Skip to main content
. 2016 Sep;22:65–81. doi: 10.1016/j.ehb.2016.03.002

Table 3.

Impact of total energy intake on change in heights and weights, Guatemala.

Total energy Distribution of total energy coefficient
sig > 0 sig < 0
# of sp. p25 p50 p75 %-Sig %-Sig
Panel A: Height
All IV 546 −0.0182 0.0099 0.0288 35 0
All over-identified IV 525 −0.0179 0.0107 0.0289 36 0
CD > 1 P-val HJ > 5 137 −0.0090 0.0034 0.0170 15 0
CD > 3 P-val HJ > 5 21 0.0009 0.0231 0.0438 57 0



Panel B: Weight
All IV 546 −0.0061 0.0059 0.0159 36 0
All over-identified IV 525 −0.0036 0.0060 0.0159 38 0
CD > 1 P-val HJ > 5 129 −0.0024 0.0050 0.0233 32 0
CD > 3 P-val HJ > 5 36 0.0142 0.0230 0.0239 83 0

CD = Robust Kleibergen-Paap F statistic, P-value, J = P-value of Hansen J stat × 100.

1st column: # of specifications that meet criteria; 2nd–4th col: percentile of distribution of estimated coefficients.

5th (6th) column: percent of estimated coefficients that are positive (negative) and significant at 5% significance level.

1st row: all specifications; 2nd row: all over-identified specifications for which # of IVs># of endogenous variables. Other rows include all specifications satisfying the indicated criteria based on the CD and HJ tests.

All specifications include breastfeeding, diarrhea, sex, age, and age squared as covariates and a seasonal dummy for the Philippines, and lagged height and lagged weight, both of which are treated as endogenous.

Height coefficients are divided by 1000 for presentation purposes.