Message/Author 

Anonymous posted on Monday, November 28, 2005  1:16 pm



I'm working on a multigroup model and I was wondering how I might hold the residual variances equal for my groups? I can't seem to find this in the user's guide. Thank you 


If y1y4 are the endogenous variables, you would say in the MODEL command: y1y4 (1); This would hold them equal to each other and equal across groups. y1 (1); y2 (2); y3 (3); y4 (4); would hold them equal only across groups. 

Helen Zhao posted on Wednesday, May 11, 2011  2:18 am



Hi professors, I wonder how could I label the residual variance of DV (everything cannot be explained by IV is explained by this residual variance)? I found the way of labeling residual variance for each indicator in UG (example 5.20 in version 6 UG) but i cannot find how to label this residual variance. Can you please help? Thank you! 


Residual variances are labeled in the same way as the variances shown in Example 5.20. Both variances and residual variances are referred to by the variable name. 

Kate Walton posted on Wednesday, October 12, 2011  8:03 am



Here is part of the input: MODEL: in by anti_in@1 alc_in nic_in sub_in; fu1 by anti_fu1@1 alc_fu1 nic_fu1 sub_fu1; fu3 by anti_fu3@1 alc_fu3 nic_fu3 sub_fu3; ext by in@1 fu1 fu3; marry on ext; anti_fu3 with marry; alc_fu3 with marry; nic_fu3 with marry; sub_fu3 with marry; This is a complex, multigroup design, using theta parameterization. The indicators are categorical, as is marry. When I added in “marry on ext”, the psi matrix contained some estimates that I don’t understand. For the first group, there are estimates for (1) ext variance, and covariance between (2) marry and anti_fu3, (3) marry and alc_fu3, (4) marry and nic_fu3, and (5) marry and sub_fu3. For the second group, these five estimates are in psi, as well as (6) anti_fu3, (7) alc_fu3, (8) nic_fu3, and (9) sub_fu3 variances. According to the manual, the psi matrix contains the variances and covariances of the continuous latent variables. Anti, alc, etc. are not continuous latent variables so I’m not sure why they are showing up in the psi matrix. Second, I want to know the correlation between marry and the residual variance of anti_fu3, between marry and the residual variance of alc_fu3, etc. Is the “with” statement giving me this? 


When variables appear in psi rather than theta, this is not a problem. It is a behind the scenes part of Mplus that has not statistical meaning or consequence. The matrix where the parameter appears does not affect its estimation. These are residual covariances. 

Sanja Franic posted on Wednesday, November 23, 2011  7:16 am



Hi, I can't seem to successfully constrain my residual variances to equal 0. I have a model in which: f1 by a1@1; f2 by a2@1; f3 by a3@1; and a1@0; a2@0; a3@0; so the residual variances should equal 0. However, I get an estimate of 1 for all residual variances. To make things weirder, it is a multigroup analyses, and in the other group (for which I specify exactly the same, it is a copypaste of model for group 1), I do get the residual variances of 0. Any advice? 


Please send the full output and your license number to support@statmodel.com. 


Apologies if I am posting this message in the wrong place. My question is about restraining the residual variance vs. not restraining. We are trying to model cannabis use across four time points. We would typically restrain the residual variances to (1) when building an unconditional parallel latent growth model but when we tried running the cannabis parallel latent growth model with the constrained variances we don’t get a good fit (RMSEA of 0.000 and CFI/TLI of 1.000/1.093 and SRMR of 0.069). However, when we don’t constrain we get a pretty good fitting model (Chi Sq p value 0.2054, RMSEA 0.029, CFI/TLI 0.908/0.890, SRMR 0.091). The residual variances themselves aren’t very different whether you constrain or not. Can you please give some insight on why this might be happening? Is it possible not to constrain the variance and still have a valid model? Also, can you please suggest any resources for help with interpreting quadratic growth trends? Thank you in advance for your time. 


Are you looking at this correctly? Having RMSEA of 0.000 and CFI/TLI of 1.000/1.093 and SRMR of 0.069 is better than having RMSEA 0.029, CFI/TLI 0.908/0.890, SRMR 0.091 The general analysis strategy question and interpretation question are better asked on SEMNET. 


Thanks for your quick reply. So you are saying that having a RMSEA of 0.000 does not indicate there is something wrong with the model? 


No, RMSEA = 0 reflects good fit. 

Back to top 