Skip to main content

Advertisement

Table 4 Percentages of selecting \(\mathcal {M}_{1}\) according to the BIC on N=500 datasets, given t v =(0,1,2,4,6,8,10,12) and \(\sigma _{0}^{2}=1.5\)

From: Item response models for the longitudinal analysis of health-related quality of life in cancer clinical trials

Parameter Scenarios  
Values AM using δ ne CM using δ fa CM using δ ne AM using δ fa
\(\sigma _{1}^{2}\) β 1 LMM AM CM LMM AM CM LMM AM CM LMM AM CM
0.2 −0.3 0 0 0 0 0 0 0 0 0 0 0 0
0.2 0.3 0 0 0 0 0 0 0 0 0 0 0 0
0 −0.5 97.7 99.3 56.49 100 94.6 93.0 100 61.3 95.7 100 99.7 89.5
0 −0.3 99.0 100 33.0 100 88.6 93.3 100 36.3 94.9 100 100 83.3
0 −0.2 100 99.6 49.3 100 94.6 93.8 100 71.7 95.8 100 99.6 79.0
0 −0.1 98.7 95.7 94.8 100 98.7 89.6 100 99.0 90.4 100 100 88.1
0 0.0 95.6 100 94.6 99.0 99.7 91.8 99.0 99.7 89.7 97.0 99.7 94.4
0 0.1 83.0 100 94.8 93.3 100 92.6 97.0 100 90.9 87.3 100 94.7
0 0.3 98.3 99.6 90.6 100 99.6 89.1 100 100 93.7 100 99.6 93.8
0 0.5 100 100 94.3 100 99.3 94.7 100 100 97.6 100 100 97.2
  1. The (adjacent,logistic, Z 1,U a ) a=1,2 models and the (cumulative,logistic, Z 1,U a ) a=1,2 models are denoted respectively by AM and CM. For the random component, U 1 if \(\sigma _{1}^{2}=0\) and U 2 if \(\sigma _{1}^{2}>0\)