**INTRUCTIONS**
This folder contains: (1) the database adopted to estimate the Latent Growth Curve Model; (2) AMOS input file adopted to perform the analysis; (3) the output of the analysis.
**1- Database**
The DatiCalling_T1_T2_T3_WideFormat_ZFS_OSF.SAV file is an SPSS file which contains data adopted to perform the analysis in AMOS.
**2- Syntax**
The file "Calling.amw" contains the AMOS syntax file. Analysis were performed in AMOS 20.
**3- Output**
The file "Calling.AmosOutput" report the output of the analysis.
**ANALYTICAL PROCEDURE**
We specified and estimated an unconditional latent growth curve model (Bollen & Curran, 2006; Little, 2013). Being a multilevel SEM, this analysis allows us to estimate the changes within persons in terms of slopes and intercepts, and at the same time it summarizes the between-person differences of individual growth trajectories. Indeed, latent variables for both the initial level (intercept) and the rate of change across time (slope) were specified. Change trajectories were modeled as a distribution across subjects. Then, the mean and the standard deviation of this distribution were estimated. The estimates of loadings for the intercept factor were all constrained at 1 by design. The shape of the trajectories was defined by constraining at 0 and 1 the loadings from the slope factor to the observed measure of calling at Times 1 and 2, respectively. The unconstrained estimation of the loading at Time 3 gave the trajectory a shape, and informed the amount of change that happens from Time 2 to Time 3 in terms of the proportion of the change that happened from Time 1 to Time 2. The final model has four predictors of change (age, gender, mean grade at Time 1, and living out a calling at Time 2) and four outcomes of change (living out a calling, mean grade, academic satisfaction, and dropout intentions at Time 3).