Skip to main content
Cross Validated

Return to Question

Notice removed Draw attention by Łukasz Deryło
Bounty Ended with Jeremy Miles's answer chosen by Łukasz Deryło
added 220 characters in body
Source Link
Łukasz Deryło
Łukasz Deryło
  • 4.1k
  • 1
  • 15
  • 29

I need to validate factor structure for existing questionnaire that was translated to another language. Items of the questionnaire are organized into 3 subscales, say D1, D2 and D3. Total score, D, can also be computed (according to manual to original version of questionnaire) as a sum of 3 subscales.

So I run confirmatory factor analysis. My question is, shoud I include latent construct for total score (with arrows pointing from it to 3 subscales on path diagram) to CFA?, like this:

enter image description here

My first reasoning was that I shouldn't because my primary question is if split into 3 subscales form original version is valid in new version. So this do not involve total score at all., like this: enter image description here

I need to validate factor structure for existing questionnaire that was translated to another language. Items of the questionnaire are organized into 3 subscales. Total score can also be computed (according to manual to original version of questionnaire) as a sum of 3 subscales.

So I run confirmatory factor analysis. My question is, shoud I include latent construct for total score (with arrows pointing from it to 3 subscales on path diagram) to CFA?

My first reasoning was that I shouldn't because my primary question is if split into 3 subscales form original version is valid in new version. So this do not involve total score at all.

I need to validate factor structure for existing questionnaire that was translated to another language. Items of the questionnaire are organized into 3 subscales, say D1, D2 and D3. Total score, D, can also be computed (according to manual to original version of questionnaire) as a sum of 3 subscales.

So I run confirmatory factor analysis. My question is, shoud I include latent construct for total score (with arrows pointing from it to 3 subscales on path diagram) to CFA, like this:

enter image description here

My first reasoning was that I shouldn't because my primary question is if split into 3 subscales form original version is valid in new version. So this do not involve total score at all, like this: enter image description here

Notice added Draw attention by Łukasz Deryło
Bounty Started worth 100 reputation by Łukasz Deryło
edited tags
Link
Łukasz Deryło
Łukasz Deryło
  • 4.1k
  • 1
  • 15
  • 29
Source Link
Łukasz Deryło
Łukasz Deryło
  • 4.1k
  • 1
  • 15
  • 29

Should I include latent construct for total score in CFA?

I need to validate factor structure for existing questionnaire that was translated to another language. Items of the questionnaire are organized into 3 subscales. Total score can also be computed (according to manual to original version of questionnaire) as a sum of 3 subscales.

So I run confirmatory factor analysis. My question is, shoud I include latent construct for total score (with arrows pointing from it to 3 subscales on path diagram) to CFA?

My first reasoning was that I shouldn't because my primary question is if split into 3 subscales form original version is valid in new version. So this do not involve total score at all.