Setting a lambda coefficient to the value of a specified variable

General
1

I’m hoping someone can help me out with an R problem. I want to run a CFA, but set the factor loading to a value obtained from another CFA. And I want to do this by using a variable name, and not actually type out the number.

So, here’s the code that does not work but describes what I want to do:

BasicCFAModel <- ' construct =~ V1 + V2 + V3 + V4'
BasicCFAModelResults <- cfa(BasicCFAModel, data=DataMatrix)
l1 <- BasicCFAModelResults@Fit@est[1]
l2 <- BasicCFAModelResults@Fit@est[2]
l3 <- BasicCFAModelResults@Fit@est[3]
l4 <- BasicCFAModelResults@Fit@est[4]

So, this works, and I get my 4 desired lambda values, called l1, l2, l3, and l4,

But this is what won’t work:

NewModel <- ' f1 =~ Lambda1V1 + Lambda2V2 + Lambda3V3 + Lambda4V4
Lambda1 == l1
Lambda2 == l2
Lambda3 == l3
Lambda4 == l4
'
NewModelResults <- cfa(NewModelResults, data=DataMatrix)

Likewise, this also doesn’t work

NewModel <- ' f1 =~ l1V1 + l2V2 + l3V3 + l4V4
'
NewModelResults <- cfa(NewModelResults, data=DataMatrix)

This will run, but it estimates the values of l1, l2, l3, l4 as normal (i.e., it does not constrain them to equal the values computed earlier).

Any suggestions?

2

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.