Dr. Emoto, Water, Intention and Magic

I’ve been recently discussing with someone the possibility that, basically, magic is a real thing that really happens in the real world. Really. In fairness, those are not the terms that they use, but nevertheless that’s the argument being presented.

The idea is that since our intentions can affect water, and humans are “70% to 90% water, depending on age”, then we can totally affect the health of other people with our thoughts. As evidence for this claim, when pushed (and it was like pulling teeth) they refer to Dr. Emoto’s work on water and intention. Ironically, Dr. Emoto appears to have done very little science on this topic, insofar as he has a total of one (1) paper published, and even then it’s in the fringe science Journal of Scientific Exploration. The paper is titled (this link goes to a PDF) “Effects of Distant Intention on Water Crystal Formation: A Triple-Blind Replication“. The rest of this post will be a breakdown of that paper.

To summarise the experiment:

1. Six bottles of water were purchased, all the same brand (Fuji).
2. The bottles were randomly assigned labels.
3. The bottles were randomly assigned into three groups (two bottles in each).
4. Four of the bottles were placed in a room, and were photographed (two of them being edited out of photo for the purposes of blinding)
5. The remaining two bottles were placed in a different room, and not photographed.
6. Some time later, three different groups in three different cities were asked to pray ‘at’ the photographed bottles, on three separate days.
7. The bottles were packaged together and shipped to another location
8. 50 drops of water were taken from each bottle, placed separately in labeled petri dishes, and frozen. (all in the same freezer, distributed randomly)
9. After they formed ice crystals, they were photographed.
10. The photographs were evaluated for “beauty” vs “interest”
11. These evaluations were contrasted according to which sample the water came from

The conclusion that the authors reached is “The comparison of main interest confirmed, weakly, that the treated water crystals were rated as more beautiful, on average, than the proximal controls (p = 0.05, one-tailed)”, which means that they basically claim that ‘our intentions affect water’. If you go back to their introduction, you’ll see that this is the foundation for the claim that ‘we can affect other humans with our minds’. So it raises the question: is any of this justified?

Methodology

On the face of it, the methodology employed here seems fairly reasonable: all bottles are from the same source in order to rule out the possibility of different mineral composition creating the effects. There’s two bottles in each group, to rule out that one bottle is just composed oddly. There’s a second pair of bottles in the same room as the ‘intentionally treated group’, to ensure that if the ‘intentionally treated group’ does have a different result that it’s not due to a differing storage space. Keeping them together while being moved and then frozen ensures that they’re not affected by differing environments. As things stand, I don’t see anything objectionable in the methodology as listed.

Results

Here’s where everything, frankly, falls apart.

Fortunately, we have the dataset, in the form of Figure 3 (reproduced below). The ‘distant control group’ is the first 100 images, the ‘proximal control group’ the next 100 (numbers 101 to 200), and the ‘intentionally treated group’ the final 100 (201 to 300).

Just by looking at the chart, we can see that there really isn’t that much of a difference between the three groups: the ratings are mainly clustered between 1 and 2, with a few images being rated between 2 and 3, and but a handful being rated above that. Notice that each image also has an error bar, indicating that none of the images has a precise score, rather they have an approximate score. The authors then removed anything that scored a 1 or less, on the basis that

“This subset of images [Brian: i.e. rated above 1] was examined in a secondary analysis because it was more likely to contain crystalline shapes, which was of main interest in this experiment.That is, the intentional hypothesis was not that more crystals would form due to intention, but rather that crystals that did form would appear to be more beautiful in the treatment condition vs. the proximal control condition.”

Let’s accept that, for the sake of discussion. But… Why 1? Why not 1.5? Why not 0.5? This value of ‘1’ appears to be entirely arbitrary. Things, however, become a little more clear if we look at our dataset with the boundary clearly marked:

Which group had the most samples rated at 1 or lower? The ‘intentionally treated group’. What happens if we remove the 1-and-lower pictures? Well, the average is going to go up. And as it just so happens, as the ‘intentionally treated group’ has the most 1-and-lower pictures, then its average will bounce up the most. See Figure 4, included below:

Notice how little the average changes for the ‘distant control group’: there’s still a huge overlap in the error bars between ‘all trials’ and ‘beauty > 1’. Likewise, there’s about a 40% overlap in the error bars for the ‘proximal control group’ too. But for the ‘intentionally treated group’? Only a small overlap (I’d eyeball about 10% or so). Bear in mind that those error lines need to be extended down as much as they go up. Weird how they didn’t include that in the figure, eh?

Returning to the quote I mentioned above:

“That is, the intentional hypothesis was not that more crystals would form due to intention, but rather that crystals that did form would appear to be more beautiful in the treatment condition vs. the proximal control condition.”

Notice here the focus on the ‘proximal group’. Now, ostensibly, this has always been the focus of the study, and the ‘distant control group’ was really there just to muddy/confuse the ‘intentions’ of the other experimenters (because their psychic powers would interfere with the experiment), but we readers don’t need to follow their lead.  The introduction to this article clearly states:

“One exception that has elevated the question about intention and water from the obscure to the infamous is the claim that water exposed to or ‘‘treated’’ by positive intentions results in frozen water crystals that are aesthetically more pleasing than similar crystals formed from ‘‘untreated’’ water”

The ‘distant control’ group is clearly “untreated” water as well (I’d argue that it’s all untreated, but let’s not go there), but now we’re narrowing our analysis to focus on the ‘proximal control group’? Why is that, I wonder? Could it be that the ‘distant control group’ and the ‘intentionally treated group’ are all but indistinguishable? That if we compare those two groups, there is zero difference between them? Moreover, when we zoom in our barchart to look at the number ranges of only 1.5 to 2.1, the difference in these bar charts look huge, when in reality the scale was from 0 to 7. The impression that this bar chart leaves us with is somewhat false.

The ‘intentionally treated group’ and the ‘proximal control group’ look hugely different. But are they really all that different? I mean, what accounts for the bulk of the difference between them?

I’ve drawn in two vertical red lines here to mark the border between the two groups a little more clearly, and a horizontal blue line at the value of ‘3’. It’s important to bear in mind that the ‘intentionally treated group’ has less members than the ‘proximal control group’, as this means that high-scoring crystals in the ‘intentionally treated group’ are going to drag up the average of that group more than they would in the ‘proximal control group’, simply by dint of having a lower denominator. Moreover, the main difference between the groups is really going to be the number of crystals scoring above the average (and I’m arbitrarily drawing the line at ‘3’). The ‘proximal control group’ has 9 crystals rated around 3 or higher. The ‘intentionally treated group’ has 15 (giving the benefit of the doubt to a couple whose error bars cross the blue line). Let’s be clear here, assuming the thesis of this experiment holds, 1900 people prayed together in order to gain 6 pretty ice crystals. Let that sink in. If we were comparing the performance of two medicines, what we would say here is that sure, there’s a statistically significant difference between these two groups, but there’s no clinical significance (i.e. the difference is real, but so tiny that it’s irrelevant’.

As it stands, there’s no significant difference between the ‘distant control group’ and the ‘intentionally treated group’.

Conclusion

While the authors of this paper managed to get the result that they wanted with a little manipulation of the images being used, and then some careful ignoring of the relevant data, there’s something else that we need to not lose focus on:

“Can one person’s intention affect another person’s health from a distance?”

That’s the opening statement of this article. That’s the whole point of this experiment: to demonstrate that human minds can affect the health of another person at a distance. I have a question to ask: what in the hell does ice-crystal formation have to do with healing people? Even if we got a clinically significant result (all 100 crystals from the ‘intentionally treated group’ were rated at 6 or above), what would that matter? We can save on air conditioning? We can use millions of people praying instead of refrigerators in supermarkets? How often does a doctor watch a patient die, gnash her teeth, put her hand to her forehead and wail “if only we had some way to form pretty (but not ugly!) ice crystals inside this patient, we would have saved them!”

I’m at a loss.

Moreover, there’s an additional nugget of nonsense buried in the introduction of this paper that should be an immediate red flag to anyone even passingly competent at reading scientific papers:

“By analogy with a quantum optics system, in which the knowledge one has of the path that photons take through a double-slit apparatus influences the behavior of those photons, we speculated that knowledge of the experimental conditions in this test might influence what was ultimately measured. Thus, to provide some control over the distant intentions in this study we required a comparison condition that was unknown to M.E., to T.K., or to the groups of ‘’distant intenders.’’”

This is a profoundly mistaken understanding (speaking charitably) of the double-slit experiments. The knowledge of the experimenter, that is the mental state of the experimenter, had no effect on the behaviour of the photons. None. Zero. Zilch. Nada. What affected the photons was that detectors were set up to measure the photons. Neither “knowledge” nor “intention” had any bearing in the double-slit experiment, at all. Ultimately, what the double-slit experiment told us is that when we measure something, we also interact with it, and that interaction affects it. When you get to a small enough scale, the lightest touch has big consequences. As we’re talking about ‘ice formation’ in this experiment, we are discussing things that are well above the sub-atomic scale, and thus this kind of nonsense has no place in this article.

This paper demonstrates nothing but the desperation and ignorance of its authors. It certainly does not demonstrate that humans have magic mind powers.

Follow Brian on Twitter!

[GARD]