Apr 03 2017

Transcendental Meditation Pseudoscience

Transcendental-meditationIt’s fun to run into such a wonderful example of pure pseudoscience. Let’s deconstruct this one: Field Effects of Consciousness and Reduction in U.S. Urban Murder Rates: Evaluation of a Prospective Quasi-Experiment. This study comes from the Maharishi University of Management.

The idea here (which, let’s be clear, is a tenet of religious faith, not a scientific theory) is that consciousness is a field, and that there is a universal field of consciousness of which we are all a part. When individuals engage in transcendental meditation (TM) they are not only affecting their own consciousness, they are affecting the entire field.

The point of this and other similar TM studies is to confirm the belief (they are not testing the belief) that if enough people put good vibrations into the universal field of consciousness, society in general will benefit. How many is enough? Well apparently they have an answer for that. It is the square root of 1% of the population. Why? Because math.

That is such an excellent example of pseudoscience, having the trappings of science without the real essence of science. Look, they use numbers and everything. Apparently there isn’t a dose-response effect, there is a threshold effect, and once you get over the magic threshold the effect kicks in. That threshold has a simple mathematical formula, the square root of 1%. There is no established theoretical reason for this, but it sounds nice, having more in common with a magic ritual than a scientific process.

This is what they claim to have done with the experiment: The tracked the murder rate in the 206 largest US cities where FBI murder statistics were available. They then compared the baseline period from 2002-2006 with the intervention period from a 2007-2010. The intervention period was when the, “Transcendental Meditation and TM-Sidhi group exceeded 1,725 participants beginning in January 2007.” So they claim that murder rates were increasing and then when participant #1725 (because 1724 was not enough) signed up, the rate started to drop.

What is really funny is the blatant frequentist nonsense they spout in the press release. They claim:

“They calculated that the probability that the reduced trend in murder rates could simply be due to chance was 1 in 10 million million.”

So they think the probability that the murder rate declined from 2007 to 2010 was one in a trillion. Right.

This study screams p-hacking and assuming causation from correlation. The authors should take a look at spurious correlations website. There are countless trends in any variable you want to look at in society, and you can find countless correlations if you look. That does not mean there is a causal relationship.

If you look at the FBI statistics there has been a general downward trend in homicides since 1991 (hey, maybe vaccines are decreasing homicides). Like all long term trends, there are likely to be ups and downs over shorter periods. The trend leveled off a bit in the early 2000’s and then resumed the previous downward trend.

There is no consensus on what is causing this downward trend in crime and homicides in the last 25 years. No one societal factor explains the trend. That is why it is also laughable that the study authors claim that no other factor explains the drop in homicides during their study period, as if this lends support to their TM hypothesis.

This study, and other similar studies touted by the authors, is entirely worthless. They are simply taking credit for general trends in society. TM has been around since the 1950s, and became popular in the US in the 1970’s. Despite TM’s popularity, the 1970s was the beginning of a two decade historic increase in crime and murder rates. They should take credit for that also.

It’s interesting that the authors call their study a “quasi-experiment” which is a sort-of admission that it is not a real experiment. There is no real control. They cannot control for how many people are using TM at any time (so how do they know their magic 1,725 figure is reliable). This is like prayer studies – how do you know people who are not part of the study aren’t praying for the subject?

Also, how do they define “society.” Does Canada count? Does the universal consciousness field obey political boundaries. Should they include the population of Canada in their square root of 1% calculation, or include Canadian crime statistics (which did not follow the same trend)? What about Mexico – does language matter? Perhaps they can use east and west of the Mississippi as a control. Or can they go state by state?

That’s the problem with studying magic. There are no real rules. This raises p-hacking to a new level, because you can arbitrarily make up whatever variables or thresholds you want. Essentially all they are doing in this and the other studies they tout is cherry picking favorable trends.

They then do useless frequentist statistics to make it seem “statistically significant,” but they are crunching the numbers as if this were a clean experiment and not just looking at societal trends. The numbers ultimately mean nothing because there is no real control group.

Applying numbers to magic does not make it science.

38 responses so far