3. Reports and Graphics › 3.3 Statistical Analyses › 3.3.7 Stepwise Regression Report

3.3.7 Stepwise Regression Report

The Stepwise Regression Report provides insight into the
relationships between one measurement item and one or more
other measurement items.  The report produces a model, or
formula, for deriving values of a measurement item from
values of one or more other measurement items (for example,
Z=aX+bY+c).  The option that produces the report provides
five methods for stepwise regression analysis.

Parameters are provided for regression technique,
computational parameters, and report format.  The parameters
allow you to do the following:

o  Choose forward or backward regression
o  Specify the minimum and maximum R-square values for
   inclusion (or exclusion)
o  Define entry significance levels and stay levels for use
   in model building
o  Force selected entries to be used in the model
o  Specify the minimum and maximum number of variables to be
   used in the model
o  Show model building statistics

An analysis of variance report is produced for each step in
the procedure.

Figures 3-25 and 3-26 show the first, third, and last pages
of a sample stepwise regression report.  The regression
produced a model for CICS CPU time (CSYCPUTM) based on the
number of I/O access method calls (CSYSIOCT), the number of
CICS output messages (CSYOMSGS), and the number of CICS
transactions processed (CSYTRANS).

To create a stepwise regression report, select the Stepwise
Regression Report option from either of the following menus:

o  Direct Inquiry Statistical Analysis Menu
o  Structured Inquiry Statistical Analysis Menu

(First page)
        STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE CSYCPUTM

STEP 1   VARIABLE CSYSIOCT ENTERED    R SQUARE = 0.34202814
                                      C(P) =    22.10532991

              DF          SUM OF SQUARES     MEAN SQUARE           F    PROB>F

REGRESSION     1           3808.14253047   3808.14253047       24.95    0.0001
ERROR         48           7325.86111002    152.62210646
TOTAL         49          11134.00364049

                 B VALUE       STD ERROR      TYPE II SS           F    PROB>F

INTERCEPT     4.32510779
CSYSIOCT      0.00778661      0.00155884   3808.14253047       24.95    0.0001











(Third page)
        STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE CSYCPUTM

STEP 3   VARIABLE CSYTRANS ENTERED    R SQUARE = 0.56032271
                                      C(P) =     3.51010231

              DF          SUM OF SQUARES     MEAN SQUARE           F    PROB>F

REGRESSION     3           6238.63511260   2079.54503753       19.54    0.0001
ERROR         46           4895.36852789    106.42105495
TOTAL         49          11134.00364049

                 B VALUE       STD ERROR      TYPE II SS           F    PROB>F

INTERCEPT     7.88105857
CSYTRANS     -0.27525872      0.06954592   1667.11809796       15.67    0.0003
CSYSIOCT      0.01014079      0.00156759   4453.53773981       41.85    0.0001
CSYOMSGS      0.32783844      0.07301867   2145.25753545       20.16    0.0001

Figure 3-25.  Sample Stepwise Regression Report (Part 1)

(Last page)
        STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE CSYCPUTM

NO OTHER VARIABLES MET THE 0.1500 SIGNIFICANCE LEVEL FOR ENTRY INTO THE MODEL.

   SUMMARY OF STEPWISE REGRESSION PROCEDURE FOR DEPENDENT VARIABLE CSYCPUTM

                    VARIABLE       NUMBER  PARTIAL   MODEL
         STEP  ENTERED   REMOVED       IN     R**2    R**2       C(P)

            1  CSYSIOCT                 1   0.3420  0.3420    22.1053
            2  CSYOMSGS                 2   0.0686  0.4106    17.0086
            3  CSYTRANS                 3   0.1497  0.5603     3.5101

               VARIABLE
    STEP  ENTERED   REMOVED           F    PROB>F  LABEL

       1  CSYSIOCT              24.9514    0.0001  ACCESS METHOD CALLS
       2  CSYOMSGS               5.4672    0.0237  OUTPUT MESSAGES
       3  CSYTRANS              15.6653    0.0003  TRANSACTIONS PROCESSED

Figure 3-26.  Sample Stepwise Regression Report (Part 2)