A mediation analysis design that examines the effect of treatment (Z) on mediator (M) and the effect of mediator (M) on outcome (Y) (given Z=0) as well as direct effect of treatment (Z) on outcome (Y) (given M=0). Analysis is implemented using an interacted regression model. Note this model is not guaranteed to be unbiased despite randomization of Z because of possible violations of sequential ignorability.

mediation_analysis_designer(N = 200, a = 1, b = 0.4, c = 0,
  d = 0.5, rho = 0)

Arguments

N

An integer. Size of sample.

a

A number. Parameter governing effect of treatment (Z) on mediator (M).

b

A number. Effect of mediator (M) on outcome (Y) when Z = 0.

c

A number. Interaction between mediator (M) and (Z) for outcome (Y).

d

A number. Direct effect of treatment (Z) on outcome (Y), when M = 0.

rho

A number in [-1,1]. Correlation between mediator (M) and outcome (Y) error terms. Non zero correlation implies a violation of sequential ignorability.

Value

A mediation analysis design.

Details

See vignette online.

Examples

# Generate a mediation analysis design using default arguments: mediation_1 <- mediation_analysis_designer() draw_estimands(mediation_1)
#> estimand_label estimand #> 1 FirstStage 0.365 #> 2 Indirect_0 0.400 #> 3 Indirect_1 0.400 #> 4 Controlled_Direct_0 0.500 #> 5 Controlled_Direct_1 0.500 #> 6 Natural_Direct_0 0.500 #> 7 Natural_Direct_1 0.500
# NOT RUN { diagnose_design(mediation_1, sims = 1000) # }
# A design with a violation of sequential ignorability and heterogeneous effects: mediation_2 <- mediation_analysis_designer(a = 1, rho = .5, c = 1, d = .75) draw_estimands(mediation_2)
#> estimand_label estimand #> 1 FirstStage 0.36 #> 2 Indirect_0 0.40 #> 3 Indirect_1 1.40 #> 4 Controlled_Direct_0 0.75 #> 5 Controlled_Direct_1 1.75 #> 6 Natural_Direct_0 1.24 #> 7 Natural_Direct_1 1.60
# NOT RUN { diagnose_design(mediation_2, sims = 1000) # }

Author

DeclareDesign Team