and Sweden) and New Zealand, between 1820 and 1964	All available data from vital registration systems, sample registration and surveillance systems, surveys and censuses from 195 countries since 1950.
Age ranges	Infant (<1 year), child (1–4 years) and under 5 (<5 years)
	Intended to reflect sex differentials in childhood mortality in the absence of substantial discrimination, but at relatively high levels of child mortality. Described by the authors as a mortality ‘standard’.	Intended to reflect actual sex ratios in mortality, regardless of the presence of gender bias. Country-specific sex ratios are presented; these results are based on the product of the expected sex ratio and a country-specific multiplier, which represents the relative advantage or disadvantage of girls to boys compared with other countries with similar total mortality rates.
Original estimates
Assessment of gender bias	Female advantage index (difference between the observed and expected female/male mortality ratios) for any given level of male mortality	Countries with outlying sex ratios. Excess female mortality expressed as the difference between the expected and estimated female mortality rates