Hi,

I am running a second-order latent growth model and searching for syntax advice.

My model has a first-order factor and two time-varying latent predictors.

The three latent factors all achieved partial strong invariance.

So, here's my question.

1) To identify the first-order factor, one indicator loading has to be set to 1 and one measurement intercept needs to be set to 0. (reference indicator model)

Do the TVCs need to be identified in the same way since they are latent factors?

(in this model, they are not assumed to have a trajectory because it is not a parallel

model)

2)Should the predictor (TVC) factor loadings and intercepts be labeled for their invariance?

(please see below)

(Mplus syntax)

Create factor for TVC1

AS1 by t1as1

t1as2 t1as3 t1as4 t1as5 (LL2-LL5);

AS2 by t2as1

t2as2 t2as3 t2as4 t2as5 (LL2-LL5);

AS3 by t3as1

t3as2 t3as3 t3as4 t3as5 (LL2-LL5);

AS4 by t4as1

t4as2 t4as3 t4as4 t4as5 (LL2-LL5);

AS5 by t5as1

t5as2 t5as3 t5as4 t5as5 (LL2-LL5);

AS6 by t6as1

t6as2 t6as3 t6as4 t6as5 (LL2-LL5);

AS9 by t9as1

t9as2 t9as3 t9as4 t9as5 (LL2-LL5);

[t1as1-t9as1@0]; !set first variable to 0 for identification

!All other intercepts are invariant (strong invariance)

[t1as2 t2as2 t3as2 t4as2 t5as2 t6as2 t9as2](II2);

[t1as3 t2as3 t3as3 t4as3 t5as3 t6as3 t9as3](II3);

[t1as4 t2as4 t3as4 t4as4 t5as4 t6as4 t9as4](II4);

[t1as5 t2as5 t3as5 t4as5 t5as5 t6as5 t9as5](II5);

Thank you in advance for your help.