The lavaan Example Link

There are four variables in the data file that need to be named. There is no "USEVARIABLES ARE" statement here because all variables are used in the LGC model. The Mplus default is to use all variables.

VARIABLE:
  NAMES ARE
  anxiety1 anxiety2 anxiety3 anxiety4;

First the fixed portion of the model that sets up the intercept and linear slope needs to be specified. The intercept is set by setting the factor loading for each time point to be equal to 1. he linear slope latent variable is set by setting the factor loading for the first time point to 0, the second to 1, the third to 2, and the forth to 3. It is important that the first time point has a loading of 0; this allows the mean of the intercept to be interpreted as the mean value at the first time point. Non-linear slopes can also be specified here (e.g., 0, 1, 4, & 9 would create a quadratic growth curve).

MODEL:
  INTCEPT BY anxiety1@1 anxiety2@1 anxiety3@1 anxiety4@1;
  SLOPE BY anxiety1@0 anxiety2@1 anxiety3@2 anxiety4@3;

Next the residual variances for the manifest variables need to be freely estimated. Here this is done with a handy Mplus shortcut, rather than have to state "anxiety1 anxiety2 anxiety3 anxiety4;" the user can give the range of the anxiety variables to be included. Next the intercepts for each manifest variable need to be set to be equal to 0. This is done to allow an estimate of the mean of the intercept latent variable.

anxiety1-anxiety4;
[anxiety1-anxiety4@0];

In the last part of this command the latent variable variances are requested, then the latent variable means, and then the covariance between the intercept and slope.

INTCEPT SLOPE;
[INTCEPT SLOPE];
INTCEPT WITH SLOPE;

Click lgc-01.out to see the output.