Thinking outside the curve, part I: modeling birthweight distribution

Table 2 Estimating Parameters in a Four-Component Mixture Model

Quantity	p₁	p₂	p₃	p₄
[average of 25 estimates]	.007	.182	.758	.052
[standard deviation of 25 estimates]	.001	.039	.037	.008
[bias adjustment]	.001	.041	.032	.009
Confidence interval	(.005, .010)	(.092, .272)	(.681, .836)	(.033, .071)
Quantity	μ₁	μ₂	μ₃	μ₄
[average of 25 estimates]	832	2772	3170	3804
[standard deviation of 25 estimates]	46	103	7	25
[bias adjustment]	34	80	9	38
Confidence interval	(741, 924)	(2565, 2979)	(3152, 3187)	(3735, 3873)
Quantity	σ₁	σ₂	σ₃	σ₄
[average of 25 estimates]	210	740	417	413
[standard deviation of 25 estimates]	28	23	10	38
[bias adjustment]	30	23	7	46
Confidence interval	(146, 274)	(688, 792)	(398, 436)	(321, 506)

Parameters in a 4-component normal mixture model for birthweight distribution are estimated, based on 25 samples of size 50000 from the population of white singletons born to heavily smoking mothers. Interval estimates are constructed using Equations (7) and (8) with C ₀ = 2.5 and φ = .2465.

ISSN: 1471-2393