Factor Analysis

https://github.com/maugavilla/well_hello_stats/blob/main/tutorials/13_EFA_cont.md

# https://bookdown.org/yihui/rmarkdown-cookbook/

#Compute EFA

fit.efa <- efa(spring2023,

ov.names=paste0(c("KD2_1","KD2_2","KD2_3","KD2_4","KD2_5")),

nfactors = 1:2,

rotation = "geomin",

baseline = T,

missing="fiml")

fit_ind <- NULL

for(j in 1:2){

temp <- c(nfac = nrow(lavInspect(fit.efa[[j]],"cor.lv")),

fitMeasures(fit.efa[[j]], c("npar","chisq","df","pvalue","cfi","tli","srmr","rmsea","aic","bic")),

moreFitIndices(fit.efa[[j]], "gammaHat"))

fit_ind <- do.call(rbind, list(fit_ind, temp))

}

fit_ind <- as.data.frame(fit_ind)

round(fit_ind, 4)

summary(fit.efa[[2]], nd = 2L, cutoff = 0.3, dot.cutoff = 0.05)

round(fit_ind[2,], 3)

#Compute reliability, alpha value

library(dplyr)

Resourceful <- spring2023 %>% select (KD2_1,KD2_2,KD2_3,KD2_4,KD2_5)

#install.packages("ltm")

#library(ltm)

ltm::cronbach.alpha(Resourceful, standardized = TRUE, CI=TRUE, na.rm=TRUE)

#Compute composite score

spring2023 = mutate(spring2023,belonging_social_expect=(Q2_2+Q2_4+Q2_6)/3,

academic_expect=(Q2_1+Q2_5+Q2_7+Q2_8)/4)

#Correlation matrix

columns <- c("belonging_social_expect","academic_expect")

cor_matrix <- cor(spring2023[,columns],use="pairwise.complete.obs")

cor_matrix

abs(cor_matrix) >0.5