attach(Coping_and_cognition_data)
#To calculate means and standard deviations of variables, shown here for age:
mean(age)
sd(age)
To calculate variable normality, shown here for age:
shapiro.test(age)
hist(age)
#To calculate the Cronbach’s alpha of a scale, each of the items of that scale were entered into a separate data frame, which was then analysed using the “alpha” function.  Shown here for facet positive emotion:
library(psych)
alpha(NEO_pos)
#
#Attention biases were calculated, and outliers removed, using a separate dataset, extracted from Gorilla Experiment Builder.  Biases calculated separately for each individual, show here for participant number 3869149:
attach(Attention)
#removing outliers
Ind <- subset(Attention, Participant_Private_ID == 3869149)
Q1 <- quantile(Reaction_Time, 0.25)
Q3 <- quantile(Reaction_Time, 0.75)
low_bench <- Q1 - 1.5*IQR(Reaction_Time)
high_bench <- Q3 + 1.5*IQR(Reaction_Time)
Ind <- subset(Ind, Reaction_Time < high_bench
              & Reaction_Time > low_bench)
#creating subsets for conditions
Sad <- subset(Ind, Target == "Sad")
Happy <- subset(Ind, Target == "Happy")
Afraid <- subset(Ind, Target == "Afraid")
Neutral_sad <- subset(Ind, randomise_blocks == "Female, sad vs neutral"
                      | randomise_blocks == "Male, sad vs neutral")
Neutral_sad <- subset(Neutral_sad, Target == "Neutral")
Neutral_happy <- subset(Ind, randomise_blocks == "Female, happy vs neutral"
                        | randomise_blocks == "Male, happy vs neutral")
Neutral_happy <- subset(Neutral_happy, Target == "Neutral")
Neutral_afraid <- subset(Ind, randomise_blocks == "Female, afraid vs neutral"
                         | randomise_blocks == "Male, afraid vs neutral")
Neutral_afraid <- subset(Neutral_afraid, Target == "Neutral")
#calculating means of RTs over trials
sad_RT <- mean(Sad$Reaction_Time)
happy_RT <- mean(Happy$Reaction_Time)
afraid_RT <- mean(Afraid$Reaction_Time)
Neutral_sad_RT <- mean(Neutral_sad$Reaction_Time)
Neutral_happy_RT <- mean(Neutral_happy$Reaction_Time)
Neutral_afraid_RT <- mean(Neutral_afraid$Reaction_Time)
#bias scores:
#sad bias:
Neutral_sad_RT - sad_RT
#happy bias
Neutral_happy_RT - happy_RT
#afraid bias
Neutral_afraid_RT - afraid_RT
#
#Factor analysis, shown here for COPE scales
attach(Data_276)
#scaling data
scale.data <- as.data.frame(scale(Data_276), centre = TRUE, scale = TRUE)
#To create a subset, only including emotion regulation variables
ER <- scale.data[c(6:11, 13:16, 18:20)]
#Principal components analysis
ER.pca <- princomp(ER, cor = TRUE, score = TRUE)
summary(ER.pca)
loadings(ER.pca)
#To view the scree plot
plot(ER.pca)
#in princomp, the eigenvalues are denoted as "standard deviation"
#Factor analysis, using 3 factors (based on the results of the principal components analysis above)
ER.fa <- factanal(ER, factors = 3, rotation = "varimax", scores = "regression")
ER.fa
head(ER.fa$scores)
#Factor scores are computer for each individual, and are written to a csv file
write.csv(ER.fa$scores, "ER_fa_scores.csv")
#
#To calculate the correlation matrix:
corr.test(Coping_and_cognition_data, method = "pearson")
cormatrix <- corr.test(Coping_and_cognition_data, method = "pearson")
print(cormatrix, short=FALSE)
write.csv(cormatrix$r, "r_matrix.csv")
write.csv(cormatrix$p, "p_matrix.csv")
#
#Multipe regression, shown here for BFAS traits explaining variance in BDI-2:
reg <- lm(BDI ~ BFAS_N + BFAS_E + BFAS_C + BFAS_A + BFAS_O)
summary(reg)
#
#Structural equation modelling, shown here for BFAS traits, AST_D, COPE, and interpretation of emotional images factors explaining variance in BDI-2:
scale.data <- as.data.frame(scale(Data_276_fa))
attach(scale.data)
#SEM:
scale.data <- as.data.frame(scale(Data_276))
attach(scale.data)
library(lavaan)
path <-  'AST_D ~ BFAS_N + BFAS_E + BFAS_C + BFAS_O
          Negative ~ BFAS_N + BFAS_A
          Neutral_and_positive ~ BFAS_N + BFAS_E + BFAS_C + BFAS_A + BFAS_O
          Adaptive_ER ~ BFAS_N + BFAS_E + BFAS_C + BFAS_O + AST_D + Neutral_and_positive
          Maladaptive_ER ~ BFAS_N + BFAS_E + BFAS_C + BFAS_A + BFAS_O + AST_D + Negative_social + Neutral_and_positive
          BDI ~ BFAS_N + BFAS_E + BFAS_C + BFAS_A + BFAS_O + Adaptive_ER + Maladaptive_ER + AST_D + Negative + Neutral_and_positive'
fit <- sem(path, data = scale.data) 
summary(fit, fit.measures = T)
fitMeasures(fit, c("cfi", "rmsea", "srmr"))
modindices(fit, minimum.value = 3.84)
#Statically and theoretically plausible regressions paths are added, based on the modification indices, and the model is tested again, to improve statistical fit