Quote:
Originally Posted by BillyMac
I'm an environmental chemical analyst. In analyzing samples, for every twenty samples in a batch we have to randomly select sample for a duplicate analysis. If the results of the actual sample and the "rep" (replicate, or duplicate) are the same, this is one way for us to confirm that our reagents, equipment, instruments, procedures, etc. are working properly, and we can then, and only then, proceed to report the results of all twenty samples in that particular batch to our clients. In our industry, if we get a "rep" result that is 20% or less, or 20% or more, than our actual sample, it is considered that we have successfully duplicated the result, and the batch of twenty samples successfully passes. I happen to work with an instrument that measures accurately down to one milligram per liter (one part per million). The minimum detectable level of my instrument is 0.02 milligram per liter (0.02 parts per million).
|
In my high school science classes I was taught that the result of any calculation is only as precise as the least precise factor used. So if you have two weights, one measured in 10ths of a gram and the other in grams, no matter how certain you are of the first measurement, the result must be rounded to the nearest gram.
So what are we saying when we allow both 1/10 second precision and also nearest second (roughly) accuracy?