Scribd
Upload
25 views

β school+u (for women) : Estimation of log (wage) =β +

The document presents results from regressions of log wages on years of schooling for women, men, and both. For women, log wages increase by 0.095 for each additional year of schooling. For men, the increase is 0.071. When pooled, log wages rise 0.095 for women and 0.071 for men per extra school year, with men earning 5.577% more than women at same education levels. The coefficients are statistically significant except effect of schooling on men's wages cannot be concluded to differ from 8%.

Uploaded by

Muthomi Timothy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
25 views

β school+u (for women) : Estimation of log (wage) =β +

The document presents results from regressions of log wages on years of schooling for women, men, and both. For women, log wages increase by 0.095 for each additional year of schooling. For men, the increase is 0.071. When pooled, log wages rise 0.095 for women and 0.071 for men per extra school year, with men earning 5.577% more than women at same education levels. The coefficients are statistically significant except effect of schooling on men's wages cannot be concluded to differ from 8%.

Uploaded by

Muthomi Timothy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Estimation of log ( wage ) = 0+ 1 school+u (for women)

. regress lwage school if male==0


Source

SS

df

MS

Model
Residual

132.661347
1132.67363

1
2789

132.661347
.406121776

Total

1265.33498

2790

.453525082

lwage

Coef.

school
_cons

.0949798
.4590116

Std. Err.
.0052552
.0691256

t
18.07
6.64

Number of obs
F( 1, 2789)
Prob > F
R-squared
Adj R-squared
Root MSE

=
=
=
=
=
=

2791
326.65
0.0000
0.1048
0.1045
.63728

P>|t|

[95% Conf. Interval]

0.000
0.000

.0846754
.323469

.1052842
.5945542

l wage=0 .4590116 +0 .0949798 school


Estimation of log ( wage ) = 0+ 1 school+u (for men)
. regress lwage school if male==1
Source

SS

df

MS

Model
Residual

92.1219323
1426.90429

1
3105

92.1219323
.459550497

Total

1519.02623

3106

.489061889

lwage

Coef.

school
_cons

.0710905
1.01669

Std. Err.
.0050211
.0640046

t
14.16
15.88

Number of obs
F( 1, 3105)
Prob > F
R-squared
Adj R-squared
Root MSE

=
=
=
=
=
=

3107
200.46
0.0000
0.0606
0.0603
.6779

P>|t|

[95% Conf. Interval]

0.000
0.000

.0612455
.8911944

.0809354
1.142186

l wage=1.01669+0 .0710905 school

Estimation of log ( wage ) = 0+ 1 school+ 2 male + 3 school male+ u(for both)


. regress lwage school male schoolXmale
Source

SS

df

MS

Model
Residual

294.141681
2559.57793

3
5894

98.047227
.434268396

Total

2853.71961

5897

.483927354

lwage

Coef.

school
male
schoolXmale
_cons

.0949798
.5576785
-.0238893
.4590116

Std. Err.
.0054342
.0947668
.0073045
.0714809

t
17.48
5.88
-3.27
6.42

Number of obs
F( 3, 5894)
Prob > F
R-squared
Adj R-squared
Root MSE

P>|t|
0.000
0.000
0.001
0.000

=
=
=
=
=
=

5898
225.78
0.0000
0.1031
0.1026
.65899

[95% Conf. Interval]


.0843267
.3719009
-.0382088
.3188828

.1056329
.7434561
-.0095699
.5991404

lwage 0.490116+ 0.0949798 school+0.5576785 male0.0238893 school male

a) The intercept parameter of the pooled model (third model) implies that, holding all other
factors constant, the log of wages will be 0.490116. This parameter is statistically
significant at all levels given that its p-value (0.000) is less than these levels of
significance (1%, 5%, 10%).
The coefficient of school (0.0949798) means that an increase in years of school
by one unit (ceteris paribus) with increase the wage rate by 9.49798%. The coefficient is
statistically significant at 1%, 5%, and 10% given that its p-value (0.000) is less than
these levels of significance.
The coefficient of male (0.5576785) means that a man earns 5.576785% more
than a female counterpart, holding years of schooling constant. This coefficient is
statistically significant at 1%, 5%, and 10% since its p-value (0.000) is less than these
levels.
The coefficient of schoolmale (-0.0238893) implies the effect of a unit increase
in schooling on wage rate is less in men than in women by 2.38893%. The coefficient is
statistically significant at all levels since its p-value (0.001) is less than 1%, 5% and 10%.
b) We can recover the coefficients of the first two models from the third model by
substituting the value of male in the third model as follows.
To obtain the first model we replace male with 0
lwage 0.490116+ 0.0949798 school+0.5576785 00.0238893 school 0
We remain with
lwage 0.490116+ 0.0949798 school
This is the model for women as earlier estimated
To obtain the 2nd model we replace male with 1 in the third model
lwage 0.490116+ 0.0949798 school+0.5576785 10.0238893 school 1
lwage (0.490116+0.5576785)+(0.09497980.0238893) school
lwage 1.0477945+0.0710905 school
c)
d) Using the stata command test, the following output was obtained
. test _b[school]=0.08
( 1)

school = .08
F(

1, 3105) =
Prob > F =

3.15
0.0761

Given that the p-value (0.0761) is greater than 5% (0.05), we fail to reject the null
hypothesis and conclude that a year of education for men increases wages by 8%.
e) Performing a test similar to the one in d) above but using the first model gives the
following output.
. test _b[school]=0.08
( 1) school = .08
F( 1, 2789) =
Prob > F =

8.13
0.0044

By observing the p-value (0.0044), we realize that it is less than 5% (0.05). For this reason, we
reject the null hypothesis and conclude that a year of education for women doesnt increase
wages by 8%.

You might also like