Skip to main content

Advertisement

Table 5 50% change of selection bias effect and linear time trend effect (η,θ) on probability of type I error for complete randomization (CR) and big stick design (BSD)

From: ERDO - a framework to select an appropriate randomization procedure for clinical trials

Randomization Selection Linear-Time Type I Error P RP (ω≤0.05)
Procedure Bias Trend Bias Probability  
[R P] η θ [mean]  
CR 0.04 0.13 0.050 0.52
CR 0.09 0.26 0.050 0.53
CR 0.14 0.39 0.051 0.56
BSD(3) 0.04 0.13 0.051 0.10
BSD(3) 0.09 0.26 0.054 0.11
BSD(3) 0.14 0.39 0.059 0.10
BSD(4) 0.04 0.13 0.050 0.32
BSD(4) 0.09 0.26 0.052 0.34
BSD(4) 0.14 0.39 0.053 0.34
BSD(5) 0.04 0.13 0.050 0.45
BSD(5) 0.09 0.26 0.051 0.46
BSD(5) 0.14 0.39 0.051 0.47
BSD(10) 0.04 0.13 0.050 0.52
BSD(10) 0.09 0.26 0.050 0.53
BSD(10) 0.14 0.39 0.050 0.57