dominance.manual: Dominance Analysis, Manually Inputting a Correlation Matrix

Description

This function conducts dominance analysis (Budescu, 1993; Azen & Budescu, 2003) based on a (model-implied) correlation matrix of the manifest or latent variables. Note that the function only provides general dominance.

Usage

dominance.manual(x, out = NULL, digits = 3, write = NULL, check = TRUE,
                 output = TRUE)

Value

Returns an object of class misty.object, which is a list with following entries:

call: function call
type: type of analysis
x: correlation matrix specified in x
args: specification of function arguments
result: results table for the general dominance

Arguments

x: a matrix or data frame with the (model-implied) correlation matrix of the manifest or latent variables. Note that column names need to represent the variables names in x.
out: a character string representing the outcome variable. By default, the first row and column represents the outcome variable.
digits: an integer value indicating the number of decimal places to be used for displaying results. Note that the percentage relative importance of predictors are printed with digits minus 1 decimal places.
write: a character string for writing the results into a Excel file naming a file with or without file extension '.xlsx', e.g., "Results.xlsx" or "Results".
check: logical: if TRUE, argument specification is checked.
output: logical: if TRUE, output is shown.

Author

Takuya Yanagida takuya.yanagida@univie.ac.at

References

Azen, R., & Budescu, D. V. (2003). The dominance analysis approach for comparing predictors in multiple regression. Psychological Methods, 8(2), 129–148. https://doi.org/10.1037/1082-989X.8.2.129

Bolker, B., Warnes, G., & Lumley, T. (2022). gtools: Various R Programming Tools. R package version 3.9.4, https://CRAN.R-project.org/package=gtools

Budescu, D. V. (1993). Dominance analysis: A new approach to the problem of relative importance of predictors in multiple regression. Psychological Bulletin, 114(3), 542–551. https://doi.org/10.1037/0033-2909.114.3.542

Gu, X. (2022). Assessing the relative importance of predictors in latent regression models. Structural Equation Modeling: A Multidisciplinary Journal, 4, 569-583. https://doi.org/10.1080/10705511.2021.2025377

Examples

Run this code

if (FALSE) {
#----------------------------
# Linear model

dat <- data.frame(x1 = c(3, 2, 4, 9, 5, 3, 6, 4, 5, 6, 3, 5),
                  x2 = c(1, 4, 3, 1, 2, 4, 3, 5, 1, 7, 8, 7),
                  x3 = c(0, 0, 1, 0, 1, 1, 1, 1, 0, 0, 1, 1),
                  y  = c(0, 1, 0, 2, 0, 1, 0, 0, 1, 2, 1, 0))

# Dominance analysis
dominance.manual(cor(dat[, c("y", "x1", "x2", "x3")]))

# Equivalent results using the dominance() function
mod <- lm(y ~ x1 + x2 + x3, data = dat)
dominance(mod)

# Outcome 'x3' predicted by 'y', 'x1', and 'x2'
dominance.manual(cor(dat[, c("y", "x1", "x2", "x3")]), out = "x3")

#----------------------------
# Structural equation modeling

library(lavaan)

#.............
# Latent variables

# Model specification
model <- '# Measurement model
          ind60 =~ x1 + x2 + x3
          dem60 =~ y1 + y2 + y3 + y4
          dem65 =~ y5 + y6 + y7 + y8
          # regressions
          ind60 ~ dem60 + dem65'

# Model estimation
fit <- sem(model, data = PoliticalDemocracy)

# Model-implied correlation matrix of the latent variables
fit.cor <- lavInspect(fit, what = "cor.lv")

# Dominance analysis
dominance.manual(fit.cor)

#.............
# Latent and manifest variables

# Model specification, convert manifest to latent variable
model <- '# Measurement model
          ind60 =~ x1 + x2 + x3
          dem60 =~ y1 + y2 + y3 + y4
          # Manifest as latent variable
          ly5 =~ 1*y5
          y5 ~~ 0*y5
          # Regressions
          ind60 ~ dem60 + ly5'

# Model estimation
fit <- sem(model, data = PoliticalDemocracy)

# Model-implied correlation matrix of the latent variables
fit.cor <- lavInspect(fit, what = "cor.lv")

# Dominance analysis
dominance.manual(fit.cor)

#----------------------------
# Multilevel modeling

# Model specification
model <- 'level: 1
            fw =~ y1 + y2 + y3
            # Manifest as latent variables
            lx1 =~ 1*x1
            lx2 =~ 1*x2
            lx3 =~ 1*x3
            x1 ~~ 0*x1
            x2 ~~ 0*x2
            x3 ~~ 0*x3
            # Regression
            fw ~ lx1 + lx2 + lx3
          level: 2
            fb =~ y1 + y2 + y3
            # Manifest as latent variables
            lw1 =~ 1*w1
            lw2 =~ 1*w2
            # Regression
            fb ~ lw1 + lw2'

# Model estimation
fit <- sem(model, data = Demo.twolevel, cluster = "cluster")

# Model-implied correlation matrix of the latent variables
fit.cor <- lavInspect(fit, what = "cor.lv")

# Dominance analysis Within
dominance.manual(fit.cor$within)

# Dominance analysis Between
dominance.manual(fit.cor$cluster)

#----------------------------
# Mplus
#
# In Mplus, the model-impied correlation matrix  of the latent variables
# can be requested by OUTPUT: TECH4 and imported into R by using the
# MplusAuomtation package, for example:

library(MplusAutomation)

# Read Mplus output
output <- readModels()

# Extract model-implied correlation matrix of the latent variables
fit.cor <- output$tech4$latCorEst
}

Run the code above in your browser using DataLab

Last chance! 50% off unlimited learning