test_mediation: (Robust) mediation analysis

Description Usage Arguments Details Value Note Author(s) References See Also Examples

View source: R/test_mediation.R

Description

Perform (robust) mediation analysis via a (fast and robust) bootstrap test or Sobel's test.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
test_mediation(data, ...)

## Default S3 method:
test_mediation(data, x, y, m, covariates = NULL,
  test = c("boot", "sobel"), alternative = c("twosided", "less",
  "greater"), R = 5000, level = 0.95, type = c("bca", "perc"),
  method = c("regression", "covariance"), robust = TRUE,
  median = FALSE, control, ...)

## S3 method for class 'fit_mediation'
test_mediation(data, test = c("boot", "sobel"),
  alternative = c("twosided", "less", "greater"), R = 5000,
  level = 0.95, type = c("bca", "perc"), ...)

robmed(..., test = "boot", method = "regression", robust = TRUE,
  median = FALSE)

indirect(..., test = "boot", method = "regression", robust = FALSE,
  median = FALSE)

Arguments

data

a data frame containing the variables. Alternatively, this can be a mediation model fit as returned by fit_mediation.

...

additional arguments to be passed down. For the bootstrap tests, those can be used to specify arguments of boot, for example for parallel computing.

x

a character string, an integer or a logical vector specifying the column of data containing the independent variable.

y

a character string, an integer or a logical vector specifying the column of data containing the dependent variable.

m

a character, integer or logical vector specifying the columns of data containing the hypothesized mediator variables.

covariates

optional; a character, integer or logical vector specifying the columns of data containing additional covariates to be used as control variables.

test

a character string specifying the test to be performed for the indirect effect. Possible values are "boot" (the default) for the bootstrap, or "sobel" for Sobel's test. Currently, Sobel's test is not implemented for more than one hypothesized mediator variable.

alternative

a character string specifying the alternative hypothesis in the test for the indirect effects. Possible values are "twosided" (the default), "less" or "greater".

R

an integer giving the number of bootstrap replicates. The default is to use 5000 bootstrap replicates.

level

numeric; the confidence level of the confidence interval in the bootstrap test. The default is to compute a 95% confidence interval.

type

a character string specifying the type of confidence interval to be computed in the bootstrap test. Possible values are "bca" (the default) for the bias-corrected and accelerated bootstrap, or "perc" for the percentile bootstrap.

method

a character string specifying the method of estimation for the mediation model. Possible values are "regression" (the default) to estimate the effects via regressions, or "covariance" to estimate the effects via the covariance matrix. Note that the effects are always estimated via regressions if more than one hypothesized mediator is supplied in m, or if control variables are specified via covariates.

robust

a logical indicating whether to perform a robust test (defaults to TRUE).

median

a logical indicating if the effects should be estimated via median regression (defaults to FALSE). This is ignored unless method is "regression" and robust is TRUE.

control

a list of tuning parameters for the corresponding robust method. For robust regression (method = "regression", robust = TRUE and median = FALSE), a list of tuning parameters for lmrob as generated by reg_control. For Huberized covariance matrix estimation (method = "covariance" and robust = TRUE), a list of tuning parameters for cov_Huber as generated by cov_control. No tuning parameters are necessary for median regression (method = "regression", robust = TRUE and median = TRUE).

Details

If method is "regression", robust is TRUE and median is FALSE (the defaults), the tests are based on robust regressions with lmrob. The bootstrap test is thereby performed via the fast and robust bootstrap.

Note that the regression estimator implemented in lmrob can be seen as weighted least squares estimator, where the weights are dependent on how much an observation is deviating from the rest. The trick for the fast and robust bootstrap is that on each bootstrap sample, first a weighted least squares estimator is computed (using those robustness weights from the original sample) followed by a linear correction of the coefficients. The purpose of this correction is to account for the additional uncertainty of obtaining the robustness weights.

If method is "regression", robust is TRUE and median is TRUE, the tests are based on median regressions with rq and the standard bootstrap (). Unlike the robust regressions described above, median regressions are not robust against outliers in the explanatory variables, and the standard bootstrap can suffer from oversampling of outliers in the bootstrap samples.

If method is "covariance" and robust is TRUE, the tests are based on a Huber M-estimator of location and scatter. For the bootstrap test, the M-estimates are used to first clean the data via a transformation. Then the standard bootstrap is performed with the cleaned data. Note that this covariance-based approach is less robust than the approach based on robust regressions described above. Furthermore, the bootstrap does not account for the variability from cleaning the data.

robmed is a wrapper function for performing robust mediation analysis via regressions and the fast and robust bootstrap.

indirect is a wrapper function for performing non-robust mediation analysis via regressions and the bootstrap (inspired by Preacher & Hayes' SPSS macro INDIRECT).

Value

An object inheriting from class "test_mediation" (class "boot_test_mediation" if test is "boot" or "sobel_test_mediation" if test is "sobel") with the following components:

ab

a numeric vector containing the point estimates of the indirect effects.

ci

a numeric vector of length two or a matrix of two columns containing the bootstrap confidence intervals for the indirect effects (only "boot_test_mediation").

reps

an object of class "boot" containing the bootstrap replicates of the effects (only "boot_test_mediation").

se

numeric; the standard error of the indirect effect according to Sobel's formula (only "sobel_test_mediation").

statistic

numeric; the test statistic for Sobel's test (only "sobel_test_mediation").

p_value

numeric; the p-value from Sobel's test (only "sobel_test_mediation").

alternative

a character string specifying the alternative hypothesis in the test for the indirect effects.

R

an integer giving the number of bootstrap replicates (only "boot_test_mediation").

level

numeric; the confidence level of the bootstrap confidence interval (only "boot_test_mediation").

type

a character string specifying the type of bootstrap confidence interval (only "boot_test_mediation").

fit

an object inheriting from class "fit_mediation" containing the estimation results for the direct effect and the total effect in the mediation model.

Note

For the fast and robust bootstrap, the simpler correction of Salibian-Barrera & Van Aelst (2008) is used rather than the originally proposed correction of Salibian-Barrera & Zamar (2002).

Author(s)

Andreas Alfons

References

Alfons, A., Ates, N.Y. and Groenen, P.J.F. (2018) A robust bootstrap test for mediation analysis. ERIM Report Series in Management, Erasmus Research Institute of Management. URL https://hdl.handle.net/1765/109594.

Preacher, K.J. and Hayes, A.F. (2004) SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, & Computers, 36(4), 717–731.

Preacher, K.J. and Hayes, A.F. (2008) Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879–891.

Salibian-Barrera, M. and Van Aelst, S. (2008) Robust model selection using fast and robust bootstrap. Computational Statistics & Data Analysis, 52(12), 5121–5135

Salibian-Barrera, M. and Zamar, R. (2002) Bootstrapping robust estimates of regression. The Annals of Statistics, 30(2), 556–582.

Sobel, M.E. (1982) Asymptotic confidence intervals for indirect effects in structural equation models. Sociological Methodology, 13, 290–312.

Yuan, Y. and MacKinnon, D.P. (2014) Robust mediation analysis based on median regression. Psychological Methods, 19(1), 1–20.

Zu, J. and Yuan, K.-H. (2010) Local influence and robust procedures for mediation analysis. Multivariate Behavioral Research, 45(1), 1–44.

See Also

fit_mediation

coef, confint, fortify and plot methods, p_value

boot, lmrob, lm, cov_Huber, cov_ML

Examples

1
2
3
4
5
6
data("BSG2014")
test <- test_mediation(BSG2014,
                       x = "ValueDiversity",
                       y = "TeamCommitment",
                       m = "TaskConflict")
summary(test)

robmed documentation built on Nov. 5, 2018, 5:06 p.m.