Common methods variance PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Tristan posted on Saturday, March 03, 2012 - 6:13 am
Hi there,

I am new to SEM and MPLUS, and attempting to model a latent factor that influences all indicators in the structural model as per the recommendations of Podsakoff et al (2003). I am getting some weird results when I compare the path coefficients in my structural models (no CMV factor versus CMV factor present). The coefficients (and overall R2 for endogenous variable) differs significantly between CMV and no CMV models, so I want to check if I am modelling CMV correctly. My code is below:

MODEL:
factor definition and path modelling omitted...

CMV by Q9ci_RecWelc-Q29_ServNeeds;
CMV@1;
CMV with RECEPT@0 ROOM@0 WAI@0 POSTAPP@0 SERVQUAL@0 Q24_Oasat@0;

Thanks so much,

Tristan
 Linda K. Muthen posted on Saturday, March 03, 2012 - 8:44 am
I see nothing wrong in what you present. To answer your question, I need to see the full outputs and your license number at support@statmodel.com.
 Jorge Walter posted on Thursday, July 03, 2014 - 10:39 am
Dear Linda,

I'm working on the same common method test described above. And while my models are working, I was wondering how exactly I could use the output of such a CFA model to partition the variance, i.e., estimate the percentage of variance in responses due to trait, method, and random error components
(Podsakoff et al., 2003; Williams, Cote, & Buckley, 1989)?

Thanks in advance for your help!
 Bengt O. Muthen posted on Thursday, July 03, 2014 - 4:21 pm
You may want to ask this general question on SEMNET. The Mplus Model Constraint command can be used for any such expression.
 Jorge Walter posted on Saturday, July 05, 2014 - 4:21 am
Thanks, Bengt, will do.
 Tunde Ogunfowora posted on Tuesday, February 03, 2015 - 12:31 pm
Hi there,

I am trying to test for common method bias effects in a 6-factor measurement model. Of the six factors, three of them are single indicators (one of which is categorical with two levels).

I have specified the following:


Model:

OrgCyn_F by ogcyn1 ogcyn4 Ogcyn2 ogcyn3 ogcyn5;
Ego_Cse_F by Egoist1 Egoist4 Egoist3 Egoist2;
Val_Cse_F by values3 values4 values2 values5;


Ego_Cse_F OrgCyn_F Val_Cse_F voluntr purchse DonatBhv with Ego_Cse_F OrgCyn_F Val_Cse_F voluntr purchse DonatBhv;


CMV_F by voluntr
purchse ogcyn1 ogcyn4 Ogcyn2 ogcyn3 ogcyn5 Egoist1 Egoist4 (a)
Egoist3 Egoist2 values3 values4 values2 values5 DonatBhv (a);

CMV_F@1 ;
CMV_F with OrgCyn_F@0 Val_Cse_F@0 Ego_Cse_F@0 ;


My understanding is that constraining all variables to load equally on the CMV factor allows one to compute the proportion of variance explained by the CMV factor. This is why I have constrained using (a).

However, I am also interested in testing changes in model fit between the 6-factor model and the 7 factor model. Constraining the factor loadings (a) gives a different chi-squared estimate compared to the unconstrained model. Can you please educate me on which of these is appropriate?
 Tunde Ogunfowora posted on Tuesday, February 03, 2015 - 12:32 pm
Also, do I have to specify anything for the single item variables? For instance, do I have to specify the following separately?:

DonatBhv;
purchse;
voluntr;

Lastly, do I have to specify correlations (equal to zero) between the CMV factor and the single item variables like i did for the latent variables? Specifically:

CMV_F with DonatBhv@0 purchse@0 voluntr@0;


Thank you for your help!

Tunde
 Bengt O. Muthen posted on Tuesday, February 03, 2015 - 4:44 pm
You want to ask these general modeling questions on SEMNET which is a more general discussion forum.

Also, we ask the no more than one window is used for Mplus Discussion.
 Tunde Ogunfowora posted on Wednesday, February 04, 2015 - 9:05 am
Hi Bengt,

I will look into SEMNET for my general SEM questions. However, can you please let me know if my MPLUS syntax is appropriate for what i am trying to achieve:

Specifically, do I have to specify the following lines for my single item factors? This is a Mplus specific coding question.

DonatBhv;
purchse;
voluntr;

Thanks,

Tunde
 Bengt O. Muthen posted on Wednesday, February 04, 2015 - 3:21 pm
No, you just use the variables in the ON or WITH statements you have in mind.
 Tunde Ogunfowora posted on Wednesday, February 04, 2015 - 3:35 pm
Thanks!
 baozhenzhou posted on Saturday, April 25, 2015 - 12:39 am
Hi there,
I'm working on the same common method test described above. However, when i put the latent cmv factor into the CFA model, the model became no convergence,and the number of interactions is exceeded.Could you tell me how can i solve this problem? Thanks very much!
 baozhenzhou posted on Saturday, April 25, 2015 - 2:04 am
Hi there,
I'm working on the same common method test described above. However, when i put the latent cmv factor into the cfa model, the model became no convergence and the number of iterations is exceeded. Could you tell me how can i solve this problem? Thanks very much!
 Bengt O. Muthen posted on Saturday, April 25, 2015 - 1:05 pm
We need to see our output to be able to say - please send to support along with your license number.
 Jeff Shao posted on Wednesday, October 05, 2016 - 2:30 pm
I want to ask a question with regard to the Tristan's (the first person in this discussion chain) and Tunde's (the sixth person in this discussion chain) codes. why restricting the variance of the method factor to 1? for example: CMV@1 and CMV_F@1?
 Jeff Shao posted on Wednesday, October 05, 2016 - 2:40 pm
In Williams, Cote, and Buckley (1989), the authors calculated the variance partitioning for traits, methods and errors. Can anyone tell me how they made the calculation, please?

In the paper, it says "for each measure, the square of the trait factor loading indicates the percentage of variance due to the trait factors, and the square of the method factor loading indicates the amount of variance due to the method factor". However, there are several factor loadings for each latent variable. How to calculate a single variance score as the authors did in the paper to indicate how much the method factor explains the variance?

Many thanks!
 Bengt O. Muthen posted on Wednesday, October 05, 2016 - 2:55 pm
Q1. It doesn't seem necessary if the first loading is fixed.

Q2. You want to ask this general analysis question on SEMNET.
 Gargi Sawhney posted on Tuesday, December 13, 2016 - 1:05 pm
Hello,

When I run a CFA model with all indicators, my model runs fine. However, the moment I add a method factor, I get the following error message:

THE STANDARD ERRORS OF THE MODEL PARAMETER ESTIMATES COULD NOT BE
COMPUTED. THE MODEL MAY NOT BE IDENTIFIED. CHECK YOUR MODEL.
PROBLEM INVOLVING THE FOLLOWING PARAMETER:
Parameter 102, METHOD

What I can do in this case? Thanks for your help!
 Bengt O. Muthen posted on Tuesday, December 13, 2016 - 6:18 pm
You should send this general question to SEMNET, but you need to describe the model fully.
 John Jeffrey Gish posted on Thursday, April 13, 2017 - 5:41 pm
I have a similar problem to baozhenzhou above.

"...when i put the latent cmv factor into the CFA model, the model became no convergence,and the number of interactions is exceeded.Could you tell me how can i solve this problem? Thanks very much!"

My CFA has good fit indices, but I'm trying to test for common method bias with a common latent factor. I get the same error message as above. Any suggestions based on the previous investigation?
 Bengt O. Muthen posted on Friday, April 14, 2017 - 3:37 pm
Try SEMNET for this one.
 Abdulrahman posted on Monday, December 18, 2017 - 3:53 pm
Hi there,

I am trying to test for common method bias effects using unmeasured latent method factor technique,

May i ask how to do that using Mplus?

Best regards
 Bengt O. Muthen posted on Tuesday, December 19, 2017 - 3:46 pm
This question is suitable for SEMNET.
 Dinah Spitzley posted on Wednesday, May 22, 2019 - 2:20 am
Hello, i am trying to test for common method bias using latent factor method with mplus. I have figured using the constraint command, but can I find the mplus commands on this somewhere in the user guide? I could not find anything on how to perfom this with mplus. I would really appreciate if someone could help me!
 Bengt O. Muthen posted on Wednesday, May 22, 2019 - 4:16 pm
I don't know what "common method bias using latent factor method" is. SEMNET might help you.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: