Residual autocorrelation PreviousNext
Mplus Discussion > Dynamic Structural Equation Modeling >
 Samuli Helle posted on Monday, April 08, 2019 - 12:16 am
Is there a way to plot residual autocorrelation (ACF, PACF) of Y in n=1 time-series analysis?
 Bengt O. Muthen posted on Monday, April 08, 2019 - 4:55 pm
Autocorrelation plots are available. I don't know what you mean by residual autocorr plots.
 Samuli Helle posted on Tuesday, April 09, 2019 - 1:06 am
If I fit a n=1 time-series model with covariates shown below I can plot the raw autocorrelation function of Y. However, is there a way to plot the autocorrelation function of the residuals of Y?

Y ON X1 X2 X3;

I red about RDSEM and tried the following code to model e.g. AR(2) model.

Y ON X1 X2 X3;
Y^ ON Y^1 Y^2;

But there is no way to obtain ACF of Y^ in order to have a clue on the autocorrelation pattern needed?
 Tihomir Asparouhov posted on Tuesday, April 09, 2019 - 7:57 pm
I would recommend to keep increasing the lag until the new autocorrelation is no longer significant. For example, if you estimate

Y ON X1 X2 X3;
Y^ ON Y^1 Y^2 Y^3;

and the coefficient in front of Y^3 is no longer significant I would conclude that lag 2 is enough. You can go a bit further to make sure that something cyclical is not happening.
 Samuli Helle posted on Wednesday, April 10, 2019 - 12:26 am
 Samuli Helle posted on Wednesday, May 22, 2019 - 2:04 pm
Is it possible to define residual autocorrelation structure for between-level response variable? I tried but got the warning below:

"Variables specified on the BETWEEN option cannot have lag and be mentioned on the LAGGED option."
 Bengt O. Muthen posted on Wednesday, May 22, 2019 - 4:10 pm
What does the Between level portray? Why is time on Between?
 Samuli Helle posted on Thursday, May 23, 2019 - 4:04 am
Yes, I know the setup is a bit awkward. I have a response variable that has been measured annually (i.e., N=1 time-series) for some 80 years (=between-level) and I'd like to know how some IVs measured several times within years (=within-level) influence the response's variation at the between-level. Can such a model be fit? I guess it would be a bit like adding a spatial autocorrelation into the model?
 Bengt O. Muthen posted on Thursday, May 23, 2019 - 4:58 pm
Maybe you could try the auto-regressive approach of UG ex6.17 and somehow apply it to the Between level.
Back to top
Add Your Message Here
Username: Posting Information:
This is a public posting area. Enter your username and password if you have an account. Otherwise, enter your full name as your username and leave the password blank. Your e-mail address is optional.
Options: Enable HTML code in message
Automatically activate URLs in message