lm interaction power

lm-interaction-power

Monte Carlo simulation of power for linear model with interactions using the superb paramtest package.

Two independent variables

Assumptions:

• statistical test: lm
• alpha: .05
• predictor effect sizes:
  • b1: .5 (sd: 1)
  • b2: .15 (sd: 1)
  • b1 * b2: .10
• residual variance: yvar <- sqrt(1 - b1^2 - b2^2 - b3^2)

Three independent variables

Assumptions:

Just look at the source code...