Fixed Effects

Figure 1. Estimates of fixed effects for fixed effects-only model
Estimates of fixed effects for fixed effects-only model

The estimate of [TIME=1] suggests that patient weights were about 8 pounds higher at the beginning of the study. However, the result has a significance value greater than 0.10, suggesting that the observed difference of 8 pounds could be due to chance, and throwing the utility of the diet into doubt. The problem is that the current model does not account for the fact that these are repeated measurements on patients. The test of the fixed effect [TIME=1] in this model is equivalent to a 2-sample t-test, and so the great amount of patient-to-patient variation in natural bodyweight obscures the effect of the diet.

Next