Fit Summary - What is d_G and d_ULS

[Newbie456]

Hi all,

I am new to this software. Can anyone tell me what these measurements are d_G and d_ULS are, I see them in my Fit Summary? Screenshot attached. Thanks!

Kind regards,
N

[cringle]

Thanks. Please take a look at the SmartPLS resources under www.smartpls.com:
https://www.smartpls.com/documentation/fit

Best regards
Christian

[raschel]

hi

i am a student from Iran in marketing management.
d_G and d_ULS, these items examine to assay the fitness of model
The bootstrap routine provide the confidence intervals of these discrepancy values. A model fits well if the difference between the correlation matrix implied by your model and the empirical correlation matrix is so small that it can be purely attributed to sampling error. Hence, the difference between the correlation matrix implied by your model and the empirical correlation matrix should be non-significant (p > 0.05). Otherwise, if the discrepancy is significant (p < 0.05), model fit has not been established.(ref>> www.smartpls.de/documentation/fit)

[cringle]

@raschel: Thanks! That's perfectly fine. d_G and d_ULS values themselves are not useful. You need to take a look at their bootstrapping results (use the option "complete" in bootstrapping). If the original value of d_G and d_ULS is smaller than the upper bound of the bootstrap confidence interval, you establish fit.

Best
CR

[mnabi]

@raschel: Thanks! That's perfectly fine. d_G and d_ULS values themselves are not useful. You need to take a look at their bootstrapping results (use the option "complete" in bootstrapping). If the original value of d_G and d_ULS is smaller than the upper bound of the bootstrap confidence interval, you establish fit.

Best
CR

Dear Professor Ringle,

1. What do you mean by the "original value"? Is it the saturated model or estimated (because the both have very different values in my research. Saturated model value shows fit, but the estimated model doesn't)!?
2. what if the model doesn't fit for an exploratory one? What can we do for modifying the model?
3. Which stage should we choose for assessing the fit indices in the two stage approach?

Thanks

[cringle]

Hi

@1: The result of the model estimation; see further explanations here: https://www.smartpls.com/documentation/ ... /model-fit

@2: Don't know - I usually do not use fit (see the note of caution at the end of the previously quoted link). The CB-SEM researchers usually delete indicators until they have a super reliable model and also add structural model relationships to improve fit.

@3: Well, if you use the results of the second stage as final results, I would look into fit here.

Best
CR

[mnabi]

Hi

@1: The result of the model estimation; see further explanations here: https://www.smartpls.com/documentation/ ... /model-fit

@2: Don't know - I usually do not use fit (see the note of caution at the end of the previously quoted link). The CB-SEM researchers usually delete indicators until they have a super reliable model and also add structural model relationships to improve fit.

@3: Well, if you use the results of the second stage as final results, I would look into fit here.

Best
CR

Thank you very much.
Just to confirm: By "the result of the model estimation", do you mean the estimated model SRMR outcome? Then, what is the saturated model SRMR outcome useful for?