NAs and its meaning in the context of Environment Exposure Analysis

I am working with Environmental Exposure data, BPA, Phenols, Phthalates etc and how they effect women health.

Most of the lab results for these chemicals are entered as 0s in the datatabase, 0.0000 mg/ML. I guess this is due to very low concetrations of these chemicals in the womens blood or urine.

During Analysis phase, I am log transforming these Environmental Exposures data. Plotting box plots or running models etc. When I log transform this data, the 0s are converted to NAs and they are excluded from the dataset.

I am wondering what is the consequence of excluding these observations because they turned to NA when they are actually 0 which indicates these women dont have any significant traces of these chemicals ?



I often use transformations such as cube root, which has no problem with zeros (if you can act as if they are true zeros).

Thanks Frank, I will test that approach.

Question Frank, why cube root ? why not fifth root ^(1/5) or tenth root ^(1/10) ?

Zero seems more logical than NA ( = excluding observation).

1 Like