0.85. 0.90. 0.95. 1.00. 0.0. 0.2. 0.4. 0.6. R. W. 2. P o wer q. 0.75. 0.80. 0.85. 0.90. 0.95. 1.00. 0.0. 0.2. 0.4. 0.6. RL. 2. P o wer. Adj.est. H00. H01. H02. Unadj.est.
1.00
0.95
0.95
Adj.est.
0.90
0.90
H00 H01 H02
Power
Power
1.00
0.85
●
●
0.80
0.80
0.75
H00 H01 H02
0.75
0.0
0.2
0.4
0.6
0.0
R2W
0.2
0.4
0.6
R2L
1.00
1.00
0.95
0.95
Adj.est.
0.90
H00 H01 H02
0.90
Power
Power
Unadj.est.
0.85
●
0.85
Unadj.est.
0.85
●
●
●
0.80
0.80
0.75 0
1
2
3
4
H00 H01 H02
0.75 100
dY (years)
200
300
400
500
accrual rate (pts/yr)
Reject H00 under (d) Reject H01 under (b) Reject H02 under (c) 1.00
dL / dY Figure 6: Change in power under simulation studies 1-5. The power of H00 /H01 /H02 is the probability to reject at least H00 /H01 /H02 under effect setting (d)/(b)/(c). Since the results corresponding to unadjusted estimator do not change as the design characteristics are varied, they are marked only once next to the vertical axis using the 35 circle, square, diamond, and triangle symbols.