#AEQ diary study LCM analysis for Moscow

#Rosseel example
# linear growth model with 4 timepoints
# intercept and slope with fixed coefficients
 i =~ 1*t1 + 1*t2 + 1*t3 + 1*t4
 s =~ 0*t1 + 1*t2 + 2*t3 + 3*t4
 model <- ' i =~ 1*t1 + 1*t2 + 1*t3 + 1*t4
           s =~ 0*t1 + 1*t2 + 2*t3 + 3*t4 '
fit <- growth(model, data=Demo.growth)
summary(fit)


#bauer & curran approach intercept only
library(lavaan)
int <- 'I =~ 1*enjoy_1 + 1*enjoy_2 + 1*enjoy_3 + 1*enjoy_4 + 1*enjoy_5 + 1*enjoy_6 + 1*enjoy_7 + 1*enjoy_8 + 1*enjoy_9
'
fit.int <- growth(int, data=AEQ_diary_study, missing="ML")
summary(fit.int, fit.measures=TRUE, estimates=FALSE)

#to get gamma hat fit index
library(semTools)
moreFitIndices(fit.int)

#to get x2/df ratio and p-value-thanks to Terry Jorgensen UvA (amsterdam)
fm <- fitMeasures(fit.int)
fm[["chisq"]] / fm[["df"]]
fm["pvalue"] 

#to get conventional correlated factors diagram
semPaths(fit.int, intercept = FALSE, whatLabel = "std",  residuals = FALSE, exoCov = TRUE) 

#bauer & curran approach LCM
lin <- 'I =~ 1*enjoy_1 + 1*enjoy_2 + 1*enjoy_3 + 1*enjoy_4 + 1*enjoy_5 + 1*enjoy_6 + 1*enjoy_7 + 1*enjoy_8 + 1*enjoy_9
S =~ 0*enjoy_1 + 1*enjoy_2 + 2*enjoy_3 + 3*enjoy_4 + 4*enjoy_5 + 5*enjoy_6 + 6*enjoy_7 + 7*enjoy_8 + 8*enjoy_9
'
fit.lin <- growth(lin, data=AEQ_diary_study, missing="ML")
summary(fit.lin, fit.measures=TRUE, rsquare=TRUE)
#to get conventional correlated factors diagram
semPaths(fit.lin, intercept = FALSE, whatLabel = "std",  residuals = FALSE, exoCov = TRUE) 

lavTestLRT(fit.lin, fit.int)

#bauer & curran approach LCM
#factor structures positive only
lin_pos <- '
pos1=~ enjoy_1 + hope_1 + proud_1 + relief_1 
pos2=~ enjoy_2 + hope_2 + proud_2 + relief_2 
pos3=~ enjoy_3 + hope_3 + proud_3 + relief_3 
pos4=~ enjoy_4 + hope_4 + proud_4 + relief_4 
pos5=~ enjoy_5 + hope_5 + proud_5 + relief_5 
pos6=~ enjoy_6 + hope_6 + proud_6 + relief_6 
pos7=~ enjoy_7 + hope_7 + proud_7 + relief_7 
pos8=~ enjoy_8 + hope_8 + proud_8 + relief_8 
pos9=~ enjoy_9 + hope_9 + proud_9 + relief_9 

I =~ 1*pos1 + 1*pos2 + 1*pos3 + 1*pos4 + 1*pos5 + 1*pos6 + 1*pos7 + 1*pos8 + 1*pos9
S =~ 0*pos1 + 1*pos2 + 2*pos3 + 3*pos4 + 4*pos5 + 5*pos6 + 6*pos7 + 7*pos8 + 8*pos9
'
fit.lin_pos <- growth(lin_pos, data=AEQ_diary_study, missing="ML")
summary(fit.lin_pos, fit.measures=TRUE, rsquare=TRUE)
#to get conventional correlated factors diagram
semPaths(fit.lin_pos, intercept = FALSE, whatLabel = "std",  residuals = FALSE, exoCov = TRUE) 


#bauer & curran approach LCM; no optimizer result
#factor structures negative only no bored
lin_neg_nobored <- '
neg1=~ angry_1 + anxious_1 + shame_1 + hopeless_1  
neg2=~ angry_2 + anxious_2 + shame_2 + hopeless_2
neg3=~ angry_3 + anxious_3 + shame_3 + hopeless_3
neg4=~ angry_4 + anxious_4 + shame_4 + hopeless_4
neg5=~ angry_5 + anxious_5 + shame_5 + hopeless_5
neg6=~ angry_6 + anxious_6 + shame_6 + hopeless_6
neg7=~ angry_7 + anxious_7 + shame_7 + hopeless_7
neg8=~ angry_8 + anxious_8 + shame_8 + hopeless_8
neg9=~ angry_9 + anxious_9 + shame_9 + hopeless_9

I =~ 1*neg1 + 1*neg2 + 1*neg3 + 1*neg4 + 1*neg5 + 1*neg6 + 1*neg7 + 1*neg8 + 1*neg9
S =~ 0*neg1 + 1*neg2 + 2*neg3 + 3*neg4 + 4*neg5 + 5*neg6 + 6*neg7 + 7*neg8 + 8*neg9
'
fit.lin_neg_nobored <- growth(lin_neg_nobored, data=AEQ_diary_study, missing="ML")
summary(fit.lin_neg_nobored, fit.measures=TRUE, rsquare=TRUE)
semPaths(fit.lin_neg_nobored, intercept = FALSE, whatLabel = "std",  residuals = FALSE, exoCov = TRUE) 

lavTestLRT(fit.lin, fit.int)

lavTestLRT(fit.lin, fit.int)


#bauer & curran approach LCM; no no optimizer result
#factor structures positive and negative 
lin <- '
pos1=~ enjoy_1 + hope_1 + proud_1 + relief_1 
pos2=~ enjoy_2 + hope_2 + proud_2 + relief_2 
pos3=~ enjoy_3 + hope_3 + proud_3 + relief_3 
pos4=~ enjoy_4 + hope_4 + proud_4 + relief_4 
pos5=~ enjoy_5 + hope_5 + proud_5 + relief_5 
pos6=~ enjoy_6 + hope_6 + proud_6 + relief_6 
pos7=~ enjoy_7 + hope_7 + proud_7 + relief_7 
pos8=~ enjoy_8 + hope_8 + proud_8 + relief_8 
pos9=~ enjoy_9 + hope_9 + proud_9 + relief_9 
neg1=~ angry_1 + anxious_1 + shame_1 + hopeless_1  
neg2=~ angry_2 + anxious_2 + shame_2 + hopeless_2
neg3=~ angry_3 + anxious_3 + shame_3 + hopeless_3
neg4=~ angry_4 + anxious_4 + shame_4 + hopeless_4
neg5=~ angry_5 + anxious_5 + shame_5 + hopeless_5
neg6=~ angry_6 + anxious_6 + shame_6 + hopeless_6
neg7=~ angry_7 + anxious_7 + shame_7 + hopeless_7
neg8=~ angry_8 + anxious_8 + shame_8 + hopeless_8
neg9=~ angry_9 + anxious_9 + shame_9 + hopeless_9

I =~ 1*pos1 + 1*pos2 + 1*pos3 + 1*pos4 + 1*pos5 + 1*pos6 + 1*pos7 + 1*pos8 + 1*pos9 + 1*neg1 + 1*neg2 + 1*neg3 + 1*neg4 + 1*neg5 + 1*neg6 + 1*neg7 + 1*neg8 + 1*neg9
S =~ 0*pos1 + 1*pos2 + 2*pos3 + 3*pos4 + 4*pos5 + 5*pos6 + 6*pos7 + 7*pos8 + 8*pos9 + 0*neg1 + 1*neg2 + 2*neg3 + 3*neg4 + 4*neg5 + 5*neg6 + 6*neg7 + 7*neg8 + 8*neg9
'
fit.lin <- growth(lin, data=AEQ_diary_study, missing="ML")
summary(fit.lin, fit.measures=TRUE, rsquare=TRUE)
lavTestLRT(fit.lin, fit.int)



# week 1 linear growth model with 3 timepoints and 2 correlated latent factors
SCOAlcmwk1 <- ' 
#factor structures
pos1=~enjoy_1 + hope_1 + proud_1 + relief_1 
neg1=~ angry_1 + anxious_1 + shame_1 + hopeless_1  
pos2=~ enjoy_2 + hope_2 + proud_2 + relief_2 
neg2=~ angry_2 + anxious_2 + shame_2 + hopeless_2
pos3=~ enjoy_3 + hope_3 + proud_3 + relief_3 
neg3=~ angry_3 + anxious_3 + shame_3 + hopeless_3

# intercept and slope with fixed coefficients
I =~ 1*neg1 + 1*neg2 + 1*neg3 + 1*pos1 + 1*pos2 + 1*pos3 
S =~ 0*neg1 + 1*neg2 + 2*neg3 + 0*pos1 + 1*pos2 + 2*pos3
	'
fit.lin_wk1 <- growth(SCOAlcmwk1, data=AEQ_diary_study, missing="ML")
summary(fit.lin_wk1, fit.measures=TRUE, rsquare=TRUE)
#to get conventional correlated factors diagram
semPaths(fit.lin_wk1, intercept = FALSE, whatLabel = "std",  residuals = FALSE, exoCov = TRUE) 

#bauer & curran approach LCM; 
#factor structures negative 1-5 only
lin_neg5 <- '
neg1=~ angry_1 + anxious_1 + shame_1 + bored_1 + hopeless_1  
neg2=~ angry_2 + anxious_2 + shame_2 + bored_2 + hopeless_2
neg3=~ angry_3 + anxious_3 + shame_3 + bored_3 + hopeless_3
neg4=~ angry_4 + anxious_4 + shame_4 + bored_4 + hopeless_4
neg5=~ angry_5 + anxious_5 + shame_5 + bored_5 + hopeless_5


I =~ 1*neg1 + 1*neg2 + 1*neg3 + 1*neg4 + 1*neg5 
S =~ 0*neg1 + 1*neg2 + 2*neg3 + 3*neg4 + 4*neg5 
'
fit.lin_neg5 <- growth(lin_neg5, data=AEQ_diary_study, missing="ML")
summary(fit.lin_neg5, fit.measures=TRUE, rsquare=TRUE)
#to get conventional correlated factors diagram
semPaths(fit.lin_wk1, intercept = FALSE, whatLabel = "std",  residuals = FALSE, exoCov = TRUE) 

#bauer & curran approach LCM; no optimizer result
#factor structures negative + positive 1-5 only
lin_negpos5 <- '
neg1=~ angry_1 + anxious_1 + shame_1 + bored_1 + hopeless_1  
neg2=~ angry_2 + anxious_2 + shame_2 + bored_2 + hopeless_2
neg3=~ angry_3 + anxious_3 + shame_3 + bored_3 + hopeless_3
neg4=~ angry_4 + anxious_4 + shame_4 + bored_4 + hopeless_4
neg5=~ angry_5 + anxious_5 + shame_5 + bored_5 + hopeless_5
pos1=~ enjoy_1 + hope_1 + proud_1 + relief_1 
pos2=~ enjoy_2 + hope_2 + proud_2 + relief_2 
pos3=~ enjoy_3 + hope_3 + proud_3 + relief_3 
pos4=~ enjoy_4 + hope_4 + proud_4 + relief_4 
pos5=~ enjoy_5 + hope_5 + proud_5 + relief_5 

I =~ 1*neg1 + 1*neg2 + 1*neg3 + 1*neg4 + 1*neg5 + 1*pos1 + 1*pos2 + 1*pos3 + 1*pos4 + 1*pos5
S =~ 0*neg1 + 1*neg2 + 2*neg3 + 3*neg4 + 4*neg5 + 0*pos1 + 1*pos2 + 2*pos3 + 3*pos4 + 4*pos5
'
fit.lin_negpos5 <- growth(lin_negpos5, data=AEQ_diary_study, missing="ML")
summary(fit.lin_negpos5, fit.measures=TRUE, rsquare=TRUE)

#bauer & curran approach LCM; no optimizer result
#factor structures negative + positive 1-4 only
lin_negpos4 <- '
neg1=~ angry_1 + anxious_1 + shame_1 + bored_1 + hopeless_1  
neg2=~ angry_2 + anxious_2 + shame_2 + bored_2 + hopeless_2
neg3=~ angry_3 + anxious_3 + shame_3 + bored_3 + hopeless_3
neg4=~ angry_4 + anxious_4 + shame_4 + bored_4 + hopeless_4
pos1=~ enjoy_1 + hope_1 + proud_1 + relief_1 
pos2=~ enjoy_2 + hope_2 + proud_2 + relief_2 
pos3=~ enjoy_3 + hope_3 + proud_3 + relief_3 
pos4=~ enjoy_4 + hope_4 + proud_4 + relief_4 

I =~ 1*neg1 + 1*neg2 + 1*neg3 + 1*neg4 + 1*pos1 + 1*pos2 + 1*pos3 + 1*pos4
S =~ 0*neg1 + 1*neg2 + 2*neg3 + 3*neg4 + 0*pos1 + 1*pos2 + 2*pos3 + 3*pos4
'
fit.lin_negpos5 <- growth(lin_negpos5, data=AEQ_diary_study, missing="ML")
summary(fit.lin_negpos5, fit.measures=TRUE, rsquare=TRUE)

#bauer & curran approach LCM; no optimizer result
#factor structures negative + positive 1-3 only
lin_negpos3 <- '
neg1=~ angry_1 + anxious_1 + shame_1 + bored_1 + hopeless_1  
neg2=~ angry_2 + anxious_2 + shame_2 + bored_2 + hopeless_2
neg3=~ angry_3 + anxious_3 + shame_3 + bored_3 + hopeless_3

pos1=~ enjoy_1 + hope_1 + proud_1 + relief_1 
pos2=~ enjoy_2 + hope_2 + proud_2 + relief_2 
pos3=~ enjoy_3 + hope_3 + proud_3 + relief_3 

I =~ 1*neg1 + 1*neg2 + 1*neg3 + 1*pos1 + 1*pos2 + 1*pos3 
S =~ 0*neg1 + 1*neg2 + 2*neg3 + 0*pos1 + 1*pos2 + 2*pos3 
'
fit.lin_negpos3 <- growth(lin_negpos3, data=AEQ_diary_study, missing="ML")
summary(fit.lin_negpos3, fit.measures=TRUE, rsquare=TRUE)

#bauer & curran approach LCM; no optimizer result probably bored doesn't load strongly enough see cfa result
#factor structures negative only
lin_neg <- '
neg1=~ angry_1 + anxious_1 + shame_1 + bored_1 + hopeless_1  
neg2=~ angry_2 + anxious_2 + shame_2 + bored_2 + hopeless_2
neg3=~ angry_3 + anxious_3 + shame_3 + bored_3 + hopeless_3
neg4=~ angry_4 + anxious_4 + shame_4 + bored_4 + hopeless_4
neg5=~ angry_5 + anxious_5 + shame_5 + bored_5 + hopeless_5
neg6=~ angry_6 + anxious_6 + shame_6 + bored_6 + hopeless_6
neg7=~ angry_7 + anxious_7 + shame_7 + bored_7 + hopeless_7
neg8=~ angry_8 + anxious_8 + shame_8 + bored_8 + hopeless_8
neg9=~ angry_9 + anxious_9 + shame_9 + bored_9 + hopeless_9

I =~ 1*neg1 + 1*neg2 + 1*neg3 + 1*neg4 + 1*neg5 + 1*neg6 + 1*neg7 + 1*neg8 + 1*neg9
S =~ 0*neg1 + 1*neg2 + 2*neg3 + 3*neg4 + 4*neg5 + 5*neg6 + 6*neg7 + 7*neg8 + 8*neg9
'
fit.lin_neg <- growth(lin_neg, data=AEQ_diary_study, missing="ML")
#correlation patterns in model; not correlating time factors with each otehr; correlation is solely in the intercept and slope
	pos1 ~~ 0*pos2 + 0*neg2 + 0*pos3 + 0*neg3 + 0*pos4 + 0*neg4 + 0*pos5 + 0*neg5 + 0*pos6 + 0*neg6 + 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	neg1 ~~ 0*pos2 + 0*neg2 + 0*pos3 + 0*neg3 + 0*pos4 + 0*neg4 + 0*pos5 + 0*neg5 + 0*pos6 + 0*neg6 + 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	pos2 ~~ 0*pos3 + 0*neg3 + 0*pos4 + 0*neg4 + 0*pos5 + 0*neg5 + 0*pos6 + 0*neg6 + 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	neg2 ~~ 0*pos3 + 0*neg3 + 0*pos4 + 0*neg4 + 0*pos5 + 0*neg5 + 0*pos6 + 0*neg6 + 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	pos3 ~~ 0*pos4 + 0*neg4 + 0*pos5 + 0*neg5 + 0*pos6 + 0*neg6 + 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	neg3 ~~ 0*pos4 + 0*neg4 + 0*pos5 + 0*neg5 + 0*pos6 + 0*neg6 + 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	pos4 ~~ 0*pos5 + 0*neg5 + 0*pos6 + 0*neg6 + 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	neg4 ~~ 0*pos5 + 0*neg5 + 0*pos6 + 0*neg6 + 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	pos5 ~~ 0*pos6 + 0*neg6 + 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	neg5 ~~ 0*pos6 + 0*neg6 + 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	pos6 ~~ 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	neg6 ~~ 0*pos7 + 0*neg7 + 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	pos7 ~~ 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	neg7 ~~ 0*pos8 + 0*neg8 + 0*pos9 + 0*neg9
	pos8 ~~ 0*pos9 + 0*neg9
	neg8 ~~ 0*pos9 + 0*neg9
	'