questions about manifest and latent variables

#1
hi
i have 2 main questions


Latent variables are unobserved variables that are measured by multiple observed items(called manifest variables).


now my questions:
1- each latent variable is an unobserved variable
are all unobserved variable called "latent"?


2- each manifest variable is an observed variable
are all observed variable called "manifest"?

In other words, are always these equations true:
1) latent variable = unobserved variable
2) manifest variable = observed variable

thanks!
 

Englund

TS Contributor
#2
I think both statements are unidirectional. I would say that a latent variable is a variable that cannot be observed directly. An unobserved variable is simply an unobserved variable. If you collect data on weight and height, then number of times I've shaved myself this week is not a latent variable, although unobserved.
 

noetsi

No cake for spunky
#3
I also think this language is largely artificial and likely to vary with the author. In the SEM courses I took, which dealt with similar topics, they never used the term manifast (they talked of observed items which measured latent variables).
 
#4
thank you both

now my first problem is solved

but second question, not yet!
2- each manifest variable is an observed variable
are all observed variable called "manifest"?

noetsi said that in SEM courses they never use the term "manifest"
but still i see this word in texts...

when they use manifest, do they mean those observed variable that measure latent variabe? or do they mean All observed variables?


thanks !
 
Last edited:

spunky

Can't make spagetti
#5
2- each manifest variable is an observed variable
are all observed variable called "manifest"?

neotsi said that in SEM courses they never use the term "manifest"
but still i see this word in texts...
yes. whether you use 'observed' or 'manifest' it doesn't matter. they both mean the same thing.

source: me (I teach this stuff).
 

noetsi

No cake for spunky
#6
I actually did not say the never use the term manifast in SEM - I said they did not use it in my SEM courses. It is entirely possible they do in some discussions of SEM. A basic issue in statistics is that the language is not standardized so different statisticians and different research areas can use very different terms for the same thing. It is common in discussions of statistics to see 'this is also called ...by some authors."