Table 4
Comparison of nuclear back scattering of light in the crystalline lens between systemic treatment with simvastatin and placebo
| Sources of variation | Degree of freedom | Mean square (GSU)2 | Expected mean square | Test statistic (Ff1:f2:0.95) |
| Treatments | 1 | 0.3327 | ς2 ε+cς2 B+bcκ2 α | 0.93 (3.84) |
| Times | 2 | 0.1635 | ς2 ε+ς2 Bγ+abκ2 γ | 10.64 (3.00) |
| Patients | 158 | 0.3582 | ς2 ε+cς2 B | |
| Int (Tr − T) | 2 | 0.0137 | ς2 ε+ς2 Bγ+bκ2 αγ | 0.89 (3.00) |
| Int (P − T) | 316 | 0.0154 | ς2 ε+ς2 Bγ |
-
Int (Tr − T) = interaction between treatments and time points. Int (P − T) = interaction between patients and time points. κ2 = a factor corresponding to the systematic shift because the indexed fixed source; α = treatment; γ = time points. ς2 = The expected variance for the indexed random source; B = patients, ε = measurement error. a = The number of treatments = 2; b = the number of patients within a treatment = 80; c = the number of time points for each patient = 3.
