Hi,
I am trying to run a hierarchical factor model with items contributing to measures that contribute to elements. The data is from a survey, and I would like not only to test the fit of the model but also to calculate scores of the measure factors and element factor. Below is my code. The CFA runs just fine but I am running into trouble with the scoring. Obviously, I need the data from the factors created by the measure to make the element but those do not exist on my original data file. I'm stuck, does anyone have any ideas?
proc corr data=GENS22015 out= cfaLTCH;
var T201a T201b T201c T201d T201e T208a T208b T208c T208d T1002
T1001 T209c T1011 T209e T1010 T209g T209h T210a T210c T210e T1004 T1004b;
run;
/*Below: f1-f4 are measures, and f5 is an element*/
proc calis data=cfaLTCH nobs = 63513 method= max pestim se corr outstat=ostat;
title 'comfirmatory factor model for leadership';
lineqs
T201a = x1*f1 + e1,
T201b = x2*f1+ e2,
T201c = x3*f1 + e3,
T201d = x4*f1 + e4,
T201e = x5*f1 + e5,
T208a = x6*f2 + e6,
T208b = x7*f2 + e7,
T208c = x8*f2 + e8,
T208d = x9*f2 + e9,
T1002 = x10*f3 + e10,
T1001 = x11*f3 + e11,
T209c = x12*f3 + e12,
T1011 = x13*f3 + e13,
T209e = x14*f3 + e14,
T1010 = x15*f3 + e15,
T209g = x16*f3 + e16,
T209h = x17*f3 + e17,
T210a = x18*f4 + e18,
/*T210c = x19*f4 + e19,*/
/*T210e = x20*f4 + e20,*/
T1004 = x19*f4 + e19,
T1004b = x20*f4 + e20,
f1 = x21*f5 + e21,
f2 = x22*f5 + e22,
f3 = x23*f5 + e23,
f4 = x24*f5 + e24;
variance
f5 = 1.,
e1-e20 = u1-u20,
e21-e24 = 4*1.;
bounds 0.<=u1-u20;
run;
proc score data=GENS22015 score = ostat out= scores;
run;