SRMR for second order construct measurement model

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
Man18
PLS User
Posts: 16
Joined: Mon Apr 18, 2016 1:55 pm
Real name and title: Manuela FR

SRMR for second order construct measurement model

Post by Man18 »

Hi,
I hope find an answer in this forum unless i'm really disapointed i really need your precious help. I read many papers before asking and still not find a clear answer specially for 2nd order construct.
SRMR should be interpreted for both model fit with second order construct measurement and structural model or just for structural one??
I used repeated indicator approach mode A since my model is type I. but unfortunately when conducting PLS algorithm under smartPLS i found bad model fit indices (SRMR =.0123) even tht AVE, rho , alpha of Cronbach, discrimant validity are all ok and also when i run bootstrap all the outer loadings and outer weights are all significant.
for structural model after using scores of LV then run PLS algorithm again SRMR is not pretty good 0.093 when i run bootstrap i found significant result for my hypothesis. I'm so confused am i doing something wrong?
any clarifications pleaaaase
Note.my model is composed by 3 LVs all second order+ all reflective (2 exo+1 endo)
Thank you all
Man18
PLS User
Posts: 16
Joined: Mon Apr 18, 2016 1:55 pm
Real name and title: Manuela FR

Re: SRMR for second order construct measurement model

Post by Man18 »

Please no one can answer me :( i need to know if i am in the wrong way pleaaase
User avatar
Hengkov
PLS Super-Expert
Posts: 1599
Joined: Sun Apr 24, 2011 10:13 am
Real name and title: Hengky Latan
Location: AMQ, Indonesia
Contact:

Re: SRMR for second order construct measurement model

Post by Hengkov »

If the goal you want to test the theory with the data, then you are using SRMR on both. But, if your goal is just to predictions, it is only on the structural model. Keep in mind, if only rely on SRMR was not accurate, in the latest version SmartPLS, has provided some measure of fit another. And also to model the 2nd, need caution in the interpretation of GoF.
Man18
PLS User
Posts: 16
Joined: Mon Apr 18, 2016 1:55 pm
Real name and title: Manuela FR

Re: SRMR for second order construct measurement model

Post by Man18 »

thank you so much for your kind reply
well i have to test 2 hypothesis for 2nd order LVs that's my goal. well if i understand and please correct me : srmr can be much better if the outer loading is ok ?
the confustion i have is why i found significant outer loading when i run bootstrap but before that (with pls algorithm) i find some outer loading between 2nd order and items of 1st order under .6 ? knowing that outer loading between items and 1st order LV are all ok before and after bootsrap.
I thought that i should eliminate one by one but when i do it srmr become little better and i was faced to eliminate sometimes a whole dimension and don't know if it's permetted for my reflective model to do!
do i continue on this way until i have may be an acceptable srmr? (it remembers me the process of modifications indices and other solutions in amos when you try to correct the model fit)
Any advice will be appreciable i'm really in vicious circle

I'm using the latest release and saw the new indices CHI 2 and NFI and others whom i never see. i don't know how to interpret them since i used repeated indicator and find also n/a for some of them. any references?

thank you again
Man18
PLS User
Posts: 16
Joined: Mon Apr 18, 2016 1:55 pm
Real name and title: Manuela FR

Re: SRMR for second order construct measurement model

Post by Man18 »

Please help me :-( no body can answer me?!
RaffaeleP
PLS Junior User
Posts: 8
Joined: Sat Aug 08, 2015 10:20 am
Real name and title: Raffaele Parise

Re: SRMR for second order construct measurement model

Post by RaffaeleP »

Hi Man18,

did you run pls twice with different inputs?
Because SmartPLS is not able to create 2nd Order Results by running only once.

Please see this video: https://www.youtube.com/watch?v=kPeUTKjMF7o and read this article http://www.sciencedirect.com/science/ar ... 6312002172.

I think the video and the paper will help you.
Especially the paper is about that 2nd Order Constructs that are refelctive are not that useful.
In addition, don't base your findings to much on SRMR because the use of it related to PLS is quite new and until now not very much has been written about it.

Best, Raffaele
Last edited by RaffaeleP on Sun May 22, 2016 9:41 am, edited 1 time in total.
RaffaeleP
PLS Junior User
Posts: 8
Joined: Sat Aug 08, 2015 10:20 am
Real name and title: Raffaele Parise

Re: SRMR for second order construct measurement model

Post by RaffaeleP »

you just got this answer:
jmbecker wrote:You are raising very good questions. The problem is that they are mostly unanswered in the literature.
Some researchers started to request these model fit indices for PLS. In my opinion, without a solid knowledgebase on their behavior. As researchers are requesting these features, we (SmartPLS) decided to provide them to fulfill the demand.
Yet, we believe that there is much more research necessary to apply them appropriately. Easy access to these indices via SmartPLS might foster the necessary research. Hence, we decided to implement them, even though they are not yet fully understood.
We are preparing some information for the resource section on smartpls.com in which we will try to summarize current findings and our position on using these new indices.

My personal recommendation would be: use them with care; experiment with them; get a feeling on their behavior; but don’t base too much weight on assessing these indices without further research.

One observation regarding your model: You are probably having repeated indicators in your model (e.g., a second-order repeated indicator model). In this case some of the indices are not properly defined (you will get in “infinite” or “n/a” value). Something else that needs further research.
Man18
PLS User
Posts: 16
Joined: Mon Apr 18, 2016 1:55 pm
Real name and title: Manuela FR

Re: SRMR for second order construct measurement model

Post by Man18 »

RaffaeleP wrote:Hi Man18,

did you run pls twice with different inputs?
Because SmartPLS is not able to create 2nd Order Results by running only once.

Please see this video: https://www.youtube.com/watch?v=kPeUTKjMF7o and read this article http://www.sciencedirect.com/science/ar ... 6312002172.

I think the video and the paper will help you.
Especially the paper is about that 2nd Order Constructs that are refelctive are not that useful.
In addition, don't base your findings to much on SRMR because the use of it related to PLS is quite new and until now not very much has been written about it.

Best, Raffaele
Hi Raffaele
thank you for your kind reply. it's the same input. I can't understand where is the fault when i am assesing measurement model or there is step i'm missing. Since i read a paper which said (for both reflective & formative) you have to delete items with low and non signicant outer loading but in my case even some are below 0.5 they are all signifcant that's makes me confused. the paper doesn't provide more informations and the application was for formative.

if i don't rely on srmr which indicator i'll use to proove my model fit??! Even the Gof is not considered as fit index and it's not provided by smartpls. i'm very upset and still in vicious circle :( any suggestions any advice
Note . unfortunatlly i can't have access for your link. i know this video since my first time to deal with SmartPLs but still not resolve my problem.
Post Reply