ComputeEffectxx: Computes direct, indirect, or total effects of xx

In longjp/mediateR: Direct and Indirect Mediation Effects with Multiple Mediators

Computes direct, indirect, or total effects of xx

Description

ComputeEffectxx computes either direct, indirect, or total effect of each xx variable on response y when transition x from xp to xpp.

Usage

ComputeEffectxx(dat, fit, effect, xp = 0, xpp = 1, risk_scale = NULL,
  mmn = FALSE, rmean = NULL)

Arguments

dat	List containing data (see SimulateData for format)
fit	List with path coefficients.
effect	Either 'direct', 'indirect', or 'total'
xp	The initial value of x.
xpp	The new value of x.
risk_scale	Either 'diff' (for risk difference) or 'ratio' (for odds). see details for more information.
mmn	Should gene set network be simulated. If FALSE assume conditionally independent mediators.

Value

Vector of effects.

Examples

params <- SimpleSim()
dat <- SimulateData(params)
fit <- ComputePath(dat)
d <- ComputeEffectxx(dat,fit,"direct")
i <- ComputeEffectxx(dat,fit,"indirect")
to <- ComputeEffectxx(dat,fit,"total")
## effects computed by estimating integrals
print(c(d,i,to))
## effects using model linearity assumption
ComputeEffectsLinear(fit) ## based on sample
## exact effects using simulation coefficients
ComputeEffectsLinear(params) ## exact

longjp/mediateR documentation built on May 24, 2023, 12:30 p.m.