# Automatically set working directory to the folder containing the source file # Load rstudioapi package library(rstudioapi) # Set working directory to source file location setwd(dirname(getActiveDocumentContext()$path)) getwd() library(readxl) library(metaSEM) library(semPlot) library(semTools) citation("metaSEM") #______________________________________________________________________________ # TL -> PE -> IWB #_______________________________________________________________________________ #_________________________STAGE_1_____________________________________________ data <- read_excel("metasem.xlsx", sheet = 1) data # Tüm veri çerçevesini sayısala çevir data <- as.data.frame(lapply(data, as.numeric)) names(data) nvar = 3 varnames = c("TL", "PE", "IWB") labels = list(varnames,varnames) cordat = list() for (i in 1:nrow(data)){ cordat[[i]] = vec2symMat(as.matrix(data[i,3:5]), diag = FALSE) dimnames(cordat[[i]]) = labels } cordat ## Sample sizes my.n <- data$N my.n ## Total sample sizes in each cell pattern.n(cordat, my.n) ## Number of studies in each cell pattern.na(cordat, show.na = FALSE) ## Stage 1 analysis: find an average correlation matrix stage1 <- tssem1(cordat, my.n) summary(stage1) stage1 <- tssem1(Cov=cordat, n=data$N, method="REM", RE.type = "Diag") summary(stage1) ## Average correlation matrix meanR <- vec2symMat(coef(stage1, select = "fixed"), diag = FALSE) dimnames(meanR) <- list(varlist, varlist) meanR ## Absolute heterogeneity variance: tau^2 tau2 <- vec2symMat(coef(stage1, select = "random"), diag = FALSE) dimnames(tau2) <- list(varlist, varlist) tau2 ## Relative heterogeneity index: I^2 I2 <- vec2symMat(summary(stage1)$I2.values[, "Estimate"], diag = FALSE) dimnames(I2) <- list(varlist, varlist) I2 ## Proposed model model <- "IWB ~ c*TL + b*PE PE ~ a*TL TL ~~ 1*TL" plot(model, color="yellow") varlist = c("TL", "PE", "IWB") RAM1 <- lavaan2RAM(model, obs.variables = varlist) RAM1 #_________________________STAGE_2_____________________________________________ # #A Matrisi – Yapısal Katsayılar (Paths) # #A[1,1] = 0 TL kendini etkilemez (döngü yok) # A[1,2] = 0.1*b21 PE'nin TL üzerindeki etkisi (b21 parametresi, 0.1 ile ölçeklenmiş) # #A[3,1:3] = 0 IWB kendisini veya diğerlerini etkilemez (modelin son değişkeni gibi) # A[1,3] = 0.1*b31 IWB'nin TL üzerindeki etkisi (b31 parametresi) # A[2,1] = 0 TL’nin PE üzerinde etkisi yok (çünkü bu ters yön) # A[2,2] = 0 PE kendini etkilemez # A[2,3] = 0.1*b32 IWB'nin PE üzerindeki etkisi (b32) # # Harmonik ortalama N hesapla # N_values <- data$N # harmonic_N <- length(N_values) / sum(1 / N_values) # # harmonic_N # # 1. Harmonik N hesapla # harmonic_N <- length(data$N) / sum(1 / data$N) # # # 2. stage1 nesnesinin içindeki toplam N değerini override et # stage1$total.n <- harmonic_N # # stage2a <- tssem2( stage1, RAM=RAM1, diag.constraints = TRUE, intervals = "LB" ) summary(stage2a) # Yol diyagramı çizimi plot_object_PE <- meta2semPlot(stage2a) # Yol grafiğini çiz semPaths(plot_object_PE, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 7, nCharNodes = 0) semPaths(plot_object_PE, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, rotation = 1, # 0 = yukarıdan aşağı, 2 = soldan sağa sizeMan = 9, nCharNodes = 0, color = list(lat = "skyblue", man = "lightgreen"), edge.color = "black", edge.label.color = "blue") # semPaths(plot_object_PE, # layout = "tree2", # whatLabels = "est", # style = "ram", # sizeMan = 8, # color = list(man = "lightblue", lat = "lightyellow"), # border.color = list(man = "darkblue", lat = "orange"), # edge.label.cex = 1.2) stage2b <- tssem2(stage1, RAM=RAM1, diag.constraints=TRUE, intervals="LB", mx.algebras = list(Indirect=mxAlgebra(a*b, name="Indirect"), Direct=mxAlgebra(c, name="Direct"))) summary(stage2b) HatVar <-0.74 R2 <- 1 - HatVar R2 ## Add a constraint on c=a*b ## Rerun to remove errors stage2b <- mxModel(stage2b, mxConstraint(c==a*b, name="constraint")) stage2b <- mxRun(stage2b) stage2b <- mxTryHard(stage2b, extraTries = 100) stage2b <- mxRun(stage2b, intervals = TRUE) summary(stage2b) anova(stage2$mx.fit, stage2b) #______________________________________________________________________________ # TL -> CSE -> IWB #_______________________________________________________________________________ #_________________________STAGE_1_____________________________________________ data <- read_excel("metasem.xlsx", sheet = 2) data # Tüm veri çerçevesini sayısala çevir data <- as.data.frame(lapply(data, as.numeric)) names(data) nvar = 3 varnames = c("TL", "CSE", "IWB") labels = list(varnames,varnames) cordat = list() for (i in 1:nrow(data)){ cordat[[i]] = vec2symMat(as.matrix(data[i,3:5]), diag = FALSE) dimnames(cordat[[i]]) = labels } cordat ## Sample sizes my.n <- data$N my.n ## Total sample sizes in each cell pattern.n(cordat, my.n) ## Number of studies in each cell pattern.na(cordat, show.na = FALSE) ## Stage 1 analysis: find an average correlation matrix stage1 <- tssem1(cordat, my.n) summary(stage1) stage1 <- tssem1(Cov=cordat, n=data$N, method="REM", RE.type = "Diag") summary(stage1) #_________________________STAGE_2_____________________________________________ # #A Matrisi – Yapısal Katsayılar (Paths) # #A[1,1] = 0 TL kendini etkilemez (döngü yok) # A[1,2] = 0.1*b21 PE'nin TL üzerindeki etkisi (b21 parametresi, 0.1 ile ölçeklenmiş) # #A[3,1:3] = 0 IWB kendisini veya diğerlerini etkilemez (modelin son değişkeni gibi) # A[1,3] = 0.1*b31 IWB'nin TL üzerindeki etkisi (b31 parametresi) # A[2,1] = 0 TL’nin PE üzerinde etkisi yok (çünkü bu ters yön) # A[2,2] = 0 PE kendini etkilemez # A[2,3] = 0.1*b32 IWB'nin PE üzerindeki etkisi (b32) A = create.mxMatrix( c( 0, 0, 0, "0.1*b21", 0, 0, "0.1*b31", "0.1*b32", 0 ), type = "Full", nrow = 3, ncol = 3, byrow = TRUE, name = "A", dimnames = list(varnames,varnames)) #S Matrisi – Hata / Artık Varyans-Kovaryansları S <- create.mxMatrix( c( 1, # TL'in varyansı sabit 1 0, "1*p22", # PE'nin varyansı serbest p22 0, 0, "1*p33" # IWB'nin varyansı serbest p33 ), type = "Symm", byrow = TRUE, name = "S", dimnames = list(varnames, varnames) ) stage2 <- tssem2(stage1, Amatrix=A, Smatrix=S, diag.constraints=TRUE, intervals="LB") summary(stage2) # Yol diyagramı çizimi plot_object_CSE <- meta2semPlot(stage2) # Yol grafiğini çiz semPaths(plot_object_CSE, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 7, nCharNodes = 0) semPaths(plot_object_CSE, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 9, nCharNodes = 0, color = list(lat = "skyblue", man = "lightgreen"), edge.color = "black", edge.label.color = "blue") stage2_direct <- tssem2(stage1, Amatrix=A, Smatrix=S, diag.constraints=TRUE, intervals="LB", mx.algebras=list( Direct_Effect = mxAlgebra(b31, name="Direct_Effect"), Indirect_Effect = mxAlgebra(b21 * b32, name="Indirect_Effect") )) summary(stage2_direct) #______________________________________________________________________________ # TL -> KS -> IWB #_______________________________________________________________________________ #_________________________STAGE_1_____________________________________________ data <- read_excel("metasem.xlsx", sheet = 3) data # Tüm veri çerçevesini sayısala çevir data <- as.data.frame(lapply(data, as.numeric)) names(data) nvar = 3 varnames = c("TL", "KS", "IWB") labels = list(varnames,varnames) cordat = list() for (i in 1:nrow(data)){ cordat[[i]] = vec2symMat(as.matrix(data[i,3:5]), diag = FALSE) dimnames(cordat[[i]]) = labels } cordat ## Sample sizes my.n <- data$N my.n ## Total sample sizes in each cell pattern.n(cordat, my.n) ## Number of studies in each cell pattern.na(cordat, show.na = FALSE) ## Stage 1 analysis: find an average correlation matrix stage1 <- tssem1(cordat, my.n) summary(stage1) stage1 <- tssem1(Cov=cordat, n=data$N, method="REM", RE.type = "Diag") summary(stage1) #_________________________STAGE_2_____________________________________________ # #A Matrisi – Yapısal Katsayılar (Paths) # #A[1,1] = 0 TL kendini etkilemez (döngü yok) # A[1,2] = 0.1*b21 PE'nin TL üzerindeki etkisi (b21 parametresi, 0.1 ile ölçeklenmiş) # #A[3,1:3] = 0 IWB kendisini veya diğerlerini etkilemez (modelin son değişkeni gibi) # A[1,3] = 0.1*b31 IWB'nin TL üzerindeki etkisi (b31 parametresi) # A[2,1] = 0 TL’nin PE üzerinde etkisi yok (çünkü bu ters yön) # A[2,2] = 0 PE kendini etkilemez # A[2,3] = 0.1*b32 IWB'nin PE üzerindeki etkisi (b32) A = create.mxMatrix( c( 0, 0, 0, "0.1*b21", 0, 0, "0.1*b31", "0.1*b32", 0 ), type = "Full", nrow = 3, ncol = 3, byrow = TRUE, name = "A", dimnames = list(varnames,varnames)) #S Matrisi – Hata / Artık Varyans-Kovaryansları S <- create.mxMatrix( c( 1, # TL'in varyansı sabit 1 0, "1*p22", # PE'nin varyansı serbest p22 0, 0, "1*p33" # IWB'nin varyansı serbest p33 ), type = "Symm", byrow = TRUE, name = "S", dimnames = list(varnames, varnames) ) stage2 <- tssem2(stage1, Amatrix=A, Smatrix=S, diag.constraints=TRUE, intervals="LB") summary(stage2) # Yol diyagramı çizimi plot_object_KS <- meta2semPlot(stage2) # Yol grafiğini çiz semPaths(plot_object_KS, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 7, nCharNodes = 0) semPaths(plot_object_KS, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 9, nCharNodes = 0, color = list(lat = "skyblue", man = "lightgreen"), edge.color = "black", edge.label.color = "blue") stage2_direct <- tssem2(stage1, Amatrix=A, Smatrix=S, diag.constraints=TRUE, intervals="LB", mx.algebras=list( Direct_Effect = mxAlgebra(b31, name="Direct_Effect"), Indirect_Effect = mxAlgebra(b21 * b32, name="Indirect_Effect") )) summary(stage2_direct) #______________________________________________________________________________ # TL -> LMX -> IWB #_______________________________________________________________________________ #_________________________STAGE_1_____________________________________________ data <- read_excel("metasem.xlsx", sheet = 4) data # Tüm veri çerçevesini sayısala çevir data <- as.data.frame(lapply(data, as.numeric)) names(data) nvar = 3 varnames = c("TL", "LMX", "IWB") labels = list(varnames,varnames) cordat = list() for (i in 1:nrow(data)){ cordat[[i]] = vec2symMat(as.matrix(data[i,3:5]), diag = FALSE) dimnames(cordat[[i]]) = labels } cordat ## Sample sizes my.n <- data$N my.n ## Total sample sizes in each cell pattern.n(cordat, my.n) ## Number of studies in each cell pattern.na(cordat, show.na = FALSE) ## Stage 1 analysis: find an average correlation matrix stage1 <- tssem1(cordat, my.n) summary(stage1) stage1 <- tssem1(Cov=cordat, n=data$N, method="REM", RE.type = "Diag") summary(stage1) #_________________________STAGE_2_____________________________________________ # #A Matrisi – Yapısal Katsayılar (Paths) # #A[1,1] = 0 TL kendini etkilemez (döngü yok) # A[1,2] = 0.1*b21 PE'nin TL üzerindeki etkisi (b21 parametresi, 0.1 ile ölçeklenmiş) # #A[3,1:3] = 0 IWB kendisini veya diğerlerini etkilemez (modelin son değişkeni gibi) # A[1,3] = 0.1*b31 IWB'nin TL üzerindeki etkisi (b31 parametresi) # A[2,1] = 0 TL’nin PE üzerinde etkisi yok (çünkü bu ters yön) # A[2,2] = 0 PE kendini etkilemez # A[2,3] = 0.1*b32 IWB'nin PE üzerindeki etkisi (b32) A = create.mxMatrix( c( 0, 0, 0, "0.1*b21", 0, 0, "0.1*b31", "0.1*b32", 0 ), type = "Full", nrow = 3, ncol = 3, byrow = TRUE, name = "A", dimnames = list(varnames,varnames)) #S Matrisi – Hata / Artık Varyans-Kovaryansları S <- create.mxMatrix( c( 1, # TL'in varyansı sabit 1 0, "1*p22", # PE'nin varyansı serbest p22 0, 0, "1*p33" # IWB'nin varyansı serbest p33 ), type = "Symm", byrow = TRUE, name = "S", dimnames = list(varnames, varnames) ) stage2 <- tssem2(stage1, Amatrix=A, Smatrix=S, diag.constraints=TRUE, intervals="LB") summary(stage2) # Yol diyagramı çizimi plot_object_LMX <- meta2semPlot(stage2) semPaths(plot_object_LMX, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 9, nCharNodes = 0, color = list(lat = "skyblue", man = "lightgreen"), edge.color = "black", edge.label.color = "blue") # Yol grafiğini çiz semPaths(plot_object_LMX, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 7, nCharNodes = 0) stage2_direct <- tssem2(stage1, Amatrix=A, Smatrix=S, diag.constraints=TRUE, intervals="LB", mx.algebras=list( Direct_Effect = mxAlgebra(b31, name="Direct_Effect"), Indirect_Effect = mxAlgebra(b21 * b32, name="Indirect_Effect") )) summary(stage2_direct) #______________________________________________________________________________ # TL -> WE -> IWB #_______________________________________________________________________________ #_________________________STAGE_1_____________________________________________ data <- read_excel("metasem.xlsx", sheet = 5) data # Tüm veri çerçevesini sayısala çevir data <- as.data.frame(lapply(data, as.numeric)) names(data) nvar = 3 varnames = c("TL", "WE", "IWB") labels = list(varnames,varnames) cordat = list() for (i in 1:nrow(data)){ cordat[[i]] = vec2symMat(as.matrix(data[i,3:5]), diag = FALSE) dimnames(cordat[[i]]) = labels } cordat ## Sample sizes my.n <- data$N my.n ## Total sample sizes in each cell pattern.n(cordat, my.n) ## Number of studies in each cell pattern.na(cordat, show.na = FALSE) ## Stage 1 analysis: find an average correlation matrix stage1 <- tssem1(cordat, my.n) summary(stage1) stage1 <- tssem1(Cov=cordat, n=data$N, method="REM", RE.type = "Diag") summary(stage1) #_________________________STAGE_2_____________________________________________ # #A Matrisi – Yapısal Katsayılar (Paths) # #A[1,1] = 0 TL kendini etkilemez (döngü yok) # A[1,2] = 0.1*b21 PE'nin TL üzerindeki etkisi (b21 parametresi, 0.1 ile ölçeklenmiş) # #A[3,1:3] = 0 IWB kendisini veya diğerlerini etkilemez (modelin son değişkeni gibi) # A[1,3] = 0.1*b31 IWB'nin TL üzerindeki etkisi (b31 parametresi) # A[2,1] = 0 TL’nin PE üzerinde etkisi yok (çünkü bu ters yön) # A[2,2] = 0 PE kendini etkilemez # A[2,3] = 0.1*b32 IWB'nin PE üzerindeki etkisi (b32) A = create.mxMatrix( c( 0, 0, 0, "0.1*b21", 0, 0, "0.1*b31", "0.1*b32", 0 ), type = "Full", nrow = 3, ncol = 3, byrow = TRUE, name = "A", dimnames = list(varnames,varnames)) #S Matrisi – Hata / Artık Varyans-Kovaryansları S <- create.mxMatrix( c( 1, # TL'in varyansı sabit 1 0, "1*p22", # PE'nin varyansı serbest p22 0, 0, "1*p33" # IWB'nin varyansı serbest p33 ), type = "Symm", byrow = TRUE, name = "S", dimnames = list(varnames, varnames) ) stage2 <- tssem2(stage1, Amatrix=A, Smatrix=S, diag.constraints=TRUE, intervals="LB") summary(stage2) # Yol diyagramı çizimi plot_object_WE <- meta2semPlot(stage2) # Yol grafiğini çiz semPaths(plot_object_WE, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 7, nCharNodes = 0) semPaths(plot_object_WE, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 9, nCharNodes = 0, color = list(lat = "skyblue", man = "lightgreen"), edge.color = "black", edge.label.color = "blue") stage2_direct <- tssem2(stage1, Amatrix=A, Smatrix=S, diag.constraints=TRUE, intervals="LB", mx.algebras=list( Direct_Effect = mxAlgebra(b31, name="Direct_Effect"), Indirect_Effect = mxAlgebra(b21 * b32, name="Indirect_Effect") )) summary(stage2_direct) #______________________________________________________________________________ # TL -> IM -> IWB #_______________________________________________________________________________ #_________________________STAGE_1_____________________________________________ data <- read_excel("metasem.xlsx", sheet = 6) data # Tüm veri çerçevesini sayısala çevir data <- as.data.frame(lapply(data, as.numeric)) names(data) nvar = 3 varnames = c("TL", "IM", "IWB") labels = list(varnames,varnames) cordat = list() for (i in 1:nrow(data)){ cordat[[i]] = vec2symMat(as.matrix(data[i,3:5]), diag = FALSE) dimnames(cordat[[i]]) = labels } cordat ## Sample sizes my.n <- data$N my.n ## Total sample sizes in each cell pattern.n(cordat, my.n) ## Number of studies in each cell pattern.na(cordat, show.na = FALSE) ## Stage 1 analysis: find an average correlation matrix stage1 <- tssem1(cordat, my.n) summary(stage1) stage1 <- tssem1(Cov=cordat, n=data$N, method="REM", RE.type = "Diag") summary(stage1) #_________________________STAGE_2_____________________________________________ # #A Matrisi – Yapısal Katsayılar (Paths) # #A[1,1] = 0 TL kendini etkilemez (döngü yok) # A[1,2] = 0.1*b21 PE'nin TL üzerindeki etkisi (b21 parametresi, 0.1 ile ölçeklenmiş) # #A[3,1:3] = 0 IWB kendisini veya diğerlerini etkilemez (modelin son değişkeni gibi) # A[1,3] = 0.1*b31 IWB'nin TL üzerindeki etkisi (b31 parametresi) # A[2,1] = 0 TL’nin PE üzerinde etkisi yok (çünkü bu ters yön) # A[2,2] = 0 PE kendini etkilemez # A[2,3] = 0.1*b32 IWB'nin PE üzerindeki etkisi (b32) A = create.mxMatrix( c( 0, 0, 0, "0.1*b21", 0, 0, "0.1*b31", "0.1*b32", 0 ), type = "Full", nrow = 3, ncol = 3, byrow = TRUE, name = "A", dimnames = list(varnames,varnames)) #S Matrisi – Hata / Artık Varyans-Kovaryansları S <- create.mxMatrix( c( 1, # TL'in varyansı sabit 1 0, "1*p22", # PE'nin varyansı serbest p22 0, 0, "1*p33" # IWB'nin varyansı serbest p33 ), type = "Symm", byrow = TRUE, name = "S", dimnames = list(varnames, varnames) ) stage2 <- tssem2(stage1, Amatrix=A, Smatrix=S, diag.constraints=TRUE, intervals="LB") summary(stage2) # Yol diyagramı çizimi plot_object_IM <- meta2semPlot(stage2) # Yol grafiğini çiz semPaths(plot_object_IM, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 7, nCharNodes = 0) semPaths(plot_object_IM, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 9, nCharNodes = 0, color = list(lat = "skyblue", man = "lightgreen"), edge.color = "black", edge.label.color = "blue") stage2_direct <- tssem2(stage1, Amatrix=A, Smatrix=S, diag.constraints=TRUE, intervals="LB", mx.algebras=list( Direct_Effect = mxAlgebra(b31, name="Direct_Effect"), Indirect_Effect = mxAlgebra(b21 * b32, name="Indirect_Effect") )) summary(stage2_direct) #------------------------------------------------------------------------------- #______________________________________________________________________________ # TL -> PS-> IWB #_______________________________________________________________________________ #_________________________STAGE_1_____________________________________________ #_________________________STAGE_1_____________________________________________ data <- read_excel("metasem.xlsx", sheet = 7) data # Tüm veri çerçevesini sayısala çevir data <- as.data.frame(lapply(data, as.numeric)) names(data) nvar = 3 varnames = c("TL", "PS", "IWB") labels = list(varnames,varnames) cordat = list() for (i in 1:nrow(data)){ cordat[[i]] = vec2symMat(as.matrix(data[i,3:5]), diag = FALSE) dimnames(cordat[[i]]) = labels } cordat ## Sample sizes my.n <- data$N my.n ## Total sample sizes in each cell pattern.n(cordat, my.n) ## Number of studies in each cell pattern.na(cordat, show.na = FALSE) ## Stage 1 analysis: find an average correlation matrix stage1 <- tssem1(cordat, my.n) summary(stage1) stage1 <- tssem1(Cov=cordat, n=data$N, method="REM", RE.type = "Diag") summary(stage1) ## Average correlation matrix varlist = c("TL", "PS", "IWB") meanR <- vec2symMat(coef(stage1, select = "fixed"), diag = FALSE) dimnames(meanR) <- list(varlist, varlist) meanR ## Absolute heterogeneity variance: tau^2 tau2 <- vec2symMat(coef(stage1, select = "random"), diag = FALSE) dimnames(tau2) <- list(varlist, varlist) tau2 ## Relative heterogeneity index: I^2 I2 <- vec2symMat(summary(stage1)$I2.values[, "Estimate"], diag = FALSE) dimnames(I2) <- list(varlist, varlist) I2 ## Proposed model model <- "IWB ~ c*TL + b*PS PS ~ a*TL TL ~~ 1*TL" plot(model, color="yellow") varlist = c("TL", "PS", "IWB") RAM1 <- lavaan2RAM(model, obs.variables = varlist) RAM1 #_________________________STAGE_2_____________________________________________ # #A Matrisi – Yapısal Katsayılar (Paths) # #A[1,1] = 0 TL kendini etkilemez (döngü yok) # A[1,2] = 0.1*b21 PE'nin TL üzerindeki etkisi (b21 parametresi, 0.1 ile ölçeklenmiş) # #A[3,1:3] = 0 IWB kendisini veya diğerlerini etkilemez (modelin son değişkeni gibi) # A[1,3] = 0.1*b31 IWB'nin TL üzerindeki etkisi (b31 parametresi) # A[2,1] = 0 TL’nin PE üzerinde etkisi yok (çünkü bu ters yön) # A[2,2] = 0 PE kendini etkilemez # A[2,3] = 0.1*b32 IWB'nin PE üzerindeki etkisi (b32) # # Harmonik ortalama N hesapla # N_values <- data$N # harmonic_N <- length(N_values) / sum(1 / N_values) # # harmonic_N # # 1. Harmonik N hesapla # harmonic_N <- length(data$N) / sum(1 / data$N) # # # 2. stage1 nesnesinin içindeki toplam N değerini override et # stage1$total.n <- harmonic_N # # stage2a <- tssem2( stage1, RAM=RAM1, diag.constraints = TRUE, intervals = "LB" ) summary(stage2a) # Yol diyagramı çizimi plot_object_PS <- meta2semPlot(stage2a) # Yol grafiğini çiz semPaths(plot_object_PS, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 7, nCharNodes = 0) semPaths(plot_object_PS, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, rotation = 1, # 0 = yukarıdan aşağı, 2 = soldan sağa sizeMan = 9, nCharNodes = 0, color = list(lat = "skyblue", man = "lightgreen"), edge.color = "black", edge.label.color = "blue") # semPaths(plot_object_PE, # layout = "tree2", # whatLabels = "est", # style = "ram", # sizeMan = 8, # color = list(man = "lightblue", lat = "lightyellow"), # border.color = list(man = "darkblue", lat = "orange"), # edge.label.cex = 1.2) stage2b <- tssem2(stage1, RAM=RAM1, diag.constraints=TRUE, intervals="LB", mx.algebras = list(Indirect=mxAlgebra(a*b, name="Indirect"), Direct=mxAlgebra(c, name="Direct"))) summary(stage2b) HatVar <-0.81 R2 <- 1 - HatVar R2 ## Add a constraint on c=a*b ## Rerun to remove errors stage2b <- mxModel(stage2b, mxConstraint(c==a*b, name="constraint")) stage2b <- mxRun(stage2b) stage2b <- mxTryHard(stage2b, extraTries = 100) stage2b <- mxRun(stage2b, intervals = TRUE) summary(stage2b) anova(stage2$mx.fit, stage2b) #------------------------------------------------------------------------------- #______________________________________________________________________________ # TL -> Trs-> IWB #_______________________________________________________________________________ #_________________________STAGE_1_____________________________________________ data <- read_excel("metasem.xlsx", sheet = 8) data # Tüm veri çerçevesini sayısala çevir data <- as.data.frame(lapply(data, as.numeric)) names(data) nvar = 3 varnames = c("TL", "TRS", "IWB") labels = list(varnames,varnames) cordat = list() for (i in 1:nrow(data)){ cordat[[i]] = vec2symMat(as.matrix(data[i,3:5]), diag = FALSE) dimnames(cordat[[i]]) = labels } cordat ## Sample sizes my.n <- data$N my.n ## Total sample sizes in each cell pattern.n(cordat, my.n) ## Number of studies in each cell pattern.na(cordat, show.na = FALSE) length(cordat) length(my.n) ## Stage 1 analysis: find an average correlation matrix stage1 <- tssem1(cordat, my.n) summary(stage1) stage1 <- tssem1(Cov=cordat, n=data$N, method="REM", RE.type = "Diag") summary(stage1) ## Average correlation matrix meanR <- vec2symMat(coef(stage1, select = "fixed"), diag = FALSE) dimnames(meanR) <- list(varlist, varlist) meanR ## Absolute heterogeneity variance: tau^2 tau2 <- vec2symMat(coef(stage1, select = "random"), diag = FALSE) dimnames(tau2) <- list(varlist, varlist) tau2 ## Relative heterogeneity index: I^2 I2 <- vec2symMat(summary(stage1)$I2.values[, "Estimate"], diag = FALSE) dimnames(I2) <- list(varlist, varlist) I2 ## Proposed model model <- "IWB ~ c*TL + b*TRS TRS ~ a*TL TL ~~ 1*TL" plot(model, color="yellow") varlist = c("TL", "TRS", "IWB") RAM1 <- lavaan2RAM(model, obs.variables = varlist) RAM1 #_________________________STAGE_2_____________________________________________ # #A Matrisi – Yapısal Katsayılar (Paths) # #A[1,1] = 0 TL kendini etkilemez (döngü yok) # A[1,2] = 0.1*b21 PE'nin TL üzerindeki etkisi (b21 parametresi, 0.1 ile ölçeklenmiş) # #A[3,1:3] = 0 IWB kendisini veya diğerlerini etkilemez (modelin son değişkeni gibi) # A[1,3] = 0.1*b31 IWB'nin TL üzerindeki etkisi (b31 parametresi) # A[2,1] = 0 TL’nin PE üzerinde etkisi yok (çünkü bu ters yön) # A[2,2] = 0 PE kendini etkilemez # A[2,3] = 0.1*b32 IWB'nin PE üzerindeki etkisi (b32) # # Harmonik ortalama N hesapla # N_values <- data$N # harmonic_N <- length(N_values) / sum(1 / N_values) # # harmonic_N # # 1. Harmonik N hesapla # harmonic_N <- length(data$N) / sum(1 / data$N) # # # 2. stage1 nesnesinin içindeki toplam N değerini override et # stage1$total.n <- harmonic_N # # stage2a <- tssem2( stage1, RAM=RAM1, diag.constraints = TRUE, intervals = "LB" ) summary(stage2a) # Yol diyagramı çizimi plot_object_TRS <- meta2semPlot(stage2a) # Yol grafiğini çiz semPaths(plot_object_TRS, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 7, nCharNodes = 0) semPaths(plot_object_TRS, whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, rotation = 1, # 0 = yukarıdan aşağı, 2 = soldan sağa sizeMan = 9, nCharNodes = 0, color = list(lat = "skyblue", man = "lightgreen"), edge.color = "black", edge.label.color = "blue") # semPaths(plot_object_PE, # layout = "tree2", # whatLabels = "est", # style = "ram", # sizeMan = 8, # color = list(man = "lightblue", lat = "lightyellow"), # border.color = list(man = "darkblue", lat = "orange"), # edge.label.cex = 1.2) stage2b <- tssem2(stage1, RAM=RAM1, diag.constraints=TRUE, intervals="LB", mx.algebras = list(Indirect=mxAlgebra(a*b, name="Indirect"), Direct=mxAlgebra(c, name="Direct"))) summary(stage2b) HatVar <-0.74 R2 <- 1 - HatVar R2 ## Add a constraint on c=a*b ## Rerun to remove errors stage2b <- mxModel(stage2b, mxConstraint(c==a*b, name="constraint")) stage2b <- mxRun(stage2b) stage2b <- mxTryHard(stage2b, extraTries = 100) stage2b <- mxRun(stage2b, intervals = TRUE) summary(stage2b) anova(stage2$mx.fit, stage2b) #______________________________________________________________________________ models <- list(plot_object_PE, plot_object_IM, plot_object_CSE, plot_object_LMX, plot_object_WE, plot_object_KS, plot_object_PS, plot_object_TRS) names <- c("PE", "IM", "CSE", "LMX", "WE", "KS", "PS","TRS") for (i in 1:length(models)) { fname <- paste0("sem_", names[i], ".png") png(fname, width = 1200, height = 800, res = 150) semPaths(models[[i]], whatLabels = "est", layout = "tree2", edge.label.cex = 1.2, sizeMan = 9, nCharNodes = 0, color = list(lat = "skyblue", man = "lightgreen"), edge.color = "black", edge.label.color = "blue") dev.off() } ################################################################################## library(magick) # PNG dosyalarını oku img1 <- image_read("sem_PE.png") img2 <- image_read("sem_CSE.png") img3 <- image_read("sem_WE.png") img4 <- image_read("sem_IM.png") img5 <- image_read("sem_LMX.png") img6 <- image_read("sem_KS.png") img7 <- image_read("sem_PS.png") img8 <- image_read("sem_TRS.png") ## Her satırı yatay olarak birleştir row1 <- image_append(c(img1, img2)) # 1. satır row2 <- image_append(c(img3, img4)) # 2. satır row3 <- image_append(c(img5, img6)) # 3. satır row4 <- image_append(c(img7, img8)) # 4. satır # Satırları dikey olarak birleştir combined <- image_append(c(row1, row2, row3, row4), stack = TRUE) # Görseli kaydet # Görseli kaydet image_write(combined, "combined_2x4_sem_paths.png")