Hans Breiter, a Northwestern medical doctor (Psychiatry) concocted one of the more anti-science scare articles outside of the World Nut Daily. He claimed to prove that casual use of marijuana could be detected in adverse MRI findings in the brain of smokers.
"Cannibis Use Is Qualitatively Associated with Nucleus Accumbens and Amygdala Abnormalities in Young Adult Recreational Users, Neurobiology of Disease, 34 (2014) 5529-5538.
The only problem is that is a collection of horse shit so dense and pervasive, that he should have timed the publishing for April Fools day, instead.
PROBLEM ONE
An honest investigator and researcher, say anyone who is not named Hans Breiter, realizes that for a test to have any significance, you have to minimize impacts that are not being studied, predicting what other factors may have an impact on your test results, and to set up a strict inclusion/exclusion criteria that makes your data the fairest and most accurate, given the parameters of your test.
For example, in placebo v. drug test, you want the diseases to be studied to be as close in progression as possible, and that your test subjects (rabbits, human inmates populating private prisons, or chimps) to be as close and as similar as possible. Otherwise, you will probably end up testing for garbage, which will clearly be reflected in your results.
One method of minimizing variation and loopy outside forces is by applying weighted "P Values," a standard statistical practice which is taught in Stats 101. If you miss this one, you flunk, because ALL OF YOUR RESULTS ARE GARBAGE.
One lesson we have learned with Problem One? Psychiatrists are really, really bad at math, and even worse at statistics. Case in point, Hans Breiter. By design, willful ignorance, or ineffable stupidity, Breiter not only relied on RAW, UNCORRECTED P Values, he hunted those findings that only supported his intended result.
Rule of Thumb on P-Values: The larger your value (corrected) the more likely that you have a mess of a test, and that your findings are gibberish. The smaller the P Value (like 0.50 compared to 0.00023) the more likely that you have data that you can hang your hat on.
Liar? Fool? Or Paid Hack?
PROBLEM TWO
Causation vs. Correlation
Correlation is what gave us today's bible beating bastards. Take your average caveman. Take an earthquake, volcano, or massive drought. Add a virgin stolen from a nearby tribe. Stir carefully. 9 times out of 10, your sacrifice of that poor, unfortunate virginal lass results in:
a. No more Earthquake
b. A shut down Volcano
c. Rain dense and sustained enough to solve your drought.
Under "Correlation," people really bad at logic, people with no mathematics training, folks with no statistical knowledge, woefully ignorant cavemen, and Hans Breiter would correlate the sequence of events, (Earthquake, Sacrifice, end of Earthquake) with their killing that poor innocent girl. In Breiter's case, I suspect he would proclaim that he created the psychiatrist's version of Earthquake prevention.
Causation is a far more peculiar, particular, and limited finding. For one thing, you need a number of studies, compared to each other, and you need to get repeated findings, time after time after time, assuming that you created the test parameters strictly enough with sufficient controls in the first place.
One result may be an outlier. 20 results not only suggest a trend, but they suggest causation, especially if all the noise factor is removed by careful data analysis and gathering. Causation, EVERY ARGUMENT about causation requires a significant amount of repeat studies that show identical (or strongly similar) results. Every time.
Boy Wonder Hans? He relies on ONE RESULT. Again, back to Stats 101, this is such a bad misuse of data that I am surprised that he dared publish this crap. Let's just agree on one thing. One test result can never ever create a statistical basis for showing "Causation." "Qualitatively Associated?" My ass, Dr. Breiter. My NU psych and Econ instructors would have reamed me a new asshole if I tried to pull some sort of garbage like that.
Liar? Fool? Or Paid Hack?
PROBLEM THREE
Frankly, there are so many problems with this study, that it is hard to know where to start. So, let us start at the very beginning, a very good place to start.
Sample Size.
If I designed a rigorous, well-designed study, and I wanted to have a confidence level of between 90-100% (most of the best and serious studies look for a confidence level of 95%), depending on my study group, there are some long standing factors that determine how big your study group has to be to have any relevance or substance to it. If I wanted to study the impact of marijuana on the brains of new smokers who had never toked up before, and I took many, many, many, many other precautions, (something that Hans admits he never even considered - because he HAND PICKED HIS SAMPLE GROUP!) I would need a sample size 5-6 times as large as Hans', and that assumes that the sample group had significant controls to prevent hand-picking the target sample.
Not only was Hans' group way, way way (should I have 5 or 6 ways?) too small for any meaningful interpretation, that idiot hand-picked them to get the result he was looking for! With a sample size so small, you get nothing. Period. Nada. Nil. Zero. There are no reliable results upon which you can base any conclusion.
Liar? Fool? Or Paid Hack?
Liar? I am sure that Hans is good to his family and students, never picks up young coeds, and is truthful to the extent that any psychiatrist can know what any truth may actually be.
Fool? Anyone with his psychiatric background has had to study long and hard, and needed to convince others that he was able to be a psychiatrist. (An MD instead of a psychologist, who may only have a masters or a PhD) He is not a fool in his field.
BUT. . . THIS. . . IS . . . NOT . . . HIS . . . FIELD . . .! ! ! ! !
With all of the problems that are clearly visible, even to a lawyer like me, that were built-in to this study, it is clear that he wanted to make a splash, make a name for himself and the very well-funded anti-pot movement (think tobacco, national police orgs, and more) and publish some scary garbage solely for self-promotion, or because of his orders from Big Tobacco.
Here's what I think happened. Hans was hired to be a hack for Big Pharma, Big Tobacco, or some anti-pot police organization that sees the writing on the wall with legalization. So, they use Hans to muddy the waters, to scare people, to confuse them with outlandish, unsupported claims, and to put a stop to a growing movement towards medical marijuana.
I would suggest that Hans Breiter should be very ashamed of himself for putting out such a bogus and ridiculous study, but that would imply that this psychiatrist is capable of feeling shame. Of that, I remain unconvinced.