While on the one hand, the ratios in the boat fire sample were indeed very skewed, and concluding that gasoline was in that sample was probably inappropriate, it raises the age old question: what is an acceptable level of skewing or distortion? Plenty of samples are positive for gasoline even when one or two ratios are slightly out. Where do you draw the line? Obviously this lab went a bridge too far...

I don't think a lack of QC documentation in this area is uncommon. With such a subjective analysis, it is hard to put into words exactly how data interpretation should be carried out. And once it gets written, lab managers have the task of deciding how skewed is too skewed, or how low in concentration is too low, etc. Those are big questions to ask, and sometimes people would prefer not to answer them.

A lot of lab managers are going to be scrambling to their Quality Manuals as a result of this.