“2021MetaPOQ”的版本间的差异

来自OBHRM百科
跳转至： 导航、 搜索
第6行： 第6行：
  
 
==论文R代码==
 
==论文R代码==
 +
<pre>
 +
#加载安装包
 +
library(psychmeta)
 +
library(tidyr)
 +
library(magrittr)
 +
library(dplyr)
 +
library(tidyverse)
 +
library(rio)
 +
library(metaSEM)
 +
library(semPlot)
 +
library(plyr)
 +
library(meta)
 +
 +
setwd("D:\\")
 +
#请将data20201209.xlsx放在D盘根目录下。
 +
#如果放在其他目录下，请在此做相应调整。
 +
##################  全部数据准备  ######################
 +
data <- import("data20201209.xlsx")
 +
attach(data)
 +
correct_main <- correct_r(correction = c("meas", "uvdrr_x", "uvdrr_y", "uvirr_x",
 +
                                        "uvirr_y", "bvdrr", "bvirr"),
 +
                          rxyi = rxy, ux = 1, uy = 1,
 +
                          rxx = rxx,
 +
                          ryy = ryy, ux_observed = TRUE, uy_observed = TRUE,
 +
                          rxx_restricted = TRUE, rxx_type = "alpha", k_items_x = NA,
 +
                          ryy_restricted = TRUE, ryy_type = "alpha", k_items_y = NA,
 +
                          sign_rxz = 1, sign_ryz = 1, n = NULL, conf_level = 0.95,
 +
                          correct_bias = FALSE)
 +
data$rtp = correct_main[["correlations"]][["rtp"]]
 +
write.csv(data, "data.csv")
 +
data  <- import("data.csv")
 +
orig.data <- subset(data.frame(data, correct_main$correlations), rtp <1)
 +
 +
k.article <- length(unique(paste(orig.data$author, orig.data$title)))
 +
k.inde.sample <- length(unique(orig.data$id))
 +
k.eff.size <- nrow(unique(data.frame(orig.data$title, orig.data$rxy, orig.data$x, orig.data$y)))
 +
sample.size <- unique(data.frame(orig.data$title, orig.data$n))
 +
n.sample.size <- as.vector(sample.size$orig.data.n)%>%sum
 +
 +
descr <- data.frame(c("文献", "独立样本", "效应值", "样本规模"),
 +
  c(k.article, k.inde.sample, k.eff.size, n.sample.size))%>%
 +
  set_colnames(c("类型", "数量"))
 +
#### 1………………直接效应分析………………####
 +
#### 失安全系数计算 ####
 +
d <- subset(orig.data, orig.data$x.frame.all == "资质过剩感")
 +
attach(d)
 +
d$zr <- 0.5*log((1+rxy)/(1-rxy))
 +
zr <- split(d$zr, d$y.coding.t)
 +
n <- split(d$n, d$y.coding.t)
 +
t <- list()
 +
for (i in 1:length(n)){
 +
  t[[i]] <- zr[[i]]*sqrt(n[[i]]-3)
 +
  i = i + 1}
 +
nfs <- vector()
 +
for (i in 1:length(zr)){
 +
  nfs[[i]] <- abs(round(sum(t[[i]])^2/2.706 - length(zr[[i]])))
 +
  i = i + 1}
 +
final.nfs = data.frame(names(zr),nfs) %>% set_colnames(c("name", "nfs"))
 +
 +
zr.t <- split(d$zr, d$y.frame.all)
 +
n.t <- split(d$n, d$y.frame.all)
 +
for (i in 1:length(n.t)){
 +
  t[[i]] <- zr.t[[i]]*sqrt(n.t[[i]]-3)
 +
  i = i + 1}
 +
nfs <- vector()
 +
for (i in 1:length(zr.t)){
 +
  nfs[[i]] <- abs(round(sum(t[[i]])^2/2.706 - length(zr.t[[i]])))
 +
  i = i + 1}
 +
final.nfs.t = data.frame(names(zr.t),nfs)%>% set_colnames(c("name", "nfs"))
 +
nfs = rbind(final.nfs, final.nfs.t)
 +
 +
order <- data.frame(c("消极情绪", "愤怒", "无聊",
 +
                      "积极自我概念", "一般效能感", "其他效能感", "一般自尊", "基于组织的自尊",
 +
                      "消极行为", "反生产行为", "离职意愿",
 +
                      "积极行为", "创新", "工作重塑", "建言", "前瞻行为",
 +
                      "知识共享", "组织公民行为", "任务绩效")) %>% set_colnames("order")
 +
 +
t.nfs = left_join(order, nfs, by=c("order"="name"))
 +
write.csv(t.nfs, "nfs.csv")
 +
 +
attach(orig.data)
 +
#### 主要变量直接效应 ####
 +
main_r <- ma_r(rxyi = rxy, n = n,
 +
              sample_id = id, citekey = id,
 +
              construct_x = orig.data$x.frame.all,
 +
              construct_y = orig.data$y.frame.all,
 +
              moderators = orig.data$y.coding.t,
 +
              data = orig.data)
 +
main_rho <- ma_r(rxyi = rtp, n = n,
 +
                sample_id = id, citekey = id,
 +
                construct_x = orig.data$x.frame.all,
 +
                construct_y = orig.data$y.frame.all,
 +
                moderators = orig.data$y.coding.t,
 +
                data = orig.data)
 +
 +
#### 异质性检验 ####
 +
hetero <- heterogeneity(main_rho)%>%get_heterogeneity
 +
data.hetero = data.frame(matrix(NA, nrow = length(hetero), ncol=2))
 +
names(data.hetero) = c("id","var")
 +
data.hetero$id = colnames(sapply(hetero, function(x) x[[1]]))
 +
data.hetero$var = sapply(hetero, function(x) x[[1]]$percent_var_accounted[[1]])
 +
 +
#### 主效应分析表制作 ####
 +
main_r <- main_r %>% summary %>% get_metatab
 +
main_rho <- main_rho %>% summary %>% get_metatab
 +
maineffects <- data.frame(main_r$analysis_id,
 +
                          main_r$construct_x,
 +
                          main_r$construct_y,
 +
                          main_r$`orig.data$y.coding.t`,
 +
                          main_r$k, main_r$N,
 +
                          round(cbind(main_r$mean_r,  main_r$sd_r,
 +
                                      main_rho$mean_r,main_rho$sd_r),2),
 +
                          paste("[", round(main_rho$CI_LL_95, 2), ", ",
 +
                                round(main_rho$CI_UL_95, 2), "]"),
 +
                          paste("[", round(main_rho$CR_LL_80, 2), ", ",
 +
                                round(main_rho$CR_UL_80, 2),"]"),
 +
                          data.hetero) %>%
 +
  set_colnames(c("id", "x", "y", "dim", "k", "n",
 +
                "r", "SD_r","rho", "SD_rho",
 +
                "95% CI", "80% CV",
 +
                "var%"))
 +
main.poq <- subset(maineffects, maineffects$x == "资质过剩感" &
 +
                    maineffects$y != "个人工作匹配"&
 +
                    maineffects$y != "个人组织匹配"&
 +
                    maineffects$y != "健康"&
 +
                    maineffects$y != "对工作的态度"&
 +
                    maineffects$y != "对组织的态度")
 +
main.poq$`5k+10` = main.poq$k * 5 + 10
 +
 +
maineffects.poq = left_join(main.poq, t.nfs, by=c("dim"="order")) %>%
 +
  left_join(t.nfs, by=c("y"="order")) %>% unique
 +
 +
mainpoq <- data.frame(maineffects.poq[, 3:12],
 +
                      round(unlist(maineffects.poq$`var%`),2),
 +
                      maineffects.poq[, 14:16])%>%set_colnames(
 +
                      c("y", "dim", "k", "n", "r",
 +
                        "SD_r", "rho", "SD_rho", "95%CI", "80%CV",
 +
                        "var", "5k+10", "nfs.dim", "nfs.all"))
 +
write.csv(mainpoq, file = "poq直接效应.csv")
 +
 +
 +
attach(maineffects)
 +
maineffects$word <- paste("(", r,", ", rho, ")","\n",
 +
                          "(", k,", ", maineffects$n, ")","\n",
 +
                          `95% CI`,"\n",
 +
                          `80% CV`)
 +
 +
 +
#### 2………………中介效应检验………………####
 +
#### 数据准备：全模型数据 ####
 +
varnames <- c("POQ","NE", "PS", "NB", "PB", "JP")
 +
attach(maineffects)
 +
poq.ne <- subset(maineffects, x == "资质过剩感" &
 +
                  y == "消极情绪" &
 +
                  dim == "All Levels")
 +
poq.ps <- subset(maineffects, x == "资质过剩感" &
 +
                  y == "积极自我概念" &
 +
                  dim == "All Levels")
 +
poq.nb <- subset(maineffects, x == "资质过剩感" &
 +
                  y == "消极行为" &
 +
                  dim == "All Levels")
 +
poq.pb <- subset(maineffects, x == "资质过剩感" &
 +
                  y == "积极行为" &
 +
                  dim == "All Levels")
 +
poq.jp <- subset(maineffects, x == "资质过剩感" &
 +
                  y == "任务绩效" &
 +
                  dim == "All Levels")
 +
 +
ne.ps <- subset(maineffects, (x == "消极情绪" &
 +
                                y == "积极自我概念" &
 +
                                dim == "All Levels")|
 +
                  (y == "消极情绪" &
 +
                    x == "积极自我概念" &
 +
                    dim == "All Levels"))
 +
ne.nb <- subset(maineffects, (x == "消极情绪" &
 +
                                y == "消极行为" &
 +
                                dim == "All Levels")|
 +
                  (y == "消极情绪" &
 +
                    x == "消极行为" &
 +
                    dim == "All Levels"))
 +
ne.pb <- subset(maineffects, (x == "消极情绪" &
 +
                                y == "积极行为" &
 +
                                dim == "All Levels")|
 +
                  (y == "消极情绪" &
 +
                    x == "积极行为" &
 +
                    dim == "All Levels"))
 +
ne.jp <- subset(maineffects, (x == "消极情绪" &
 +
                                y == "任务绩效" &
 +
                                dim == "All Levels")|
 +
                  (y == "消极情绪" &
 +
                    x == "任务绩效" &
 +
                    dim == "All Levels"))
 +
 +
ps.nb <- subset(maineffects, (x == "积极自我概念" &
 +
                                y == "消极行为" &
 +
                                dim == "All Levels")|
 +
                  (y == "积极自我概念" &
 +
                    x == "消极行为" &
 +
                    dim == "All Levels"))
 +
ps.pb <- subset(maineffects, (x == "积极自我概念" &
 +
                                y == "积极行为" &
 +
                                dim == "All Levels")|
 +
                  (y == "积极自我概念" &
 +
                    x == "积极行为" &
 +
                    dim == "All Levels"))
 +
ps.jp <- subset(maineffects, (x == "积极自我概念" &
 +
                                y == "任务绩效" &
 +
                                dim == "All Levels")|
 +
                  (y == "积极自我概念" &
 +
                    x == "任务绩效" &
 +
                    dim == "All Levels"))
 +
 +
nb.pb <- subset(maineffects, (x == "消极行为" &
 +
                                y == "积极行为" &
 +
                                dim == "All Levels")|
 +
                  (y == "消极行为" &
 +
                    x == "积极行为" &
 +
                    dim == "All Levels"))
 +
nb.jp <- subset(maineffects, (x == "消极行为" &
 +
                                y == "任务绩效" &
 +
                                dim == "All Levels")|
 +
                  (y == "消极行为" &
 +
                    x == "任务绩效" &
 +
                    dim == "All Levels"))
 +
 +
pb.jp <- subset(maineffects, (x == "积极行为" &
 +
                                y == "任务绩效" &
 +
                                dim == "All Levels")|
 +
                  (y == "积极行为" &
 +
                    x == "任务绩效" &
 +
                    dim == "All Levels"))
 +
all.matrix <- vec2symMat(c(poq.ne$rho, poq.ps$rho, poq.nb$rho, poq.pb$rho, poq.jp$rho,
 +
                          ne.ps$rho, ne.nb$rho, ne.pb$rho, ne.jp$rho,
 +
                          ps.nb$rho, ps.pb$rho, ps.jp$rho,
 +
                          nb.pb$rho, nb.jp$rho,
 +
                          pb.jp$rho),
 +
                        diag = FALSE, byrow = FALSE) %>%
 +
  set_colnames(varnames) %>%
 +
  set_rownames(varnames)
 +
 +
all_dat <- list(all.matrix)
 +
is.pd(all_dat)
 +
 +
 +
attach(orig.data)
 +
main_in_r <- ma_r(rxyi = rxy, n = n,
 +
                  sample_id = id, citekey = id,
 +
                  construct_x = orig.data$x.frame.all,
 +
                  construct_y = orig.data$y.frame.all,
 +
                  moderators = orig.data$indi,
 +
                  data = orig.data)
 +
main_in_rho <- ma_r(rxyi = rtp, n = n,
 +
                    sample_id = id, citekey = id,
 +
                    construct_x = orig.data$x.frame.all,
 +
                    construct_y = orig.data$y.frame.all,
 +
                    moderators = orig.data$indi,
 +
                    data = orig.data)
 +
hetero_in <- heterogeneity(main_in_rho)%>%get_heterogeneity
 +
data.hetero_in = data.frame(matrix(NA, nrow = length(hetero_in), ncol=2))
 +
names(data.hetero_in) = c("id","var")
 +
data.hetero_in$id = colnames(sapply(hetero_in, function(x) x[[1]]))
 +
data.hetero_in$var = sapply(hetero_in, function(x) x[[1]]$percent_var_accounted[[1]])
 +
 +
 +
main_in_r <- main_in_r %>% summary %>% get_metatab %>% na.omit
 +
main_in_rho <- main_in_rho %>% summary %>% get_metatab %>% na.omit
 +
 +
main_indi <- data.frame(main_in_r$analysis_id,
 +
                        main_in_r$construct_x,
 +
                        main_in_r$construct_y,
 +
                        main_in_r$`orig.data$indi`,
 +
                        main_in_r$k, main_in_r$N,
 +
                        round(cbind(main_in_r$mean_r,  main_in_r$sd_r,
 +
                                    main_in_rho$mean_r,main_in_rho$sd_r),2),
 +
                        paste("[", round(main_in_rho$CI_LL_95, 2), ", ",
 +
                              round(main_in_rho$CI_UL_95, 2), "]"),
 +
                        paste("[", round(main_in_rho$CR_LL_80, 2), ", ",
 +
                              round(main_in_rho$CR_UL_80, 2),"]")
 +
) %>%
 +
  set_colnames(c("id", "x", "y", "indi", "k", "n",
 +
                "r", "SD_r","rho", "SD_rho",
 +
                "95% CI", "80% CV"))
 +
 +
#### 数据准备：低个人主义（高集体主义）数据 ####
 +
attach(main_indi)
 +
poq.ne.l <- subset(main_indi, x == "资质过剩感" &  y == "消极情绪" & indi == "L")
 +
 +
poq.ps.l <- subset(main_indi, x == "资质过剩感" &
 +
                    y == "积极自我概念" &
 +
                    indi == "L")
 +
poq.nb.l <- subset(main_indi, x == "资质过剩感" &
 +
                    y == "消极行为" &
 +
                    indi == "L")
 +
poq.pb.l <- subset(main_indi, x == "资质过剩感" &
 +
                    y == "积极行为" &
 +
                    indi == "L")
 +
poq.jp.l <- subset(main_indi, x == "资质过剩感" &
 +
                    y == "任务绩效" &
 +
                    indi == "L")
 +
 +
ne.ps.l <- subset(main_indi, (x == "消极情绪" &
 +
                                y == "积极自我概念" &
 +
                                indi == "L")|
 +
                    (y == "消极情绪" &
 +
                      x == "积极自我概念" &
 +
                      indi == "L"))
 +
ne.nb.l <- subset(main_indi, (x == "消极情绪" &
 +
                                y == "消极行为" &
 +
                                indi == "L")|
 +
                    (y == "消极情绪" &
 +
                      x == "消极行为" &
 +
                      indi == "L"))
 +
ne.pb.l <- subset(main_indi, (x == "消极情绪" &
 +
                                y == "积极行为" &
 +
                                indi == "L")|
 +
                    (y == "消极情绪" &
 +
                      x == "积极行为" &
 +
                      indi == "L"))
 +
ne.jp.l <- subset(main_indi, (x == "消极情绪" &
 +
                                y == "任务绩效" &
 +
                                indi == "L")|
 +
                    (y == "消极情绪" &
 +
                      x == "任务绩效" &
 +
                      indi == "L"))
 +
 +
ps.nb.l <- subset(main_indi, (x == "积极自我概念" &
 +
                                y == "消极行为" &
 +
                                indi == "L")|
 +
                    (y == "积极自我概念" &
 +
                      x == "消极行为" &
 +
                      indi == "L"))
 +
ps.pb.l <- subset(main_indi, (x == "积极自我概念" &
 +
                                y == "积极行为" &
 +
                                indi == "L")|
 +
                    (y == "积极自我概念" &
 +
                      x == "积极行为" &
 +
                      indi == "L"))
 +
ps.jp.l <- subset(main_indi, (x == "积极自我概念" &
 +
                                y == "任务绩效" &
 +
                                indi == "L")|
 +
                    (y == "积极自我概念" &
 +
                      x == "任务绩效" &
 +
                      indi == "L"))
 +
 +
nb.pb.l <- subset(main_indi, (x == "消极行为" &
 +
                                y == "积极行为" &
 +
                                indi == "L")|
 +
                    (y == "消极行为" &
 +
                      x == "积极行为" &
 +
                      indi == "L"))
 +
nb.jp.l <- subset(main_indi, (x == "消极行为" &
 +
                                y == "任务绩效" &
 +
                                indi == "L")|
 +
                    (y == "消极行为" &
 +
                      x == "任务绩效" &
 +
                      indi == "L"))
 +
 +
pb.jp.l <- subset(main_indi, (x == "积极行为" &
 +
                                y == "任务绩效" &
 +
                                indi == "L")|
 +
                    (y == "积极行为" &
 +
                      x == "任务绩效" &
 +
                      indi == "L"))
 +
all.matrix.l <- vec2symMat(c(poq.ne.l$rho, poq.ps.l$rho, poq.nb.l$rho, poq.pb.l$rho, poq.jp.l$rho,
 +
                            ne.ps.l$rho, ne.nb.l$rho, ne.pb.l$rho, ne.jp.l$rho,
 +
                            ps.nb.l$rho, ps.pb.l$rho, ps.jp.l$rho,
 +
                            nb.pb.l$rho, nb.jp.l$rho,
 +
                            pb.jp.l$rho),
 +
                          diag = FALSE, byrow = FALSE) %>%
 +
  set_colnames(varnames) %>%
 +
  set_rownames(varnames)
 +
 +
all_dat.l <- list(all.matrix.l)
 +
is.pd(all_dat.l)
 +
#### 数据准备：高个人主义（低集体主义）数据 ####
 +
attach(main_indi)
 +
poq.ne.h <- subset(main_indi, x == "资质过剩感" &
 +
                    y == "消极情绪" &
 +
                    indi == "H")
 +
poq.ps.h <- subset(main_indi, x == "资质过剩感" &
 +
                    y == "积极自我概念" &
 +
                    indi == "H")
 +
poq.nb.h <- subset(main_indi, x == "资质过剩感" &
 +
                    y == "消极行为" &
 +
                    indi == "H")
 +
poq.pb.h <- subset(main_indi, x == "资质过剩感" &
 +
                    y == "积极行为" &
 +
                    indi == "H")
 +
poq.jp.h <- subset(main_indi, x == "资质过剩感" &
 +
                    y == "任务绩效" &
 +
                    indi == "H")
 +
 +
ne.ps.h <- subset(main_indi, (x == "消极情绪" &
 +
                                y == "积极自我概念" &
 +
                                indi == "H")|
 +
                    (y == "消极情绪" &
 +
                      x == "积极自我概念" &
 +
                      indi == "H"))
 +
ne.nb.h <- subset(main_indi, (x == "消极情绪" &
 +
                                y == "消极行为" &
 +
                                indi == "H")|
 +
                    (y == "消极情绪" &
 +
                      x == "消极行为" &
 +
                      indi == "H"))
 +
ne.pb.h <- subset(main_indi, (x == "消极情绪" &
 +
                                y == "积极行为" &
 +
                                indi == "H")|
 +
                    (y == "消极情绪" &
 +
                      x == "积极行为" &
 +
                      indi == "H"))
 +
ne.jp.h <- subset(main_indi, (x == "消极情绪" &
 +
                                y == "任务绩效" &
 +
                                indi == "H")|
 +
                    (y == "消极情绪" &
 +
                      x == "任务绩效" &
 +
                      indi == "H"))
 +
 +
ps.nb.h <- subset(main_indi, (x == "积极自我概念" &
 +
                                y == "消极行为" &
 +
                                indi == "H")|
 +
                    (y == "积极自我概念" &
 +
                      x == "消极行为" &
 +
                      indi == "H"))
 +
ps.pb.h <- subset(main_indi, (x == "积极自我概念" &
 +
                                y == "积极行为" &
 +
                                indi == "H")|
 +
                    (y == "积极自我概念" &
 +
                      x == "积极行为" &
 +
                      indi == "H"))
 +
ps.jp.h <- subset(main_indi, (x == "积极自我概念" &
 +
                                y == "任务绩效" &
 +
                                indi == "H")|
 +
                    (y == "积极自我概念" &
 +
                      x == "任务绩效" &
 +
                      indi == "H"))
 +
 +
nb.pb.h <- subset(main_indi, (x == "消极行为" &
 +
                                y == "积极行为" &
 +
                                indi == "H")|
 +
                    (y == "消极行为" &
 +
                      x == "积极行为" &
 +
                      indi == "H"))
 +
nb.jp.h <- subset(main_indi, (x == "消极行为" &
 +
                                y == "任务绩效" &
 +
                                indi == "H")|
 +
                    (y == "消极行为" &
 +
                      x == "任务绩效" &
 +
                      indi == "H"))
 +
 +
pb.jp.h <- subset(main_indi, (x == "积极行为" &
 +
                                y == "任务绩效" &
 +
                                indi == "H")|
 +
                    (y == "积极行为" &
 +
                      x == "任务绩效" &
 +
                      indi == "H"))
 +
all.matrix.h <- vec2symMat(c(poq.ne.h$rho, poq.ps.h$rho, poq.nb.h$rho, poq.pb.h$rho, poq.jp.h$rho,
 +
                            ne.ps.h$rho, ne.nb.h$rho, ne.pb.h$rho, ne.jp.h$rho,
 +
                            ps.nb.h$rho, ps.pb.h$rho, ps.jp.h$rho,
 +
                            nb.pb.h$rho, nb.jp.h$rho,
 +
                            pb.jp.h$rho),
 +
                          diag = FALSE, byrow = FALSE) %>%
 +
  set_colnames(varnames) %>%
 +
  set_rownames(varnames)
 +
 +
all_dat.h <- list(all.matrix.h)
 +
is.pd(all_dat.h)
 +
#### 调和平均数 ####
 +
library(psych)
 +
all_n <- c(poq.ne$n, poq.ps$n, poq.nb$n, poq.pb$n, poq.jp$n,
 +
          ne.ps$n, ne.nb$n, ne.pb$n, ne.jp$n,
 +
          ps.nb$n, ps.pb$n, ps.jp$n,
 +
          nb.pb$n, nb.jp$n,
 +
          pb.jp$n) %>% harmonic.mean %>% round
 +
n_l <- c(poq.ne.l$n, poq.ps.l$n, poq.nb.l$n, poq.pb.l$n, poq.jp.l$n,
 +
        ne.ps.l$n, ne.nb.l$n, ne.pb.l$n, ne.jp.l$n,
 +
        ps.nb.l$n, ps.pb.l$n, ps.jp.l$n,
 +
        nb.pb.l$n, nb.jp.l$n,
 +
        pb.jp.l$n) %>% harmonic.mean %>% round
 +
n_h <- c(poq.ne.h$n, poq.ps.h$n, poq.nb.h$n, poq.pb.h$n, poq.jp.h$n,
 +
        ne.ps.h$n, ne.nb.h$n, ne.pb.h$n, ne.jp.h$n,
 +
        ps.nb.h$n, ps.pb.h$n, ps.jp.h$n,
 +
        nb.pb.h$n, nb.jp.h$n,
 +
        pb.jp.h$n) %>% harmonic.mean %>% round
 +
 +
####模型设置 ####
 +
 +
model_all <- 'NB ~  POQ + PS + NE
 +
  PB ~  POQ + PS + NE
 +
  JP ~  POQ + PS + NE
 +
 +
  NE ~ POQ
 +
  PS ~ POQ
 +
 +
  NE ~~ 1*NE
 +
  PS ~~ 1*PS
 +
  POQ ~~ 1*POQ
 +
 +
  NE ~~ 0.1*PS
 +
 +
  NB ~~ 0.1*NB
 +
  PB ~~ 0.1*PB
 +
  JP ~~ 0.1*JP
 +
 +
  NB ~~ 0.1*PB
 +
  NB ~~ 0.1*JP
 +
  JP ~~ 0.1*PB'
 +
 +
plot(model_all, col="yellow")
 +
 +
RAM_all <- lavaan2RAM(model_all, obs.variables = varnames)
 +
RAM_all
 +
 +
 +
#### SEM全部数据及调和平均数 ####
 +
fin_dat <- list(all_dat, all_dat.l, all_dat.h)
 +
fin_n <- c(all_n, n_l, n_h)
 +
 +
 +
stage1 <- list()
 +
stage2 <- list()
 +
stage2_LB <- list()
 +
stage2_Z <- list()
 +
ind_LB <- list()
 +
ind_Z <- list()
 +
fitness <- list()
 +
path <- list()
 +
ind <- list()
 +
sem <- list()
 +
name <- c("all", "low individualism", "high individualism")
 +
for (i in 1:length(fin_dat)){
 +
  #### stage1 ####
 +
  stage1[[i]] <- tssem1(Cov = fin_dat[[i]], n = fin_n[i], method = "FEM",
 +
                        RE.type = "Diag", RE.startvalues=0.1,
 +
                        acov = "weighted")
 +
 
 +
  summary(stage1[[i]])
 +
  #### stage2 ####
 +
  stage2[[i]] <- tssem2(stage1[[i]], Amatrix = RAM_all$A, Smatrix = RAM_all$S)
 +
  summary(stage2[[i]])
 +
 
 +
  stage2_LB[[i]] <- tssem2(stage1[[i]], Amatrix = RAM_all$A, Smatrix = RAM_all$S,
 +
                          diag.constraints = TRUE, intervals.type = "LB",
 +
                          mx.algebras = list(
 +
                            #POQ-NE-Y                       
 +
                            IndNE_NB = mxAlgebra(NEONPOQ*NBONNE, name = "IndNE_NB"),
 +
                            IndNE_PB = mxAlgebra(NEONPOQ*PBONNE, name = "IndNE_PB"),
 +
                            IndNE_JP = mxAlgebra(NEONPOQ*JPONNE, name = "IndNE_JP"),
 +
                            #POQ-PS-Y                         
 +
                            IndPS_NB = mxAlgebra(PSONPOQ*NBONPS, name = "IndPS_NB"),
 +
                            IndPS_PB = mxAlgebra(PSONPOQ*PBONPS, name = "IndPS_PB"),
 +
                            IndPS_JP = mxAlgebra(PSONPOQ*JPONPS, name = "IndPS_JP"),
 +
                            #POQ-NE+PS-Y                         
 +
                            Indall_NB = mxAlgebra(NEONPOQ*NBONNE + PSONPOQ*NBONPS, name = "Indall_NB"),
 +
                            Indall_PB = mxAlgebra(NEONPOQ*PBONNE + PSONPOQ*PBONPS, name = "Indall_PB"),
 +
                            Indall_JP = mxAlgebra(NEONPOQ*JPONNE + PSONPOQ*JPONPS, name = "Indall_JP"),
 +
                            #POQ-NE+PS-Y  PLUS POQ-Y                         
 +
                            Total_NB = mxAlgebra(NEONPOQ*NBONNE + PSONPOQ*NBONPS + NBONPOQ, name = "Total_NB"),
 +
                            Total_PB = mxAlgebra(NEONPOQ*PBONNE + PSONPOQ*PBONPS + PBONPOQ, name = "Total_PB"),
 +
                            Total_JP = mxAlgebra(NEONPOQ*JPONNE + PSONPOQ*JPONPS + JPONPOQ, name = "Total_JP"),
 +
                            #Diff POQ-(NE-PS)-Y                         
 +
                            DIF_NEminusPS_NB= mxAlgebra(NEONPOQ*NBONNE - PSONPOQ*NBONPS, name = "DIF_NEminusPS_NB"),
 +
                            DIF_NEminusPS_PB= mxAlgebra(NEONPOQ*PBONNE - PSONPOQ*PBONPS, name = "DIF_NEminusPS_PB"),
 +
                            DIF_NEminusPS_JP= mxAlgebra(NEONPOQ*JPONNE - PSONPOQ*JPONPS, name = "DIF_NEminusPS_JP")))
 +
  stage2_LB[[i]] <- rerun(stage2_LB[[i]])
 +
  ind_LB[[i]] <- summary(stage2_LB[[i]])
 +
 
 +
  stage2_Z[[i]] <- tssem2(stage1[[i]], Amatrix = RAM_all$A, Smatrix = RAM_all$S,
 +
                          diag.constraints = FALSE, intervals.type = "z",
 +
                          mx.algebras = list(
 +
                            IndNE_NB = mxAlgebra(NEONPOQ*NBONNE, name = "IndNE_NB"),
 +
                            IndNE_PB = mxAlgebra(NEONPOQ*PBONNE, name = "IndNE_PB"),
 +
                            IndNE_JP = mxAlgebra(NEONPOQ*JPONNE, name = "IndNE_JP"),
 +
                           
 +
                            IndPS_NB = mxAlgebra(PSONPOQ*NBONPS, name = "IndPS_NB"),
 +
                            IndPS_PB = mxAlgebra(PSONPOQ*PBONPS, name = "IndPS_PB"),
 +
                            IndPS_JP = mxAlgebra(PSONPOQ*JPONPS, name = "IndPS_JP"),
 +
                           
 +
                            Indall_NB = mxAlgebra(NEONPOQ*NBONNE + PSONPOQ*NBONPS, name = "Indall_NB"),
 +
                            Indall_PB = mxAlgebra(NEONPOQ*PBONNE + PSONPOQ*PBONPS, name = "Indall_PB"),
 +
                            Indall_JP = mxAlgebra(NEONPOQ*JPONNE + PSONPOQ*JPONPS, name = "Indall_JP"),
 +
                           
 +
                            Total_NB = mxAlgebra(NEONPOQ*NBONNE + PSONPOQ*NBONPS + NBONPOQ, name = "Total_NB"),
 +
                            Total_PB = mxAlgebra(NEONPOQ*PBONNE + PSONPOQ*PBONPS + PBONPOQ, name = "Total_PB"),
 +
                            Total_JP = mxAlgebra(NEONPOQ*JPONNE + PSONPOQ*JPONPS + JPONPOQ, name = "Total_JP"),
 +
                           
 +
                            DIF_NEminusPS_NB= mxAlgebra(NEONPOQ*NBONNE - PSONPOQ*NBONPS, name = "DIF_NEminusPS_NB"),
 +
                            DIF_NEminusPS_PB= mxAlgebra(NEONPOQ*PBONNE - PSONPOQ*PBONPS, name = "DIF_NEminusPS_PB"),
 +
                            DIF_NEminusPS_JP= mxAlgebra(NEONPOQ*JPONNE - PSONPOQ*JPONPS, name = "DIF_NEminusPS_JP")))
 +
  stage2_Z[[i]] <- rerun(stage2_Z[[i]])
 +
  ind_Z[[i]] <- summary(stage2_Z[[i]])
 +
  #### 拟合系数 ####
 +
  fitness[[i]] <- ind_LB[[i]][["stat"]]%>%as.data.frame%>%round(2)%>%
 +
    set_colnames("Goodness-of-fit indices")
 +
  #### 路径系数 ####
 +
  path[[i]] <- data.frame(round(ind_Z[[i]][["coefficients"]][,1], 2),
 +
                          paste("[", round(ind_Z[[i]][["coefficients"]][,3], 2),",",
 +
                                round(ind_Z[[i]][["coefficients"]][,4], 2), "]"),
 +
                          round(ind_Z[[i]][["coefficients"]][,2], 2),
 +
                          round(ind_Z[[i]][["coefficients"]][,5], 2),
 +
                          round(ind_Z[[i]][["coefficients"]][,6], 2)) %>%
 +
    set_colnames(c("β", "95% CI", "SE", "z", "p")) %>%
 +
    set_rownames(c("JPONNE", "JPONPOQ", "JPONPS",
 +
                  "NBONNE", "NBONPOQ", "NBONPS",
 +
                  "NEONPOQ", "PBONNE", "PBONPOQ",
 +
                  "PBONPS", "PSONPOQ"))
 +
  #### 中介效应 ####
 +
  ind[[i]] <- data.frame(round(ind_LB[[i]][["mx.algebras"]][,2],2),
 +
                        paste("[", round(ind_LB[[i]][["mx.algebras"]][,1],2), ",",
 +
                              round(ind_LB[[i]][["mx.algebras"]][,3],2), "]")) %>%
 +
    set_colnames(c("β", "95%CI"))%>%
 +
    set_rownames(c("IndNE_NB", "IndNE_PB", "IndNE_JP",
 +
                  "IndPS_NB", "IndPS_PB", "IndPS_JP",
 +
                  "Indall_NB", "Indall_PB", "Indall_JP",
 +
                  "Total_NB", "Total_PB", "Total_JP",
 +
                  "DIF_NEminusPS_NB", "DIF_NEminusPS_PB", "DIF_NEminusPS_JP"))
 +
  sem[[i]] <- list(name = name[i],
 +
                  fit=fitness[[i]], path = path[[i]], ind = ind[[i]])
 +
  sem[[i]]
 +
 
 +
  i = i + 1}
 +
sem
 +
#### 结构方程分析结果表格 ####
 +
final_fit <- cbind(sem[[1]][["fit"]][c(1,7,4),],
 +
                  sem[[2]][["fit"]][c(1,7,4),],
 +
                  sem[[3]][["fit"]][c(1,7,4),])%>%
 +
  set_rownames(c("n", "Chi", "p"))%>%
 +
  set_colnames(c("全样本", "高集体主义", "低集体主义"))
 +
final_path <- rbind(c("全样本", " "," ", " ", " ",
 +
                      "高集体主义", " "," ", " ", " ",
 +
                      "低集体主义", " "," ", " ", " "),
 +
                    rep(c("β", "95%CI", "SE", "z", "p"), 3),
 +
                    cbind(rbind(sem[[1]][["path"]]),
 +
                          rbind(sem[[2]][["path"]]),
 +
                          rbind(sem[[3]][["path"]])))%>%
 +
  set_colnames(c(1:15))
 +
final_ind <- rbind(rep(c("β", "95%CI"),3),
 +
                  cbind(sem[[1]][["ind"]],
 +
                        sem[[2]][["ind"]],
 +
                        sem[[3]][["ind"]]))%>%
 +
  set_colnames(c(1:6))
 +
write.csv(final_ind, "ind.csv")
 +
#### 相关系数文本 ####
 +
sem.matrix <- cbind(c(poq.ne$word, poq.ps$word, poq.nb$word, poq.pb$word, poq.jp$word),
 +
                    c(" ", ne.ps$word, ne.nb$word, ne.pb$word, ne.jp$word),
 +
                    c(" ", " ", ps.nb$word, ps.pb$word,ps.jp$word),
 +
                    c(" ", " ", " ", nb.pb$word, nb.jp$word),
 +
                    c(" ", " "," ", " ", pb.jp$word))%>%
 +
  set_colnames(c("资质过剩感", "消极情绪", "积极自我概念", "消极角色外行为", "积极角色外行为"))%>%
 +
  set_rownames(c("消极情绪", "积极自我概念", "消极角色外行为", "积极角色外行为", "任务绩效"))
 +
sem.matrix
 +
 +
#### 3………………概念冗余度检验………………####
 +
#### 相关矩阵构建 ####
 +
attach(maineffects)
 +
matrix <- data.frame(x, y, dim,
 +
                    paste( "(", k, ",", maineffects$n, ")" ) ,
 +
                    paste( "(", r, ",", rho, ")" ) ,
 +
                    `95% CI`, `80% CV`, rho)%>%
 +
  set_colnames(c("x", "dim", "y", "(k, n)", "(r, ρ)", "95% CI", "80% CV", "rho"))
 +
 +
matrix$all = paste(matrix$`(k, n)`, "\n", matrix$`(r, ρ)`, "\n",
 +
                  matrix$`95% CI`, "\n", matrix$`80% CV`)
 +
 +
 +
attach(matrix)
 +
poq_pj <- subset(matrix, x == "资质过剩感" & y == "个人工作匹配")
 +
poq_po <- subset(matrix, x == "资质过剩感" & y == "个人组织匹配")
 +
poq_y <- subset(matrix, (y == "压力（反向）"|
 +
                      y == "组织承诺"|
 +
                      y == "工作满意度"|
 +
                      y == "创新"|
 +
                      y == "组织公民行为"|
 +
                      y == "反生产行为"|
 +
                      y == "离职意愿"|
 +
                      y == "任务绩效")&
 +
                      x == "资质过剩感")
 +
pj_po <- subset(matrix, x == "个人工作匹配" &
 +
                      dim == "个人组织匹配")
 +
pj_y<- subset(matrix, (x == "个人工作匹配"&
 +
                      (y == "压力（反向）"|
 +
                        y == "组织承诺"|
 +
                        y == "工作满意度"))|
 +
                      (dim == "个人工作匹配"&
 +
                      (y == "创新"|
 +
                        y == "组织公民行为"|
 +
                        y == "反生产行为"|
 +
                        y == "离职意愿"|
 +
                        y == "任务绩效")))
 +
po_y <- subset(matrix, (x == "个人组织匹配"&
 +
                          (y == "压力（反向）"|
 +
                          y == "组织承诺"|
 +
                          y == "工作满意度"))|
 +
                      (dim == "个人组织匹配"&
 +
                          (y == "创新"|
 +
                          y == "组织公民行为"|
 +
                          y == "反生产行为"|
 +
                          y == "离职意愿"|
 +
                          y == "任务绩效")))
 +
 +
#### 相关矩阵表格 ####
 +
ynames <- c("创新", "组织公民行为", "反生产行为", "离职意愿",
 +
            "任务绩效", "压力（反向）", "工作满意度", "组织承诺")
 +
rwa.text <- data.frame(ynames, poq_y= poq_y$all, pj_y = pj_y$all, po_y = po_y$all,
 +
                      poq_pj = poq_pj$all, poq_po = poq_po$all,
 +
                      pj_po = pj_po$all)
 +
 +
varnames <- list()
 +
lables <- list()
 +
text.mx <- list()
 +
for (i in 1:nrow(rwa.text)){
 +
  varnames[[i]] <- c(ynames[[i]], "资质过剩感", "个人工作匹配", "个人组织匹配")
 +
  lables[[i]] <- list (varnames[[i]], varnames[[i]])
 +
  text.mx[[i]] <- vec2symMat(as.matrix(rwa.text[i,2:ncol(rwa.text)]),
 +
                            diag = FALSE, byrow = FALSE)
 +
  dimnames(text.mx[[i]]) <- lables[[i]]
 +
  i = i+1}
 +
as.vector(poq_y$y)
 +
poqfit.text.all <- t(data.frame(text.mx[[1]][2:4,2:4], text.mx[[1]][2:4,1],
 +
                                text.mx[[2]][2:4,1], text.mx[[3]][2:4,1],
 +
                                text.mx[[4]][2:4,1], text.mx[[5]][2:4,1],
 +
                                text.mx[[6]][2:4,1], text.mx[[7]][2:4,1],
 +
                                text.mx[[8]][2:4,1])) %>%
 +
                    set_rownames(c("资质过剩感", "个人工作匹配", "个人组织匹配",
 +
                                  "创新", "组织公民行为", "反生产行为", "离职意愿",
 +
                                "任务绩效", "压力（反向）", "工作满意度", "组织承诺"))
 +
write.csv(poqfit.text.all, file = "poq_fit相对矩阵文本.csv")
 +
 +
 +
#### 数据准备：相对权重分析 ####
 +
data.rwa <- data.frame(ynames, poq_y= poq_y$rho, pj_y = pj_y$rho, po_y = po_y$rho,
 +
                      poq_pj = poq_pj$rho, poq_po = poq_po$rho,
 +
                      pj_po = pj_po$rho)
 +
 +
varnames <- c("y", "poq", "pj", "po")
 +
lables <- list(varnames, varnames)
 +
cordat <- list()
 +
for (i in 1:nrow(data.rwa)){
 +
  cordat[[i]] <- vec2symMat(as.matrix(data.rwa[i,2:ncol(data.rwa)]),
 +
                            diag = FALSE, byrow = FALSE)
 +
  dimnames(cordat[[i]]) <- lables}
 +
 +
#### 相对权重分析函数 ####
 +
#注：函数来源参考https://relativeimportance.davidson.edu/
 +
#根据本研究做相应调整
 +
multRegress<-function(mydata){
 +
  numVar<<-NCOL(mydata)
 +
  Variables<<- names(mydata)[2:numVar]
 +
  RXX<-mydata[2:numVar,2:numVar]
 +
  RXY<-mydata[2:numVar,1]
 +
  RXX.eigen<-eigen(RXX)
 +
  D<-diag(RXX.eigen$val)
 +
  delta<-sqrt(D)
 +
  lambda<-RXX.eigen$vec%*%delta%*%t(RXX.eigen$vec)
 +
  lambdasq<-lambda^2
 +
  beta<-solve(lambda)%*%RXY
 +
  rsquare<<-sum(beta^2)
 +
  RawWgt<-lambdasq%*%beta^2
 +
  import<-(RawWgt/rsquare)*100
 +
  result<-data.frame(Variables, Raw.W=round(RawWgt,2),
 +
                      R.perc=round(import,2),
 +
                      R2 = round(rsquare,2))
 +
}
 +
 +
#### POQ fit 相对权重计算 ####
 +
rawdata <- list()
 +
thedata <- list()
 +
Labels <- list()
 +
a <- list()
 +
title <- list()
 +
R_square <- list()
 +
rwa.results <- list()
 +
 +
for (i in 1:length(cordat))
 +
{
 +
rawdata[[i]]<- data.frame(cordat[[i]])
 +
attach(rawdata[[i]])
 +
thedata[[i]] <- data.frame(y, poq, pj, po)
 +
Labels[[i]] <- names(thedata[[i]])[2:length(thedata[[i]])]
 +
a[[i]] <- multRegress(thedata[[i]]) %>% set_colnames(c("变量", "W", "%R", "R2"))
 +
title[[i]] <- data.frame("", ynames[[i]], "", "") %>% set_colnames(c("变量", "W", "%R", "R2"))
 +
R_square[[i]] <- data.frame("R2", unique(a[[i]][,4]), "", "") %>% set_colnames(c("变量", "W", "%R", "R2"))
 +
rwa.results[[i]] <- rbind(a[[i]], R_square[[i]], title[[i]],
 +
                            data.frame("变量", "W", "%R", "R2") %>%
 +
                              set_colnames(c("变量", "W", "%R", "R2")))[, 1:3]
 +
rwa.results[[i]] <- data.frame(id = c(1:nrow(rwa.results[[i]])), rwa.results[[i]])
 +
i = i + 1
 +
}
 +
 +
rwa.results
 +
 +
total<-merge(rwa.results[[1]],
 +
            rwa.results[2:length(rwa.results)], by="id")%>%t%>%unique%>%t
 +
     
 +
total.rwa <- rbind(total[5:6,], total[1:4,])
 +
write.csv(total.rwa, file = "poq_fit相对权重分析结果.csv")
 +
</pre>
  
 
==纳入元分析文献==
 
==纳入元分析文献==

2021年1月1日 (五) 20:28的版本

论文基本信息

杨伟文 李超平. 资质过剩感对个体绩效的作用效果及机制：基于情绪-认知加工系统与文化情境的元分析研究. 心理学报, 53(4)，100-120

论文数据

论文R代码

#加载安装包
library(psychmeta) 
library(tidyr)
library(magrittr)
library(dplyr)
library(tidyverse)
library(rio)
library(metaSEM)
library(semPlot)
library(plyr)
library(meta)

setwd("D:\\")
#请将data20201209.xlsx放在D盘根目录下。
#如果放在其他目录下，请在此做相应调整。
##################  全部数据准备  ######################
data <- import("data20201209.xlsx")
attach(data)
correct_main <- correct_r(correction = c("meas", "uvdrr_x", "uvdrr_y", "uvirr_x",
                                         "uvirr_y", "bvdrr", "bvirr"), 
                          rxyi = rxy, ux = 1, uy = 1, 
                          rxx = rxx,
                          ryy = ryy, ux_observed = TRUE, uy_observed = TRUE,
                          rxx_restricted = TRUE, rxx_type = "alpha", k_items_x = NA,
                          ryy_restricted = TRUE, ryy_type = "alpha", k_items_y = NA,
                          sign_rxz = 1, sign_ryz = 1, n = NULL, conf_level = 0.95,
                          correct_bias = FALSE) 
data$rtp = correct_main[["correlations"]][["rtp"]] 
write.csv(data, "data.csv")
data  <- import("data.csv")
orig.data <- subset(data.frame(data, correct_main$correlations), rtp <1) 

k.article <- length(unique(paste(orig.data$author, orig.data$title)))
k.inde.sample <- length(unique(orig.data$id))
k.eff.size <- nrow(unique(data.frame(orig.data$title, orig.data$rxy, orig.data$x, orig.data$y)))
sample.size <- unique(data.frame(orig.data$title, orig.data$n))
n.sample.size <- as.vector(sample.size$orig.data.n)%>%sum

descr <- data.frame(c("文献", "独立样本", "效应值", "样本规模"),
  c(k.article, k.inde.sample, k.eff.size, n.sample.size))%>%
  set_colnames(c("类型", "数量"))
#### 1………………直接效应分析………………####
#### 失安全系数计算 ####
d <- subset(orig.data, orig.data$x.frame.all == "资质过剩感")
attach(d)
d$zr <- 0.5*log((1+rxy)/(1-rxy))
zr <- split(d$zr, d$y.coding.t)
n <- split(d$n, d$y.coding.t)
t <- list()
for (i in 1:length(n)){
  t[[i]] <- zr[[i]]*sqrt(n[[i]]-3)
  i = i + 1}
nfs <- vector()
for (i in 1:length(zr)){
  nfs[[i]] <- abs(round(sum(t[[i]])^2/2.706 - length(zr[[i]])))
  i = i + 1}
final.nfs = data.frame(names(zr),nfs) %>% set_colnames(c("name", "nfs"))

zr.t <- split(d$zr, d$y.frame.all)
n.t <- split(d$n, d$y.frame.all)
for (i in 1:length(n.t)){
  t[[i]] <- zr.t[[i]]*sqrt(n.t[[i]]-3)
  i = i + 1}
nfs <- vector()
for (i in 1:length(zr.t)){
  nfs[[i]] <- abs(round(sum(t[[i]])^2/2.706 - length(zr.t[[i]])))
  i = i + 1}
final.nfs.t = data.frame(names(zr.t),nfs)%>% set_colnames(c("name", "nfs"))
nfs = rbind(final.nfs, final.nfs.t)

order <- data.frame(c("消极情绪", "愤怒", "无聊",
                      "积极自我概念", "一般效能感", "其他效能感", "一般自尊", "基于组织的自尊",
                      "消极行为", "反生产行为", "离职意愿", 
                      "积极行为", "创新", "工作重塑", "建言", "前瞻行为",
                      "知识共享", "组织公民行为", "任务绩效")) %>% set_colnames("order") 

t.nfs = left_join(order, nfs, by=c("order"="name"))
write.csv(t.nfs, "nfs.csv")

attach(orig.data)
#### 主要变量直接效应 ####
main_r <- ma_r(rxyi = rxy, n = n,
               sample_id = id, citekey = id,
               construct_x = orig.data$x.frame.all,
               construct_y = orig.data$y.frame.all,
               moderators = orig.data$y.coding.t,
               data = orig.data) 
main_rho <- ma_r(rxyi = rtp, n = n,
                 sample_id = id, citekey = id,
                 construct_x = orig.data$x.frame.all,
                 construct_y = orig.data$y.frame.all,
                 moderators = orig.data$y.coding.t,
                 data = orig.data)

#### 异质性检验 ####
hetero <- heterogeneity(main_rho)%>%get_heterogeneity
data.hetero = data.frame(matrix(NA, nrow = length(hetero), ncol=2))
names(data.hetero) = c("id","var") 
data.hetero$id = colnames(sapply(hetero, function(x) x[[1]]))
data.hetero$var = sapply(hetero, function(x) x[[1]]$percent_var_accounted[[1]])

#### 主效应分析表制作 ####
main_r <- main_r %>% summary %>% get_metatab
main_rho <- main_rho %>% summary %>% get_metatab
maineffects <- data.frame(main_r$analysis_id,
                          main_r$construct_x,
                          main_r$construct_y,
                          main_r$`orig.data$y.coding.t`,
                          main_r$k, main_r$N,
                          round(cbind(main_r$mean_r,  main_r$sd_r,
                                      main_rho$mean_r,main_rho$sd_r),2),
                          paste("[", round(main_rho$CI_LL_95, 2), ", ",
                                round(main_rho$CI_UL_95, 2), "]"),
                          paste("[", round(main_rho$CR_LL_80, 2), ", ",
                                round(main_rho$CR_UL_80, 2),"]"),
                          data.hetero) %>%
  set_colnames(c("id", "x", "y", "dim", "k", "n",
                 "r", "SD_r","rho", "SD_rho", 
                 "95% CI", "80% CV",
                 "var%"))
main.poq <- subset(maineffects, maineffects$x == "资质过剩感" &
                     maineffects$y != "个人工作匹配"&
                     maineffects$y != "个人组织匹配"&
                     maineffects$y != "健康"&
                     maineffects$y != "对工作的态度"&
                     maineffects$y != "对组织的态度")
main.poq$`5k+10` = main.poq$k * 5 + 10

maineffects.poq = left_join(main.poq, t.nfs, by=c("dim"="order")) %>% 
  left_join(t.nfs, by=c("y"="order")) %>% unique

mainpoq <- data.frame(maineffects.poq[, 3:12], 
                      round(unlist(maineffects.poq$`var%`),2),
                      maineffects.poq[, 14:16])%>%set_colnames(
                      c("y", "dim", "k", "n", "r",
                        "SD_r", "rho", "SD_rho", "95%CI", "80%CV",
                        "var", "5k+10", "nfs.dim", "nfs.all"))
write.csv(mainpoq, file = "poq直接效应.csv") 


attach(maineffects)
maineffects$word <- paste("(", r,", ", rho, ")","\n",
                          "(", k,", ", maineffects$n, ")","\n",
                          `95% CI`,"\n",
                          `80% CV`)


#### 2………………中介效应检验………………####
#### 数据准备：全模型数据 ####
varnames <- c("POQ","NE", "PS", "NB", "PB", "JP")
attach(maineffects)
poq.ne <- subset(maineffects, x == "资质过剩感" &
                   y == "消极情绪" &
                   dim == "All Levels")
poq.ps <- subset(maineffects, x == "资质过剩感" &
                   y == "积极自我概念" &
                   dim == "All Levels")
poq.nb <- subset(maineffects, x == "资质过剩感" &
                   y == "消极行为" &
                   dim == "All Levels")
poq.pb <- subset(maineffects, x == "资质过剩感" &
                   y == "积极行为" &
                   dim == "All Levels")
poq.jp <- subset(maineffects, x == "资质过剩感" &
                   y == "任务绩效" &
                   dim == "All Levels")

ne.ps <- subset(maineffects, (x == "消极情绪" &
                                y == "积极自我概念" &
                                dim == "All Levels")|
                  (y == "消极情绪" &
                     x == "积极自我概念" &
                     dim == "All Levels"))
ne.nb <- subset(maineffects, (x == "消极情绪" &
                                y == "消极行为" &
                                dim == "All Levels")|
                  (y == "消极情绪" &
                     x == "消极行为" &
                     dim == "All Levels"))
ne.pb <- subset(maineffects, (x == "消极情绪" &
                                y == "积极行为" &
                                dim == "All Levels")|
                  (y == "消极情绪" &
                     x == "积极行为" &
                     dim == "All Levels")) 
ne.jp <- subset(maineffects, (x == "消极情绪" &
                                y == "任务绩效" &
                                dim == "All Levels")|
                  (y == "消极情绪" &
                     x == "任务绩效" &
                     dim == "All Levels")) 

ps.nb <- subset(maineffects, (x == "积极自我概念" &
                                y == "消极行为" &
                                dim == "All Levels")|
                  (y == "积极自我概念" &
                     x == "消极行为" &
                     dim == "All Levels")) 
ps.pb <- subset(maineffects, (x == "积极自我概念" &
                                y == "积极行为" &
                                dim == "All Levels")|
                  (y == "积极自我概念" &
                     x == "积极行为" &
                     dim == "All Levels")) 
ps.jp <- subset(maineffects, (x == "积极自我概念" &
                                y == "任务绩效" &
                                dim == "All Levels")|
                  (y == "积极自我概念" &
                     x == "任务绩效" &
                     dim == "All Levels")) 

nb.pb <- subset(maineffects, (x == "消极行为" &
                                y == "积极行为" &
                                dim == "All Levels")|
                  (y == "消极行为" &
                     x == "积极行为" &
                     dim == "All Levels")) 
nb.jp <- subset(maineffects, (x == "消极行为" &
                                y == "任务绩效" &
                                dim == "All Levels")|
                  (y == "消极行为" &
                     x == "任务绩效" &
                     dim == "All Levels")) 

pb.jp <- subset(maineffects, (x == "积极行为" &
                                y == "任务绩效" &
                                dim == "All Levels")|
                  (y == "积极行为" &
                     x == "任务绩效" &
                     dim == "All Levels")) 
all.matrix <- vec2symMat(c(poq.ne$rho, poq.ps$rho, poq.nb$rho, poq.pb$rho, poq.jp$rho,
                           ne.ps$rho, ne.nb$rho, ne.pb$rho, ne.jp$rho, 
                           ps.nb$rho, ps.pb$rho, ps.jp$rho,
                           nb.pb$rho, nb.jp$rho, 
                           pb.jp$rho), 
                         diag = FALSE, byrow = FALSE) %>%
  set_colnames(varnames) %>%
  set_rownames(varnames)

all_dat <- list(all.matrix)
is.pd(all_dat)


attach(orig.data)
main_in_r <- ma_r(rxyi = rxy, n = n,
                  sample_id = id, citekey = id,
                  construct_x = orig.data$x.frame.all,
                  construct_y = orig.data$y.frame.all,
                  moderators = orig.data$indi,
                  data = orig.data)
main_in_rho <- ma_r(rxyi = rtp, n = n,
                    sample_id = id, citekey = id,
                    construct_x = orig.data$x.frame.all,
                    construct_y = orig.data$y.frame.all,
                    moderators = orig.data$indi,
                    data = orig.data) 
hetero_in <- heterogeneity(main_in_rho)%>%get_heterogeneity
data.hetero_in = data.frame(matrix(NA, nrow = length(hetero_in), ncol=2))
names(data.hetero_in) = c("id","var") 
data.hetero_in$id = colnames(sapply(hetero_in, function(x) x[[1]]))
data.hetero_in$var = sapply(hetero_in, function(x) x[[1]]$percent_var_accounted[[1]])


main_in_r <- main_in_r %>% summary %>% get_metatab %>% na.omit
main_in_rho <- main_in_rho %>% summary %>% get_metatab %>% na.omit

main_indi <- data.frame(main_in_r$analysis_id,
                        main_in_r$construct_x,
                        main_in_r$construct_y,
                        main_in_r$`orig.data$indi`,
                        main_in_r$k, main_in_r$N,
                        round(cbind(main_in_r$mean_r,  main_in_r$sd_r,
                                    main_in_rho$mean_r,main_in_rho$sd_r),2),
                        paste("[", round(main_in_rho$CI_LL_95, 2), ", ",
                              round(main_in_rho$CI_UL_95, 2), "]"),
                        paste("[", round(main_in_rho$CR_LL_80, 2), ", ",
                              round(main_in_rho$CR_UL_80, 2),"]")
) %>%
  set_colnames(c("id", "x", "y", "indi", "k", "n",
                 "r", "SD_r","rho", "SD_rho", 
                 "95% CI", "80% CV"))

#### 数据准备：低个人主义（高集体主义）数据 ####
attach(main_indi)
poq.ne.l <- subset(main_indi, x == "资质过剩感" &  y == "消极情绪" & indi == "L")

poq.ps.l <- subset(main_indi, x == "资质过剩感" &
                     y == "积极自我概念" &
                     indi == "L")
poq.nb.l <- subset(main_indi, x == "资质过剩感" &
                     y == "消极行为" &
                     indi == "L")
poq.pb.l <- subset(main_indi, x == "资质过剩感" &
                     y == "积极行为" &
                     indi == "L")
poq.jp.l <- subset(main_indi, x == "资质过剩感" &
                     y == "任务绩效" &
                     indi == "L")

ne.ps.l <- subset(main_indi, (x == "消极情绪" &
                                y == "积极自我概念" &
                                indi == "L")|
                    (y == "消极情绪" &
                       x == "积极自我概念" &
                       indi == "L"))
ne.nb.l <- subset(main_indi, (x == "消极情绪" &
                                y == "消极行为" &
                                indi == "L")|
                    (y == "消极情绪" &
                       x == "消极行为" &
                       indi == "L"))
ne.pb.l <- subset(main_indi, (x == "消极情绪" &
                                y == "积极行为" &
                                indi == "L")|
                    (y == "消极情绪" &
                       x == "积极行为" &
                       indi == "L")) 
ne.jp.l <- subset(main_indi, (x == "消极情绪" &
                                y == "任务绩效" &
                                indi == "L")|
                    (y == "消极情绪" &
                       x == "任务绩效" &
                       indi == "L")) 

ps.nb.l <- subset(main_indi, (x == "积极自我概念" &
                                y == "消极行为" &
                                indi == "L")|
                    (y == "积极自我概念" &
                       x == "消极行为" &
                       indi == "L")) 
ps.pb.l <- subset(main_indi, (x == "积极自我概念" &
                                y == "积极行为" &
                                indi == "L")|
                    (y == "积极自我概念" &
                       x == "积极行为" &
                       indi == "L")) 
ps.jp.l <- subset(main_indi, (x == "积极自我概念" &
                                y == "任务绩效" &
                                indi == "L")|
                    (y == "积极自我概念" &
                       x == "任务绩效" &
                       indi == "L")) 

nb.pb.l <- subset(main_indi, (x == "消极行为" &
                                y == "积极行为" &
                                indi == "L")|
                    (y == "消极行为" &
                       x == "积极行为" &
                       indi == "L")) 
nb.jp.l <- subset(main_indi, (x == "消极行为" &
                                y == "任务绩效" &
                                indi == "L")|
                    (y == "消极行为" &
                       x == "任务绩效" &
                       indi == "L")) 

pb.jp.l <- subset(main_indi, (x == "积极行为" &
                                y == "任务绩效" &
                                indi == "L")|
                    (y == "积极行为" &
                       x == "任务绩效" &
                       indi == "L")) 
all.matrix.l <- vec2symMat(c(poq.ne.l$rho, poq.ps.l$rho, poq.nb.l$rho, poq.pb.l$rho, poq.jp.l$rho,
                             ne.ps.l$rho, ne.nb.l$rho, ne.pb.l$rho, ne.jp.l$rho, 
                             ps.nb.l$rho, ps.pb.l$rho, ps.jp.l$rho,
                             nb.pb.l$rho, nb.jp.l$rho, 
                             pb.jp.l$rho), 
                           diag = FALSE, byrow = FALSE) %>%
  set_colnames(varnames) %>%
  set_rownames(varnames)

all_dat.l <- list(all.matrix.l)
is.pd(all_dat.l)
#### 数据准备：高个人主义（低集体主义）数据 ####
attach(main_indi)
poq.ne.h <- subset(main_indi, x == "资质过剩感" &
                     y == "消极情绪" &
                     indi == "H")
poq.ps.h <- subset(main_indi, x == "资质过剩感" &
                     y == "积极自我概念" &
                     indi == "H")
poq.nb.h <- subset(main_indi, x == "资质过剩感" &
                     y == "消极行为" &
                     indi == "H")
poq.pb.h <- subset(main_indi, x == "资质过剩感" &
                     y == "积极行为" &
                     indi == "H")
poq.jp.h <- subset(main_indi, x == "资质过剩感" &
                     y == "任务绩效" &
                     indi == "H")

ne.ps.h <- subset(main_indi, (x == "消极情绪" &
                                y == "积极自我概念" &
                                indi == "H")|
                    (y == "消极情绪" &
                       x == "积极自我概念" &
                       indi == "H"))
ne.nb.h <- subset(main_indi, (x == "消极情绪" &
                                y == "消极行为" &
                                indi == "H")|
                    (y == "消极情绪" &
                       x == "消极行为" &
                       indi == "H"))
ne.pb.h <- subset(main_indi, (x == "消极情绪" &
                                y == "积极行为" &
                                indi == "H")|
                    (y == "消极情绪" &
                       x == "积极行为" &
                       indi == "H")) 
ne.jp.h <- subset(main_indi, (x == "消极情绪" &
                                y == "任务绩效" &
                                indi == "H")|
                    (y == "消极情绪" &
                       x == "任务绩效" &
                       indi == "H")) 

ps.nb.h <- subset(main_indi, (x == "积极自我概念" &
                                y == "消极行为" &
                                indi == "H")|
                    (y == "积极自我概念" &
                       x == "消极行为" &
                       indi == "H")) 
ps.pb.h <- subset(main_indi, (x == "积极自我概念" &
                                y == "积极行为" &
                                indi == "H")|
                    (y == "积极自我概念" &
                       x == "积极行为" &
                       indi == "H")) 
ps.jp.h <- subset(main_indi, (x == "积极自我概念" &
                                y == "任务绩效" &
                                indi == "H")|
                    (y == "积极自我概念" &
                       x == "任务绩效" &
                       indi == "H")) 

nb.pb.h <- subset(main_indi, (x == "消极行为" &
                                y == "积极行为" &
                                indi == "H")|
                    (y == "消极行为" &
                       x == "积极行为" &
                       indi == "H")) 
nb.jp.h <- subset(main_indi, (x == "消极行为" &
                                y == "任务绩效" &
                                indi == "H")|
                    (y == "消极行为" &
                       x == "任务绩效" &
                       indi == "H")) 

pb.jp.h <- subset(main_indi, (x == "积极行为" &
                                y == "任务绩效" &
                                indi == "H")|
                    (y == "积极行为" &
                       x == "任务绩效" &
                       indi == "H")) 
all.matrix.h <- vec2symMat(c(poq.ne.h$rho, poq.ps.h$rho, poq.nb.h$rho, poq.pb.h$rho, poq.jp.h$rho,
                             ne.ps.h$rho, ne.nb.h$rho, ne.pb.h$rho, ne.jp.h$rho, 
                             ps.nb.h$rho, ps.pb.h$rho, ps.jp.h$rho,
                             nb.pb.h$rho, nb.jp.h$rho, 
                             pb.jp.h$rho), 
                           diag = FALSE, byrow = FALSE) %>%
  set_colnames(varnames) %>%
  set_rownames(varnames)

all_dat.h <- list(all.matrix.h)
is.pd(all_dat.h)
#### 调和平均数 ####
library(psych)
all_n <- c(poq.ne$n, poq.ps$n, poq.nb$n, poq.pb$n, poq.jp$n,
           ne.ps$n, ne.nb$n, ne.pb$n, ne.jp$n, 
           ps.nb$n, ps.pb$n, ps.jp$n,
           nb.pb$n, nb.jp$n, 
           pb.jp$n) %>% harmonic.mean %>% round
n_l <- c(poq.ne.l$n, poq.ps.l$n, poq.nb.l$n, poq.pb.l$n, poq.jp.l$n,
         ne.ps.l$n, ne.nb.l$n, ne.pb.l$n, ne.jp.l$n, 
         ps.nb.l$n, ps.pb.l$n, ps.jp.l$n,
         nb.pb.l$n, nb.jp.l$n, 
         pb.jp.l$n) %>% harmonic.mean %>% round
n_h <- c(poq.ne.h$n, poq.ps.h$n, poq.nb.h$n, poq.pb.h$n, poq.jp.h$n,
         ne.ps.h$n, ne.nb.h$n, ne.pb.h$n, ne.jp.h$n, 
         ps.nb.h$n, ps.pb.h$n, ps.jp.h$n,
         nb.pb.h$n, nb.jp.h$n, 
         pb.jp.h$n) %>% harmonic.mean %>% round

####模型设置 ####

model_all <- 'NB ~  POQ + PS + NE
			  PB ~  POQ + PS + NE
			  JP ~  POQ + PS + NE

			  NE ~ POQ
			  PS ~ POQ

			  NE ~~ 1*NE
			  PS ~~ 1*PS
			  POQ ~~ 1*POQ

			  NE ~~ 0.1*PS

			  NB ~~ 0.1*NB
			  PB ~~ 0.1*PB
			  JP ~~ 0.1*JP

			  NB ~~ 0.1*PB
			  NB ~~ 0.1*JP
			  JP ~~ 0.1*PB'

plot(model_all, col="yellow")

RAM_all <- lavaan2RAM(model_all, obs.variables = varnames)
RAM_all


#### SEM全部数据及调和平均数 ####
fin_dat <- list(all_dat, all_dat.l, all_dat.h)
fin_n <- c(all_n, n_l, n_h)


stage1 <- list()
stage2 <- list()
stage2_LB <- list()
stage2_Z <- list()
ind_LB <- list()
ind_Z <- list()
fitness <- list()
path <- list()
ind <- list()
sem <- list()
name <- c("all", "low individualism", "high individualism")
for (i in 1:length(fin_dat)){
  #### stage1 ####
  stage1[[i]] <- tssem1(Cov = fin_dat[[i]], n = fin_n[i], method = "FEM", 
                        RE.type = "Diag", RE.startvalues=0.1,
                        acov = "weighted")
  
  summary(stage1[[i]])
  #### stage2 ####
  stage2[[i]] <- tssem2(stage1[[i]], Amatrix = RAM_all$A, Smatrix = RAM_all$S)
  summary(stage2[[i]])
  
  stage2_LB[[i]] <- tssem2(stage1[[i]], Amatrix = RAM_all$A, Smatrix = RAM_all$S,
                           diag.constraints = TRUE, intervals.type = "LB",
                           mx.algebras = list(
                             #POQ-NE-Y                         
                             IndNE_NB = mxAlgebra(NEONPOQ*NBONNE, name = "IndNE_NB"),
                             IndNE_PB = mxAlgebra(NEONPOQ*PBONNE, name = "IndNE_PB"),
                             IndNE_JP = mxAlgebra(NEONPOQ*JPONNE, name = "IndNE_JP"),
                             #POQ-PS-Y                          
                             IndPS_NB = mxAlgebra(PSONPOQ*NBONPS, name = "IndPS_NB"),
                             IndPS_PB = mxAlgebra(PSONPOQ*PBONPS, name = "IndPS_PB"),
                             IndPS_JP = mxAlgebra(PSONPOQ*JPONPS, name = "IndPS_JP"),
                             #POQ-NE+PS-Y                          
                             Indall_NB = mxAlgebra(NEONPOQ*NBONNE + PSONPOQ*NBONPS, name = "Indall_NB"),
                             Indall_PB = mxAlgebra(NEONPOQ*PBONNE + PSONPOQ*PBONPS, name = "Indall_PB"),
                             Indall_JP = mxAlgebra(NEONPOQ*JPONNE + PSONPOQ*JPONPS, name = "Indall_JP"),
                             #POQ-NE+PS-Y  PLUS POQ-Y                          
                             Total_NB = mxAlgebra(NEONPOQ*NBONNE + PSONPOQ*NBONPS + NBONPOQ, name = "Total_NB"),
                             Total_PB = mxAlgebra(NEONPOQ*PBONNE + PSONPOQ*PBONPS + PBONPOQ, name = "Total_PB"),
                             Total_JP = mxAlgebra(NEONPOQ*JPONNE + PSONPOQ*JPONPS + JPONPOQ, name = "Total_JP"),
                             #Diff POQ-(NE-PS)-Y                          
                             DIF_NEminusPS_NB= mxAlgebra(NEONPOQ*NBONNE - PSONPOQ*NBONPS, name = "DIF_NEminusPS_NB"),
                             DIF_NEminusPS_PB= mxAlgebra(NEONPOQ*PBONNE - PSONPOQ*PBONPS, name = "DIF_NEminusPS_PB"),
                             DIF_NEminusPS_JP= mxAlgebra(NEONPOQ*JPONNE - PSONPOQ*JPONPS, name = "DIF_NEminusPS_JP")))
  stage2_LB[[i]] <- rerun(stage2_LB[[i]])
  ind_LB[[i]] <- summary(stage2_LB[[i]])
  
  stage2_Z[[i]] <- tssem2(stage1[[i]], Amatrix = RAM_all$A, Smatrix = RAM_all$S,
                          diag.constraints = FALSE, intervals.type = "z",
                          mx.algebras = list(
                            IndNE_NB = mxAlgebra(NEONPOQ*NBONNE, name = "IndNE_NB"),
                            IndNE_PB = mxAlgebra(NEONPOQ*PBONNE, name = "IndNE_PB"),
                            IndNE_JP = mxAlgebra(NEONPOQ*JPONNE, name = "IndNE_JP"),
                            
                            IndPS_NB = mxAlgebra(PSONPOQ*NBONPS, name = "IndPS_NB"),
                            IndPS_PB = mxAlgebra(PSONPOQ*PBONPS, name = "IndPS_PB"),
                            IndPS_JP = mxAlgebra(PSONPOQ*JPONPS, name = "IndPS_JP"),
                            
                            Indall_NB = mxAlgebra(NEONPOQ*NBONNE + PSONPOQ*NBONPS, name = "Indall_NB"),
                            Indall_PB = mxAlgebra(NEONPOQ*PBONNE + PSONPOQ*PBONPS, name = "Indall_PB"),
                            Indall_JP = mxAlgebra(NEONPOQ*JPONNE + PSONPOQ*JPONPS, name = "Indall_JP"),
                            
                            Total_NB = mxAlgebra(NEONPOQ*NBONNE + PSONPOQ*NBONPS + NBONPOQ, name = "Total_NB"),
                            Total_PB = mxAlgebra(NEONPOQ*PBONNE + PSONPOQ*PBONPS + PBONPOQ, name = "Total_PB"),
                            Total_JP = mxAlgebra(NEONPOQ*JPONNE + PSONPOQ*JPONPS + JPONPOQ, name = "Total_JP"),
                            
                            DIF_NEminusPS_NB= mxAlgebra(NEONPOQ*NBONNE - PSONPOQ*NBONPS, name = "DIF_NEminusPS_NB"),
                            DIF_NEminusPS_PB= mxAlgebra(NEONPOQ*PBONNE - PSONPOQ*PBONPS, name = "DIF_NEminusPS_PB"),
                            DIF_NEminusPS_JP= mxAlgebra(NEONPOQ*JPONNE - PSONPOQ*JPONPS, name = "DIF_NEminusPS_JP")))
  stage2_Z[[i]] <- rerun(stage2_Z[[i]])
  ind_Z[[i]] <- summary(stage2_Z[[i]])
  #### 拟合系数 ####
  fitness[[i]] <- ind_LB[[i]][["stat"]]%>%as.data.frame%>%round(2)%>%
    set_colnames("Goodness-of-fit indices")
  #### 路径系数 ####
  path[[i]] <- data.frame(round(ind_Z[[i]][["coefficients"]][,1], 2),
                          paste("[", round(ind_Z[[i]][["coefficients"]][,3], 2),",",
                                round(ind_Z[[i]][["coefficients"]][,4], 2), "]"),
                          round(ind_Z[[i]][["coefficients"]][,2], 2),
                          round(ind_Z[[i]][["coefficients"]][,5], 2),
                          round(ind_Z[[i]][["coefficients"]][,6], 2)) %>% 
    set_colnames(c("β", "95% CI", "SE", "z", "p")) %>% 
    set_rownames(c("JPONNE", "JPONPOQ", "JPONPS",
                   "NBONNE", "NBONPOQ", "NBONPS",
                   "NEONPOQ", "PBONNE", "PBONPOQ",
                   "PBONPS", "PSONPOQ"))
  #### 中介效应 ####
  ind[[i]] <- data.frame(round(ind_LB[[i]][["mx.algebras"]][,2],2),
                         paste("[", round(ind_LB[[i]][["mx.algebras"]][,1],2), ",",
                               round(ind_LB[[i]][["mx.algebras"]][,3],2), "]")) %>%
    set_colnames(c("β", "95%CI"))%>%
    set_rownames(c("IndNE_NB", "IndNE_PB", "IndNE_JP", 
                   "IndPS_NB", "IndPS_PB", "IndPS_JP", 
                   "Indall_NB", "Indall_PB", "Indall_JP", 
                   "Total_NB", "Total_PB", "Total_JP", 
                   "DIF_NEminusPS_NB", "DIF_NEminusPS_PB", "DIF_NEminusPS_JP"))
  sem[[i]] <- list(name = name[i], 
                   fit=fitness[[i]], path = path[[i]], ind = ind[[i]])
  sem[[i]]
  
  i = i + 1}
sem
#### 结构方程分析结果表格 ####
final_fit <- cbind(sem[[1]][["fit"]][c(1,7,4),],
                   sem[[2]][["fit"]][c(1,7,4),],
                   sem[[3]][["fit"]][c(1,7,4),])%>% 
  set_rownames(c("n", "Chi", "p"))%>%
  set_colnames(c("全样本", "高集体主义", "低集体主义"))
final_path <- rbind(c("全样本", " "," ", " ", " ",
                      "高集体主义", " "," ", " ", " ",
                      "低集体主义", " "," ", " ", " "),
                    rep(c("β", "95%CI", "SE", "z", "p"), 3),
                    cbind(rbind(sem[[1]][["path"]]),
                          rbind(sem[[2]][["path"]]),
                          rbind(sem[[3]][["path"]])))%>%
  set_colnames(c(1:15))
final_ind <- rbind(rep(c("β", "95%CI"),3),
                   cbind(sem[[1]][["ind"]],
                         sem[[2]][["ind"]],
                         sem[[3]][["ind"]]))%>%
  set_colnames(c(1:6))
write.csv(final_ind, "ind.csv")
#### 相关系数文本 ####
sem.matrix <- cbind(c(poq.ne$word, poq.ps$word, poq.nb$word, poq.pb$word, poq.jp$word),
                    c(" ", ne.ps$word, ne.nb$word, ne.pb$word, ne.jp$word),
                    c(" ", " ", ps.nb$word, ps.pb$word,ps.jp$word),
                    c(" ", " ", " ", nb.pb$word, nb.jp$word),
                    c(" ", " "," ", " ", pb.jp$word))%>%
  set_colnames(c("资质过剩感", "消极情绪", "积极自我概念", "消极角色外行为", "积极角色外行为"))%>%
  set_rownames(c("消极情绪", "积极自我概念", "消极角色外行为", "积极角色外行为", "任务绩效"))
sem.matrix 

#### 3………………概念冗余度检验………………####
#### 相关矩阵构建 ####
attach(maineffects)
matrix <- data.frame(x, y, dim, 
                     paste( "(", k, ",", maineffects$n, ")" ) ,
                     paste( "(", r, ",", rho, ")" ) ,
                     `95% CI`, `80% CV`, rho)%>%
  set_colnames(c("x", "dim", "y", "(k, n)", "(r, ρ)", "95% CI", "80% CV", "rho"))

matrix$all = paste(matrix$`(k, n)`, "\n", matrix$`(r, ρ)`, "\n",
                   matrix$`95% CI`, "\n", matrix$`80% CV`)


attach(matrix)
poq_pj <- subset(matrix, x == "资质过剩感" & y == "个人工作匹配")
poq_po <- subset(matrix, x == "资质过剩感" & y == "个人组织匹配")
poq_y <- subset(matrix, (y == "压力（反向）"|
                      y == "组织承诺"|
                      y == "工作满意度"|
                      y == "创新"|
                      y == "组织公民行为"|
                      y == "反生产行为"|
                      y == "离职意愿"|
                      y == "任务绩效")&
                      x == "资质过剩感")
pj_po <- subset(matrix, x == "个人工作匹配" & 
                       dim == "个人组织匹配")
pj_y<- subset(matrix, (x == "个人工作匹配"&
                       (y == "压力（反向）"|
                        y == "组织承诺"|
                        y == "工作满意度"))|
                      (dim == "个人工作匹配"&
                       (y == "创新"|
                        y == "组织公民行为"|
                        y == "反生产行为"|
                        y == "离职意愿"|
                        y == "任务绩效")))
po_y <- subset(matrix, (x == "个人组织匹配"&
                          (y == "压力（反向）"|
                           y == "组织承诺"|
                           y == "工作满意度"))|
                       (dim == "个人组织匹配"&
                          (y == "创新"|
                           y == "组织公民行为"|
                           y == "反生产行为"|
                           y == "离职意愿"|
                           y == "任务绩效")))

#### 相关矩阵表格 ####
ynames <- c("创新", "组织公民行为", "反生产行为", "离职意愿",
            "任务绩效", "压力（反向）", "工作满意度", "组织承诺")
rwa.text <- data.frame(ynames, poq_y= poq_y$all, pj_y = pj_y$all, po_y = po_y$all, 
                       poq_pj = poq_pj$all, poq_po = poq_po$all,
                       pj_po = pj_po$all)

varnames <- list()
lables <- list()
text.mx <- list()
for (i in 1:nrow(rwa.text)){
  varnames[[i]] <- c(ynames[[i]], "资质过剩感", "个人工作匹配", "个人组织匹配")
  lables[[i]] <- list (varnames[[i]], varnames[[i]])
  text.mx[[i]] <- vec2symMat(as.matrix(rwa.text[i,2:ncol(rwa.text)]), 
                             diag = FALSE, byrow = FALSE) 
  dimnames(text.mx[[i]]) <- lables[[i]]
  i = i+1}
as.vector(poq_y$y)
poqfit.text.all <- t(data.frame(text.mx[[1]][2:4,2:4], text.mx[[1]][2:4,1],
                                text.mx[[2]][2:4,1], text.mx[[3]][2:4,1],
                                text.mx[[4]][2:4,1], text.mx[[5]][2:4,1],
                                text.mx[[6]][2:4,1], text.mx[[7]][2:4,1],
                                text.mx[[8]][2:4,1])) %>% 
                    set_rownames(c("资质过剩感", "个人工作匹配", "个人组织匹配",
                                   "创新", "组织公民行为", "反生产行为", "离职意愿",
                                "任务绩效", "压力（反向）", "工作满意度", "组织承诺"))
write.csv(poqfit.text.all, file = "poq_fit相对矩阵文本.csv")


#### 数据准备：相对权重分析 ####
data.rwa <- data.frame(ynames, poq_y= poq_y$rho, pj_y = pj_y$rho, po_y = po_y$rho, 
                       poq_pj = poq_pj$rho, poq_po = poq_po$rho,
                       pj_po = pj_po$rho)

varnames <- c("y", "poq", "pj", "po")
lables <- list(varnames, varnames)
cordat <- list()
for (i in 1:nrow(data.rwa)){
  cordat[[i]] <- vec2symMat(as.matrix(data.rwa[i,2:ncol(data.rwa)]), 
                            diag = FALSE, byrow = FALSE) 
  dimnames(cordat[[i]]) <- lables}

#### 相对权重分析函数 ####
#注：函数来源参考https://relativeimportance.davidson.edu/
#根据本研究做相应调整
multRegress<-function(mydata){
  numVar<<-NCOL(mydata)
  Variables<<- names(mydata)[2:numVar]
  RXX<-mydata[2:numVar,2:numVar]
  RXY<-mydata[2:numVar,1]
  RXX.eigen<-eigen(RXX)
  D<-diag(RXX.eigen$val)
  delta<-sqrt(D)
  lambda<-RXX.eigen$vec%*%delta%*%t(RXX.eigen$vec)
  lambdasq<-lambda^2
  beta<-solve(lambda)%*%RXY
  rsquare<<-sum(beta^2)
  RawWgt<-lambdasq%*%beta^2
  import<-(RawWgt/rsquare)*100
  result<-data.frame(Variables, Raw.W=round(RawWgt,2), 
                      R.perc=round(import,2), 
                      R2 = round(rsquare,2))
}

#### POQ fit 相对权重计算 ####
rawdata <- list()
thedata <- list()
Labels <- list()
a <- list()
title <- list()
R_square <- list()
rwa.results <- list()

for (i in 1:length(cordat))
	{
		rawdata[[i]]<- data.frame(cordat[[i]]) 
		attach(rawdata[[i]])
		thedata[[i]] <- data.frame(y, poq, pj, po)
		Labels[[i]] <- names(thedata[[i]])[2:length(thedata[[i]])]
		a[[i]] <- multRegress(thedata[[i]]) %>% set_colnames(c("变量", "W", "%R", "R2"))
		title[[i]] <- data.frame("", ynames[[i]], "", "") %>% set_colnames(c("变量", "W", "%R", "R2"))
		R_square[[i]] <- data.frame("R2", unique(a[[i]][,4]), "", "") %>% set_colnames(c("变量", "W", "%R", "R2"))
		rwa.results[[i]] <- rbind(a[[i]], R_square[[i]], title[[i]], 
                            data.frame("变量", "W", "%R", "R2") %>% 
                              set_colnames(c("变量", "W", "%R", "R2")))[, 1:3]
		rwa.results[[i]] <- data.frame(id = c(1:nrow(rwa.results[[i]])), rwa.results[[i]])
		i = i + 1
	}

rwa.results

total<-merge(rwa.results[[1]],
             rwa.results[2:length(rwa.results)], by="id")%>%t%>%unique%>%t
       
total.rwa <- rbind(total[5:6,], total[1:4,])
write.csv(total.rwa, file = "poq_fit相对权重分析结果.csv")

纳入元分析文献

取自“http://www.obhrm.net/index.php?title=2021MetaPOQ&oldid=8115”