Marijuana and Intelligence

Does long term use of cannabis make you dumber?

Posted Mar 19, 2013

The ideal study would be a conducted in the lab, on a large sample of subjects randomly assigned to smoking versus no smoking, and then measured again on intelligence tests several times. Because this is almost impossible, researchers rely on correlational studies with high attrition rates and small volunteer samples. This is problematic because correlation does not imply causation and there may be many confounding variables and alternative explanations for the results. Furthermore, psychological and statistical tests contribute another level of error. And,   finally, even the best psychological study should be replicated before it is considered conclusive. Oh, did I forget to mention that negative results don’t get published? So, if I found no decline in IQ or a non-significant one, we would never even know it!

Given all the difficulties studying the long-term relationship between cannabis and intelligence, it is surprising that anyone would even attempt it. However, one study has recently been getting a lot of media attention and some rather outrageous claims.  According to Scientific American, March 2013, heavy use of marijuana among adolescents can do permanent damage. According to Neurology Today (Oct 4 2012), smoking marijuana as a teen could lead to lower IQ as an adult according to a decades long study of more than 1000 New Zealanders. And, the authors themselves title their work, “Persistent cannabis users show neuropsychological decline from childhood to midlife.” Hmm..this is interesting - a large, longitudinal sample. If this is true, they have some pretty convincing evidence. I better look into this. What are their findings? What is the evidence?

I contacted the author, Madeline H. Meier, at the Center for Child and Family Policy, and asked for a reference to the original research.  She directed me to an article titled, “Persistent cannabis users show neuropsychological decline from childhood to midlife” and published in The Proceedings of the National Academy of Sciences (PNAS) 2012. I read this article very carefully and here is what I learned. To measure IQ, the authors used a very good psychological test. They tested children on the Wechsler Intelligence Scale for Children-Revised (WISC-R) and repeat tested a small sample of adults (n=38) on the Wechsler Adult Intelligence Scale (WAIS-IV). The group of regular users did indeed show a statistically significant decline in test scores on Vocabulary, Similarities, Information and Full Scale IQ which is heavily correlated with these verbal tests. However, the much more surprising thing is that the non-verbal scales, such as Arithmetic, Block design, and Picture Completion showed no statistically significant decline. If cannabis causes cognitive decline shouldn’t the ability to do arithmetic slow down? Shouldn’t block design and other non-verbal measures of intelligence be affected? One possible explanation for the discrepancy in their findings is that the first testing of the subjects was when they were in school and reading a lot of books, and taking a lot of vocabulary and spelling tests. The subjects were used to taking verbal tests. Out of school for at least twenty years and tested again at age 38, they may have stopped reading. Certainly, they are no longer taking vocabulary tests. The subjects were simply out of practice.This is just one possible alternative explanation that cannot be ruled out due to the correlational nature of the data.

Furthermore, Roeberg (PNAS, 2013) and Daly (PNAS, 2013) present rival alternative explanations for the decline in IQ scores. Roeberg states that the decline in IQ is explained by an alternative model of the effects of socio-economic status on IQ. Daly proposes that it may be explained by personality differences. Furthermore, Meir’s causal  inference (i.e. that cannabis causes permanent neuropsychological decline) is premature and most likely overestimated.  My own opinion is that much more research should be conducted before such exaggerated claims are used to make public policy. Given the many limitations of the current study, including its correlational nature, the small adult (n=38) sample, no statistically significant decline in arithmetic and other non-verbal intelligence scores, the many possible confounding variables, other alternative explanations, and the need for replication, the authors and the media, too, should be much more cautious in their conclusions.