The regression results for individual states of India are
given in Table 4. It is clear from the table that: (a) what
proximate factors are important for low birth weight varies
by state; and (b) the magnitude of the effect of low BMI of
mothers varies by state. The critical relevance of mothers
nutritional status is underscored by the fact that in almost
all the states of India mothers who were underweight
increased the risk of low birth weight of their infants by at
least 30% (in 13 out of 17 states), although the effect was
statistically significant only in five states. When we
included all births born during the 3 years prior to the
survey date the effect was statistically significant in 13
states (results not shown). This indicates that the lack of
statistical significance even when the estimated impact is
over 30% (it was over 50% in the case of Punjab) is likely
to be mainly due to small sample sizes for individual states