Skip to main content
. 2019 Feb 8;4(1):e001175. doi: 10.1136/bmjgh-2018-001175

Table 1.

Sample comparison

Initial sample Main analysis sample Adjusted sample
(N=599 834) (N=416 181) (N=167 523)
Interview year: before 2000 3.7% 4.4% 0.0%
Interview year: 2000–2004 2.8% 3.6% 0.0%
Interview year: 2005–2009 18.0% 20.3% 22.7%
Interview year: 2010 or later 68.9% 65.2% 73.0%
Sub-Saharan Africa 56.6% 49.5% 60.2%
East Asia and Pacific 3.8% 4.3% 4.4%
Europe and Central Asia 3.8% 4.7% 3.1%
Latin America and the Caribbean 14.5% 18.3% 16.8%
Middle East and North Africa 8.3% 9.8% 0.0%
South Asia 13.1% 13.4% 15.6%
Low-income countries 50.7% 45.2% 58.0%
Child is a girl 49.2% 49.2% 48.9%
Child born in single birth 97.6% 97.7% 98.6%
Children ever born to mother 3.6 (SD=2.3) 3.6 (SD=2.3) 3.5 (SD=2.4)
Births in last 5 years 1.7 (SD=0.7) 1.6 (SD=0.7) 1.5 (SD=0.6)
Child is stunted n/a 33.1% 33.7%
Child age in months n/a 28.6 (SD=16.9) 22.6 (SD=15.4)
Household in poorest quartile n/a 30.5% 28.0%
Household in second poorest quartile n/a 24.3% 24.4%
Household in second richest quartile n/a 23.4% 24.1%
Household in richest quartile n/a 21.7% 23.5%

No data on stunting, age and household SES is presented for the initial sample due to a high number of missing values. The share of children living in households of each SES quartile differs from 25% as birth rates tend to be higher in low SES households and ties in household asset scores can occur.

“n/a” = not applicable (statistics on stunting, child age and household SES not calculated in initial sample due to missing values)

SES, socioeconomic status.