Issue with T-scores PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
Message/Author
 Sarah Lowe posted on Tuesday, June 27, 2017 - 6:25 pm
Hello!

I am working with an integrated dataset and am trying to run a LCGA with t-scores. Unfortunately, I am getting the following error:

The number of fixed time scores is not sufficient for model identification in the following growth process: I S Q

For your reference, here is my model:

******

USEVARIABLES ARE v1 v2 v3 v4
TIME1 TIME2 TIME3 TIME4;
TSCORES = TIME1 TIME2 TIME3 TIME4;
CLASSES = C(1);


ANALYSIS: TYPE = RANDOM MIXTURE;
ESTIMATOR IS MLR;
STARTS = 500 20;
STITERATIONS = 20;

MODEL:


%OVERALL%

i s q | v1 v2 v3 v4 @
TIME1 TIME2 TIME3 TIME4;

i s q WITH i s q; i; s; q;
v1; v2; v3; v4;

%C#1%
[i s q];
i s q WITH i s q;
i; s; q;
v1; v2; v3; v4;

****

I am wondering if this has to do with the large variability in timing within each t-score, as well as the extent of missing data? Here are the ranges for the t-scores:

Time 1: 0 to 61 days (Median = 18)
Time 2: 62 to 183 days (Median = 110)
Time 3: 184 to 301 days (Median = 207)
Time 4: 305 to 487 days (Median = 376)


The between-time covariance coverage ranges from 0.126 to 0.557

Any insight you have on this matter would be greatly appreciated!

Thanks so much,
Sarah
 Linda K. Muthen posted on Wednesday, June 28, 2017 - 6:14 am
Try the following:

i s q | v1 - v4 at TIME1 - TIME4;
 Sarah Lowe posted on Wednesday, June 28, 2017 - 5:27 pm
Thanks for the help - changing the @ to AT worked!
 Mayra Galvis posted on Wednesday, April 24, 2019 - 5:51 am
Dear Mplus team,
I am interested in identifying trajectories of adaptation to spinal cord injury during the rehabilitation time using LGMM.
I have 240 participants and 3 measurement time points which are assessed as follows:
T1: approximately 1 month after injury
T2: approximately 3 months after injury
T3: at rehabilitation discharge
When exploring the raw data, I realized that there is great variability in the time when individuals where assessed at T3 (some individuals were discharged earlier and therefore assessed soon after 3 months, while others were discharged 6 or more months later).
I was then wondering what would be the best approach to handle such variability in assessment time.
Would the use of individually varying times of observation (TSCORES) be a suitable option?
Is there any other approach that could be used to account for the inconsistency of assessment time?
I appreciate a lot your insights regarding these issues.
 Bengt O. Muthen posted on Wednesday, April 24, 2019 - 4:58 pm
TSCORES is the way to go. You can also do 2-level modeling in long format with time as a within-level covariate - but that's the same thing.
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: