Meta-analysis


Note: Input values must be separated by tabs. Copy and paste from Excel/Numbers.

Your data needs to have exactly the same header (variable names) in the first row.

Option:


                
                

Effect size and sampling variance


                

Fixed effects model


                

Random effects model


                

[Criteria for checking heterogeneity]

I^2 (How much effect sizes across studies differ)
25–50: Little different
50–75: Quite different
75–100: Considerably different

Test for Heterogeneity: p-val < .05 (not homogeneous)

H (sqrt(H^2)) > 1: There is unexplained heterogeneity.



Forest plot (Fixed effects model)


Forest plot (Random effects model)


Funnel plot (Fixed effects model)

Open circles (if any) on the right side show missing NULL studies estimated with the trim-and-fill method, added in the funnel plot.


Funnel plot (Random effects model)

Open circles (if any) on the right side show missing NULL studies estimated with the trim-and-fill method, added in the funnel plot.




                

Fail-safe N is the number of nonsignificant studies necessary to make the result nonsignificant. "When the fail-safe N is high, that is interpreted to mean that even a large number of nonsignificant studies may not influence the statistical significance of meta-analytic results too greatly. Although ... it is not a very precise measure of publication bias" (Oswald & Plonsky, 2010, p. 92) .



Moderator (subgroup) analysis


                

Categorical moderator graph (Fixed effects model)


Categorical moderator graph (Random effects model)



R session info

              

Note: Input values must be separated by tabs. Copy and paste from Excel/Numbers.

Your data needs to have exactly the same header (variable names) in the first row.


Mean Differences (n, M, SD)


                
                

Mean Differences (n, Effect size d)


                
                

Correlations (n, r)


                
                
Note

This web application is developed with Shiny.


List of Packages Used
library(shiny)
library(shinyAce)
library(metafor)
library(meta)
library(MAd)
library(MAc)
library(quantreg)
library(ggplot2)

Code

Source code for this application is based on "The handbook of Research in Foreign Language Learning and Teaching" (Takeuchi & Mizumoto, 2012).

The code for this web application is available at GitHub.

If you want to run this code on your computer (in a local R session), run the code below:
library(shiny)
runGitHub("meta","mizumot")


Acknowledgment

I thank Dr. Luke Plonsky and Dr. Yo In'nami for their support and feedback to create this web application.


Citation in Publications

Mizumoto, A. (2015). Langtest (Version 1.0) [Web application]. Retrieved from http://langtest.jp


Article

Mizumoto, A., & Plonsky, L. (2015). R as a lingua franca: Advantages of using R for quantitative research in applied linguistics. Applied Linguistics, Advance online publication. doi:10.1093/applin/amv025


Recommended

To learn more about R, I suggest this excellent and free e-book (pdf), A Guide to Doing Statistics in Second Language Research Using R, written by Dr. Jenifer Larson-Hall.

Also, if you are a cool Mac user and want to use R with GUI, MacR is defenitely the way to go!


Author

Atsushi MIZUMOTO, Ph.D.
Professor of Applied Linguistics
Faculty of Foreign Language Studies /
Graduate School of Foreign Language Education and Research,
Kansai University, Osaka, Japan



Code for "Meta-analysis"
by Atsushi Mizumoto

show with app
library(shiny)
library(shinyAce)
library(meta)
library(metafor)
library(MAd)
library(MAc)
library(quantreg)
library(ggplot2)


shinyServer(function(input, output) {
    
    options(warn=-1)
    
    
# First calculation to be used later
    W.data <- reactive({
        
        dat <- read.csv(text=input$text, sep="\t")


        if (input$type == "mdms") {
            
            dat <- escalc(measure="SMD", n1i=N1, n2i=N2,
            m1i=M1, m2i=M2,
            sd1i=SD1, sd2i=SD2,
            data=dat, append=TRUE)
            
            dat$ES <- dat$yi
            dat$yi <- NULL
            dat$SV <- dat$vi # SV=sampling variances
            dat$vi <- NULL
            
            list(dat = dat) # To be used later
        }
        
        
        else if (input$type == "mdes") {
            
            df <- (dat$N1 + dat$N2) - 2
            j <- 1 - (3/(4 * df - 1))
            g <- j * dat$d
            dat$ES <- g
            
            dat$SV <- (((dat$N1+dat$N2)/(dat$N1*dat$N2))+((dat$ES*dat$ES)/(2*(dat$N1+dat$N2))))
            
            list(dat = dat) # To be used later
        }
        
        
        else if (input$type == "cor") {
            
            dat <- escalc(measure="ZCOR", ni=N, ri=r, data=dat, append=TRUE)
            dat$FZ <- dat$yi
            dat$yi <- NULL
            dat$SV <- dat$vi # SV=sampling variances
            dat$vi <- NULL

            list(dat = dat) # To be used later
            
        }
    })
    
    
    


# Fixed effects model to be used later(精度を高めるため小数点以下丸めない)
    FE.est <- reactive({
        
        if (input$type == "mdms") {
            
            dat <- read.csv(text=input$text, sep="\t")
            
            dat <- escalc(measure="SMD", n1i=N1, n2i=N2,
            m1i=M1, m2i=M2,
            sd1i=SD1, sd2i=SD2,
            data=dat, append=TRUE)
            
            dat$ES <- dat$yi # 小数点以下丸めるのを外している
            dat$yi <- NULL
            dat$SV <- dat$vi # 小数点以下丸めるのを外している
            dat$vi <- NULL

            FE.res <- rma(ES, SV, method="FE", data=dat, slab=paste(Study))
            
            list(FE.res = FE.res) # To be used later
        }
        
        else if (input$type == "mdes") {
            
            dat <- read.csv(text=input$text, sep="\t")

            df <- (dat$N1 + dat$N2) - 2
            j <- 1 - (3/(4 * df - 1))
            g <- j * dat$d
            dat$ES <- g  # 小数点以下丸めるのを外している
            
            dat$SV <- (((dat$N1+dat$N2)/(dat$N1*dat$N2))+((dat$ES*dat$ES)/(2*(dat$N1+dat$N2)))) # 小数点以下丸めるのを外している
            
            FE.res <- rma(ES, SV, method="FE", data=dat, slab=paste(Study))
            
            list(FE.res = FE.res) # To be used later
        }
        
        
        else if (input$type == "cor") {
            
            dat <- read.csv(text=input$text, sep="\t")
            
            dat <- escalc(measure="ZCOR", ni=N, ri=r, data=dat, append=TRUE)
            dat$FZ <- dat$yi
            dat$yi <- NULL
            dat$SV <- dat$vi # SV=sampling variances
            dat$vi <- NULL

            FE.res <- rma(FZ, SV, data=dat, method = "FE", slab=paste(Study))
            
            list(FE.res = FE.res) # To be used later
        }
    })
    
    
    
    
    
# Random effects model to be used later
    RE.est  <- reactive({

        if (input$type == "mdms") {
            
            dat <- read.csv(text=input$text, sep="\t")
            
            dat <- escalc(measure="SMD", n1i=N1, n2i=N2,
            m1i=M1, m2i=M2,
            sd1i=SD1, sd2i=SD2,
            data=dat, append=TRUE)
            
            dat$ES <- dat$yi # 小数点以下丸めるのを外している
            dat$yi <- NULL
            dat$SV <- dat$vi # 小数点以下丸めるのを外している
            dat$vi <- NULL

            RE.res <- rma(ES, SV, method="REML", data=dat, slab=paste(Study))
            
            list(RE.res = RE.res) # To be used later
        }
        
        
        else if (input$type == "mdes") {
            
            dat <- read.csv(text=input$text, sep="\t")
            
            df <- (dat$N1 + dat$N2) - 2
            j <- 1 - (3/(4 * df - 1))
            g <- j * dat$d
            dat$ES <- g  # 小数点以下丸めるのを外している
            
            dat$SV <- (((dat$N1+dat$N2)/(dat$N1*dat$N2))+((dat$ES*dat$ES)/(2*(dat$N1+dat$N2)))) # 小数点以下丸めるのを外している

            RE.res <- rma(ES, SV, method="REML", data=dat, slab=paste(Study))
            
            list(RE.res = RE.res) # To be used later
        }
        
        
        else if (input$type == "cor") {
            
            dat <- read.csv(text=input$text, sep="\t")
            
            dat <- escalc(measure="ZCOR", ni=N, ri=r, data=dat, append=TRUE)
            dat$FZ <- dat$yi
            dat$yi <- NULL
            dat$SV <- dat$vi # SV=sampling variances
            dat$vi <- NULL

            RE.res <- rma(FZ, SV, data=dat, method = "REML", slab=paste(Study))

            list(RE.res = RE.res) # To be used later
            
        }
    })
    
    
    
    

# -- Displaying the first calculation --

    data <- reactive({
    
        dat <- read.csv(text=input$text, sep="\t")
    
    
        if (input$type == "mdms") {
        
            dat <- escalc(measure="SMD", n1i=N1, n2i=N2,
                            m1i=M1, m2i=M2,
                            sd1i=SD1, sd2i=SD2,
                            data=dat, append=TRUE)
            
            dat$ES <- round(dat$yi, 3)
            dat$yi <- NULL
            dat$SV <- round(dat$vi, 3) # SV=sampling variances
            dat$vi <- NULL

            cat("\n","ES = Effect size [Hedges's g]", "\n",
                "SV = Sampling variance [sqrt(SV) = Std err]", "\n", "\n"
                ) # ," W = Inverse variance weight", "\n", "\n"
                cat("---","\n")
        
            print(dat)
        }
    
    
        else if (input$type == "mdes") {
        
            df <- (dat$N1 + dat$N2) - 2
            j <- 1 - (3/(4 * df - 1))
            g <- j * dat$d
            dat$ES <- round(g, 3)
            
            dat$SV <- round((((dat$N1+dat$N2)/(dat$N1*dat$N2))+((dat$ES*dat$ES)/(2*(dat$N1+dat$N2)))),3)
        
            cat("\n","ES = Effect size [Hedges's g]", "\n",
                "SV = Sampling variance [sqrt(SV) = Std err]", "\n", "\n"
                ) # , " W = Inverse variance weight", "\n", "\n"
                cat("---","\n")
        
            print(dat)
        }
    
    
        else if (input$type == "cor") {
        
            dat <- escalc(measure="ZCOR", ni=N, ri=r, data=dat, append=TRUE)
            dat$FZ <- round(dat$yi,3)
            dat$yi <- NULL
            dat$SV <- round(dat$vi, 3) # SV=sampling variances
            dat$vi <- NULL

            cat("\n","FZ = Fisher's Z", "\n",
                "SV = Sampling variance [sqrt(SV) = Std err]", "\n", "\n")
                cat("---","\n")
        
            print(dat)
            
        }
    })





################################################
# -- FE & RE model result
################################################

    fe <- reactive({
        
        if (input$type == "mdms") {
            
            FE.res <- FE.est()$FE.res

            cat("The FE model is a description of the K studies (Kovalchik, 2013).","\n")
            cat("---","\n")
            
            FE.res
        }
        
        
        else if (input$type == "mdes") {
            
            FE.res <- FE.est()$FE.res
            
            cat("The FE model is a description of the K studies (Kovalchik, 2013).","\n",
            "---","\n")
            
            FE.res
            
        }
        
        
        else if (input$type == "cor") { # Using different function here.
            
            dat <- read.csv(text=input$text, sep="\t")

            FE.res <- metacor(dat$r, dat$N)
            
            FE.res
        }
    })
    
    
    
    
    
    re <- reactive({
        
        if (input$type == "mdms") {
            
            RE.res <- RE.est()$RE.res
            
            cat("The RE model regards the K studies as a sample of","\n")
            cat(" a larger universe of studies (Kovalchik, 2013).","\n")
            cat("---","\n")
            
            RE.res
        }
        
        
        else if (input$type == "mdes") {
            
            RE.res <- RE.est()$RE.res

            cat("The RE model regards the K studies as a sample of","\n")
            cat(" a larger universe of studies (Kovalchik, 2013).","\n")
            cat("---","\n")
            
            RE.res
            
        }
        
        
        else if (input$type == "cor") {
            
            cat("Both FE and RE model results are reported above.","\n","\n")
            
            cat("---","\n")
            
            cat("The FE model is a description of the K studies.","\n")
            cat("The RE model regards the K studies as a sample of","\n")
            cat(" a larger universe of studies (Kovalchik, 2013).","\n")
            
        }
    })
    
    
    


################################################
# Forest plot
################################################
    
    makefePlot <- function(){
        
        if (input$type == "mdms") {
            
            FE.res <- FE.est()$FE.res
            
            forest(FE.res)
        }
        
        
        else if (input$type == "mdes") {
            
            FE.res <- FE.est()$FE.res

            forest(FE.res)
        }
        
        
        else if (input$type == "cor") {
            
            FE.res <- FE.est()$FE.res

            forest(FE.res, transf=transf.ztor)
            
        }
    }
    
    
    output$fePlot <- renderPlot({
        print(makefePlot())
    })
    
    
    
    
    
    makerePlot <- function(){
        
        if (input$type == "mdms") {
            
            RE.res <- RE.est()$RE.res

            forest(RE.res)
            
        }
        
        
        else if (input$type == "mdes") {
            
            RE.res <- RE.est()$RE.res
            
            forest(RE.res)
            
        }
        
        
        else if (input$type == "cor") {
            
            RE.res <- RE.est()$RE.res

            forest(RE.res, transf=transf.ztor)
            
        }
    }
    
    
    output$rePlot <- renderPlot({
        print(makerePlot())
    })
    
    
    
    
    
################################################
# Funnel plot
################################################

    makeFunFixPlot <- function(){
        
        if (input$type == "mdms") {
            
            FE.res <- FE.est()$FE.res
            
            metafor::funnel(trimfill(FE.res))
            
        }
        
        
        else if (input$type == "mdes") {
            
            FE.res <- FE.est()$FE.res
            
            metafor::funnel(trimfill(FE.res))
            
        }
        
        
        else if (input$type == "cor") {
            
            FE.res <- FE.est()$FE.res
            
            metafor::funnel(trimfill(FE.res))
            
        }
    }
    
    
    output$FunFixPlot <- renderPlot({
        print(makeFunFixPlot())
    })
    
    
    
    
    
    makeFunRandPlot <- function(){
        
        if (input$type == "mdms") {
            
            RE.res <- RE.est()$RE.res
            
            metafor::funnel(trimfill(RE.res))
            
        }
        
        
        else if (input$type == "mdes") {
            
            RE.res <- RE.est()$RE.res

            metafor::funnel(trimfill(RE.res))
            
        }
        
        
        else if (input$type == "cor") {
            
            RE.res <- RE.est()$RE.res

            metafor::funnel(trimfill(RE.res))
            
        }
    }
    
    
    output$FunRandPlot <- renderPlot({
        print(makeFunRandPlot())
    })
    
    
    


################################################
# Test of asymmetry & Fail-safe N
################################################

    asy <- reactive({
        
        dat <- read.csv(text=input$text, sep="\t")
        
        
        if (input$type == "mdms") {
            
            RE.res <- RE.est()$RE.res

            regt <- regtest(RE.res, model="lm")
            value <- fsn(y = RE.res$yi, v = RE.res$vi)
            
            return(list('No publication bias if p > .05 (Nonsignificant)' = regt,
            'File drawer analysis' = value))
        }
        
        
        else if (input$type == "mdes") {
            
            RE.res <- RE.est()$RE.res
            
            regt <- regtest(RE.res, model="lm")
            value <- fsn(y = RE.res$yi, v = RE.res$vi)
            
            return(list('No publication bias if p > .05 (Nonsignificant)' = regt,
            'File drawer analysis' = value))
            
        }
        
        
        else if (input$type == "cor") {
            
            RE.res <- RE.est()$RE.res
            
            regt <- regtest(RE.res, model="lm")
            value <- fsn(y = RE.res$yi, v = RE.res$vi)
            
            return(list('No publication bias if p > .05 (Nonsignificant)' = regt,
            'File drawer analysis' = value))
        }
    })
    
    
    
    

################################################
# Moderator analysis
################################################

    modAnalysis <- reactive({
        
        #if (input$moderator == 1) {
            
            
            if (input$type == "mdms") {
                
                dat <- read.csv(text=input$text, sep="\t")
                
                dat <- escalc(measure="SMD", n1i=N1, n2i=N2,
                m1i=M1, m2i=M2,
                sd1i=SD1, sd2i=SD2,
                data=dat, append=TRUE)
                
                dat$ES <- dat$yi # 小数点以下を丸めない
                dat$yi <- NULL
                dat$SV <- dat$vi # 小数点以下を丸めない
                dat$vi <- NULL
                
                fixed <- MAd::macat(ES, SV, mod = Moderator, data=dat, method= "fixed")
                random <- MAd::macat(ES, SV, mod = Moderator, data=dat, method= "random")
                
                cat("---", "\n", "Fixed effects model:", "\n")
                print(fixed)
                
                cat("\n", "\n", "---", "\n", "Random effects model:", "\n")
                print(random)
                
            }
            
            
            else if (input$type == "mdes") {
                
                dat <- read.csv(text=input$text, sep="\t")
                
                df <- (dat$N1 + dat$N2) - 2
                j <- 1 - (3/(4 * df - 1))
                g <- j * dat$d
                dat$ES <- g # 小数点以下を丸めない
                
                dat$SV <- (((dat$N1+dat$N2)/(dat$N1*dat$N2))+((dat$ES*dat$ES)/(2*(dat$N1+dat$N2)))) # 小数点以下を丸めない

                fixed <- MAd::macat(ES, SV, mod = Moderator, data=dat, method= "fixed")
                random <- MAd::macat(ES, SV, mod = Moderator, data=dat, method= "random")
                
                cat("---", "\n", "Fixed effects model:", "\n")
                print(fixed)
                
                cat("\n", "\n", "---", "\n", "Random effects model:", "\n")
                print(random)
                
            }
            
            
            else if (input$type == "cor") {
                
                dat <- read.csv(text=input$text, sep="\t")
                
                dat <- escalc(measure="ZCOR", ni=N, ri=r, data=dat, append=TRUE)
                dat$FZ <- dat$yi # 小数点以下を丸めない
                dat$yi <- NULL
                dat$SV <- dat$vi # 小数点以下を丸めない
                dat$vi <- NULL
                
                
                dat$var.z <- var_z(dat$N) # 正確な値を計算するために追加
                
                # Fixed effects
                fixed <- MAc::macat(FZ, var.z, mod = Moderator, data=dat, ztor = TRUE, method= "fixed")
                z.fixed <- MAc::macat(FZ, var.z, mod = Moderator, data=dat, ztor = FALSE, method= "fixed") # Accurate z and p
                
                # Random effects
                random <- MAc::macat(FZ, var.z, mod = Moderator, data=dat, ztor = TRUE, method= "random")
                z.random <- MAc::macat(FZ, var.z, mod = Moderator, data=dat, ztor = FALSE, method= "random") # Accurate z and p
                
                
                cat("---", "\n", "Fixed effects model:", "\n")
                print(fixed)
                
                cat("\n", "Accurate z and p values:", "\n")
                print(z.fixed$Model[8:9])
                
                
                cat("\n", "\n", "---", "\n", "Random effects model:", "\n")
                print(random)
                
                cat("\n", "Accurate z and p values:", "\n")
                print(z.random$Model[8:9])
                
                
            }
            
            #} else {
            
            #cat("No moderator (subgroup) analysis is conducted.","\n")
            
            #}
        
    })
    
    
    
    

################################################
# Categorical Moderator Graph
################################################

    ModFixGraph <- function(){
    
        if (input$type == "mdms") {
        
            dat <- W.data()$dat
        
            MAd::plotcat(ES, SV, mod = Moderator, data = dat, method= "fixed", modname= "Moderator")
        
        }
    
    
        else if (input$type == "mdes") {
        
            dat <- W.data()$dat

            MAd::plotcat(ES, SV, mod = Moderator, data = dat, method= "fixed", modname= "Moderator")
            
        }
    
    
        else if (input$type == "cor") {
        
            dat <- W.data()$dat

            MAd::plotcat(FZ, SV, mod = Moderator, data = dat, method= "fixed", modname= "Moderator")
        
        }
    
    }


    output$ModFixGraph <- renderPlot({
        print(ModFixGraph())
    })





    ModRandGraph <- function(){
    
        if (input$type == "mdms") {
        
            dat <- W.data()$dat
        
            MAd::plotcat(ES, SV, mod = Moderator, data = dat, method= "random", modname= "Moderator")
        
        }
    
    
        else if (input$type == "mdes") {
        
            dat <- W.data()$dat
        
            MAd::plotcat(ES, SV, mod = Moderator, data = dat, method= "random", modname= "Moderator")
        
        }
    
    
        else if (input$type == "cor") {
        
            dat <- W.data()$dat
        
            MAd::plotcat(FZ, SV, mod = Moderator, data = dat, method= "random", modname= "Moderator")
        
        }
    
    }


    output$ModRandGraph <- renderPlot({
        print(ModRandGraph())
    })





################################################
# R session info
################################################

    info <- reactive({
        info1 <- paste("This analysis was conducted with ", strsplit(R.version$version.string, " \\(")[[1]][1], ".", sep = "")# バージョン情報
        info2 <- paste("It was executed on ", date(), ".", sep = "")# 実行日時
        cat(sprintf(info1), "\n")
        cat(sprintf(info2), "\n")
    })
    
    
    
    
    
################################################
# server.R and ui.R connection
################################################

    output$info.out <- renderPrint({
        info()
    })
    
    
    
    
    
    output$data.out <- renderPrint({
        data()
    })
    
    output$fe.out <- renderPrint({
        fe()
    })
    
    output$re.out <- renderPrint({
        re()
    })
    
    output$asy.out <- renderPrint({
        asy()
    })
    
    output$modAnalysis.out <- renderPrint({
        modAnalysis()
    })
    
})
library(shiny)
library(shinyAce)



shinyUI(bootstrapPage(


    headerPanel("Meta-analysis"),


    sidebarPanel(

        radioButtons("type", strong("Analysis and data input type:"),
                    list("Mean Differences (n, M, SD)" = "mdms",
                         "Mean Differences (n, Effect size d)" = "mdes",
                         "Correlations (n, r)" = "cor"
                        ),
        ),

        br()

    ),




mainPanel(


    tabsetPanel(

        tabPanel("Main",

            p('Note: Input values must be separated by tabs. Copy and paste from Excel/Numbers.'),

            p(HTML("<b><div style='background-color:#FADDF2;border:1px solid black;'>Your data needs to have exactly the same header (variable names) in the first row.</div></b>")),

            strong('Option:'),
            checkboxInput("moderator", label = strong("The data contains a categorical moderator (subgroup) variable."), value = T),

            br(),

            aceEditor("text", value="Study\tModerator\tN1\tM1\tSD1\tN2\tM2\tSD2\nStudy 01\tJH\t30\t72.97\t13.23\t30\t51.57\t16.5\nStudy 02\tUNI\t24\t81.63\t14.42\t23\t75.09\t23.01\nStudy 03\tSH\t81\t35.38\t16.13\t83\t30.08\t14.29\nStudy 04\tSH\t21\t3.48\t0.68\t21\t2.95\t1.28\nStudy 05\tSH\t15\t60.47\t17.37\t15\t53.8\t17.4\nStudy 06\tSH\t7\t27.3\t4.1\t7\t15.7\t4.1\nStudy 07\tSH\t28\t33.2\t15.65\t28\t27.9\t9.57\nStudy 08\tUNI\t40\t19.23\t9.55\t40\t17.53\t8.87\nStudy 09\tUNI\t17\t29.92\t16.67\t18\t11.86\t13.24\nStudy 10\tUNI\t25\t27.98\t16.52\t21\t29.76\t16\nStudy 11\tUNI\t26\t9.65\t2.99\t26\t8.23\t3.59\nStudy 12\tUNI\t48\t16\t3.47\t49\t13.71\t4.07\nStudy 13\tUNI\t27\t5.9\t1.4\t27\t2.8\t1.7\nStudy 14\tSH\t34\t11.03\t1.78\t41\t10.05\t2.52\nStudy 15\tUNI\t57\t4.26\t1.61\t58\t3.62\t1.79\nStudy 16\tSH\t63\t8.82\t2.5\t60\t7.36\t2.8\nStudy 17\tUNI\t15\t12.27\t4.95\t15\t5.93\t3.55\nStudy 18\tJH\t142\t17.53\t4.34\t37\t13.68\t3.68\nStudy 19\tJH\t54\t12.98\t7.67\t27\t3.3\t2.3\nStudy 20\tJH\t39\t12.36\t7.68\t35\t5.49\t3.88\nStudy 21\tJH\t34\t12.44\t5.66\t32\t5.81\t3.14\nStudy 22\tJH\t60\t18.18\t4.09\t62\t17.84\t4.09\nStudy 23\tSH\t39\t13.72\t5.32\t39\t8.77\t5\nStudy 24\tSH\t39\t79.8\t9.5\t213\t59.8\t15.3\nStudy 25\tUNI\t42\t16\t2.05\t34\t14.32\t2.79\nStudy 26\tUNI\t56\t78.17\t9.94\t77\t70.85\t11.74\nStudy 27\tUNI\t28\t85.06\t23.52\t28\t80.83\t22.47\nStudy 28\tUNI\t36\t25.02\t3.36\t33\t25.38\t4.71\nStudy 29\tUNI\t66\t0.93\t0.59\t66\t0.45\t0.29",
                mode="r", theme="cobalt"),

            br(),

            h3("Effect size and sampling variance"),

            verbatimTextOutput("data.out"),

            br(),

            h3("Fixed effects model"),
            verbatimTextOutput("fe.out"),

            br(),

            h3("Random effects model"),
            verbatimTextOutput("re.out"),

            p('[Criteria for checking heterogeneity]',
            br(),
            br(),
            'I^2 (How much effect sizes across studies differ)', br(),
            '25–50: Little different', br(),
            '50–75: Quite different', br(),
            '75–100: Considerably different', br(),
            br(),
            'Test for Heterogeneity: p-val < .05 (not homogeneous)', br(),
            br(),
            'H (sqrt(H^2)) > 1: There is unexplained heterogeneity.'
            ),

            br(),
            br(),


            h3("Forest plot (Fixed effects model)"),

            plotOutput("fePlot", height = "550px"),

            br(),

            h3("Forest plot (Random effects model)"),

            plotOutput("rePlot", height = "550px"),

            br(),

            h3("Funnel plot (Fixed effects model)"),

            plotOutput("FunFixPlot"),
            p('Open circles (if any) on the right side show missing NULL studies estimated with the trim-and-fill method, added in the funnel plot.'),

            br(),

            h3("Funnel plot (Random effects model)"),

            plotOutput("FunRandPlot"),
            p('Open circles (if any) on the right side show missing NULL studies estimated with the trim-and-fill method, added in the funnel plot.'),
            br(),
            br(),

            verbatimTextOutput("asy.out"), # regression tests for funnel plot asymmetry
            p('Fail-safe N is the number of nonsignificant studies necessary to make the result nonsignificant. "When the fail-safe N is high, that is interpreted to mean that even a large number of nonsignificant studies may not influence the statistical significance of meta-analytic results too greatly. Although ... it is not a very precise measure of publication bias"',
            a('(Oswald & Plonsky, 2010, p. 92)', href='http://dx.doi.org/10.1017/S0267190510000115', target="_blank"), '.'),

            br(),

            br(),

            # Display this only if "moderator" is checked
            conditionalPanel(condition = "input.moderator == true",
                h3("Moderator (subgroup) analysis"),
                verbatimTextOutput("modAnalysis.out")
            ),

            br(),

            # Display this only if "moderator" is checked
            conditionalPanel(condition = "input.moderator == true",
                h4("Categorical moderator graph (Fixed effects model)"),
                plotOutput("ModFixGraph")
            ),

            br(),

            # Display this only if "moderator" is checked
            conditionalPanel(condition = "input.moderator == true",
                h4("Categorical moderator graph (Random effects model)"),
                plotOutput("ModRandGraph")
            ),

            br(),
            br(),

            strong('R session info'),
            verbatimTextOutput("info.out")
            ),



    tabPanel("Input Examples",

            p('Note: Input values must be separated by tabs. Copy and paste from Excel/Numbers.'),

            p(HTML("<b><div style='background-color:#FADDF2;border:1px solid black;'>Your data needs to have exactly the same header (variable names) in the first row.</div></b>")),

            br(),

            p(strong("Mean Differences (n, M, SD)")),
            aceEditor("text1", value="Study\tModerator\tN1\tM1\tSD1\tN2\tM2\tSD2\nStudy 01\tJH\t30\t72.97\t13.23\t30\t51.57\t16.5\nStudy 02\tUNI\t24\t81.63\t14.42\t23\t75.09\t23.01\nStudy 03\tSH\t81\t35.38\t16.13\t83\t30.08\t14.29\nStudy 04\tSH\t21\t3.48\t0.68\t21\t2.95\t1.28\nStudy 05\tSH\t15\t60.47\t17.37\t15\t53.8\t17.4\nStudy 06\tSH\t7\t27.3\t4.1\t7\t15.7\t4.1\nStudy 07\tSH\t28\t33.2\t15.65\t28\t27.9\t9.57\nStudy 08\tUNI\t40\t19.23\t9.55\t40\t17.53\t8.87\nStudy 09\tUNI\t17\t29.92\t16.67\t18\t11.86\t13.24\nStudy 10\tUNI\t25\t27.98\t16.52\t21\t29.76\t16\nStudy 11\tUNI\t26\t9.65\t2.99\t26\t8.23\t3.59\nStudy 12\tUNI\t48\t16\t3.47\t49\t13.71\t4.07\nStudy 13\tUNI\t27\t5.9\t1.4\t27\t2.8\t1.7\nStudy 14\tSH\t34\t11.03\t1.78\t41\t10.05\t2.52\nStudy 15\tUNI\t57\t4.26\t1.61\t58\t3.62\t1.79\nStudy 16\tSH\t63\t8.82\t2.5\t60\t7.36\t2.8\nStudy 17\tUNI\t15\t12.27\t4.95\t15\t5.93\t3.55\nStudy 18\tJH\t142\t17.53\t4.34\t37\t13.68\t3.68\nStudy 19\tJH\t54\t12.98\t7.67\t27\t3.3\t2.3\nStudy 20\tJH\t39\t12.36\t7.68\t35\t5.49\t3.88\nStudy 21\tJH\t34\t12.44\t5.66\t32\t5.81\t3.14\nStudy 22\tJH\t60\t18.18\t4.09\t62\t17.84\t4.09\nStudy 23\tSH\t39\t13.72\t5.32\t39\t8.77\t5\nStudy 24\tSH\t39\t79.8\t9.5\t213\t59.8\t15.3\nStudy 25\tUNI\t42\t16\t2.05\t34\t14.32\t2.79\nStudy 26\tUNI\t56\t78.17\t9.94\t77\t70.85\t11.74\nStudy 27\tUNI\t28\t85.06\t23.52\t28\t80.83\t22.47\nStudy 28\tUNI\t36\t25.02\t3.36\t33\t25.38\t4.71\nStudy 29\tUNI\t66\t0.93\t0.59\t66\t0.45\t0.29", mode="r", theme="solarized_light"),


            br(),
            p(strong("Mean Differences (n, Effect size d)")),
            aceEditor("text2", value="Study\tModerator\tN1\tN2\td\nStudy 01\tJH\t30\t30\t1.431\nStudy 02\tUNI\t24\t23\t0.3423\nStudy 03\tSH\t81\t83\t0.3481\nStudy 04\tSH\t21\t21\t0.5171\nStudy 05\tSH\t15\t15\t0.3837\nStudy 06\tSH\t7\t7\t2.8293\nStudy 07\tSH\t28\t28\t0.4086\nStudy 08\tUNI\t40\t40\t0.1845\nStudy 09\tUNI\t17\t18\t1.2039\nStudy 10\tUNI\t25\t21\t-0.1093\nStudy 11\tUNI\t26\t26\t0.4298\nStudy 12\tUNI\t48\t49\t0.605\nStudy 13\tUNI\t27\t27\t1.9907\nStudy 14\tSH\t34\t41\t0.4422\nStudy 15\tUNI\t57\t58\t0.3758\nStudy 16\tSH\t63\t60\t0.5508\nStudy 17\tUNI\t15\t15\t1.4719\nStudy 18\tJH\t142\t37\t0.9136\nStudy 19\tJH\t54\t27\t1.5079\nStudy 20\tJH\t39\t35\t1.111\nStudy 21\tJH\t34\t32\t1.4368\nStudy 22\tJH\t60\t62\t0.0831\nStudy 23\tSH\t39\t39\t0.9588\nStudy 24\tSH\t39\t213\t1.3729\nStudy 25\tUNI\t42\t34\t0.6976\nStudy 26\tUNI\t56\t77\t0.6642\nStudy 27\tUNI\t28\t28\t0.1839\nStudy 28\tUNI\t36\t33\t-0.0886\nStudy 29\tUNI\t66\t66\t1.0326", mode="r", theme="solarized_light"),


            br(),
            p(strong("Correlations (n, r)")),
            aceEditor("text3", value="Study\tN\tr\tModerator\nIzumi (2000)\t175\t0.78\tcollege\nYu (2009)\t53\t0.38\tJS high\nThuy (1996)\t250\t0.69\tcollege\nOckey (2002)\t90\t0.89\tcollege\nAraru (2005)\t86\t0.52\tJS high\nWee (1997)\t182\t0.98\tcollege\nOzoda (2007)\t591\t0.91\tcollege\nHala (2004)\t30\t0.95\tcollege\nTapio (2008)\t37\t0.47\tJS high\nAndarani (2008)\t107\t0.84\tcollege\nDavis (1999)\t74\t0.99\tcollege\nPlonsky (2002)\t217\t0.86\tcollege\nGassel (1993)\t203\t0.99\tcollege",mode="r", theme="solarized_light"),

            br()

            ),






        tabPanel("About",

            strong('Note'),
            p('This web application is developed with',
            a("Shiny.", href="http://www.rstudio.com/shiny/", target="_blank"),
            ''),

            br(),

            strong('List of Packages Used'), br(),
            code('library(shiny)'),br(),
            code('library(shinyAce)'),br(),
            code('library(metafor)'),br(),
            code('library(meta)'),br(),
            code('library(MAd)'),br(),
            code('library(MAc)'),br(),
            code('library(quantreg)'),br(),
            code('library(ggplot2)'),br(),


            br(),

            strong('Code'),
            p('Source code for this application is based on',
            a('"The handbook of Research in Foreign Language Learning and Teaching" (Takeuchi & Mizumoto, 2012).', href='http://mizumot.com/handbook/', target="_blank")),

            p('The code for this web application is available at',
            a('GitHub.', href='https://github.com/mizumot/meta', target="_blank")),

            p('If you want to run this code on your computer (in a local R session), run the code below:',
            br(),
            code('library(shiny)'),br(),
            code('runGitHub("meta","mizumot")')
            ),

            br(),

            strong('Acknowledgment'),
            p('I thank',
            a("Dr. Luke Plonsky", href="http://oak.ucc.nau.edu/ldp3/", target="_blank"), 'and',
            a("Dr. Yo In'nami", href="https://sites.google.com/site/yoinnami/", target="_blank"),
            'for their support and feedback to create this web application.'),

            br(),

            strong('Citation in Publications'),
            p('Mizumoto, A. (2015). Langtest (Version 1.0) [Web application]. Retrieved from http://langtest.jp'),

            br(),

            strong('Article'),
            p('Mizumoto, A., & Plonsky, L. (2015).', a("R as a lingua franca: Advantages of using R for quantitative research in applied linguistics.", href='http://applij.oxfordjournals.org/content/early/2015/06/24/applin.amv025.abstract', target="_blank"), em('Applied Linguistics,'), 'Advance online publication. doi:10.1093/applin/amv025'),

            br(),

            strong('Recommended'),
            p('To learn more about R, I suggest this excellent and free e-book (pdf),',
            a("A Guide to Doing Statistics in Second Language Research Using R,", href="http://cw.routledge.com/textbooks/9780805861853/guide-to-R.asp", target="_blank"),
            'written by Dr. Jenifer Larson-Hall.'),

            p('Also, if you are a cool Mac user and want to use R with GUI,',
            a("MacR", href="https://sites.google.com/site/casualmacr/", target="_blank"),
            'is defenitely the way to go!'),

            br(),

            strong('Author'),
            p(a("Atsushi MIZUMOTO,", href="http://mizumot.com", target="_blank"),' Ph.D.',br(),
            'Professor of Applied Linguistics',br(),
            'Faculty of Foreign Language Studies /',br(),
            'Graduate School of Foreign Language Education and Research,',br(),
            'Kansai University, Osaka, Japan'),

            br(),

            a(img(src="http://i.creativecommons.org/p/mark/1.0/80x15.png"), target="_blank", href="http://creativecommons.org/publicdomain/mark/1.0/"),

            p(br())

)
)
)
))
Code license: GPL-3