Use the data in the BWGHT.txt for this question. The variable descriptions are given in
the file BWGHT_variables.txt. There are 1388 observations in this dataset.
(a) Run the regression
0 1 BWGHT ? ? ?? CIGS ?? .
Report and interpret your results.
(b) To what extent can you view this regression as capturing a causal relationship between
the child’s birthweight and the mother’s smoking habits? Are there any other variables
that may be correlated with CIGS that may explain BWGHT ?
(c) What is the predicted BWGHT when CIGS ? 0 ? What is the predicted BWGHT when
CIGS ? 20? Include standard errors in your predictions.
(d) Test the hypothesis that
(i) 0 1 H :? ? 0 vs 1 : 0 A H ? ?
(ii) 0 1 H :? ? ?1 vs 1 : 1 A H ? ? ?
(iii) 0 1 H :? ? 0 vs 1 : 0 A H ? ?
For part (ii), show how you can compute the t-statistic from the parameter estimate and
the parameter standard error. Show also how you can reparameterize the equation so that
EVIEWS will generate the desired t-statistic automatically. Verify that both methods give
you the same t-statistic.