Package {medmod}

medmod: Simple Mediation and Moderation Analysis

Description

This toolbox allows you to do simple mediation and moderation analysis. Models are estimated with the 'lavaan' package by Rosseel (2012) doi:10.18637/jss.v048.i02; standard errors for the mediation estimates are computed with the delta method following Sobel (1982) doi:10.2307/270723 or by bootstrapping. It is also available as a module for 'jamovi' (see https://www.jamovi.org for more information). You can find an in depth tutorial on the 'lavaan' model syntax used for this package on https://lavaan.ugent.be/tutorial/index.html.

Details

Author(s)

Maintainer: Ravi Selker selker.ravi@gmail.com

Authors:

• Ravi Selker selker.ravi@gmail.com

See Also

Useful links:

• https://github.com/raviselker/medmod

• Report bugs at https://github.com/raviselker/medmod/issues

Center a vector

Description

center returns a numeric vector with centered values.

Usage

center(x)

Arguments

Draw a path diagram

Description

drawPathDiagram renders boxes and labelled arrows as a ggplot. Arrows run between box borders; an edge whose to is not a node name ends at its explicit toX/toY point (used to point at a path).

Usage

drawPathDiagram(nodes, edges, ggtheme, theme, sigCaption = TRUE)

Arguments

Extract parameter rows from lavaan estimates

Description

lavaanRow returns the rows of a lavaan::parameterestimates() data frame whose columns equal the given values, e.g. lavaanRow(est, label = 'ab') or lavaanRow(est, lhs = dep, rhs = pred). It errors when the expected parameter is absent instead of silently returning zero rows.

Usage

lavaanRow(est, ...)

Arguments

Mediation

Description

Simple mediation analysis. Estimates the indirect, direct and total effects of a model in which a predictor influences the dependent variable through a mediator. The model is fitted with the lavaan package (Rosseel, 2012); standard errors are computed with the delta method (equivalent to the Sobel test for the indirect effect) or by bootstrapping. Optionally provides the individual path estimates, an estimate plot, and an annotated path diagram of the model.

Usage

med(
  data,
  dep,
  med,
  pred,
  estMethod = "standard",
  bootstrap = 1000,
  test = TRUE,
  ci = FALSE,
  ciWidth = 95,
  pm = FALSE,
  paths = FALSE,
  label = FALSE,
  estPlot = FALSE,
  pathDiagram = FALSE,
  pathDiagramLabel = TRUE,
  pathDiagramEst = TRUE,
  pathDiagramSig = TRUE
)

Arguments

Value

A results object containing:

Tables can be converted to data frames with asDF or as.data.frame. For example:

results$med$asDF

as.data.frame(results$med)

Examples

set.seed(1234)
X <- rnorm(100)
M <- 0.5*X + rnorm(100)
Y <- 0.7*M + rnorm(100)
dat <- data.frame(X=X, M=M, Y=Y)

med(dat, dep = "Y", pred = "X", med = "M")

#
#  Mediation Estimates
#  -----------------------------------------------------
#    Effect      Estimate    SE        Z        p
#  -----------------------------------------------------
#    Indirect      0.3736    0.0920    4.059    < .001
#    Direct        0.0364    0.1044    0.348     0.728
#    Total         0.4100    0.1247    3.287     0.001
#  -----------------------------------------------------
#
#

Moderation

Description

Simple moderation analysis. Estimates a model in which the effect of a predictor on the dependent variable depends on the value of a moderator, testing the main effects of predictor and moderator and their interaction. All variables are mean-centered before the model is fitted with the lavaan package (Rosseel, 2012); standard errors are computed with the delta method or by bootstrapping. Optionally provides a simple slope analysis with a corresponding plot, and an annotated path diagram of the model.

Usage

mod(
  data,
  dep,
  mod,
  pred,
  estMethod = "standard",
  bootstrap = 1000,
  label = FALSE,
  test = TRUE,
  ci = FALSE,
  ciWidth = 95,
  pathDiagram = FALSE,
  pathDiagramLabel = TRUE,
  pathDiagramEst = TRUE,
  pathDiagramSig = TRUE,
  pathDiagramMainEffect = FALSE,
  simpleSlopeEst = FALSE,
  simpleSlopePlot = FALSE
)

Arguments

Value

A results object containing:

Tables can be converted to data frames with asDF or as.data.frame. For example:

results$mod$asDF

as.data.frame(results$mod)

Examples

set.seed(1234)
X <- rnorm(100)
M <- rnorm(100)
X_M <- X*M
Y <- 0.7*X + 0.1*M + 4.2*X_M + rnorm(100)
dat <- data.frame(X=X, M=M, Y=Y)

mod(dat, dep = "Y", pred = "X", mod = "M")

#
#  Moderation Estimates
#  --------------------------------------------------
#             Estimate    SE        Z        p
#  --------------------------------------------------
#    X           0.951    0.0965     9.86    < .001
#    M          -0.471    0.0923    -5.10    < .001
#    X:M         4.185    0.1009    41.50    < .001
#  --------------------------------------------------
#
#

Format a path coefficient annotation

Description

pathLabel builds the arrow annotation for a path diagram as a plotmath expression, e.g. 'b[1] == "0.42***"', so subscripted path names render as real subscripts on any graphics device.

Usage

pathLabel(name, est, p, showName = TRUE, showEst = TRUE, showSig = TRUE)

Arguments

Shorten labels for display

Description

shortenLabel truncates long labels with an ellipsis so they fit inside path diagram boxes.

Usage

shortenLabel(x, n = 16)

Arguments

Significance stars for p values

Description

sigStars converts p values to the conventional significance stars: * p < .05, ** p < .01, *** p < .001.

Usage

sigStars(p)

Arguments