Table 3. The percentage of cases (out of 1000 simulations) in which the given distributions were found to be different at each time point with simulated data.

Predictive check 1 uses a sample size equal to the number of samples in the original data set at each time point. A sample size of 500 at each time point was used for predictive check 2.

Predictive check 1
Time (weeks)	0	1	2	3	4	6	8	10	12
Data vs. log exp	32.3	4.9	2.9	2.7	8.4	7.1	6.9	0.2	42
Data vs. log bi-exp	5.9	2.9	3.3	2.6	5.0	0.8	14.4	0.8	12.7
Log exp vs. log bi-exp	11.8	4.7	5.6	6.2	8.7	10.1	5.3	3.6	11.9
Predictive check 2
Time (weeks)	0	1	2	3	4	6	8	10	12
Data vs. log exp	98.8	0.1	0	0	0.6	0.8	0	0	100
Data vs. log bi-exp	0	0	0	0	0.4	0	2.3	0	15.4
Log exp vs. log bi-exp	99.5	36.3	10.8	61.9	89.2	92.2	40.3	14.2	97.5