Scribd
Upload
41 views

Statistical Inference 1. Determinant of Salary of Law School Graduates

This document summarizes a statistical analysis of the determinants of law school graduate salaries based on ranking, GPA, and LSAT scores. It provides the results of a linear regression model finding that: 1) Ranking, GPA, and LSAT scores all significantly predict salary, with higher rankings and scores associated with higher salaries. 2) Hypothesis tests show the effects of both ranking and LSAT scores on salary are statistically significant based on the regression results.

Uploaded by

Toulouse18
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
41 views

Statistical Inference 1. Determinant of Salary of Law School Graduates

This document summarizes a statistical analysis of the determinants of law school graduate salaries based on ranking, GPA, and LSAT scores. It provides the results of a linear regression model finding that: 1) Ranking, GPA, and LSAT scores all significantly predict salary, with higher rankings and scores associated with higher salaries. 2) Hypothesis tests show the effects of both ranking and LSAT scores on salary are statistically significant based on the regression results.

Uploaded by

Toulouse18
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Statistical Inference

1. Determinant of salary of law school graduates


Suppose that the determinants of salary of law school graduates are given by a linear model with
three explanatory variables ranking, GPA, and LSAT
.
Least squares estimates using STATA are give by
. regress salary rank GPA LSAT

Source | SS df MS Number of obs = 136


-------------+------------------------------ F( 3, 132) = 148.41
Model | 1.5774e+10 3 5.2580e+09 Prob > F = 0.0000
Residual | 4.6766e+09 132 35428563.7 R-squared = 0.7713
-------------+------------------------------ Adj R-squared = 0.7661
Total | 2.0451e+10 135 151485812 Root MSE = 5952.2

------------------------------------------------------------------------------
salary | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
rank | -145.5202 15.93291 -9.13 0.000 -177.037 -114.0033
GPA | 12975.26 4343.839 2.99 0.003 4382.716 21567.8
LSAT | 448.5722 190.3353 2.36 0.020 72.07024 825.0742
_cons | -62994.16 26301.64 -2.40 0.018 -115021.4 -10966.91
------------------------------------------------------------------------------
Note that STATA automatically computes the t-statistics (where unknown is assumed). The
calculated is given by SSR/n-k = 35428563.7. Note that the variance of each estimated coefficient
is given by

To verify this is true, obtain the total sum of squares of each explanatory variable and as follows:
. summarize salary rank GPA LSAT

Variable | Obs Mean Std. Dev. Min Max


-------------+--------------------------------------------------------
salary | 136 39444.88 12307.96 24900 78325
rank | 136 79.75735 49.07417 1 174
GPA | 136 3.309632 .1972305 2.8 3.82
LSAT | 136 158.5074 4.599511 140 171

. reg rank GPA LSAT

Source | SS df MS Number of obs = 136


-------------+------------------------------ F( 2, 133) = 88.42
Model | 185556.229 2 92778.1143 Prob > F = 0.0000
Residual | 139560.764 133 1049.32905 R-squared = 0.5707
-------------+------------------------------ Adj R-squared = 0.5643
Total | 325116.993 135 2408.27402 Root MSE = 32.393

------------------------------------------------------------------------------
rank | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
GPA | -91.56549 22.26711 -4.11 0.000 -135.609 -47.52201
LSAT | -4.631608 .9548302 -4.85 0.000 -6.520226 -2.742991
_cons | 1116.949 105.3987 10.60 0.000 908.475 1325.424
------------------------------------------------------------------------------

. reg GPA rank LSAT


Source | SS df MS Number of obs = 136
-------------+------------------------------ F( 2, 133) = 119.49
Model | 3.37386894 2 1.68693447 Prob > F = 0.0000
Residual | 1.87761273 133 .014117389 R-squared = 0.6425
-------------+------------------------------ Adj R-squared = 0.6371
Total | 5.25148166 135 .038899864 Root MSE = .11882

------------------------------------------------------------------------------
GPA | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
rank | -.0012319 .0002996 -4.11 0.000 -.0018244 -.0006393
LSAT | .0236893 .0031963 7.41 0.000 .0173671 .0300114
_cons | -.3470367 .524166 -0.66 0.509 -1.383817 .6897432
------------------------------------------------------------------------------

. reg LSAT rank GPA

Source | SS df MS Number of obs = 136


-------------+------------------------------ F( 2, 133) = 127.71
Model | 1878.04634 2 939.023172 Prob > F = 0.0000
Residual | 977.946303 133 7.35297972 R-squared = 0.6576
-------------+------------------------------ Adj R-squared = 0.6524
Total | 2855.99265 135 21.1555011 Root MSE = 2.7116

------------------------------------------------------------------------------
LSAT | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
rank | -.0324551 .0066908 -4.85 0.000 -.0456893 -.019221
GPA | 12.33845 1.66478 7.41 0.000 9.045577 15.63132
_cons | 120.2602 5.901972 20.38 0.000 108.5863 131.934
------------------------------------------------------------------------------

The computed total sum of squares of rank, GPA, and LSAT is given by
. display ((49.07417)^2)*135
325117.01

. display ((.1972305)^2)*135
5.2514825

. display ((4.599511)^2)*135
2855.9927

Together with the estimated residuals and computed of rank, GPA, and LSAT, the computed
standard error of the least squares estimator is
. display sqrt(35428563.7/(325117.01*(1-0.5707)))
15.932227

. display sqrt(35428563.7/(5.2514825*(1-0.6425)))
4344.0797

. display sqrt(35428563.7/(2855.9927*(1-0.6576)))
190.34056

2. Hypothesis testing
Recall the estimation output given by
. regress salary rank GPA LSAT

Source | SS df MS Number of obs = 136


-------------+------------------------------ F( 3, 132) = 148.41
Model | 1.5774e+10 3 5.2580e+09 Prob > F = 0.0000
Residual | 4.6766e+09 132 35428563.7 R-squared = 0.7713
-------------+------------------------------ Adj R-squared = 0.7661
Total | 2.0451e+10 135 151485812 Root MSE = 5952.2

------------------------------------------------------------------------------
salary | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
rank | -145.5202 15.93291 -9.13 0.000 -177.037 -114.0033
GPA | 12975.26 4343.839 2.99 0.003 4382.716 21567.8
LSAT | 448.5722 190.3353 2.36 0.020 72.07024 825.0742
_cons | -62994.16 26301.64 -2.40 0.018 -115021.4 -10966.91
------------------------------------------------------------------------------
The probability P(T>|t|) under the null hypothesis that the true parameter is zero is given in the fifth
column. Hence we can reject the null hypothesis that the effect of rank on salary is zero at 0.01 level
of significance. Similarly, we can reject the null hypothesis that the effect of LSAT on salary is zero at
0.05 level of significance.

You might also like