NEW YORK (AP) — A new analysis is challenging a report that suggests regular marijuana smoking during the teen years can lead to a long-term drop in IQ. The analysis says the statistical analysis behind that conclusion is flawed.

The original study, reported last August, included more than 1,000 people who’d been born in the town of Dunedin, New Zealand. Their IQ was tested at ages 13 and 38, and they were asked about marijuana use periodically between those ages.

Researchers at Duke University and elsewhere found that participants who’d reported becoming dependent on pot by age 18 showed a drop in IQ score between ages 13 and 38. The findings suggest pot is harmful to the adolescent brain, the researchers said.

Not so fast, says an analysis published online Monday by the Proceedings of the National Academy of Sciences.

Ole Rogeberg of the Ragnar Frisch Centre for Economic Research in Oslo, says the IQ trend might have nothing to do with pot. Rather, it may have emerged from differences among the study participants in socioeconomic status, or SES, which involves factors like income, education and occupation, he says.

He based his paper on a computer simulation. It traced what would happen to IQ scores over time if they were affected by differences in SES in ways suggested by other research, but not by smoking marijuana. He found patterns that looked just like what the Duke study found.

In an interview, Rogeberg said he’s not claiming that his alternative explanation is definitely right, just that the methods and evidence in the original study aren’t enough to rule it out. He suggested further analyses the researchers could do.

The Duke scientists, who learned of Rogeberg’s analysis late last week, say they conducted new statistical tests to assess his proposed explanation. Their verdict: It’s wrong. Rogeberg says they need to do still more work to truly rule it out.

Experts unconnected to the two papers said the Rogeberg paper doesn’t overturn the original study. It “raises some interesting points and possibilities,” but provides “speculation” rather than new data based on real people, said Dr Duncan Clark, who studies alcohol and drug use in adolescents at the University of Pittsburgh.

Dr Nora Volkow, director of the National Institute on Drug Abuse, said observational studies of people like the Duke work can’t definitively demonstrate that marijuana cause irreversible effects on the brain. In an email, she said Rogeberg’s paper “looks sound” but doesn’t prove that his alternative explanation is correct.