Re: heavy lab tests

I worked as an analytical chemist in a petrochemical laboratory for eight years. One thing about machines, they do not always tell the truth. We would always run a standard sample every so many runs of a particular analyte, and often the instruments would be found to be out of calibration, for no apparent reason. After some routine troubleshooting, we would usually get the device back into specification. And, for some elements, the number you get depends on the test ran. For example, phosphorous is very difficult to analyze for, you get two different values from two different tests. So anytime I see numbers that are high or low out of the normal range, I typically insist on first repeating the test and if out of spec again, run it a third time. If still out of range, run the standard, then the sample, then the standard again, so as to bracket the suspect value between the results of two known samples. Do they do this in the example given ? We tend to just trust numbers from machines, but it is not uncommon for machines to give inaccurate values. Tweet