There’s a perception among cannabis consumers that higher THC content in cannabis products is synonymous with value. Yasha Kahn, co-founder of MCR Labs, explained a recent episode of the Noid Knowledge podcast (1–3) that consumers are more likely to purchase products with high THC content because they’re generally thinking in terms of cents per milligram of THC. “If you have an option of 30% THC flower or 20% THC flower, you’re more likely to go for the 30%. And you’re more likely to pay more for that 30% and be happy with it,” he explained.

This perception breeds the incentive for industry to cultivate and sell high-THC products. That potency is verified though laboratory testing, but if a cultivator tests their product at two different labs and the results of one of the labs shows higher THC content, the cultivator is more likely to use those results because that will help them sell to the retailer who is looking for higher THC product. While, in isolation, differences in the results from competing labs do not necessarily indicate manipulation of test results, on a larger scale, there is an observable pattern in which labs feel pressure from clients to provide preferable results. If they don’t, someone else will. This has been dubbed “lab shopping,” and it’s something Khan has experienced first-hand.

“We had a lab in Pennsylvania, an excellent lab performing fast and accurate testing. Then some of our clients, cultivators, would come to us and say, you failed us, but this other lab offered not to fail us or didn’t fail us for anything,” explained Kahn. Instead of trying to figure out the discrepancies or remediate any issues such as mold, it was easier for those cultivators to just use the labs that passed them. Similarly, Kahn said that in Massachusetts, clients began coming to them demanding better potency results. “[They] would say that another lab is giving them much higher potency results. If we don’t do the same, if we can’t match the potency results that they get at the other lab, they’re going to leave us for that lab,” said Kahn. At the end of the day, while THC inflation is deceptive, there is also a larger public health concern in which some labs are passing products that are in excess of the threshold for contaminants such as mold.

This anecdotal experience is being verified with data. In the case of THC inflation, a 2023 study published in PlosOne (4) found that when testing the THC potency of 23 samples purchased from 10 different Colorado dispensaries, the tested potency was substantially lower than the label claims. The researchers found that on average, the observed THC potency was 23.1% lower than the lowest values reported on the label, and 35.6% lower than the highest values reported on the label. Seventy percent of the samples overall were greater than 15% lower in THC than label claims. While the researchers acknowledge that the exact source of the discrepancies is difficult to pin down, they explain that there are a number of factors at play, including economic incentives for high-THC products, as well as a lack of standardized testing protocols and limited regulatory oversight.

Jeff Rawson, president of the Institute of Cannabis Science, explained in a presentation (5) titled, “Market Audits Combat Cannabis Misinformation” for ASTM International’s D37 Virtual Symposium in 2023, that a big problem is the lack of analysis of compliance testing. An article of the same name co-authored by Rawson and Kahn was published in the Journal of Testing and Evaluation in July of 2024 (6). “Compliance testing happens at a different stage than consumption. It’s one step of quality control, but it’s not a quality assurance program,” Rawson explained. “Furthermore, the compliance testing that’s done is not evaluated systematically, or rarely. I don’t see a lot of people doing big data analysis on it right now, and products from the retail space aren’t measured at all. This is an engineering principle: You don’t control what you don’t measure. So, if you’re never measuring the consumer experience by checking products in the marketplace, then you’re not really controlling the quality of those products.”

Rawson offered Washington state as an example, in which seven of the 25 licensed labs have been suspended or cited for some kind of infraction, and all these infractions, he says, leave signature in the data they produce. “In fact, some of these infractions were discovered by actively monitoring compliance testing data and noticing that some of the labs had discrepant practices,” he explained. For example, among these Washington labs with discrepant practices, Rawson pointed out a pattern in which labs with above-average THC values saw a monthly increase in market share, indicating that more cultivators were going to this lab each month for preferable results. This illustrates how lab shopping happens and hurts honest labs. Some of the labs shown by Rawson to have increased their market share with high THC values were eventually shut down.

Knowing the importance of data and what it can reveal, Kahn and MCR Labs used Freedom of Information Act (FOIA) requests to acquire testing lab data from 20 of the 39 states that have cannabis regulations. This compilation of data includes information from 2.4 million flower samples, 144 labs, and has nearly 68 million data points. To give credit to policy makers and regulators, this is all thanks to the fact when they drafted regulations, they not only made testing mandatory but also required third party labs to submit all their test results to the state. That makes it public information. Now, to what extent these states monitor and use this data is another story.

Figure 1: Analysis of cannabinoids from biological fluids using HPLC.

Data has the ability to reveal an inconvenient truth. Regulators set limits on factors like total yeast and mold, for example, but these limits are not relevant when you’re dealing with result manipulation. “There are labs you can find that will pass you in almost every state, not every, but almost every state. And so…limits are just not a relevant topic up until the result manipulation issue [is addressed],” says Kahn. “Once we deal with that problem…then we should address action limits. And there needs to be public health officials that make that decision on what the limits should be. I can see in the data, in labs that test honestly, around 12 to 16% of flowers will fail at the 10,000 colony forming units action limit. Around 4 % will fail at the 100,000 colony forming units action limit.”

Figure 2: Analysis of cannabinoids from biological fluids using HPLC

Dealing with recalls not only costs money for stakeholders, but also tax revenue, which is another layer of incentive that can explain how labs and cultivators get away with result manipulation. Kahn explains that when filing FOIA request, some states were easier to work with than others. “The states that I had to sue are the ones with typically the worst data despite robust data. As in, clearly they know that there’s something to hide and they do not want this out in the open,” said Kahn. “I think it’s also that regulators are in a position where they will be held responsible if problems are found. Even if it’s not their fault. Maybe they’re not given the ability, the resources, or the jurisdiction to act on some of those things, and it’s better to like, you know, stay silent, let’s hope this blows over and then move on…In other states they really are doing as good a job as they can under the circumstances and are just not allowed to do anything. The market’s making money for the state, don’t touch it.”

Table 1: Aspegillus detection rates in Mississippi testing facilities.

As an example of the type of information these FOIA request provided, MCR Labs provided Cannabis Science and Technology with figures demonstrating the purported discrepancies between labs. These particular figures focus on Mississippi. Figures 1 and 2 show a distribution of total THC measured by two different testing facilities. In Figure 1, we see a pretty standard bell curve, while in Figure 2 we see a significant cliff in the number of products testing above 30% THC. Why such a steep cliff at 30% THC? Well, Mississippi law states that the potency of cannabis flower and trim cannot exceed 30% total THC. Looking at Aspergillus, Table 1 depicts the detection rate for Aspergillus from three different testing facilities. Out of 1,636 samples, facility 1 has a 0% detection rate. Facility 2 had a 0.23% detection rate out of 2,151 samples, while facility 3 had a 2.50% detection rate out of 320 samples tested. That means that statewide, Mississippi’s detection rate is 0.32% out 4,107 samples. Compared to other states, Mississippi has among the lowest detection rates for Aspergillus (Figure 3). There’s so much more data that can be dug into and interpreted to find patterns of manipulation, and finding these discrepancies is a great first step in solving the problem.

Figure 3: Aspergillus detection rate per state.

On the bright side, some states have already taken action to address manipulation. In the case of Mississippi, regulators did shut down one of the biggest offenders, and pulled product tested by that company off the shelf (7). California’s Department of Cannabis Control, for example, implemented Business and Professions Code section 26100(f)(2) that established standard cannabinoid test methods, “including standardized operating procedures, that shall be utilized by all licensed testing laboratories in California.” That means that January 1, 2024, all testing labs had to use this method to test for cannabinoids and there has been a noticeable dip at the retail level in the THC levels of cannabis products (8). While this does create a more level playing field for labs, this action does not address other factors such as total yeast and mold, which can affect the quality of products and pose a safety risk to consumers. There’s definitely more work that needs to be done to prevent result manipulation, and a data driven approach may be the best way to develop effective policies.

References

1. Noid Knowledge. Ep 24, Part I: Data Transparency in Cannabis Testing with Yasha Khan. Cannabis Science and Technology. November 21, 2024. https://www.cannabissciencetech.com/view/ep-24-part-i-data-transparency-in-cannabis-testing-with-yasha-khan (accessed 2025-04-16).
2. Noid Knowledge. Ep 24, Part II: Data Transparency in Cannabis Testing with Yasha Khan. Cannabis Science and Technology. December 12, 2024. https://www.cannabissciencetech.com/view/ep-24-part-ii-data-transparency-in-cannabis-testing-with-yasha-khan (accessed 2025-04-16).
3. Noid Knowledge. Ep 24, Part III: Data Transparency in Cannabis Testing with Yasha Khan. Cannabis Science and Technology. December 26, 2024. https://www.cannabissciencetech.com/view/ep-24-part-iii-data-transparency-in-cannabis-testing-with-yasha-khan (accessed 2025-04-16).
4. Schwabe, A.L.; Johnson, V.; Harrelson, J.; McGlaughlin, M.E. Uncomfortably high: Testing reveals inflated THC potency on retail Cannabis labels. PloS One. 2023, 18 (4), e0282396. DOI: 10.1371/journal.pone.0282396
5. Jeff Rawson. Market Audits Combat Cannabis Misinformation, Jeff Rawson, ASTM D37 Virtual Symposium 2023. October 19, 2024. https://www.youtube.com/watch?v=3axXcu8KpCE. (accessed 2025-04-16).
6. Turkington, V.; Kahn, Y.; Keller, L.; Willis, J.; Hamilton, A.; Utech, H.; Toth, J.; Rawson, J. Market Audits Combat Cannabis Misinformation. J. Test. Eval. 2024, 52 (6), 3203-3215. DOI: 10.1520/JTE20230665
7. Saler, C. Mississippi medical cannabis companies to receive $6.3M from lawsuit against Rapid Analytics. June 4, 2024. https://www.supertalk.fm/mississippi-medical-cannabis-companies-to-receive-6-3m-from-lawsuit-against-rapid-analytics/ (accessed 2025-05-01).
8. Divine, J. THC scores coming back down to reality in California in 2024. February 27, 2024https://www.leafly.com/news/industry/thc-testing-rules-bite-down-california (accessed 2025-05-01).