Sample Size Determination for Outcomes from Two-Level Trials 159
between intervention arms the difference in the slopes or trends in the outcome
over time. There are two ways to model slope variations across subjects: fixed
and random slope effect models. For the former, subjects are assumed to have
a common slope (Section 5.4.1) and for the latter, each subject is assumed
to have his/her own trend in outcome overtime (Section 5.4.2). In this case,
the variability in outcomes is greater than those in the fixed slope models,
especially at larger time points. We compare the required sample sizes with
fixed and random slope models in Section 5.4.3 and discuss the testing of main
effects at the end of study in Section 5.4.4. We note that Chapters 3 and 4 also ...