Figure 1

load(p0(.loc,'/Anal/WIC/data/data_fig1.rda'))

.l=0.005; .s=46
f1=ggplot(ds, aes(x,y)) + geom_point(size=.l, shape=.s,color='grey60') + 
  facet_wrap(~var) +  coord_cartesian(ylim = c(-10,17)) + 
  scale_y_continuous(breaks=c(-8,-5,0,3,5,10,15)) + 
  theme_df()  + labs(x='\nStandard Normal Z-Score', y='Observed Z-Score\n')

.s=3; .c='black'; .l=.8
f2=f1 +  geom_point(data=dshb,size=.l,shape=.s,color=.c) +
  geom_point(data=dshw,size=.l,shape=.s,color=.c) + 
  geom_point(data=dshh,size=.l,shape=.s, color=.c)

.s=3; .l=1
(f3=f2 +  geom_point(data=dslb,shape=.s,color=.c, size=.l) +
  geom_point(data=dslw,shape=.s,color=.c, size=.l) + 
  geom_point(data=dslh,shape=.s, color=.c, size=.l))

QQ plots of weightMZ (left), heightMZ, and BMIMZ (right) among 25,000 randomly selected children in WIC. The black ‘+’ symbols represent BIVs, and the grey points represent data for values that are considered biologically plausible based on the WHO cut-points. If the data were normally distributed, all points would fall on a diagonal line. Observed z-scores below -8 were recoded to -8 and values >16 were recoded as 16 so that the focus would be on the regions of interest. All body size measures had heavy tails at both ends of the distribution.