Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

# Table 6 Attrition probits to construct weights used in Table 5, reweighting for attrition bias (N = 2,392)

Model 1 Model 2
Covariates (1) if in sample (1) if in sample
Male −0.109 (0.019) −0.121 (0.021)
Age at interview in 2002–2004 −1.780 (0.168) −1.690 (0.162)
Age at interview in 2002–2004 squared 0.020 (0.002) 0.019 (0.002)
Twins −0.187 (0.081) −0.235 (0.082)
Student–teacher ratio at age 7 0.002 (0.003) 0.002 (0.002)
Lower secondary school available at age 7 0.044 (0.045) 0.047 (0.045)
Born in the two years prior to 1976 earthquake 0.097 (0.061) 0.043 (0.062)
Mother or father had died by age 18 0.091 (0.046) 0.163 (0.042)
Ln Salary in manufacturing at age 18 −4.031 (0.263) −3.817 (0.264)
Exposure to intervention 0–36 months 0.039 (0.069) 0.027 (0.071)
Exposure to intervention 0–36 months × atole 0.023 (0.061) −0.013 (0.058)
Child lived with both mother and father in 1975 0.077 (0.038)
Child lived with both mother and father in 1987 0.058 (0.027)
Mother alive in 2002 0.254 (0.042)
Father alive in 2002 0.120 (0.031)
Mother living in original community in 2002 0.094 (0.037)
Father living in original community in 2002 0.046 (0.031)
Number of siblings in survey −0.002 (0.002)
Whether any sibling reinterviewed in 2002–2004 0.269 (0.043)
χ2 statistic on variables in model 2 only 312.6 [<0.01]
Model χ2 statistic 328.9 [<0.01] 618.5 [<0.01]
Pseudo-R2 0.22 0.30
1. Sample consists of all 2,392 individuals who were exposed to the supplementation intervention between 1969 and 1977, 1,448 of whom are included in the main analyses. Standard errors calculated allowing for clustering at the birth-year-community level are shown in parentheses below coefficient estimates, which are in bold if significant at 5 % or lower; p values are in brackets (StataCorp 2011). Derivatives evaluated at the mean (dP/dx) presented