Skip to main content

Table 1 The effect of number of groups, group size, and ICC on the relative bias ( θ ^ θ θ × 100 MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaadaWcaaqaaGGaciqb=H7aXzaajaGaeyOeI0Iae8hUdehabaGae8hUdehaaiabgEna0kabigdaXiabicdaWiabicdaWaaa@37BB@ ) of estimates.

From: A simulation study of sample size for multilevel logistic regression models

Number of groups Group size ICC % converged γ 00 γ 01 γ 10 γ 11 σ 0 σ 1
30 5 0.04 56 8.77 11.12 15.85 13.26 174.04 55.49
   0.17 68 4.75 11.56 10.70 14.93 24.25 54.55
   0.38 76 3.94 12.22 5.91 14.89 15.82 54.02
  30 0.04 94 0.07 1.09 -1.93 3.57 -7.07 -6.81
   0.17 100 -0.08 3.70 -1.60 3.44 -5.89 -7.02
   0.38 100 -0.18 5.74 -1.72 5.31 -3.55 -6.92
  50 0.04 99 -0.39 0.69 -0.18 5.32 -8.47 -7.71
   0.17 100 -0.43 2.65 -1.70 5.74 -6.25 -7.16
   0.38 100 -0.39 4.62 -2.85 4.49 -5.05 -7.30
50 5 0.04 71 3.82 9.32 4.93 5.88 110.80 35.84
   0.17 86 1.44 8.40 2.34 4.73 11.95 25.55
   0.38 90 1.00 5.96 3.50 6.68 6.80 28.11
  30 0.04 99 0.09 0.21 2.01 2.72 -5.16 -2.60
   0.17 100 -0.32 0.28 2.06 3.69 -2.90 -2.55
   0.38 100 -0.37 -0.56 1.89 3.46 -3.24 -2.81
  50 0.04 100 -0.23 0.08 1.25 2.31 -6.18 -3.54
   0.17 100 -0.33 0.51 1.39 2.56 -3.76 -3.59
   0.38 100 -0.53 0.83 1.29 1.73 -3.19 -3.23
100 5 0.04 87 1.64 0.14 2.42 1.55 47.87 7.64
   0.17 98 0.47 -0.50 1.11 0.89 1.84 3.46
   0.38 98 0.95 0.25 2.25 0.41 2.23 8.82
  30 0.04 100 -0.02 -0.12 0.31 0.41 -5.13 -1.25
   0.17 100 -0.11 0.36 -0.06 0.96 -2.14 -1.17
   0.38 100 -0.21 1.04 -0.96 1.07 -2.06 -1.30
  50 0.04 100 0.03 0.35 -0.09 0.48 -3.28 -2.01
   0.17 100 -0.06 0.71 0.05 0.63 -2.12 -1.34
   0.38 100 -0.14 0.62 0.50 1.18 -1.73 -1.52