**********************************************************************
*                                                                    *
*     PATHSAS 1.0                                                    *
*     Program for Analysis of Path Coefficients Using SAS            *
*     C.S. Cramer, T.C. Wehner and S.B. Donaghy                      *
*     November 18, 1997                                              *
*                                                                    *
**********************************************************************


**********************************************************************
*                                                                    *
*     readme.txt                                                     *
*     information for running PATHSAS macros                         *
*                                                                    *
*********************************************************************;

Appendix 1 - Instructions for running PATHSAS macro

	In order to utilize the PATHSAS macro, the file, pathsas.sas,
must be available.  The file, jackboot.sas, is also needed if a 
bootstrap analysis is desired.  The two files are normally stored in 
the same location or directory on the operating system.  The user 
configures their SAS code and builds a SAS dataset in the normal 
manner.  The PATHSAS macro is invoked after the SAS dataset is 
created.  The user specifies macro parameters and values for the name 
of the data set to be analyzed, the names of independent, dependent 
and any by-variables.  If desired, the user also can specify a value 
of "yes" for macro parameters for regression analysis output, direct 
and indirect effects output, simple correlation analysis output, and 
bootstrap analysis.  If a bootstrap analysis is requested, the user 
specifies a macro parameter for a random number seed and the number 
of re-samples for the bootstrap analysis.  The macro, jackboot.sas, 
must be downloaded from the SAS Institute world wide web page (http://
www.sas.com/service/techsup/faq/stat_macro/jackboot.html).

	The macro code or a %INCLUDE (alias is %INC) statement is 
needed to define the macro to the SAS system.  The user may include 
the macro into the program editor or use a %INC statement, such as 
%inc 'pathsas.sas'.  The %INC statement specifies the physical name 
of an external file where the macro is stored.  The physical name is 
the name by which the host system recognizes the file.  Depending on 
the host system and location of the file, the entire file name may 
need to be specified.

Examples:  %inc 'c:\mysas\pathsas.sas';
           %inc '~/sasmacro/pathsas.sas';

	The file, pathsas.sas, contains the SAS macro PATHSAS.  File 
names, such as pathsas.sas, usually have an extension of sas if the 
file is a SAS program or a SAS macro.  Once the macro is defined to 
SAS, the macro can be invoked.  To invoke the macro, specify the %, 
the macro name, pathsas, and the required parameters in parenthesis.

The PATHSAS macro has 12 parameters:
    DATA     - name of the SAS dataset to analyze
    INDEP    - list of independent variables
    DEP0     - primary dependent variable-for total correlation
    DEP      - other dependent variables-for total correlation
    BYLIST   - list of BY variables
    PRINTREG - print regression output? default is no,
               but user may specify a value of "yes" if desire output
    PRINTOUT - print results of direct and indirect effects? default 
               is no, but user may specify a value of "yes" if 
               desire output
    CORRIND  - print correlation coefficients for independent    
               variables?
               default is no, but user may specify a value of    
              "yes" if desire output
    CORRDEP  - print correlation coefficients for dependent      
               variables?
               default is no, but user may specify a 
               value of "yes" if desire output
    BOOT     - compute bootstrap confidence intervals for total  
               correlation variables?
               default is no, but user may specify a value of    
              "yes" if desire bootstrap analysis
    RANDOM   - seed for pseudo random numbers for bootstrap      
               analysis
    SAMPLES  - number of re-samples to calculate for bootstrap   
               default should be at least 1000 for 90%           
               confidence and greater than 1000 for higher 
               confidence value or if BC method is specified as  
               the parameter value for METHOD in the bootstrap
               macro.  The PATHSAS macro specifies the BC       
               method in the bootstrap analysis.
 
 Example:
 %pathsas(data=abc,
          indep=branplan nodebran perfenod frtset,
          dep0=frtplatn,
          dep=markplan earplan,
          bylist=density popn cyn,
          printreg=yes,
          printout=yes,
          corrind=yes,
          corrdep=yes,
          boot=yes,
          random=4571903,
          samples=1000
          );