Tag Archives: sleep disorders

Data dredging, p-hacking and motivated discussion in anti-fluoride paper

Image credit: Quick Data Lessons: Data Dredging

Oh dear – another scientific paper claiming evidence of toxic effects from fluoridation. But a critical look at the paper shows evidence of p-hacking, data dredging and motivated reasoning to derive their conclusions. And it was published in a journal shown to be friendly to such poor science.

The paper is:

Cunningham, J. E. A., Mccague, H., Malin, A. J., Flora, D., & Till, C. (2021). Fluoride exposure and duration and quality of sleep in a Canadian population-based sample. Environmental Health, 1–10.

Data dredging

This study used data from a Canadian database – the Canadian Health Measures Survey. Databases with large numbers of variables tempt researchers to dredge for data or relationships which confirm their biases. Despite the loss of statistical significance in this approach data dredging or data mining is quite common in epidemiological studies.

Cunningham et al (2021) looked for relationships using two separate measure of fluoride exposure and four different measures of possible sleep disturbance. They found a “statistically significant (p<0.05) relationship between lower sleep duration and water fluoride. But no relationships for higher sleep duration, trouble sleeping or daytime sleepiness with either water fluoride or urinary fluoride. Their results for logical regression analysis are summarised in this figure. (Error bars crossing an Odds Ratio value of 1.0 indicate that the relationship is not statistically significant and p<0.05).

Of the 8 relationships investigated only 1 was statistically significant.


I discussed the problem of p-hacking in Statistical manipulation to get publishable results.

With a large dataset, one can inevitably find relationships that satisfy the p<0.05 criteria – because this p-value value is meaningless when multiple relationships are considered. One can even find such “statistically significant relationships” when random datasets are investigated (see Science is often wrong – be criticalI don’t “believe” in science – and neither should you, The promotion of weak statistical relationships in science  and Can we trust science). Once multiple relationships are investigated the chance of finding accidental relationships is much greater than 1 in 20 signified by the p<0.05 value.

So, one of the 8 relationships above satisfied the p<0.05 criteria when considered alone. But as part of multiple investigations, the chance of finding such a relationship by chance is much greater than 1 in 20.

Motivated reasoning

This paper smacks of motivated reasoning. The authors obviously have a commitment to the concept that fluoride causes problems with the pineal gland and drag up anything they can find in the literature to support this – without critically assessing the quality of the cited work or even mentioning the fact that the cited studies were made at much higher fluoride concentration on non-human animals. In effect, they are attempting to convert very weak results, obtained by data dredging and p-hacking, to a fact. They are attempting to make a purse out of a sow’s ear.

This research group is not new to this game. I commented on this in my critique of another sleep disorder paper from the group (see ).

Many of the same researchers are listed as authors on both papers – yet Cummingham et al (2021 ) cite the previous paper as if it was an independent study. They say “As far as we are aware, this is only the second human
study investigating the effects of fluoride exposure on sleep outcomes” which is simply disingenuous considering the involvement of the same researchers in both papers.

Both these papers were also published in the same journal – Environmental Health – a pay-to publish-journal that is known to be friendly to anti-fluoride researchers and uses very sympathetic peer reviewers (see ). The Chief editor, Philippe Grandjean, is well known for his opposition to fluoridation. I commented on his refusal to consider a paper of mine that critiqued an anti-fluoride paper published in his journal (see Fluoridation not associated with ADHD – a myth put to rest).


Yet another very weak study, published in an anti-fluoride friendly pay-to-publish journal with poor peer review. Despite the weaknesses due to data dredging, p-hacking and motivated reasoning, anti-fluoride activists will cite the single “statistically significant” result as gospel and ignore the 7 relationships that are not significant. As for inadequate consideration of confounders or other risk-modifying factors, this study ignores completely the fact that city size and geographic factors have a strong effect on both sleep patterns and water fluoride concentrations (see Perrott 2018). Such inadequate consideration of confounders is another common problem in epidemiological studies.

Oh, well, we are not a rational species. More a rationalising one. And in such areas motivated rationalisation and confirmation bias is rife.

Similar studies

Sleep disorders and fluoride: dredging data to confirm a bias

Sleep disorders are common and have many causes. But anti-fluoride activists will now be blaming them on community fluoridation. Image credit: Sleep Disorders and Problems

There is a pattern to the recent research aimed at finding a link between fluoridation and cognitive deficits, ADHD prevalence or possible thyroid problems. These researchers are simply using large databases from recent health surveys (ELEMENT in Mexico City, MIREC in Canada, INMA in Spain, and NHANES in the USA). Connecting these up with available measurements of drinking water fluoride, their own measurements of urinary or blood plasma fluoride using stored samples from the surveys, or even their own unvalidated estimates of fluoride dietary intake, they then search for statically significant (p < 0.05) relationships.

This gives them a large amount of data to search for effects – and as the p-hacking tool in my recent article, Statistical manipulation to get publishable results, shows – they will, of course, find them. They might have to use different fluoride measure to get a statistically significant result – but they have several to choose from: drinking water F, community water fluoridation, urinary F for the individual or his/her mother, blood plasma F and their own subjective estimate of dietary F intake.

In my last article, Some fluoride-IQ researchers seem to be taking in each other’s laundry, I discussed the biased peer-review process used in a new paper from the research and commented:

“One might expect that the need to use an open-access journal like Environmental Health and to choose “in-house” peer-reviewers indicates that the quality of the paper might not be the best.”

I was correct – this is another poor quality paper on fluoride and health effects which make unwarranted claims – and which will be used by anti-fluoride activists in their campaign against community water fluoridation.

This is the paper citation:

Malin, A. J., Bose, S., Busgang, S. A., Gennings, C., Thorpy, M., Wright, R. O., … Arora, M. (2019). Fluoride exposure and sleep patterns among older adolescents in the United States : a cross-sectional study of NHANES 2015 – 2016. Environmental Health, 1–9.

Multiple parameters to dredge

The researchers had a list of parameters to work with. Sleep duration, sleep apnea symptoms, snoring, daytime sleepiness (subdivided into rarely, sometimes, often and almost always), trouble sleeping, bedtime and wake time. Ten sleep disorder measures. The authors searched for statistically significant relationships of these with two fluoride measures: household tap water fluoride and blood plasma fluoride.

See the problem here? If not have a look at Statistical manipulation to get publishable results, and have a go with the p-hacking tool.

Blood plasma fluoride

In a bit of special pleading involving subgroups divided by gender they report:

“Among males, higher plasma fluoride concentrations were associated with higher odds of reporting sleep apnea symptoms, although this did not reach the threshold for statistical significance (uncorrected p = 0.17).”

And added:

“Plasma fluoride concentrations were not significantly associated with any of the other sleep outcome measures examined herein”

So nothing here. Despite using ten different sleep disorder measures and looking at subgroups there were no statistically significant relationships with blood plasma F.

Water fluoride

The tap water fluoride concentrations were mostly below 0.7 mg/L with mean and median values of 0.35 and 0.29 mg/L. There was no differentiation between community water fluoridated and non-fluoridated areas, but water fluoridation covers a high proportion of US citizens.

The only statistically significant relations shown by regression analyses were for sleep apnea, snoring, bedtime and waketime. Two out of ten sleep disorder measures or four out of ten if one counts different bedtime and waketime as disorders.

Bedtime and wake time

The paper reports that:

“fluoride exposure may be associated with shifts in the sleep-wake cycle, as higher water fluoride concentrations were associated with later weekday bedtime and wake time, but not sleep duration. Specifically, for each 0.52 mg/L increase in adolescents’ water fluoride concentrations, they tended to report going to bed 24-min later and getting out of bed 26-min later. “


Why should a different bedtime or wake time be considered sleep disorders – especially as no change in sleep duration occurs?

I think they are indulging in special pleading by attempting to find a reason for this in calcification of the pineal gland. This idea rests on an old observation that calcified pineal glands taken from elderly cadavers are high in fluoride. This is easily explained by the fact that fluoride is attracted to active calcified tissue. Calcification is caused by old age, calcium and phosphorus – not by fluoride. Fluoride is adsorbed by calcified tissue after calcification.

They do acknowledge as a limitation of their work that:

“participants were older adolescents who may be prone to sleep disruptions for various reasons, including playing video games, studying, working at jobs or having social influences, for example.”

Well, yes. And these social influences, jobs, etc., will be more common for adolescents living in cities which are more likely to have community water fluoridation than in rural and small-town areas.

Why were such “sleep disturbances” even included in their study? And why indulge in such fanciful reasoning to “explain” the result.


The authors report that:

“each 0.52 mg/L increase in household tap water fluoride concentration was associated with a 38% reduction in the likelihood of male adolescents reporting snoring.”

Maybe young men living in cities and having an active social life are less likely to admit to snoring than their counterparts in rural areas. But the authors again indulge in fanciful reasoning by speculating:

“that our findings may point to a role of fluoride exposure in disrupting this deep sleep stage, thereby reducing opportunities for snoring.”

Or, alternatively”

“another possibility is that the gains in oral health from consumption of fluoridated water may protect against tonsillar infections that can contribute to snoring .”

Perhaps surprising that they are discussing a possibility of beneficial effects of fluoridation but they concede that:

” Future studies are needed to explore potential mechanisms by which fluoride exposure may reduce self-reported snoring.”

Perhaps a more reasonable future study will find absolutely no effect of fluoridation on reported snoring if it includes more relevant factors in its multiple regressions. Remember how Malin & Till (2015) reported a significant positive relationship between fluoridation extent and ADHD prevalence in the USA – yet when more relevant factors where included in the multiple regression the relationship with fluoridation disappear (see Perrott 2018).

Sleep apnea

Only about 10% of the participants reported symptoms of sleep apnea at least once a week. Yet the paper report they found:

“that each 0.52 mg/L increase in household tap water fluoride concentration was associated with a 1.97 times higher likelihood of adolescents reporting having experienced symptoms suggestive of sleep apnea at least once per week.”

Well, there was a very large spread in the data with a confidence interval of 1.27 to 3.05.

On this basis they argue:

“This suggests that fluoride exposure at population-relevant levels may be a risk factor for sleep disturbances; however, additional studies are needed to explore this possibility, given the scarcity of data on this topic.”

Again, perhaps a more reasonable future study will find self-reported sleep apnea is related to living in a city or some other factor rather than fluoride. The results found in this study should not be used to argue that sleep apnea is caused by fluoridation. But, of course, that will not stop the anti-fluoridation activists from doing so.

Speculation without action is arrogant

I submitted a paper to a journal once where I speculated on mechanisms which could explain the associations I had found. One of the peer-reviewers pointed out that speculation was worthless in itself and that I should actually do some measurements to test the proposed mechanism before publication.

The reviewer was quite right – it was arrogant of me to think that my speculation had any scientific worth when it was not supported by data. I was simply resting on an assumed authority or credibility. But most proposed ideas in science turn out to be wrong. Speculation only has value when it is converted to a hypothesis and tested.

I did the experiment to test my speculated hypothesis, included it in my revised paper which was then published (and turned out to be a more valuable contribution). Perhaps this sleep disorder paper would have had more significance if one of the peer reviewers had made a similar comment and the authors had then set out to test some of their speculations. But fat chance of that happening when the peer-reviewers were selected from colleagues who already have a bias for finding similar effects of fluoride (see Some fluoride-IQ researchers seem to be taking in each other’s laundry).

The problem with this sleep disorder paper and other recent papers reporting relationships between fluoride and cognitive effects is that they are only reporting fishing expeditions. They simply report the results of searching through data sets containing a whole range of parameters to find statistically significant relationships. They put all their faith in the p-value so that the relationship appears important even when the effect size is small and explains a minuscule amount of the variability. In itself, a p-value can say absolutely nothing about the cause of an observed relationship or be used to claim an effect. That requires further work.

Nothing wrong with statistical fishing exhibitions like this. I also enjoy searching through data looking for relationships. But that is only the start. Identification of relationship can suggest research – experiments or survey aimed at identifying causes.

I don’t think there is any value in simply reporting the results of fishing expeditions without further research. Such papers only serve as an outlet for unwarranted and unsupported speculation – and as I say above that is arrogant. Why should anyone else take such speculation as evidence or identification of a cause?

Trouble is when one has a bias to confirm one can eagerly clutch at this sort of speculation and promote it as a real effect. When speculation like this is promoted by anti-vaccination or anti-fluoridation activists it can end up undermining effective social health policies – and that is bad.

Similar articles