You are on page 1of 6

1.

Regression Results from the estimated equation


sleep =0 + 1totwrk + 2educ + 3age + 4age2 + male + e:

Variable Parameter Estimates

Standard Error

P-value

Constant

3840.83

235.1087

0.000*

TotWrk

-.1634232

.0181321

0.000*

Educ

-11.71332

5.866886

0.046*

Age

-8.696676

11.20746

0.438

Agesq

.1284354

.1338954

0.338

Male

87.75243

34.32616

0.011*

N=706, AdjRsquare=0.1165 *Significant at 5% level of Significance

a) The coefficient associated with Male is 87.75 (positive) and it is significant at 5% level of
significance. It implies Men (Male=1) sleep more than women and the evidence is strong as the
estimated coefficient is statistically significant.
b) The estimated coefficient of the variable TotWork is -0.163 and it is statistically significant at
5% level of significance, indicating that there Sleep and total no of work time is negatively
related to each other. More time spent in work implies less time to sleep.
c)The regression equation includes both linear as well as quadratic term of age as independent
variables. Interestingly both the coefficients are statistically insignificant. In order to purely
identify the null hypothesis that age has no effect on the amount of sleep, we need to run as
regression without the agesq term.

2.The Regression Result from female sample


Variable Parameter Estimates

Standard Error

P-value

Constant

4238.729

384.8923

0.000 *

TotWrk

-.1399495

Educ

-10.20514

9.588848

0.288

Age

-30.35657

18.53091

0.102

Agesq

.3679406

.2233398

0.101

Yngkid

-118.2826

93.18757

0.205

.0276594

0.000*

N=306, AdjRsquare=0.0826*Significant at 5% level of Significance


Variable Parameter Estimates

Standard Error

P-value

Constant

3648.208

310.0393

0.000*

TotWrk

-.1821232

.0244855

0.000 *

Educ

-13.05238

7.414218

0.079**

Age

7 .156591

14.32037

0.618

Agesq

-.0447674

.1684053

0.791

59.02278

0.307

Yngkid

60.38021

N=400, AdjRsquare=0.1455
*Significant at 5% level of Significance, **Significant at 5% level of Significance
b) The notable difference between the results is in the education variable. The estimated
coefficient of education is non significant in the female sample, however, in the male sample it is
significant at 10 % level of significance. Although, the effect of Yngkid is insignificant in both
the samples, but, in case off male it is negative, while for male sample it is positive.

3.
Variable Parameter Estimates

Standard Error

P-value

Hsize 19.3

sat

Coef.

Std. Err.

P>t

[95% Conf.

Interval]

hsize

19.29706

3.832298

5.04

0.000

11.78369

26.81043

hsizesq

-2.194828

.5271635

-4.16

0.000

-3.228353

-1.161304

female

-45.09145

4.291063

-10.51

0.000

-53.50424

-36.67866

black

-169.8126

12.71315

-13.36

0.000

-194.7373

-144.888

febl

62.30636

18.15417

3.43

0.001

26.71442

97.89829

_cons

1028.097

6.290167

163.45

0.000

1015.765

1040.429

a) The estimated coefficient of hsizesq is -2.19 and it is statistically significant at 5% level


of significance, indicating that there is strong evidence to include this variable into the
model.
b) Optimal class size can be derived from solving the following equation:
19.3
h 0,, 19.32.2.19 hsize=0,4.38 hsize=19.3,, hsize=
=4.41
4.38
sat

Holding hsize _xed, what is the estimated di_erence in SAT score between nonblack
a) females and nonblack males? How statistically signi_cant is this estimated
di_erence?

4.

5.
a) The regression result is presented in the table below
Parameter
Standard PVariable Estimates
Error
value
Constant
24199.65
1078.42 0.000*
GPA
1643.27
352.29 0.000*
Metrics
5033.08
456.31 0.000*
N=50. Adj-Rsquare=0.725. *Significance at 5% level of significance
Both GPA and taking econometrics positively affects the starting salary. Both the estimated
coefficients are statistically significant at 5%level of significance. An additional GPA point
increases salary by $1643.27, whereas, taking econometrics increases salary by $5033.08.
b) When Women is introduced in the model , we have the following resuts.
Variabl Parameter
e
Estimates
Standard Error P-value
Constan
t
24241.72
1090.68
0.000*
GPA
1657.76
356.42
0.000*
Metrics
5023.51
460.47
0.000*
Female
-204.80
419.6
0.628
N=50. Adj-Rsquare=0.72. *Significance at 5% level of significance
Here, we find that both GPA and taking econometrics positively affects the starting salary. Both
the estimated coefficients are statistically significant at 5%level of significance. An additional
GPA point increases salary by $1657.76, whereas, taking econometrics increases salary by
$5023.51. However, if the individual is a female, then the salary drops by $204.80, however, this
effect is not statistically significant.
c) We run the regression using separate samples for male and female.
For men, the regression estimations are
Variable
Constant
GPA

Parameter
Standard PEstimates
Error
value
25217.24
1372.75 0.000*
1331.17
459.81 0.000*

Metrics
4919.78
611.45 0.000*
N=30. Adj-Rsquare=0.71. *Significance at 5% level of significance

For female the regression estimations are


Parameter
Standard PVariable Estimates
Error
value
Constant
21500.46
1842.1 0.000*
GPA
2459.69
582.48 0.001*
Metrics
5551.05
713.96 0.000*
N=20. Adj-Rsquare=0.76. *Significance at 5% level of significance
We find that the value of econometrics is not same for men and women, the estimated value of
taking econometrics is $5551.05 is higher for female than male ($4919.78)

6.

You might also like