As often happens, displaying the intervals as a table of numbers is not very informative. We find it much more effective to plot these intervals using > plot( intervals(fm2Orth.lis) )
# Figure 4.5
The individual confidence intervals in Figure 4.5 give a clear indication that a random effect is needed to account for subject-to-subject variability in the intercept. Except for subject M13, all confidence intervals for the slope overlap, so perhaps this parameter can be regarded as a fixed effect in the mixed-effects model. We will explore these questions in 4.2.1, while describing the lme function. To further illustrate the capabilities of lmList, we consider data on radioimmunoassays of the protein Insulin-like Growth Factor (IGF-I) presented in Davidian and Giltinan (1995, 3.2.1, p. 65). The data are from quality control radioimmunoassays for ten different lots of a radioactive tracer used
