Lunacy

[sesquiterpene]

The q value represents the statistical probability that this is a false positive. The actual result is definitive, the probability this result is false quite low therefore. It perfectly fulfills Fisherian criteria, so on the grounds of Evidence Based Medicine, quite solid. Better than antidepressants vs placebo, for instance.

You are joking, aren't you? A "study" with 15 subjects, plus two controls(??) who were plucked from the literature of the 1920s-1930s, is ridiculous (and reeks of p-hacking). I'd consider the results to be utterly meaningless, not definitive.

 

[Quid est Veritas?]

You are joking, aren't you? A "study" with 15 subjects, plus two controls(??) who were plucked from the literature of the 1920s-1930s, is ridiculous (and reeks of p-hacking). I'd consider the results to be utterly meaningless, not definitive.

Have you read the study? These aren't just a few records here and there, but meticulous ones kept by many of the patients themselves, giving about 37 years worth of data for these individuals. Again too, it is in a small subset of rapid-cycling bipolars with strict criteria. This is high quality data in a tiny subset of the population.

It would be ludicrous if they had been followed for one or two cycles or such, but by sheer amount, we are looking at something quite different. I've seen it juxtaposed to the Maudsley psychotic twin study, a classic study with about 80 participants, although we see much closer alignment here. Ordinarily yes, such a tiny sample could be dismissed, but any group followed for many many years, suffering a narrowly defined pathological process, would give very good data even if such low numbers. It is commonly so in rare illnesses or ones with limited subject availability. And again, we were talking about within that population group, not generalised to all.

Besides, this is well reviewed, and old data is not less useful in Medicine than new data, especially when tracking the same things. Our premiere hypothermia studies were conducted in the 1940s for instance.

Anyway, to quote the study on this point, we are looking at a specific criterion, so we can actually adjust for statistical significance:

"Qp value, which estimates the fraction of significant results that are null, is a multiple testing correction to P-values that is commonly used in genomics and proteomics. It provides more statistical power than a Bonferroni correction, but is still conservative. If the χ2 periodogram analysis looks for a specific periodicity, as opposed to any periodicity, corrections for multiple comparisons may not be necessary."

Don't just fall into the silly trap of 'more subjects is better', one must look at the quality, not just quantity, and the nature of the questions under investigation.

 

[sesquiterpene]

Have you read the study? These aren't just a few records here and there, but meticulous ones kept by many of the patients themselves, giving about 37 years worth of data for these individuals. Again too, it is in a small subset of rapid-cycling bipolars with strict criteria. This is high quality data in a tiny subset of the population.

It would be ludicrous if they had been followed for one or two cycles or such, but by sheer amount, we are looking at something quite different. I've seen it juxtaposed to the Maudsley psychotic twin study, a classic study with about 80 participants, although we see much closer alignment here. Ordinarily yes, such a tiny sample could be dismissed, but any group followed for many many years, suffering a narrowly defined pathological process, would give very good data even if such low numbers. It is commonly so in rare illnesses or ones with limited subject availability. And again, we were talking about within that population group, not generalised to all.

Besides, this is well reviewed, and old data is not less useful in Medicine than new data, especially when tracking the same things. Our premiere hypothermia studies were conducted in the 1940s for instance.

Anyway, to quote the study on this point, we are looking at a specific criterion, so we can actually adjust for statistical significance:

"Qp value, which estimates the fraction of significant results that are null, is a multiple testing correction to P-values that is commonly used in genomics and proteomics. It provides more statistical power than a Bonferroni correction, but is still conservative. If the χ2 periodogram analysis looks for a specific periodicity, as opposed to any periodicity, corrections for multiple comparisons may not be necessary."

Don't just fall into the silly trap of 'more subjects is better', one must look at the quality, not just quantity, and the nature of the questions under investigation.

Well, i re-read the study - and it didn't look any better the second time around.

Yes, he analyzed the long-term records of a subset of bipolar patients and found that they had - to a statistically significant degree - periodicities in their manic/depressive cycles. The problem is - as essentialsaltes pointed out - they were different periodicities. Indeed, he claims he found five different patterns in a sample of 17 subjects - and yet, with some vigorous hand-waving, he claims each of these patterns is due to a single cause! This is very thin gruel - almost homeopathic - to base his claim that the patterns are due to lunar cycles.

To reiterate, the statistical significance of his findings applies only to the claim that there were regular cycles in his bipolar patients. With so few patients, and so many different results, the results don't support the claim that the cycles were any way consistent with lunar cycles rather than a set of totally random intervals in these cycles. To establish that the length of the periods he found were significant he's going to need a lot more data, i.e., subjects.

Using data that might be almost 100 years old is a bit bizarre, but I guess you don't need to worry about Hipaa regulations and IRBs.

What reviews were you looking at ?

 

[Quid est Veritas?]

The problem is - as essentialsaltes pointed out - they were different periodicities

Doesn't matter, they are all correlated to the length of lunar cycles, even if different. The lunar cycle is such an odd period, especially biologically, such correlations are unlikely to be from something else.

To reiterate, the statistical significance of his findings applies only to the claim that there were regular cycles in his bipolar patients.

Regular cycles the length of a lunar cycle, though.

Using data that might be almost 100 years old is a bit bizarre, but I guess you don't need to worry about Hipaa regulations and IRBs.

Using such old data is quite normal in Medicine. The human organism hasn't changed that much. The strongest evidence for the Neurotransmitter hypothesis in depression is a 1955 study; and our best data for untreated Hypertension is derived from studies from 1928 up till the Framingham study in 1974, seeing that we were ethically bound to treat hypertension thereafter; for instance. Then of course, our continued use of Hypothermia trials conducted by the Nazis, or such.

What reviews were you looking at ?

Wehr has written this up a few times. Here is a review from the NEJM, as an example:
NEJM Journal Watch: Summaries of and commentary on original medical and scientific articles from key medical journals

Regardless, my Psychiatric colleagues seem quite taken with this, from discussions I have had. Obviously more studies will inevitably be conducted, so more data will follow in due course. This is not something one can simply dismiss off-hand to my mind, but the Medical establishment is luckily not doing so.

If this is nothing as you say, EBM will eventually demonstrate as much. Currently though, with our current available data, I don't think we can make this assertion.

 

[FrumiousBandersnatch]

Where does the paper say that any of it is 'conclusive' or 'definitive'?

The abstract hedges its bets: "...suggest a biological mechanism through which transits of one of the moon’s semi-diurnal gravimetric tides might have driven the patients’ bipolar cycles..."

The discussion is also fairly restrained: "The synchronies between lunar orbital cycles and bipolar mood cycles raise the possibility that the lunar cycles entrained or even generated the mood cycles. Although the present study did not address causality, several observations suggest that the hypothesis of a causal relationship between lunar cycles and mood cycles is plausible"

I don't think that bears out the thread claims.

 

[Quid est Veritas?]

Where does the paper say that any of it is 'conclusive' or 'definitive'?

The abstract hedges its bets: "...suggest a biological mechanism through which transits of one of the moon’s semi-diurnal gravimetric tides might have driven the patients’ bipolar cycles..."

The discussion is also fairly restrained: "The synchronies between lunar orbital cycles and bipolar mood cycles raise the possibility that the lunar cycles entrained or even generated the mood cycles. Although the present study did not address causality, several observations suggest that the hypothesis of a causal relationship between lunar cycles and mood cycles is plausible"

I don't think that bears out the thread claims.

You haven't read many medical studies I take it? They are more often than not, couched in such language.

Anyway, the statistics are fairly clear, which is why the articles reviewing this study, for instance the one I posted above, are not doubting his findings, but hypothesising what mechanism may be the cause, as far as I can see. Even the BBC article in the OP states: "none of the scientists contacted for this article dispute Wehr’s basic finding: that his bipolar patients mood swings are rhythmic, and that these rhythms appear to correlate with certain gravitational cycles of the Moon."

I am done defending this, which I originally just posted because I thought it interesting. You can believe the statistical analysis and in effect 43 year's worth of patient data if you want to, or you don't. On Evidence-Based methodology it is sound, so until such time as further data changes this, I am going to continue to see it as valid. This knee-jerk doubt of statistical correlation seems very much akin to Antivaxxers and Homeopathy supporters, which is company one should be wary to keep.

 

[FrumiousBandersnatch]

You haven't read many medical studies I take it? They are more often than not, couched in such language.

I had a career in medical research, so I'm pretty familiar with medical studies

I am done defending this, which I originally just posted because I thought it interesting. You can believe the statistical analysis and in effect 43 year's worth of patient data if you want to, or you don't. On Evidence-Based methodology it is sound, so until such time as further data changes this, I am going to continue to see it as valid. This knee-jerk doubt of statistical correlation seems very much akin to Antivaxxers and Homeopathy supporters, which is company one should be wary to keep.

On the contrary, I agree that it's interesting, and it's possible that some biological clocks/rhythms have a degree of lunar synchronisation; but there have been many larger studies into this over the years, and no consistent results, and it's misleading to call the results of such a small study 'definitive' or 'conclusive', especially when the authors themselves are so tentative in their analysis.

 

[Quid est Veritas?]

I had a career in medical research, so I'm pretty familiar with medical studies


On the contrary, I agree that it's interesting, and it's possible that some biological clocks/rhythms have a degree of lunar synchronisation; but there have been many larger studies into this over the years, and no consistent results, and it's misleading to call the results of such a small study 'definitive' or 'conclusive', especially when the authors themselves are so tentative in their analysis.

Only 'tentative' as to causation; not correlation with Lunar cycles, which is what the study indicated and I was talking about. They are quite firm that these patients' mood cycles are correlated to lunar ones. This may therefore indicate some other factor at play, but thus suggests the former though, which they thus 'tentatively' propose.

Further, they have done larger studies in rapidly cycling Bipolar mood disorder? Of course not. Once again, as I keep need to repeat, I am not addressing external validity or generality, but internal in this specific subset of Bipolars. Further, if you read the study, they have addressed it statistically that a larger patient group is not really required either, based on Q values and that they tried to correlate a specific cycle instead of looking for cyclical correlations in general.

Read the discussion again, as it assumes the "synchronies between lunar orbital cycles and bipolar mood cycles" in this study is a done deal, and then carefully discuss potential mechanisms for the obvious suggested cause, though of course not proven to be. I find it odd that you would make such a mistake if you are familiar with medical studies, and conflate this.

 

[FrumiousBandersnatch]

Only 'tentative' as to causation; not correlation with Lunar cycles, which is what the study indicated and I was talking about. They are quite firm that these patients' mood cycles are correlated to lunar ones. This may therefore indicate some other factor at play, but thus suggests the former though, which they thus 'tentatively' propose.

Further, they have done larger studies in rapidly cycling Bipolar mood disorder? Of course not. Once again, as I keep need to repeat, I am not addressing external validity or generality, but internal in this specific subset of Bipolars. Further, if you read the study, they have addressed it statistically that a larger patient group is not really required either, based on Q values and that they tried to correlate a specific cycle instead of looking for cyclical correlations in general.

Read the discussion again, as it assumes the "synchronies between lunar orbital cycles and bipolar mood cycles" in this study is a done deal, and then carefully discuss potential mechanisms for the obvious suggested cause, though of course not proven to be. I find it odd that you would make such a mistake if you are familiar with medical studies, and conflate this.

The point is that your OP is misleading.

 

[Quid est Veritas?]

The point is that your OP is misleading.

My apologies, was not meant to be. I have edited it already.

 

[sesquiterpene]

You haven't read many medical studies I take it? They are more often than not, couched in such language.

Lol. It is your language that's being criticized. Calling this study "definitive" or "conclusive" is rather unscientific given it's scope.

On Evidence-Based methodology it is sound, so until such time as further data changes this. .

Nonsense. This is a nonblinded, retrospective, uncontrolled experiment with a small sample size, which hasn't been reproduced. On each of these criteria, it's falls on the "less rigorous" side of Evidenced-Based medicine. Only single case studies are lower on the hierarchy. That was where Wehr started, so the hypothesis has made some progression.

I am going to continue to see it as valid.

This is part of the criticism. By evidenced-based criteria I consider it "very tentative".

This knee-jerk doubt of statistical correlation seems very much akin to Antivaxxers and Homeopathy supporters, which is company one should be wary to keep.

Bizarre insult noted. It is you, however, who has been unwilling or unable to evaluate the weaknesses of this article. It is incredibly easy to get false correlations in small sample sizes. Ask any epidemiologist for examples

 

[Quid est Veritas?]

Nonsense. This is a nonblinded, retrospective, uncontrolled experiment with a small sample size, which hasn't been reproduced. On each of these criteria, it's falls on the "less rigorous" side of Evidenced-Based medicine. Only single case studies are lower on the hierarchy. That was where Wehr started, so the hypothesis has made some progression.

Perhaps you should read up a bit about EBM. It is not a system of excluding evidence, but a heuristic technique to suggest best practice, via grading evidence by quality and then applying statistical analysis.

So would it be possible to do such a study blinded? Or randomised? Of course not. We could prospectively attempt one, but that would be confounded by bias. So actually, this is about as good evidence as one could get for this specific hypothesis, on Evidence-Based means. To think that something is not supporting evidence or that evidence that falls lower on the hierarchy can be ignored, is patently incorrect usage of EBM principles. To ignore or denigrate evidence because it is only based on lower tier studies, when higher tier studies are non-existent or impossible to do, is abuse or misapprehension of EBM. This is especially true if done on a priori grounds, such as expecting a larger sample size when the statistical analysis showed it significant regardless - the very purpose of which is to help exclude studies with insufficient data epidemiologically.

This is part of the criticism. By evidenced-based criteria I consider it "very tentative".

EBM can never be 'tentative'. EBM is a deductive method, not an inductive one; so either something is supported, not supported, or there is insufficient data. Nothing is 'tentative'. The whole point of EBM is to try and get around biases and such, which is why the EBM research cycle starts by reducing a query to a form that can be answered via empirically obtained data - reducing thus the hypothesis to an answerable question (incidentally thus it does not use Scientific Method, which at heart is a system of Induction). This is why statistically analysis plays such a large part thereof, to suggest how valid the obtained empiric data actually is. This study does well in that regard.

Guyat et al. defined it in the first place as a Kuhnian paradigm that falls outside the pitfalls of Induction (which is something that some outliers, like the Science-Based Medicine people critique) for exactly these types of studies - those that people are loathe to acknowledge for whatever reason. Certainly more data would be welcome, but what we have is sound and statistically supported as significant, so it would merely either shift the EBM Best Evidence if not supporting it, or confirm it further. EBM doesn't trade in hypotheses but Evidence; so as it stands, EBM can't be 'tentative' for results that failed to show significance statistically - this did though, so is valid evidence, not 'tentative', though lower tier it is true.

Bizarre insult noted. It is you, however, who has been unwilling or unable to evaluate the weaknesses of this article. It is incredibly easy to get false correlations in small sample sizes. Ask any epidemiologist for examples

I have acknowledged the weaknesses. I stated it has no external validity that can be shown for instance; however, its internal validity is not weak. Nor is 43 years of data a 'small sample size', something it only gets in participants - but again, this study is in a very limited type of Bipolar, so that is an expected limitation. It does however pass Epidemiological muster, as its Q values indicate, which you continue to ignore. One can have massive studies, with thousands of participants, whose conclusions can be inconclusive - that is the very reason we do forrest plots and p and q values and the ilk.

Lol. It is your language that's being criticized. Calling this study "definitive" or "conclusive" is rather unscientific given it's scope.

I have been at pains to delineate in what way it is 'definitive' and 'conclusive', but everyone keeps ignoring it. So essentially you are just erecting a Strawman to tilt at. To reiterate, Internally valid and sound, as indicated by its Q values.


I am sorry, but you really are suffering under a whole slew of misunderstandings of EBM practice - unfortunately all too common. It is heuristic, not exclusionary; it is Evidence-based, not about hypothesis, as represented by its statistical usages; lower tier evidence is not ignored if higher tier is not available or impossible to obtain; etc.

 

[essentialsaltes]

EBM can never be 'tentative'.

Come now. I'm sure we've all seen stories about coffee causing cancer or preventing cancer, or having no effect on cancer rates. Does your opinion change as each study is produced? Or do we eye it all with some skepticism in search of better or more conclusive results before offering more than tentative credence? As the researcher says in the story of conflicting studies: ""The disparity in those findings would suggest more research is needed to clarify if there is any relationship between colorectal cancer and coffee," he said."

Although all medicine based on science has some degree of empirical support, EBM goes further, classifying evidence by its epistemologic strength and requiring that only the strongest types (coming from meta-analyses, systematic reviews, and randomized controlled trials) can yield strong recommendations; weaker types (such as from case-control studies) can yield only weak recommendations.

I don't think the current study is any of those three things.

 

[Quid est Veritas?]

Come now. I'm sure we've all seen stories about coffee causing cancer or preventing cancer, or having no effect on cancer rates. Does your opinion change as each study is produced? Or do we eye it all with some skepticism in search of better or more conclusive results before offering more than tentative credence? As the researcher says in the story of conflicting studies: ""The disparity in those findings would suggest more research is needed to clarify if there is any relationship between colorectal cancer and coffee," he said."

Although all medicine based on science has some degree of empirical support, EBM goes further, classifying evidence by its epistemologic strength and requiring that only the strongest types (coming from meta-analyses, systematic reviews, and randomized controlled trials) can yield strong recommendations; weaker types (such as from case-control studies) can yield only weak recommendations.

I don't think the current study is any of those three things.

Pop Science journalism does not equal EBM.

As I said, it is a deductive method. If one study contradicts another, both sets of data are put into a meta-analysis, and the resulting pooled data is statistically evaluated for which conclusion is more valid. Usually we utilise confidence intervals to determine this. EBM assumes non-falsifiability, so that all data can be evaluated, in this manner - another difference from scientific method.

Your article merely illustrates further the common misunderstandings of EBM practice. Most notably the erroneous idea that it ever discards data it terms from weaker evidence classes. Besides, it says weaker recommendations, not only strong ones, and we have no stronger data on the question of Lunar cycles in rapidly cycling bipolars. Still, nothing is termed 'tentative' - it is, it isn't, or insufficient data requiring more study.

EBM is a specific methodology, that differs markedly from standard Scientific Method. If you don't know what it is, such mistaken ideas occur regularly - even amongst doctors. I suggest ypu read more if interested. I can supply some articles.

 

[essentialsaltes]

Most notably the erroneous idea that it ever discards data it terms from weaker evidence classes.

Well, I remain underwhelmed by this weak-but-definitive study.

 

[Quid est Veritas?]

Well, I remain underwhelmed by this weak-but-definitive study.

If I did a retrospective patient series on whether Arsenic was poisonous, I'll also find it definitive, but a weak class of evidence. Doesn't change the conclusion. I wouldn't be allowed to test it with prospective studies or randomised trials.

Fair enough, though.

 

[Rubiks]

What we now call astrology was once (and should properly be) called astromancy. It has not been validated by moon/tide cycles in any way.

"Astrology" was an actual word used by the ancient Greeks (technically "astrologia"). Terms like "biology," "anthropology" do not go back to ancient times, but were coined for scientific purposes.

A prescriptivist could argue either way, here.

 

[sesquiterpene]

You are quite right - I am approaching this from a scientifiic viewpoint (retired chemist) and reading Science-Based Medicine blog, and attributing qualities to Evidenced-based Medicine which it doesn't possess - I apologise.

Perhaps you should read up a bit about EBM. It is not a system of excluding evidence, but a heuristic technique to suggest best practice, via grading evidence by quality and then applying statistical analysis.

So would it be possible to do such a study blinded? Or randomised? Of course not. We could prospectively attempt one, but that would be confounded by bias.

What biases would a prospective trial have that the current trial doesn't? And just as importantly, what biases does the current trial have that a prospective one wouldn't?

So actually, this is about as good evidence as one could get for this specific hypothesis, on Evidence-Based means. To think that something is not supporting evidence or that evidence that falls lower on the hierarchy can be ignored, is patently incorrect usage of EBM principles. To ignore or denigrate evidence because it is only based on lower tier studies, when higher tier studies are non-existent or impossible to do, is abuse or misapprehension of EBM. This is especially true if done on a priori grounds, such as expecting a larger sample size when the statistical analysis showed it significant regardless - the very purpose of which is to help exclude studies with insufficient data epidemiologically.


EBM can never be 'tentative'. EBM is a deductive method, not an inductive one; so either something is supported, not supported, or there is insufficient data. Nothing is 'tentative'. The whole point of EBM is to try and get around biases and such, which is why the EBM research cycle starts by reducing a query to a form that can be answered via empirically obtained data - reducing thus the hypothesis to an answerable question (incidentally thus it does not use Scientific Method, which at heart is a system of Induction). This is why statistically analysis plays such a large part thereof, to suggest how valid the obtained empiric data actually is. This study does well in that regard.

Guyat et al. defined it in the first place as a Kuhnian paradigm that falls outside the pitfalls of Induction (which is something that some outliers, like the Science-Based Medicine people critique) for exactly these types of studies - those that people are loathe to acknowledge for whatever reason. Certainly more data would be welcome, but what we have is sound and statistically supported as significant, so it would merely either shift the EBM Best Evidence if not supporting it, or confirm it further. EBM doesn't trade in hypotheses but Evidence; so as it stands, EBM can't be 'tentative' for results that failed to show significance statistically - this did though, so is valid evidence, not 'tentative', though lower tier it is true.


I have acknowledged the weaknesses. I stated it has no external validity that can be shown for instance; however, its internal validity is not weak. Nor is 43 years of data a 'small sample size', something it only gets in participants - but again, this study is in a very limited type of Bipolar, so that is an expected limitation. It does however pass Epidemiological muster, as its Q values indicate, which you continue to ignore. One can have massive studies, with thousands of participants, whose conclusions can be inconclusive - that is the very reason we do forrest plots and p and q values and the ilk.



I have been at pains to delineate in what way it is 'definitive' and 'conclusive', but everyone keeps ignoring it. So essentially you are just erecting a Strawman to tilt at. To reiterate, Internally valid and sound, as indicated by its Q values.


I am sorry, but you really are suffering under a whole slew of misunderstandings of EBM practice - unfortunately all too common. It is heuristic, not exclusionary; it is Evidence-based, not about hypothesis, as represented by its statistical usages; lower tier evidence is not ignored if higher tier is not available or impossible to obtain; etc.

I think that most of the disagreement here is due to the difference between EBM and SBM. For instance, one of main criticism coming from SBM is that EBM suffers from an undue reliance on clinical studies, and (this criticism come from several sources, I believe) a slavish devotion to p<.05.

I can see the latter in your repeated descriptions of statistically significant findings as 'definitive' and 'conclusive', and lately 'pass[ing] Epidemiological muster'.

This simply isn't true. There are many different ways that even statistically significant data can be misleading or even flat out wrong. As Kimball Atwood remarked on the SBM blog:

Compared to laboratory investigations, clinical trials are necessarily less powered and more prone to numerous other sources of error: biases, whether conscious or not, causing or resulting from non-comparable experimental and control groups, cuing of subjects, post-hoc analyses, multiple testing artifacts, unrecognized confounding of data due to subjects’ own motivations, non-publication of results, inappropriate statistical analyses, conclusions that don’t follow from the data, inappropriate pooling of non-significant data from several, small studies to produce an aggregate that appears statistically significant, fraud, and more. Most of those problems are not apparent in primary reports.

There a several scenarios in which Wehr's statistically significant data might be false. He provides analyses for individual patients but doesn't pool the data from those 43 years. Is there any statistical significance left? You can't tell by reading the article.

You mention that the length of tidal periods are unusual in biology, but a 14 day period is ubiquitous in human society and could easily make their way into the data. Perhaps patients 1-8 just really don't like Mondays, and patients 9-17 have enormous antipathy to Nurse Ratched who fills in on alternate weekends. 14-day periods differ slightly from the tidal periods {edit: and they diverge over time}, but Wehr doesn't try see if the data is good enough to differentiate between the two. You can't tell by reading the article. In fact, Wehr doesn't discuss possible confounders or sources of error at all. ( that alone probably invalidates the 'pass[ing] Epidemiological muster' label ).

My initial comment was spurred by your [mis]use of of Statistically significant as 'definitive' and 'conclusive', and yes that is probably due to the differences between EBM and a more scientific approach. I'll stick with science.

 

[zippy2006]

This knee-jerk doubt of statistical correlation seems very much akin to Antivaxxers and Homeopathy supporters, which is company one should be wary to keep.

At times the irrational biases of atheist/agnostic types emerge with remarkable clarity. It seems that this is one of those cases. Apparently lunar correlation contradicts atheistic dogma. Who knew?

 

[Ophiolite]

At times the irrational biases of atheist/agnostic types emerge with remarkable clarity. It seems that this is one of those cases. Apparently lunar correlation contradicts atheistic dogma. Who knew?

At times the irrational biases of theist types emerge with remarkable clarity. It seems that this is one of those cases. Apparently well founded scepticism leading to healthy discussions on scientific methodology and statistical techniques can be seized upon as an opportunity to push an agenda. Who knew?