Built-in models:Mediation analysis

Model

A mediation model can be written as

\[ m_i = a_0 + a_1 x_i + e_{1i} \] \[ y_i = b_0 + b_1 m_i + b_2 x_i + e_{2i} \]

The mediation effect or indirect effect is \(a_1b_1\).

Code

model{
  for (i in 1:N){
      m[i] ~ dnorm(mum[i], pre.s1)
      mum[i] <- a0 + a1*x[i]

      y[i] ~ dnorm(muy[i], pre.s2)
      muy[i] <- b0 + b1*m[i] + b2*x[i]
  }
  a0 ~ dnorm(0, 1.0E-6)
  a1 ~ dnorm(0, 1.0E-6)
  b0 ~ dnorm(0, 1.0E-6)
  b1 ~ dnorm(0, 1.0E-6)
  b2 ~ dnorm(0, 1.0E-6)
  pre.s1 ~ dgamma(.001, .001) 
  pre.s2 ~ dgamma(.001, .001) 
  ab <- a1*b1
}