A batch of material is sampled (selecting some portion) and that specimen is sealed in a container for submission (in CA, this is done by the lab customer). The selected specimen is submitted to the lab where it is homogenized (ground and mixed well) and sampled (analyst selects a portion of the ground, homogenized specimen), then extracted with solvent, followed by dilution and filtering to produce a small vial of liquid to be tested on the HPLC machine.
If the same final small vial is tested repeatedly (several separate injections/analyses), the HPLC results are within about ~1% relative standard deviation. (I.e., a vial with a concentration representing 20.0% THCA would come up between 19.8% and 20.2% almost every time.) This represents the variation associated with the HPLC machine itself.
If the homogenized specimen is re-sampled by trained laboratory personnel, then extracted, diluted and filtered again prior to re-analysis, then repeats from the same specimen generally fall within ~5% relative standard deviation. (I.e., a specimen with a real concentration of 20.0% THCA in a flower would come up between 19% and 21% almost every time.) This represents a combination of the sampling/extraction/dilution/filtering process variation PLUS the HPLC machine variation.
Collecting multiple specimens from the same large batch of material can yield results that vary considerably more than that, depending on the variation of different buds/regions in the batch, and the amount of material submitted for homogenization. For flowers, small specimens from batches where the material comes from many different plants (even if started from "clone" cuttings), can lead to variations of 25% relative or even more! Differences in growing conditions or top bud/bottom bud position on the plant, etc. can make different buds in the same batch vary by 25% or more. In this case, the results from two specimens drawn from the same batch can, therefore, differ by more than 25%. (I.e., a batch with a real average concentration of 20.0% THCA in a flower could come up even less than 15% or more than 25% for different specimens from the same batch.)
Harvested, cured plant material has a great deal of variability, no matter what. Assessing the variability within a batch would require multiple samples and multiple tests, from which, good estimates of both the batch average and batch variability could be determined. From a single test of a single specimen, only an estimate of the average value for the specimen is possible.
For some purposes, the added expense of multiple specimens and multiple tests from the same batch may be worth the cost. Please contact us at email@example.com and ask to speak to one of our experts about sampling and range estimates if you have further questions.