ccb60/LCensMeans: Calculate Condtional Means of Left Censored Data

This tiny package exposes functions for estimating the conditional mean of censored values for left censored environmental data. It is still common practice in envrionmental data analysis to replace "non-detects" with essentially arbitrary values, usually 0, one half of the detection limit, or the detection limit. None of these approaches rests on any statistical foundation, and they ignore the information available from knowing that a sample concentration was below a certain value. The goal of this package is to replace non-detects with a maximum likelihood-derived estimator of the mean of the unobserved (censored) observations. It estimates those means by calculating maximum likelihood estimates of parameters of a log normal distribution that could have produced the observed censored data, simulating a draw from that distribution, and calculating an average of values that would have been censored.

install.packages("remotes")
remotes::install_github("ccb60/LCensMeans")