The contents of cannabinoid compounds in cannabis flower and cannabis-derived products is the basis of their value, but many people distrust claims of cannabis potency. The standardization of these contents requires their accurate analytical measurement, a process that uses references such as Certified Reference Materials (CRMs) to establish concentrations of cannabinoids. We report the largest collaborative study to investigate the variability between CRMs from multiple vendors and between cannabis testing labs in multiple US states and Canada, contextualized within the results of shelf testing and data from cannabis testing labs. Our presentation compares the sources of variance in the testing of cannabis for cannabinoids, which has been shown to differ from their labeled content by up to 50%, with the discrepancies in CRM concentrations.

32 cannabinoid CRM products from five companies were analyzed at five independent ISO 17025 accredited testing labs. We arbitrarily selected particular lots of CRMs to use as calibration standards, and labs prepared diluted solutions of each CRM to analyze. Errors from the expected concentrations of these prepared solutions derive from the variations between CRMs, convoluted with error from preparation and analysis of solutions.

The five labs returned 127 deduced concentrations with tight precision, with an average standard deviation of 2.8% of the determined value. The concentrations we deduced for the majority of the solutions agreed closely with their certificates, with the root of mean of squared errors equal to 2.73% of the certified values. We discovered that consistency can be achieved between CRM products, and between measurements performed in multiple labs that use different operating procedures. We also found several larger discrepancies in the concentrations of important cannabinoids, including THCA, that unscrupulous labs might exploit.