Info

Random: level-1 variance = 0.22. -2 log-likelihood(ML) = 625.4.

Random: level-1 variance = 0.22. -2 log-likelihood(ML) = 625.4.

Predicted growth curves for 26 boys

Predicted growth curves for 26 boys

Age -12.25 years

Figure 5.

Age -12.25 years

Figure 5.

and some who have been measured several times at irregular intervals. This flexibility, first noted by Laird and Ware [10], means that the multilevel approach to fitting repeated measures data is to be preferred to previous methods based upon a multivariate formulation assuming a common set of fixed occasions [19,20].

In these models it is assumed that the level-1 residual terms are independently distributed. We may relax this assumption, however, and in the case of repeated measures data this may be necessary, for example where measurements are taken very close together in time. Suppose we wish to fit a model that allows for correlations between the level-1 residuals, and to start with for simplicity let us assume that these correlations are all equal. This is easily

Table II. Height modelled as a fourth-degree polynomial on age, including a seasonal effect and serial correlation. REML estimates.

Fixed effects

Estimate

Standard error

Intercept

Was this article helpful?

0 0

Post a comment