Академический Документы
Профессиональный Документы
Культура Документы
Shang Wu
shang.wu@unsw.edu.au
School of Economics
University of New South Wales
Outline
Review questions
Week 11 problems
Review questions
Review questions
Outline
Review questions
Week 11 problems
Problem 1: 9.1
(.375 .353)/2
= 2.97.
(1 .375)/(177 8)
Problem 2: 9.3
Table:
Regressors
Model (1)
Model (2)
log(expend )
11.13
(3.3)
-
7.75
(3.04)
-0.324
(0.036)
lnchprg
Table:
Regressors
log(enroll )
lnchprg
R2
Model (1)
Model (2)
0.022
(0.615)
0.0297
-1.26
(0.58)
-0.324
(0.036)
0.1893
Table:
Regressors
log(enroll )
lnchprg
R2
Model (1)
Model (2)
0.022
(0.615)
0.0297
-1.26
(0.58)
-0.324
(0.036)
0.1893
In (1), very little (less than 3%) of the variation in math pass
rates is explained. In (2), we are explaining almost 19%.
Clearly most of the variation in math10 is explained by
variation in lnchprg . This is a common nding in studies of
school performance: family income (or related factors) are
much more important in explaining student performance than
are spending per student or other school characteristics.
Problem 3: 9.5
Problem 4: C9.3
Data: JTRAIN.dta.
(i) For log (scrap ) = 0 + 1 grant 88 + , where scrap is the
rm scrap rate and grant is a dummy indicating whether a
rm received a training grant. Any reasons why might be
correlated with grant ?
(ii) Estimate the model using data for 1988. Does receiving a
grant lower scrap rate?
log (scrap
[ ) = .409 + .057grant 88, n=54, R 2 =.0004
(.241) (.406)
log (scrap
[ ) = .021 .254grant + .831log (scrap 87),
(.089)
(.147) (.044) n=54, R 2 =.873
test lscrap87 = 1.
(v) Repeat parts (iii) and (iv) using hetero-robust s.e., report
any notable dierences.
11