Reading a Cannabis Lab Report: What a Measured Number Is—and What It Is Not

A cannabis certificate of analysis describes the sample that was tested. Its reported number is not certainty about every unit made or sold under the same product name.

By Cannible Editorial

2026-08-08

A cannabis certificate of analysis describes the sample that was tested. Its reported number is not certainty about every unit made or sold under the same product name. As of August 20, 2026, the clearest way to read a report is to separate three questions: what the laboratory measured in that sample, what the method could distinguish from blanks, and what—if anything—the result says about the product beyond the tested sample. That position can change as testing methods, data, and rules change. 1 2 3

This article is general information, not individualized legal, medical, business, or compliance advice. Whether any of it applies to a particular person, product or business depends on the facts and on applicable state law; consult a qualified professional.

What a detection limit means

The method detection limit is the minimum measured concentration of a substance that can be reported with 99% confidence as distinguishable from method blank results. In practical terms, it is a statistical threshold built around the method’s background results, not a biological or physical boundary at which the substance becomes absent. 1

The threshold belongs to a defined analytical method and includes the laboratory’s sample-processing steps. The procedure can use spiked samples and method blanks, and the initial detection limit is the greater of the value based on spiked samples or the value based on blanks. A matrix-specific detection limit may also be determined for a particular sample matrix. 1

“Not detected” does not mean zero

A result reported as “not detected” is not a measurement of zero concentration. The evidence distinguishes a numerical result below the current detection limit from an “ND” result: numerical results can include positive or negative values below the detection limit, while “ND” is not a numerical result and is commonly used when a chromatographic peak is not present. 1

That distinction matters when reading a report. “Not detected” means the reported test did not provide a reportable detection under that method and result format. It does not establish that the substance is absent at every concentration, nor does a result below the detection limit establish a precise concentration. 1

What label-accuracy studies actually found

One study tested 23 cannabis samples from 10 dispensaries in Colorado’s Front Range using high-performance liquid chromatography and compared measured THC potency with the potency printed on the packages. The average observed THC potency was 14.98% with a 2.23 percentage-point standard deviation. In that set, the average observed potency was 23.1% lower than the lowest label-reported values and 35.6% lower than the highest label-reported values; about 70% of samples were more than 15% below the label numbers, and three samples had only half of the reported maximum THC potency. These findings apply to the 23 tested samples, not automatically to all Colorado cannabis or every unit of those products. 2

A separate study purchased and analyzed 277 products from 52 Colorado dispensaries: 178 flower products and 99 concentrates. It classified a product as accurately labeled when observed THC was within plus or minus 15% of the labeled THC. By that study’s threshold, 96.0% of concentrates and 56.7% of flower products fell within the range. Observed THC was significantly lower than labeled THC in both categories, and the study reported that nearly all tested concentrates met the threshold while flower products frequently did not. Those percentages describe the products in that Colorado sample. 3

The two studies used different samples and report different accuracy summaries, so their figures should not be combined into one universal accuracy rate. Together, they show why a printed potency figure and a laboratory measurement should be treated as claims about a tested product sample, with the study’s product type, location, sample size, and accuracy rule kept in view. 2 3

Potency is not effect

A laboratory potency number should not be converted into a prediction of subjective experience. In a controlled within-person comparison, 574 current cannabis users who reported using both marijuana and concentrates answered questions about the subjective effects of each. The survey covered affect, cognitive function, psychotic-like experiences, physiological effects, and reduced consciousness. Participants reported using marijuana five to six times per week and concentrates slightly more than once per month. 4

In that study’s within-person comparison, participants rated marijuana as producing greater overall positive effects than concentrates: the reported means were 5.6 for marijuana and 4.5 for concentrates, with Cohen’s d of 0.75, paired t(561) of 14.67, and p less than .001. The study also reported greater positive affect and enhanced cognitive function for marijuana, minimal negative effects for both, and marijuana selected as the preferred type by 77.5% of participants. These are the study’s reported group-level subjective-effect findings, not a formula linking a COA percentage to an individual’s experience. 4

How regulation fits into the report

In California, the Department of Cannabis Control creates regulations for cannabis businesses, and those regulations specify testing that each product must pass before it can be sold. California’s regulations also include packaging requirements intended to inform consumers about what is inside products. A California testing requirement is a rule for products sold in California; it does not establish that the same requirement applies in another jurisdiction. 5

When reading a report, keep the jurisdiction, product type, tested sample, analyte, units, detection limit, and reported result together. If a real purchase or product-specific decision turns on how to interpret a disputed number, take the report and the product information to a qualified professional who can evaluate the method and the particular result. 1 3