How to specifiy significance level to be 0.5 in ProcGLMSelect

[LABRADOR]

I am trying to build a stepwise regression model using debt as the dependent variable, and all of the other variables in the data file as a predictor variables.

Here is the data:

/* --------------------------------------------------------------------
   Code generated by a SAS task
   
   Generated on Tuesday, March 22, 2022 at 6:55:23 PM
   By task:     Import Data Wizard
   
   Source file: P:\MS_3313\Exam_1\GDP.xlsx
   Server:      Local
   
   Output data: WORK.GDP
   Server:      Local
   -------------------------------------------------------------------- */

/* --------------------------------------------------------------------
   This DATA step reads the data values from a temporary text file
   created by the Import Data wizard. The values within the temporary
   text file were extracted from the Excel source file.
   -------------------------------------------------------------------- */

DATA WORK.GDP;
    LENGTH
        Year               8
        GDP                8
        Budget             8
        Surplus            8
        'Relative Surplus'n   8
        Debt               8
        Interest           8
        'Relative Interest'n   8 ;
    FORMAT
        Year             BEST12.
        GDP              BEST12.
        Budget           BEST12.
        Surplus          BEST12.
        'Relative Surplus'n BEST12.
        Debt             BEST12.
        Interest         BEST12.
        'Relative Interest'n BEST12. ;
    INFORMAT
        Year             BEST12.
        GDP              BEST12.
        Budget           BEST12.
        Surplus          BEST12.
        'Relative Surplus'n BEST12.
        Debt             BEST12.
        Interest         BEST12.
        'Relative Interest'n BEST12. ;
    INFILE 'C:\Users\ssd506\AppData\Roaming\SAS\EnterpriseGuide\EGTEMP\SEG-3756-4b181ba3\contents\GDP-d0419501d7d1440499323245eac8b347.txt'
        LRECL=36
        ENCODING="WLATIN1"
        TERMSTR=CRLF
        DLM='7F'x
        MISSOVER
        DSD ;
    INPUT
        Year             : BEST32.
        GDP              : BEST32.
        Budget           : BEST32.
        Surplus          : BEST32.
        'Relative Surplus'n : BEST32.
        Debt             : BEST32.
        Interest         : BEST32.
        'Relative Interest'n : BEST32. ;
RUN;

Here is my  code:

proc contents varnum data=work.gdp;
	ods select position;
run;
proc glmselect data=work.gdp plots=all;

	class Debt Budget GDP Budget Interest 'Relative Interest'n 'Relative Surplus'n Surplus Year;
	model Debt = Budget GDP Budget Interest 'Relative Interest'n 'Relative Surplus'n Surplus Year
	/ details=all stats=all selection=stepwise;
	
run;
ods graphics off;

[ballardw · Accepted solution]

The MODELAVERAGE statement has an Alpha= option and default is Alpha=0.5. So may investigate that statement.

[LABRADOR]

Thanks you!

[LABRADOR]

Does this look correct?

proc contents varnum data=work.gdp;
	ods select position;
run;
proc glmselect data=work.gdp plots=all;
	modelaverage alpha=0.5;
	class Debt Budget GDP Budget Interest 'Relative Interest'n 'Relative Surplus'n Surplus Year;
	model Debt  = Budget GDP Budget Interest 'Relative Interest'n 'Relative Surplus'n Surplus Year
	/ details=all stats=all selection=stepwise;
	
run;
ods graphics off;