Skip to main content

Table 3 Confidence Interval Coverage Probabilities in Simulation Study

From: Thinking outside the curve, part II: modeling fetal-infant mortality

C Population Size Bias adjustment included Bias adjustment omitted
   Number & Percentage of Intervals Containing Targets (mortality risks) Number & Percentage of Intervals Containing Targets (mortality risks)
2.0 200,000 69 (57.5) 26 (21.7)
  1,000,000 92 (76.7) 26 (21.7)
  Infinite 92 (76.7) 43 (35.8)
2.5 200,000 78 (65.0) 29 (24.2)
  1,000,000 100 (83.3) 37 (30.8)
  Infinite 96 (80.0) 47 (39.2)
3.0 200,000 84 (70.0) 32 (26.7)
  1,000,000 106 (88.3) 44 (36.7)
  Infinite 102 (85.0) 56 (46.7)
3.5 200,000 89 (74.2) 35 (29.2)
  1,000,000 111 (92.5) 54 (45.0)
  Infinite 110 (91.7) 60 (50.0)
4.0 200,000 93 (77.5) 44 (36.7)
  1,000,000 116 (96.7) 63 (52.5)
  Infinite 114 (95.0) 65 (54.2)
4.5 200,000 97 (80.8) 48 (40.0)
  1,000,000 117 (97.5) 72 (60.0)
  Infinite 115 (95.8) 76 (63.3)
5.0 200,000 102 (85.0) 57 (47.5)
  1,000,000 117 (97.5) 79 (65.8)
  Infinite 115 (95.8) 90 (75.0)
  1. The row with "C" = 2 and "Population size" = 200,000 identifies the numbers and percentages of confidence intervals containing their targets of mortality risks at selected birthweights (three for each of four mixture components), based on 10 repetitions in each of which 25 samples of size 50,000 were simulated from a 4-component normal mixture. Results under the heading of "Bias adjustment included" are based on Equation (6) with C = 2. Results under the heading of "Bias adjustment omitted" are based on Equation (5) with C = 2. The 25 samples of size 50,000 had overlap consistent with a population size of 200,000. Other rows correspond to different choices of C and/or population sizes.