What question did this study set out to answer?

To evaluate statistical methods for estimating the mean concentration and standard deviation of contaminants when a proportion of measurements fall below the limit of detection.

January 1, 1990Applied Occupational and Environmental Hygiene

Estimation of Average Concentration in the Presence of Nondetectable Values

View Full Paper

Ask AI

Bookmark

Authors

RHRichard HornungCincinnati Children's Hospital Medical Center LRLaurence D. ReedBritish Broadcasting Corporation (United Kingdom)

Discussion

Member takes

Implication

Simulation study evaluates estimation methods for left-censored lognormal contaminant samples, demonstrating that maximum likelihood and tailored substitution reliably recover mean concentrations.

Key Points

To evaluate statistical methods for estimating the mean concentration and standard deviation of contaminants when a proportion of measurements fall below the limit of detection.
Simulated left-censored data via computer models under the assumption of a lognormal distribution over time.
Evaluated three estimation approaches: a maximum likelihood estimation method and two substitution techniques based on the detection limit (L).
Maximum likelihood estimation produced unbiased estimates of both the mean and standard deviation across varied simulation conditions, but required complex and laborious computations.
The newly proposed substitution technique yielded more accurate estimates of the mean and standard deviation than the L/2 substitution method when data were not highly skewed.
The L/2 substitution method proved preferable when contaminant data exhibited high skewness, characterized by a geometric standard deviation of approximately 3.0 or greater.

Cite This Study

Hornung et al. (1990) studied this question.

synapsesocial.com/papers/69d9aebc387cf7069868515e https://doi.org/10.1080/1047322x.1990.10389587

View Full Paper

Ask AI

Bookmark

Implication

Simulation study evaluates estimation methods for left-censored lognormal contaminant samples, demonstrating that maximum likelihood and tailored substitution reliably recover mean concentrations.

Estimation of Average Concentration in the Presence of Nondetectable Values

Discussion

Member takes

Implication

Key Points

Cite This Study

Estimation of Average Concentration in the Presence of Nondetectable Values

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider