Skip to main content

Advertisement

Table 1 Description of 24 simulation experiments where the response variable is distributed according a beta distribution with the following mean and dispersion parameters in each respective group (or alternatively parameterized in terms of its two shape parameters – p and q – in each group)

From: A Monte Carlo simulation study comparing linear regression, beta regression, variable-dispersion beta regression and fractional logit regression at recovering average difference measures in a two sample design

  μ0 φ 0 μ1 φ 1 p0 q0 p1 q1
Type-1 error experiments 0.50 5 0.50 5 2.5 2.5 2.5 2.5
  0.50 5 0.50 10 2.5 2.5 5 5
  0.50 100 0.50 100 50 50 50 50
  0.50 100 0.50 200 50 50 100 100
  0.25 5 0.25 5 1.25 3.75 1.25 3.75
  0.25 5 0.25 10 1.25 3.75 2.50 7.50
  0.25 100 0.25 100 25 75 25 75
  0.25 100 0.25 200 25 75 50 150
  0.05 5 0.05 5 0.25 4.75 0.25 4.75
  0.05 5 0.05 10 0.25 4.75 0.50 9.50
  0.05 100 0.05 100 5 95 5 95
  0.05 100 0.05 200 5 95 10 190
Power experiments 0.5 5 0.525 5 2.5 2.5 2.625 2.375
  0.5 5 0.525 10 2.5 2.5 5.25 4.75
  0.5 100 0.525 100 50 50 52.5 47.5
  0.5 100 0.525 200 50 50 105 95
  0.25 5 0.275 5 1.25 3.75 1.375 3.625
  0.25 5 0.275 10 1.25 3.75 2.75 7.25
  0.25 100 0.275 100 25 75 27.5 62.5
  0.25 100 0.275 200 25 75 55 145
  0.05 5 0.075 5 0.25 4.75 0.375 4.625
  0.05 5 0.075 10 0.25 4.75 0.75 9.25
  0.05 100 0.075 100 5 95 7.5 92.5
  0.05 100 0.075 200 5 95 15 185