Modeling observations with a detection limit using a truncated normal distribution with censoring

Table 2 Expected Percentages of Truncation and Censoring in Single Mean Simulation Study

μ	σ	Expected Censoring %	Expected Truncated %	Censoring:Truncation Ratio ^∗
1.1	0.50	15.17%	1.39%	10.91
	0.45	13.18%	0.73%	18.05
	0.40	10.76%	0.30%	35.87
1.0	0.50	19.95%	2.28%	8.75
	0.45	18.23%	1.31%	13.92
	0.40	15.96%	0.62%	25.74
0.9	0.50	25.42%	3.59%	7.08
	0.45	24.24%	2.28%	10.63
	0.40	22.47%	1.22%	18.42
0.8	0.50	31.44%	5.48%	5.74
	0.45	31.04%	3.77%	8.23
	0.40	30.15%	2.28%	13.22
0.7	0.50	37.84%	8.08%	4.68
	0.45	38.38%	5.99%	6.41
	0.40	38.64%	4.01%	9.64

Note that censoring threshold set at ν=0.61 and truncation value at a=0.
^*This column compares the expected percent of censoring to the expected percent of truncation, e.g., when μ = 1.1 and σ = 0.50 we expect 10.91 times more censoring than truncation.

ISSN: 1471-2288