Most Americans live in states that permit some people to sell, buy and consume cannabis, despite ongoing federal prohibition. Twenty-six states have medical marijuana programs, and nine states explicitly allow recreational cannabis — though cannabis storefronts haven't opened yet in all of them.

And each election cycle, voters expand the number of states that allow marijuana possession and use. Apart from Vermont, where legislators passed a limited law legalizing use, states have changed their marijuana laws by popular ballot initiatives. To most voters, initiatives offer a binary, yes/no question: legalize it or not?

But behind that lie a host of regulatory follow-up questions — including questions of how the product's potency, consistency and safety should be regulated. Of course, black-market producers do not test marijuana for contaminants or offer potency labels for their consumers. Presumably, states will want to do more.

And cannabis-testing labs in states with legal marijuana are not yet using uniform standards. Our research shows that could be a problem. Here's why.

Regulators must cope with a rapidly evolving market featuring THC-infused edibles, beverages, tinctures and concentrates, alongside the traditionally heated and inhaled marijuana flower. In doing so, they may wish to consider testing for the following:

  1. THC, or tetrahydrocannabinol. This is the primary intoxicant in marijuana. Consumers who want to know how potent their purchase is — like liquor drinkers who wish to know whether they are buying an 80-proof or 100-proof beverage — will want to know how much THC is in a flower or concentrate. However, cannabis "potency" is not determined solely by THC levels. Other compounds matter, too.

  2. CBD, or cannabidiol. This is the second-most-prevalent cannabinoid in marijuana. While not intoxicating, CBD has various medical uses, including treating certain forms of epilepsy. Many medical marijuana strains contain very high levels of CBD and very low levels of THC. Research indicates that CBD modulates the effects of THC, reducing anxiety and paranoia; consumers might prefer certain ratios of THC to CBD.

  3. Terpenes (fragrant oils in cannabis and other plants) and other minor cannabinoids. States generally do not require testing for these, but they might contribute to different psychoactive effects associated with marijuana strains.

  4. Contaminants, such as microorganisms like E. coli, pesticides, solvent byproducts used to synthesize marijuana oils, or heavy metals.

Who will do the testing?

Conventional testing facilities have stayed away from marijuana, given federal prohibition — and an ad hoc industry of cannabis-specific testing labs has emerged. States often regulate testing labs in other settings, such as drinking water, but do not rigorously monitor the marijuana testing industry.

Testing presents some complicated questions. Precision is costly, and states haven't set standards about how much precision is required. Labs have incentives to charge growers lower prices — and, therefore, deliver less precise answers — and to report more THC and CBD, since cannabinoid-rich products sell for more. The questions include:

  • Many different laboratory procedures are currently used to quantify THC and other compounds. Which should testing labs use? Different procedures yield different results — and labs that inflate their results can claim they were simply using different procedures.

  • What are acceptable limits for pesticides, heavy metals, moisture content (indicative of microorganisms) and residual solvents?

  • What qualifications should states require for cannabis lab operators?

  • What organization (if any) should audit and accredit cannabis testing labs?

Answers to these questions are all over the map. The state of Washington requires that labs receive accreditations from RJ Lee Group. California and some others require more rigorous accreditation from the International Organization for Standardization (ISO). Oregon's labs are accredited by the state's environmental agency. Nevada alone requires that labs follow the recommendations of an American Herbal Association document called "Cannabis Fluorescence."

My colleague Nick Jikomes and I turned to the state of Washington's testing data to evaluate the consistency of results across labs. Individual sellers in the state get to choose who does the testing from the labs licensed by the state's Liquor and Cannabis Board. Labs must quantify total THC and CBD content and ensure that the plant does not contain harmful pesticides, microbes or heavy metals.

We found that results vary widely among labs. Some labs systematically report more THC than others, even after statistically controlling for grower, strain name and product type. That means consumers may have unpredictable experiences, because potency labels aren't consistent. What's more, public health investigations will lack reliable data on trends in cannabis potency.

Labs also report much different levels of CBD. Most marijuana sold in recreational markets contains high THC and low CBD. Small amounts of CBD can be difficult to quantify, so labs sometimes report that products contain zero CBD. But the percentage of "zero CBD" results varies widely among labs. Many report this for about 1 percent of samples — but we found that one lab, which also consistently reported the highest THC levels, said 70 percent of its flower samples have zero CBD.

States take various approaches to testing. Arizona, which allows medical marijuana, doesn't require any testing. California's ISO-accredited labs not only quantify potency, but also check for permissible levels of more than 50 pesticides and other contaminants. Colorado and Connecticut require testing for heavy metals, but with very different limits on what is acceptable.

As Julianne Nassif of the Association of Public Health Laboratories told the Portland Press Herald, "Some states are regulating it as if it's a pharmaceutical, and some states are regulating it as if it's an agricultural product."

Other states do not regulate the testing industry significantly more than Washington does — so there is no reason to believe that the labs are more consistent. Scholars testing marijuana products sold in stores find they often have inflated potency labels.

A different administration in Washington could shift these dynamics by permitting more research and suggesting national regulations for marijuana, as exist for alcohol. With Attorney General Jeff Sessions at the helm, such federal intervention looks unlikely. Until that changes, states will continue to be the laboratories for regulating this emerging industry — and the devil is in the details.

Michael Zoorob is a PhD student in the Department of Government at Harvard University.

Nick Jikomes is Leafly’s principal research scientist and holds a PhD in neuroscience from Harvard University.