THU 3/11/99 12:31:52 PM

SYSTAT VERSION 7.0.1
COPYRIGHT (C) 1997, SPSS INC.

Welcome to SYSTAT!

>USE 'C:\SYSTAT7\DATA\LONGLEY.SYD'
SYSTAT Rectangular file C:\SYSTAT7\DATA\LONGLEY.SYD,
created Tue Feb 18, 1986 at 13:56:00, contains variables:
 DEFLATOR     GNP          UNEMPLOY     ARMFORCE     POPULATN     TIME
 TOTAL
The Longley data set is provided as part of the SYSTAT package.  It is a highly
collinear data set used to test the accuracy of regression programs.  The dependent
variable TOTAL is total employment.
>mglh

>model total = constant + deflator + gnp + unemploy + armforce + populatn + time

>estimate
 
Dep Var: TOTAL   N: 16   Multiple R: 0.997737   Squared multiple R: 0.995479
 
Adjusted squared multiple R: 0.992465   Standard error of estimate: 304.854074
 
Effect         Coefficient    Std Error     Std Coef Tolerance     t   P(2 Tail)
 
CONSTANT      -3.48226E+06  8.90420E+05     0.0        .       -3.91080  0.00356
DEFLATOR         15.061872    84.914926     0.046282  0.007378  0.17738  0.86314
GNP              -0.035819     0.033491    -1.013746  0.000559 -1.06952  0.31268
UNEMPLOY         -2.020230     0.488400    -0.537543  0.029745 -4.13643  0.00254
ARMFORCE         -1.033227     0.214274    -0.204741  0.278635 -4.82199  0.00094
POPULATN         -0.051104     0.226073    -0.101221  0.002505 -0.22605  0.82621
TIME           1829.151465   455.478499     2.479664  0.001318  4.01589  0.00304
 
                             Analysis of Variance
 
Source             Sum-of-Squares   df  Mean-Square     F-ratio       P
 
Regression           1.84172E+08     6  3.06954E+07  330.285339    0.000000
Residual             8.36424E+05     9  9.29360E+04
-------------------------------------------------------------------------------
 
 
Durbin-Watson D Statistic     2.559
First Order Autocorrelation  -0.348


Example 1 - Coefficient simultaneously (jointly) = 0
>hypothesis

>specify deflator=0; gnp=0; populatn=0

>test
Hypothesis.
 
A Matrix
 
                         1           2           3           4           5
              1       0.0        -1.000000    0.0         0.0         0.0
              2       0.0         0.0        -1.000000    0.0         0.0
              3       0.0         0.0         0.0         0.0         0.0
 
 
                         6           7
              1       0.0         0.0
              2       0.0         0.0
              3      -1.000000    0.0
 
D Matrix
 
              1       0.0
              2       0.0
              3       0.0
 
Test of Hypothesis
 
       Source       SS        df       MS              F              P
 
  Hypothesis  4.86937E+05     3  1.62312E+05       1.746495       0.227032
  Error       8.36424E+05     9  9.29360E+04
 


Example 2 - Difference of 2 coefficients = 0  >hypothesis >specify gnp-populatn=0 >test
Hypothesis.
 
A Matrix
 
                         1           2           3           4           5
                      0.0         0.0        -1.000000    0.0         0.0
 
                         6           7
                      1.000000    0.0
Null hypothesis value for D
                      0.0
Test of Hypothesis
 
       Source       SS        df       MS              F              P
 
  Hypothesis   334.822931     1   334.822931       0.003603       0.953449
  Error       8.36424E+05     9  9.29360E+04
 


Example 3 - Coefficient of one variable 3 times coefficient of another  >hypothesis >specify unemploy=3*armforce >test
Hypothesis.
 
A Matrix
 
                         1           2           3           4           5
                      0.0         0.0         0.0        -1.000000    3.000000
 
                         6           7
                      0.0         0.0
Null hypothesis value for D
                      0.0
Test of Hypothesis
 
       Source       SS        df       MS              F              P
 
  Hypothesis  4.11199E+05     1  4.11199E+05       4.424544       0.064754
  Error       8.36424E+05     9  9.29360E+04
 


Example 4 - Coefficients of 2 variables have specific nonzero values  >hypothesis >specify populatn=0.5; time=1000 >test
Hypothesis.
 
A Matrix
 
                         1           2           3           4           5
              1       0.0         0.0         0.0         0.0         0.0
              2       0.0         0.0         0.0         0.0         0.0
 
 
                         6           7
              1      -1.000000    0.0
              2       0.0        -1.000000
 
D Matrix
 
              1      -0.500000
              2   -1000.000000
 
Test of Hypothesis
 
       Source       SS        df       MS              F              P
 
  Hypothesis  1.38981E+06     2  6.94907E+05       7.477265       0.012214
  Error       8.36424E+05     9  9.29360E+04
 


Example 5 - Difference between 2 coefficients has a specific nonzero value 

>hypothesis

>specify deflator-unemploy=20

>test
Hypothesis.
 
A Matrix
 
                         1           2           3           4           5
                      0.0        -1.000000    0.0         1.000000    0.0
 
                         6           7
                      0.0         0.0
Null hypothesis value for D
                    -20.000000
Test of Hypothesis
 
       Source       SS        df       MS              F              P
 
  Hypothesis   109.038029     1   109.038029       0.001173       0.973423
  Error       8.36424E+05     9  9.29360E+04

Example 6 - Same as 5, but specifying a less plausible (more extreme) difference

>hypothesis

>specify deflator-unemploy=500

>test
Hypothesis.
 
A Matrix
 
                         1           2           3           4           5
                      0.0        -1.000000    0.0         1.000000    0.0
 
                         6           7
                      0.0         0.0
Null hypothesis value for D
                   -500.000000
Test of Hypothesis
 
       Source       SS        df       MS              F              P
 
  Hypothesis  2.98665E+06     1  2.98665E+06      32.136635       0.000306
  Error       8.36424E+05     9  9.29360E+04
 
Note that this time the test is "significant": one would conclude H1 and reject the null hypothesis

Last modified 11 March 1999