A Duke study published last year correlating marijuana use to a decrease in IQ is being challenged by a more recent study on the matter.

The four-decade Duke study followed 1,000 people from the time they were born and found that those who used marijuana regularly suffered on average an eight point decline in IQ by the time they reached 38. The new study, led by Ole Rogeberg—a researcher at the Ragnar Frisch Centre for Economic Research in Oslo, Norway—used a mathematical model to produce results that reflect a similar IQ drop.

Rogeberg said the results suggest that socioeconomic status, rather than neuropsychological impairment, could be a significant confounding factor that was underestimated in the original study. Low socioeconomic status makes people less likely to have a nurturing home environment, which could correlate to effects including antisocial behavior.

“If you are inclined to exhibit antisocial behavior, have low self-control, low ambitions, etc., then these risk factors will both tend to raise your risk of adolescent cannabis use and dependence,” Rogeberg wrote in an email Monday. “[These risk factors] will make you more likely to get poorer grades, less schooling, less intellectual friends, less challenging jobs [and] more jailtime [and] these latter consequences will make your IQ decline relative to that of others.”

The two groups of researchers have only exchanged cursory statements, and both sides continue to defend their respective results. The Duke researchers have analyzed these confounding factors and maintain that their results are still consistent with the original findings, even after controlling for socioeconomic status.

Rogeberg’s mathematical model uses past empirical research to look at how IQ in childhood and adulthood is linked and the potential role socioeconomic status could play in this correlation. Examining the IQ drop that the Duke study linked to regular adolescent pot use, the study found that a similar drop in IQ could be accredited to socioeconomic status, thereby calling into question whether the results found in the Duke study could be attributed to marijuana use alone.

After hearing about Rogeberg’s study, the Duke researchers have re-examined their data. They said that so far, they have not found any evidence suggesting that socioeconomic status undermines their original claim.

“One of the assumptions that Rogeberg’s hypothesis hinges upon is that a person with high socioeconomic status will have a smaller drop in IQ than one with low socioeconomic status,” said Madeline Meier, a post-doctoral researcher and lead author of the Duke study.

She added that the study’s participants were stratified by their socioeconomic status and that she did not see any differential IQ change between the groups.

“We did not see that the low socioeconomic group showed any differential IQ change compared to, for example, the middle or high socioeconomic groups,” Meier said. “In fact, the low, middle and high socioeconomic groups show parallel trajectories of IQ change across time.”

Meier also noted that whereas her study looked at over 1,000 participants over a decade-long period, Rogeberg’s study relies on mathematical simulations.

Rogeberg said the simulations, however, allow him to look at other factors that could cause a drop in IQ, giving him insight into the robustness of the Duke study’s claims.

“This model allows me to use the Duke study’s methods on a data set where I know what is actually going on,” Rogeberg said. “When those methods give very similar results, it shows that this is an alternative explanation that could be important in the [Duke] data—either causing the methods to exaggerate the effect of cannabis on IQ development, or by causing them to find an effect that isn’t actually there.”

Cynthia Kuhn, a professor of pharmacology and cancer biology at the School of Medicine, said that although we know marijuana use impairs cognitive function in people, there are no definitive answers on whether marijuana exposure causes permanent effects on IQ.

“This debate highlights the limits of studies on humans—even high-quality, well-described studies like Meier et. al—they are always complicated by all the other factors that affect human beings,” Kuhn wrote in an email Tuesday. “No one study can prove causality, as the authors themselves have acknowledged.”

Meier noted that she has concerns about how Rogeberg’s study will be received by the public, and adolescents in particular.

“Our original message was that adolescent onset users experienced some IQ decline as a result of cannabis use, and that was an important message…. Right now, the last word is that no, it’s socioeconomic status that’s responsible, and I think that is a real shame.”