Help using Base SAS procedures

Including reliabilities in SAS PROC CALIS

Reply
New Contributor
Posts: 3

Including reliabilities in SAS PROC CALIS

I wish to includ known reliabilities for an SEM analysis. The relaibilities are from an external source.

Below is my code. THe results don't make sense  (-ve coefficients, some very large standardized coefficients etc.)

And this error message.

 

Could someone please tell me where my code is incorrect!  

 

title "Fixed Effects Analysis of Numeracy";
title "Including Reliabilities";

lineqs
ses08= 1 * fses + es1, ses10= lambda1 * fses + es2, ses12= lambda2 * fses + es3,
num08 = 0.87 * fnum08 + ee1,num10 = 0.87 * fnum10 + ee2, num12= 0.92 * fnum12 + ee3,
rdg08 = 0.89 * frdg08 + ee4,rdg10 = 0.85 * frdg10 + ee5, rdg12= 0.88 * frdg12 + ee6,
wrt08 = 0.75 * fwrt08 + ee7,wrt10 = 0.75 * fwrt10 + ee8, wrt12= 0.75 * fwrt12 + ee9,
spl08 = 0.92 * fspl08 + ee10,spl10 = 0.92 * fspl10 + ee11, spl12= 0.93 * fspl12 + ee12,
grm08 = 0.78 * fgrm08 + ee13,grm10 = 0.75 * fgrm10 + ee14, grm12= 0.82 * fgrm12 + ee15,

fnum08 = tn1 * intercept + betan1 * fses + betan31 * sesx08 + e1,
fnum10 = tn2 * intercept + betan1 * fses + betan21 * fnum08 + betan32 * sesx10 + betan41 * factorx08_10 + e2,
fnum12 = tn3 * intercept + betan1 * fses + betan22 * fnum10 + betan33 * sesx12 + betan42 * factorx10_12 + e3,
frdg08 = tr1 * intercept + betan1 * fses + betar31 * sesx08 + e4,
frdg10 = tr2 * intercept + betar1 * fses + betar21 * frdg08 + betar32 * sesx10 + betar41 * factorx08_10 + e5,
frdg12 = tr3 * intercept + betar1 * fses + betar22 * frdg10 + betar33 * sesx12 + betar42 * factorx10_12 + e6,
fwrt08 = tw1 * intercept + betaw1 * fses + betaw31 * sesx08 + e7,
fwrt10 = tw2 * intercept + betaw1 * fses + betaw21 * fwrt08 + betaw32 * sesx10 + betaw41 * factorx08_10 + e8,
fwrt12 = tw3 * intercept + betaw1 * fses + betaw22 * fwrt10 + betaw33 * sesx12 + betaw42 * factorx10_12 + e9,
fspl08 = ts1 * intercept + betas1 * fses + betas31 * sesx08 + e10,
fspl10 = ts2 * intercept + betas1 * fses + betas21 * fspl08 + betas32 * sesx10 + betas41 * factorx08_10 + e11,
fspl12 = ts3 * intercept + betas1 * fses + betas22 * fspl10 + betas33 * sesx12 + betas42 * factorx10_12 + e12,
fgrm08 = tg1 * intercept + betas1 * fses + betag31 * sesx08 + e13,
fgrm10 = tg2 * intercept + betag1 * fses + betag21 * fgrm08 + betag32 * sesx10 + betag42 * factorx08_10 + e14,
fgrm12 = tg3 * intercept + betag1 * fses + betag22 * fgrm10 + betag33 * sesx12 + betag42 * factorx10_12 + e15
;
variance
es1-es3= ess1-ess3,
e1-e15 = s1-s15,
ee1-ee15= ss1-ss15
;
ODS output fit=num_fit LINEQSEffects=all_ESt LINEQSEffectsStd=all_EST_std SqMultCorr=All_SMC;
run;

Frequent Contributor
Posts: 98

Re: Including reliabilities in SAS PROC CALIS

[ Edited ]

You mention an error message, but you don't list it. What was the error?

 

It's difficult to tell whether or not your system of equations is correctly specified without knowing what the variables represent or what exactly you are trying to model. It would also help if you reformatted it a bit, to put each equation on its own line, right now it is hard for me to tease anything apart.

 

Are you sure you have specified the right variances? As I said, the LINEQS statement is just a wall of text so I am having a hard time figuring out how many parameters you have in there, but you might want to double-check your variance statement corresponds to the lineqs statement.

Ask a Question
Discussion stats
  • 1 reply
  • 168 views
  • 0 likes
  • 2 in conversation