RATS 11.1
RATS 11.1

Procedures /

REGPARTCORR Procedure

Home Page

← Previous Next →

@REGPARTCORR is a regression post-processor which computes and prints the partial correlations between the regressors and the dependent variable. It must be preceded by a LINREG instruction.

@REGPARTCORR      (no options or parameters)

Example

*

* Greene, Econometric Analysis, 7th Edition

* Example 3.1 from page 38-39

* Results differ slightly because of use of rounded data in the text

*

open data tablef3-1.txt

calendar(a) 1968

data(format=prn,org=columns) 1968:01 1982:01 year gnp invest cpi interest

set realgnp = .1*gnp/cpi

set realinv = .1*invest/cpi

set trend   = t

*

* The inflation series needs to be defined from 1968 for later analysis.

* The value from 1968 can't be computed from the data set as is, so is

* provided directly. The FIRST option on SET uses a separate formula for

* the first entry.

*

set(first=4.40) inflation = (cpi/cpi{1}-1)*100.0

*

linreg realinv

# constant trend realgnp interest inflation

*

* @RegPartCorr is a procedure which computes the partial correlations for

* the most recent regression.

*

@regpartcorr

Sample Output

This shows the LINREG output followed by the @REGPARTCORR output. The latter shows the partial R-squared and the standardized coefficients.
 

Linear Regression - Estimation by Least Squares

Dependent Variable REALINV

Annual Data From 1968:01 To 1982:01

Usable Observations                        15

Degrees of Freedom                         10

Centered R^2                        0.9734972

R-Bar^2                             0.9628961

Uncentered R^2                      0.9993222

Mean of Dependent Variable       0.2034337158

Std Error of Dependent Variable  0.0341144915

Standard Error of Estimate       0.0065712644

Sum of Squared Residuals         0.0004318152

Regression F(4,10)                    91.8296

Significance Level of F             0.0000001

Log Likelihood                        57.1326

Durbin-Watson Statistic                1.9636
 

    Variable                        Coeff      Std Error      T-Stat      Signif

************************************************************************************

1.  Constant                     -0.509068377  0.053933271     -9.43886  0.00000269

2.  TREND                        -0.016589625  0.001929415     -8.59827  0.00000623

3.  REALGNP                       0.670303475  0.053799929     12.45919  0.00000021

4.  INTEREST                     -0.002428091  0.001193776     -2.03396  0.06934437

5.  INFLATION                     0.000063879  0.001318846      0.04844  0.96232303

 

Partial Correlations and Standardized Coefficients

Label     Partial R^2 Std. Coefficient

Constant        NA             0.00000

TREND        -0.93854         -2.17477

REALGNP       0.96927          3.14969

INTEREST     -0.54096         -0.20016

INFLATION     0.01531          0.00365

 


Copyright © 2026 Thomas A. Doan