Basic Econometrics, Gujarati and Porter
15.11
a) Although the results are not uniform, in several cases the logit
(c) As you can see, if you take all the matriculants, all the
coefficients are highly statistically significant. But this is not the
15.12
(a) To make the error term homoscedastic, the weight should
be the inverse of the standard error of the disturbance term u
i
.
The weight in the present case is:
(b) The weights and the transformed data are as follows:
Probability Weight (
)
*
=
*
=
0.24 0.086 – 8.113 92.717
0.35 0.140 – 2.708 92.636
0.51 1.991 0.015 10.044
0.66 0.168 2.388 179.124
0.80 0.095 8.820 420.000
(c) The weighted least-squares results are: