UCLA Academic Technology Services HomeServicesClassesContactJobs
Search

SAS Textbook Examples
Applied Regression Analysis by John Fox
Chapter 9: Statistical Theory for Linear Models

Linear contrasts on data file friendly on page 209. The option /E on the contrast statement gives the entire vector.
proc means data=friendly;
  class cond;
  var correct;
run;
proc glm data=friendly;
  class cond;
  model correct=cond;
  contrast 'SRF vs. others' cond  -0.5 -0.5 1 /E;
  contrast 'B vs. M'  cond  1 -1 0 /E;
run;
quit;

The MEANS Procedure

                            Analysis Variable : correct

              N
cond        Obs     N            Mean         Std Dev         Minimum         Maximum
-------------------------------------------------------------------------------------
Before       10    10      36.6000000       5.3374984      24.0000000      40.0000000

Meshed       10    10      36.6000000       3.0258149      30.0000000      40.0000000

SFR          10    10      30.3000000       7.3340909      21.0000000      39.0000000
-------------------------------------------------------------------------------------

The GLM Procedure

         Class Level Information

Class         Levels    Values

cond               3    Before Meshed SFR

Number of observations    30

The GLM Procedure

Coefficients for Contrast SRF vs. others

                           Row 1

Intercept                      0

cond      Before            -0.5
cond      Meshed            -0.5
cond      SFR                  1

The GLM Procedure

Coefficients for Contrast B vs. M

                           Row 1

Intercept                      0

cond      Before               1
cond      Meshed              -1
cond      SFR                  0

The GLM Procedure

Dependent Variable: correct

                                        Sum of
Source                      DF         Squares     Mean Square    F Value    Pr > F

Model                        2      264.600000      132.300000       4.34    0.0232

Error                       27      822.900000       30.477778

Corrected Total             29     1087.500000

R-Square     Coeff Var      Root MSE    correct Mean

0.243310      16.00194      5.520668        34.50000

Source                      DF       Type I SS     Mean Square    F Value    Pr > F

cond                         2     264.6000000     132.3000000       4.34    0.0232

Source                      DF     Type III SS     Mean Square    F Value    Pr > F

cond                         2     264.6000000     132.3000000       4.34    0.0232

Contrast                    DF     Contrast SS     Mean Square    F Value    Pr > F

SRF vs. others               1     264.6000000     264.6000000       8.68    0.0065
B vs. M                      1       0.0000000       0.0000000       0.00    1.0000

Calculation on page 214 and calculation on page 222 using data file duncan.

data inpt;
  set duncan;
  incpt=1;
run;
proc iml;
  use inpt;
  read all;
  x =  incpt || income || educ ;
  b=INV(x`*x)*x`*prestige;
  print b; /*regression coefficients*/
 
  r=prestige-x*b;
  v=(r`*r)/42;
  V2=v*INV(x`*x);
  d=diag(V2);
  a=sqrt(d);
  print a; /*estimated standard errors */
quit;

    B

-6.064663
0.5987328
0.5458339

              A

4.2719412         0         0
        0 0.1196673         0
        0         0 0.0982526

How to cite this page

Report an error on this page

UCLA Researchers are invited to our Statistical Consulting Services
We recommend others to our list of Other Resources for Statistical Computing Help
These pages are Copyrighted (c) by UCLA Academic Technology Services


The content of this web site should not be construed as an endorsement of any particular web site, book, or software product by the University of California