Mediation analysis A Monte Carlo simulation method to assess mediation based on Selig & Preacher (2008).

mediation(
  model_med,
  model_y,
  model_med2 = NULL,
  x,
  med,
  med2 = NULL,
  mod = NULL,
  mod_stage = NULL,
  mod_level = NULL,
  conf = 95,
  rep = 20000,
  verbose = TRUE,
  digits = 3
)

Arguments

model_med: a fitted model object for mediator.
model_y: a fitted model object for outcome
model_med2: a fitted model object for the second mediator for serial mediation
x: a character string indicating the name of the independent variable used in the models.
med: a character string indicating the name of the mediator used in the models.
med2: a character string indicating the name of the second mediator used in the models (for serial mediations)
mod: a character string indicating the name of the moderator used in the models.
mod_stage: a character string specifying the stage at which the moderating effect occurs. For instance, in a first-stage moderated mediation, where the moderator influences the effect of X on the mediator (Med), set this to "model_med". In a second-stage moderated mediation, where the moderator affects the relationship between the mediator (Med) and the outcome variable (Y), set this to "model_y".#'
mod_level: The default is -1 SD and +1 SD for a continuous variable, and it is the two levels for a binary variable.
conf: level of the returned two-sided confidence intervals. Default is to return the 2.5 and 97.5 percentiles of the simulated quantities (i.e., 95%).
rep: number of Monte Carlo draws
verbose: deafult is TRUE.
digits: number of digits to round to

Value

Nothing to return. Print the indirect effect.

References

Selig, J. P., & Preacher, K. J. (2008, June). Monte Carlo method for assessing mediation: An interactive tool for creating confidence intervals for indirect effects. http://quantpsy.org/.

Examples

new_dat = iris %>% 
  dplyr::rename(x = Petal.Length) %>% 
  dplyr::rename(m = Sepal.Length) %>% 
  dplyr::rename(moderator = Sepal.Width) %>% 
  dplyr::rename(y = Petal.Width)

model_1 = lm(data = new_dat, m ~ x)
model_2 = lm(data = new_dat, y ~ x*moderator + m)


mediation(model_med = model_1,
          model_y = model_2,
          rep = 20000,
          x = 'x', 
          med = 'm',
          mod = 'moderator',
          mod_stage = 'model_y',
          digits = 3)
#> Model Summary
#> Model Type = Mediation Analysis
#> Monte Carlo simulation = 20000
#> CI = 95%
#> Condition Indirect Effect = Low is 2.621 and High is 3.493
#> ───────────────────────────────────────────────────────────────────────────────────────────────────────────
#>   indirect_effect_low  indirect_effect_high          CI_low        CI_high  index_mod_med  CI_index_mod_med
#> ───────────────────────────────────────────────────────────────────────────────────────────────────────────
#>                -0.048                -0.036  [-0.118,0.023]  [-0.122,0.05]          0.013    [-0.007,0.033]
#> ───────────────────────────────────────────────────────────────────────────────────────────────────────────
#> You can drag and resize the R console to view the entire table