Test for autocorrelation PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 Belinda Needham posted on Thursday, August 06, 2009 - 9:06 pm
I am running a parallel process growth model for BMI and depression. A reviewer suggested that autocorrelation might be a problem, particularly for BMI. Can you tell me how to test for autocorrelation? Based on your reply to a previous post, I tried the following, but I'm not sure if it's correct in this case.

MODEL:ibmi sbmi | bmi05@0 bmi10@1 bmi15* bmi20*;
idep sdep | depsum05@0 depsum10@1 depsum15* depsum20*;
ibmi on ex3_age female nocoll05;
sbmi on ex3_age female nocoll05;
idep on ex3_age female nocoll05;
sdep on ex3_age female nocoll05;
sbmi on idep;
sdep on ibmi;
idep with ibmi;
idep with sdep;
ibmi with sbmi;
sdep with sbmi;
bmi05 with bmi10;
bmi10 with bmi15;
bmi15 with bmi20;
depsum05 with depsum10;
depsum10 with depsum15;
depsum15 with depsum20;

If the correlations are significant, does that indicate autocorrelation? If one or more of the correlations is not significant, should I remove the non-significant ones from the model? Thank you for your help.
 Linda K. Muthen posted on Friday, August 07, 2009 - 9:47 am
Example 6.17 shows auto-correlated residuals. You have correlations between adjacent time points only. If a correlation is significant, then that indicates an autocorrelation.
 Belinda Needham posted on Friday, August 07, 2009 - 11:35 am
I followed Example 6.17 and got the following error message:

*** ERROR in Model Constraint command
Unknown parameter label in MODEL CONSTRAINT: NEW(CORR)
in assignment: NEW(CORR) =

I'm not sure what I did wrong. This was my input:

DATA:file is C:\Data\CARDIA\cardia529.raw;
format is free;
type is individual;

VARIABLE:names are ex3_age c20wst1 e20wst f20wst1 g20wst
depsum05 depsum10 depsum15 depsum20 bmi05 bmi10 bmi15
bmi20 female black nocoll05 dsm_bed hidep05 hibmi05
obese05 unmard05 exsum05;

missing are .;
usevariables are c20wst1 e20wst f20wst1 g20wst;

ANALYSIS:type=missing H1;

MODEL:i s q | c20wst1@0 e20wst@1 f20wst1@2 g20wst@3;
c20wst1 e20wst f20wst1 g20wst (resvar);
c20wst1 e20wst f20wst1 PWITH e20wst f20wst1 g20wst (p1);
c20wst1 e20wst PWITH f20wst1 g20wst (p2);
c20wst1 WITH g20wst (p3);

new (corr);
p1 = resvar*corr;
p2 = resvar*corr**2;
p3 = resvar*corr**3;

 Linda K. Muthen posted on Friday, August 07, 2009 - 2:57 pm
If you look at the example, you will see that a list function is used in, for example,

c20wst1-g20wst (resvar);

and the other statements. The way you have specified it, resvar applies only to g20wst.

If you can't resolve this, please send your input, data, output, and license number to support@statmodel.com.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message