Hello,

I would like to ask three technical questions about my MPlus code.

My team and I have conducted a cross-lagged model of variables in two waves (Wave 1, CM, IC, EB, ED, ER, and Wave 2, XCM, XIC, XEB, XED, XER) comparing females and males and controlling 4 sociodemographic variables (Age, region, habitat, Uni).

GROUPING IS Gender (0=male 1=female);

ANALYSIS:

TYPE = GENERAL;

ESTIMATOR IS ML;

ITERATIONS = 1000;

CONVERGENCE = 0.00001;

MODEL:

!Step 1 : Adjusting basal predictors by controlling variables (Age UNI HABITAT Region)

CM IC EB ED ER on Age UNI HABITAT Region;

!Step 2: MODEL. Testing if W2 variables are predicted by W1 independent variables (including autocorrelations)

XCM on CM IC EB ED ER;

XIC on CM IC EB ED ER;

XEB on CM IC EB ED ER;

XED on CM IC EB ED ER;

XER on CM IC EB ED ER;

!Step 3: Adjusting correlations between variables

XCM WITH XIC;

XCM WITH XEB;

XCM WITH XED;

XCM WITH XER;

XIC WITH XEB;

XIC WITH XED;

XIC WITH XER;

XEB WITH XED;

XEB WITH XER;

XED WITH XER;

IC WITH CM;

EB WITH CM;

EB WITH IC;

ED WITH CM;

ED WITH IC;

ED WITH EB;

ER WITH CM;

ER WITH IC;

ER WITH EB;

ER WITH ED;

OUTPUT: SAMPSTAT STANDARDIZED MODINDICES;

My three questions are:

1. To compare the unconstrained model with the constrained model (constraining males and females models), is needed to constrain/fix all the paths of the code (Step 1, 2 and 3), or just the cross-lagged paths (Step 2)?

2. To determine if there is a significant difference between the constrained (i.e., constraining females and males paths to be equal) and unconstrained model, it is needed to just compare the general constrained and unconstrained models calculated separately, or we have to include in the code a section for the males and females models separately constraining the paths equally?

3. If the difference between Chi-Square and DF of both constrained and unconstrained models is significant (p < .05), that means the models are not equivalent and there are differences between males and females models? And that would mean that every difference between men and women paths are significant, or we should calculate which paths are significantly different between genders and which not?

Thank you so much!