Prac Factor and Proc Score

[Shakti_Sourav]

Hy,

I want to calculate by formula instead of using Proc Factor and Proc Score in SAS EG.

I want CSAT score from avbailable Varibales i.e positive behaviour, positive bribe and positive service,

i have attached one examnple by the name of Example1 and Example 2

help me what will be the formula to get this score

Thank You

ShaktiExample1Example2my dataset like this, and i have to calculate CSAT Score by taking of these 3 variables and CSAT score must be under 80%

[PaigeMiller]

What exactly is the problem with the code you are showing? Why isn't PROC SCORE sufficient in this case?

[Shakti_Sourav]

Proc factor and Proc score both are not accessible in my system.

[Shakti_Sourav]

I just want formula instead of proc score and Proc factor

[PaigeMiller]

So what is wrong with the formula you typed?

[Shakti_Sourav]

which formula calculate the factor. I have used that Varibales-(mean(Variables)/Std). is it calculating the Factor/Score?

[Zard]

You probably need to center the raw data as well. See if this helps:

title 'PROC FACTOR';
proc factor data=sashelp.class method=prin nfact=2 outstat=outstat out=out ;
var height weight;
run;
proc print data=out;
run;

title 'PROC SCORE';
/*The PROC SCORE statement does not contain the NOSTD option,
so the data in the input data set are standardized before scoring.*/
proc score data=sashelp.class score=outstat out=scored;
var height weight;
run;
proc print data=scored;
run;

title 'BY HAND';
proc print data=outstat;
run;
/*
Obs _TYPE_ _NAME_ Height Weight
1 MEAN 62.3368 100.026
2 STD 5.1271 22.774
3 N 19.0000 19.000
4 CORR Height 1.0000 0.878
5 CORR Weight 0.8778 1.000
6 COMMUNAL 1.0000 1.000
7 PRIORS 1.0000 1.000
8 EIGENVAL 1.8778 0.122
9 PATTERN Factor1 0.9690 0.969
10 PATTERN Factor2 0.2472 -0.247
11 SCORE Factor1 0.5160 0.516
12 SCORE Factor2 2.0227 -2.023
*/

data scores_by_hand;
set sashelp.class;
MeanHeight= 62.3368 ;
MeanWeight=100.026 ;
StdHeight=5.1271;
StdWeight= 22.774 ;
F1_StdScoringCoeff_Height=0.51601467;
F1_StdScoringCoeff_Weight=0.51601467;
F2_StdScoringCoeff_Height=2.02266093;
F2_StdScoringCoeff_Weight=-2.02266093;
Factor1=F1_StdScoringCoeff_Height*(Height-MeanHeight)/StdHeight +
F1_StdScoringCoeff_Weight*(Weight-MeanWeight)/StdWeight ;
Factor2=F2_StdScoringCoeff_Height*(Height-MeanHeight)/StdHeight +
F2_StdScoringCoeff_Weight*(Weight-MeanWeight)/StdWeight ;
drop MeanHeight MeanWeight StdHeight StdWeight
F1_StdScoringCoeff_Height F1_StdScoringCoeff_Weight
F2_StdScoringCoeff_Height F2_StdScoringCoeff_Weight ;
run;
proc print data=scores_by_hand;
run;

[Shakti_Sourav]

Thank You

But How to calculate by formula of F1_StdScoringCoeff_Height and F1_StdScoringCoeff_Weight ??

[Shakti_Sourav]

I don't have access for SAS/STAT. that's why I am Unable to use Proc Factor and Proc score. So I want to do by Formula instead of Proc factor and proc score. 

[SteveDenham]

I don't understand.  Your first post shows output from PROC FACTOR, so someplace along the line it was available.  Could you explain your situation further?

SteveDenham

[Shakti_Sourav]

Hy,

other system have access Proc factor. but in my system don't have an access of SAS/STAT. so I want to calculate by formula..

Thank you

[Zard]

I misunderstood the question. Let me regroup...

[Zard]

If you want to derive the factor scoring coefficients outside of SAS, you would still need a matrix algebra software package. Here is an example (which I did not write) using PROC IML:

proc corr data=sashelp.class outp=corr;
var age height weight;
run;

proc factor data=corr method=prin;
ods select eigenvalues;
run;

proc iml;
R= {1.0000 0.8114 0.7409,
0.8114 1.0000 0.8778,
0.7409 0.8778 1.000};
priors={1 1 1};
d=diag(priors);
r1 = R-i(3) +d;
print r r1;
call eigen(e1,e2,r1);
print e1;
quit;