Skip to main content
. Author manuscript; available in PMC: 2016 May 10.
Published in final edited form as: Stat Med. 2015 Jan 23;34(10):1621–1633. doi: 10.1002/sim.6414

Table 3.

Last Time Point Treatment Effect Testing with Baseline as a Covariatea

Data Model Interaction Statisticsb Power Type I Error
Full (CS) Strong ζ t 0.531 0.059
ζ ANCOVA 0.573 0.053
ζ FUN 0.579 0.053
ζ MSL 0.951 0.042
Moderate ζ t 0.531 0.059
ζ ANCOVA 0.573 0.053
ζ FUN 0.579 0.053
ζ MSL 0.988 0.042
Weak ζ t 0.531 0.059
ζ ANCOVA 0.573 0.053
ζ FUN 0.579 0.053
ζ MSL 0.988 0.042
Main Effects (CS) None ζ t 0.531 0.059
ζ ANCOVA 0.573 0.053
ζ FUN 0.579 0.053
ζ MSL 0.942 0.042
a

results based on 1000 simulations with 100 subjects at each treatment arm

b

model selection via BIC with 200 bootstrap resampling

ζt: t-test for last time point effect

ζANCOVA: test last time point effect via ANCOVA model

ζFUN: test last time point effect via full mixed effect model with UN covariance

ζMSL: model selection test statistics for testing last time point effect