Did you try controlling for age? Because of the menopause, I expect you'd see a lot more older women in the non-PMS cohort. If older women are (for whatever reason) inherently less likely to believe in the supernatural, it might look like a correlation between PMS and supernatural beliefs. Ditto for neuroticism.
(I'd probably have expected older generations to have *more* belief in the supernatural, but who knows. It does seem plausible that younger women are more neurotic.)
(I ask because women generally believe in the supernatural at a higher rate than men, right? So I have to wonder if there could be a common cause between that and PMS...)
Maybe it would be (or have been) useful to ask the same women if they had the same strange beliefs before they had their period. One could imagine that PMS leads to neuroticism (as you suggested) and that neuroticism then leads to strange beliefs.
This seems like confounders galore here, of people who report PMS symptoms; of people who believe in supernatural; and of people who self-report as neurotic.
"This maybe slightly contradicts a suggestibility hypothesis (lots of people say Donald Trump is bad; wouldn’t that convince more suggestible people that Donald Trump is bad?)"
'Suggestibility' here is a tricky concept that is probably *not* best understood as "being suggestible" like "believing things other people tell you". Think something like: how much can you *believe* in something? Can you make tulpas, be hypnotized, have intense reactions to psychoaffective drugs, spot Patterns, daydream for hours, dive deep into fiction? The correlation between these things and "believing what other people tell you" is quite possibly negative. I...have no idea how this should interact with PMS, though, which seems to be suggestible in the other sense in Aella's framework. (In all likelihood, there are multiple axises of "how, phenomenologically, someone is a member of the believing-in-things cluster" that are covered up by studying this in largeish quantitative studies and are, amongst other things, gender-skewed".)
Probably two factors. People who have more easily perterbed emotions in general may see stronger shifts from PMS hormones and have a stronger reaction to their own strange experiences, leading to belief in the occult.
People who don't experience strong mood swings as a general rule may simply be more level-headed and come up with more rational explations for odd experiences.
I'm confused, are people entertaining the possibility that PMS is not real? As in, that there is a significant fraction of (menstruating, cis) women who are miserable for several days before their period? Is it some "growth mindset"-like wishful thinking?
My theory would be that if you have more problems (especially like PMS which makes it clear you are not in total control of your mind) you tend more towards supernatural explanations because they are useful to keep you functioning through the problems, and you also have more neuroticism because bad things do in fact regularly happen to you...
The claim about hormones seems more plausible to me. Most people I've heard about who believe in the supernatural will cite having weird sensations as one cause of believe, and unusual sensations also feature prominently in most supernatural stories. Hormones can fuck with your perceptions and mood quite a lot, so that connection would be rather straightforward, and that PMS is connected to hormones is already the mainstream position anyway.
Though also voting for "likely to be spurious correlation" as well.
Isn't it generally agreed that conditions like PCOS affect mental health, particularly stress responses and anxiety. And hormonal contraception is sometimes prescribed to help with that. So would give precedent for the hormonal explanation. If people feel unexplained anxiety they will subconsciously look for causes and that may influence belief in supernatural
Might be interesting to include question on those kind of conditions in the next survey?
Your number one hypothesis should always be: DEMOGRAPHICS. Is the correlation robust to controlling for age? For country/state? Education level?
9 times out of 10 such things end up being some boring demographic thing like "young women get PMS and believe more in the supernatural".
If it's not some simple demographic thing, perhaps skeptical women are vaguely offended by the existence of PMS and try not to attribute their moods to it. But before jumping to such explanations I want to see demographic checks.
Seems like a proxy for proclivity to see causal patterns. If you only need one or two cases to accept that PMS is (above the threshold you impute the question to present) causing an effect maybe you only need one or 2 cases to conclude prayer of grandma's locket caused something.
I don't claim to know anything about anything but recently I've been thinking about the idea that a lot of stuff people refer to as "body-mind connection" (and a lot of experiential emotion stuff more generally) has to do with like, patterns of muscle tension that people do when having various feelings/experiences. like, I have noticed that when feeling stressed/anxious/afraid I will tend to semi-volitionally tense a lot of muscles, especially abdominal muscles; I think I also do a version of this (maybe differently?) when assigning myself a task I want to not forget? this then has further effects on my felt experience, and if it's a chronic thing then it likely also has effects on very physical body things (e.g. if you are chronically stressed and chronically tense your ab muscles this probably affects how you move and may end up affecting other muscles and stuff in the area). also I think the connection goes both ways somewhat - you can mimic a emotional state by messing with bodily signals.
anyway periods certainly do stuff to muscle tension patterns and it makes some sense to me that this might end up interacting with emotions sometimes and maybe interacting with how people tend to experience emotions in general (though I'm struggling to formulate the exact mechanism of the latter - maybe something like, if someone tends to feel a lot of anxiety with accompanying muscle tension they're likelier to interpret other muscle tension as anxiety or have that sensation trigger anxiety since it's a very familiar state for them? I guess like, sometimes when I've been sufficiently anxious I've given myself actual abdominal pain as a result, and if someone has that experience a lot and then they have period-related abdominal pain that might feel like anxiety or trigger anxiety?)
...but also, if someone does get PMS for say, hormonal reasons, I think one of the emotions it may cause *is anxiety*, in which case it is obviously true that people who get PMS have more anxiety than people who don't
(and also, people who get anxiety more in general will also have more total instances of "anxiety AND near period" and thus more material to infer spurious causality from and are not necessarily *that* careful about checking base rates)
I am a young-ish woman who doesn't take hormonal contraception and I get very strong PMS in the sense that if I know I feeling psychologically particularly unwell I know I will get my period the next few days. I have OCD so there is a very concrete ability to feel my mental stability in realising how strong my obsessions are. I also get my period very unregularly so it's not like I would be able to deduce my period coming from the passed days, so I never doubted PMS's reality. I know many mentally ill women who feel the same. So something I could see from anecdotal evidence is that maybe PMS amplifies preexisting mental conditions and that maybe neurotic people are more likely to believe in ghosts.
Next year ask if people experience bloating before their periods, because that is believed to be a result of hormones as well. Of course bloating could be making people feel 'fat', and that might be effecting their mood independently of any hormonal effects.
This makes 100% sense to me if we look at pms less as a fully psychosomatic condition and more as something that will be mediated by sensitivity (physical, cognitive, some neuropsychological in-between) to pain, discomfort and even just change. The best measure for that I can think of is Big5 N -- called neuroticism but in fact related more to how intensely people experience "negative" emotions, not necessarily "being neurotic" in the psychopathology sense.
It seems to me though that self reported "being neurotic" and "proper actual N" will be correlated (maybe highly) because humans as such are oversensitive to threats.
So -- still hypothesising -- regardless of "actual" intensity of PMS (assuming it's even a thing), women with higher N will be more likely to notice it, and report it.
The idea that higher N will also influence beliefs in the supernatural feels (irrationally and intuitively) plausible, if we accept the premise that delusional beliefs (beliefs in things that don't exist NOT SHARED BY OTHERS IN THE COMMUNITY/SOCIETY - that's the take used for example in WHO psychosis protocol), paranoia, anxiety, other fear-related conditions are all part of the same bundle related to higher sensitivity to threat signals, then it also sticks.
Obligatory introspective anecdata: when asked if I experienced PMS when menstruating "naturally", I had trouble even remembering if I had it. I suspect it would have needed to be very very bad for me to notice and register. Then my brain broke post-menopause as a result of one traumatic experience that sensitised me to body-malfunction and now I notice every twinge and ache and any change. Was I still menstruating, I'd probably notice subtle changes I'd not have even registered 10-15 years ago.
This is a terrible question for self-reporting, as it’s going to pick up noise from different women’s theory of the self.
Anecdotally, I’ve encountered women who clearly experience mood/behavioural changes when they’re pre-menstrual but insist they don’t (in one instant, only realising after it was pointed out to them). If they’d answered the survey, they’d have been no||no.
I think the root of this is lots of people (especially men, but plenty of women) have a worldview in which the mind isn’t influenced by the condition of the body. This very easily leads to people thinking they don’t get PMS because it would never occur to them that this was a thing. This could very easily apply to whole cultures, explaining away the culture-bound theory if the idea isn’t floating around in that culture.
The real comparison questions should be “Do you react differently to non-food related situations when you’re hungry?” or “How much are your thoughts affected by chemical processes in your body?”
The second of these correlations isn’t surprising, as if you accept you’re “neurotic” you accept your cognition isn’t purely rational. Belief in ghosts is more of a stretch, but does suggest a massive gulf in worldview which should come up all over the place.
"This is bad practice and for exploratory purposes only"
One thing I've been thinking a lot about is *why* it's bad practice. Obviously there's the p-hacking/multiple comparisons issue, but the alternative to testing lots of correlations is to use solid theory to decide which correlations to test, and that requires a place for the theory to come from. If that place doesn't ultimately ground out in correlations, then where does it ground out?
Over time, I've come up with some answers to this:
* The unit of analysis is a big issue. If you correlate over people, then you shatter any connections that might have been observed at a smaller unit of analysis, e.g. if the periods coincide with some other thing.
* Root causes of phenomena often have high skewness (because skewness tends to diffuse away as you go down through the causal chains, so non-root-causes tend to have lower skewness). If you have a fixed set of quantitative or semi-quantitative questions in a survey, then it's hard to observe information with high skewness, because the highly skewed variables that are relevant vary a lot from person to person. On the other hand, if you looked qualitatively at a single individual, it would be more obvious.
* Correlation adjusts for the variance in both the predictor and the outcome, but it is easier for causality to flow from variables with high variance to variables with low variance than vice versa. Thus one should focus the analysis on variables with especially high variance.
I'm currently working on an alternative survey method that might take these sorts of things better into account, by allowing respondents to define sub-units within their responses, and allowing respondents to rank the relative importance of answers within each sub-unit. Obviously by passing more of the analytic work to respondents, this would make the surveys more dependent on the respondents' reasoning than on my reasoning, but given the weakness of correlation-based research, that might be a compromise we're going to have to accept.
I'm surprised that 38% of (cis-female) ACX readers believe in the supernatural, I would have assumed that number to be way lower.
Is this a case of this community having a higher standard for what counts as non-supernatural than the general population or something?
Edit: Checked the full results in the survey. 28% has 2+ in the believes in the supernatural question 31% is some level of religious or spiritual so I guess this community is just more religiously/spiritually diverse than I was expecting it to be.
More severe PMS coincides with higher adrenaline levels, and is linked to trauma history. This also presents physiological symptoms that people are able to track and observe with common wearable devices. This might be a causal factor towards other correlates.
I suspect that one person might classify as 'mild', another person might classify as 'severe' and vice versa and perhaps it says as much about as how somebody perceives such symptoms as to the actual symptoms themselves. I wonder if some women pay more attention to the timings of their mood changes than others. Some might see mood differences as more random uncorrelated fluctuations while others might seek an 'explanation' or pattern more.
I'm a little annoyed that post-menopausal women aren't included. I really mean a little annoyed, I'm not planning to go social justice on your ass. Are there just too few in the survey populations to be worth mentioning?
Anyway, moderate but real symptoms of pain and mood swings when I had periods, nothing debilitating. If I remember correctly, I answered the ACX survey as though I still had periods.
I'm pretty neurotic.
I believe in energy, but not so much in non-physical entities, and not at all in ideas about destiny or in intention controlling reality. I believe in rather naturalistic supernaturalism. Speaking of, I once did an energy manipulation which made PMS drain way and dissolve, but I never got it to work again.
The common denominator is that both phenomena are obviously true.
Of course getting your period affects your mood. Even if it doesn’t have hormonal effects on mood, it would be astonishing if regular bleeding and cramps didn’t exert at least a small pressure towards moodiness (the same way that I would be inclined to be cranky if I was unable to remove a stone from my shoe). Admitting this shows intellectual humility; denying it shows inability or unwillingness to connect obvious evidence out of intellectual pride.
Likewise, there is a spiritual reality to the universe, which exerts such a profound impact on the human soul (giving us beauty, morality and natural law, charity and grace, etc.) that denying its existence says more than anything that your self-knowledge is lacking and you would deny the hand in front of your face in order to assert your intellectual independence.
Of course this is only a theory, the intellectual pride thing; there could be other reasons why there is a group that fails to miss both obvious insights.
I strongly suspect some of this is due to reporting bias. When studies of PMS and PMDD are conducted, it's apparently best practice (I found this out just now, investigating a hypothesis of my own) to use prospective rather than retrospective reporting, because the latter leads to substantial inflation. In one paper (https://doi.org/10.53260/GREM.212016) that looked at this specifically, "The MDQ [Menstrual Distress Questionnaire] total scores in the retrospective trial were significantly greater compared with those recorded in the prospective late-luteal trial (p < 0.001). The average value of the overestimation was 23.7 ± 35.0%." This is among currently menstruating women; among women who have to reach further back in their memory (due to menopause, pregnancy, breastfeeding, contraception, sterilization, or disease) it's probably higher. And this includes less subjective measures than mood. On a scale from 0 - 5 (more on the MEDI-Q here https://doi.org/10.1080/09513590.2023.2227275), mood swings specifically dropped from a mean of 3.00 to 2.15. This study has limitations that would lead you to expect some underreporting in the prospective phase but the problem seems well-established, insofar as lots of people agree that it exists.
Being aware and accurately assessing that your mood is subject to hormonal influence due to your menstrual cycle is not the same thing as believing and reporting that your mood is subject to hormonal influence, and there are probably type II-ish errors as well--women who do get more irritable or sad in before they menstruate, but who haven't really noticed or thought about it much. All of the reasons people might pay more or less attention to, reflect on, track, or put labels on their own emotions are potential confounds.
Fertility counselor here. Most women experience PMS between and before their period, not when their period arrives. This can vary (even in the same woman) from happening right after ovulation, a week after ovulation, or right before their periods arrive - that is a solid 2 week swing that is NOT a woman socially conditioned to "feel bad around when their period comes".
Many women also have irregular menstrual cycles at least part of the time, and do not know when their period will arrive. Yet they still experience mood swings even when they do not know their period is coming. (One time I didn't have my period for many months due to breastfeeding, and one day out of the blue I felt literally insane - jittery, upset, anxious over everything, and I had no idea why - and my period arrived that afternoon. I had no expectation of that happening, I just felt weird and upset for no explainable reason.) Again, this does not support social conditioning because women with irregular cycles are not expecting their periods to come at a particular place in time, but ARE experiencing largely negative mood swings regardless of any expectation.
If PMS arrives unpredictably for most women all the time, and for ALL women at least some of the time, not concurrent with the period arriving but beforehand, and in a way that is not directly linked to a calendar date, then there is simply no way that PMS is psychosomatic.
Doctors have found that many health issues are linked to "neuroticism" including many that are provably true. Some speculate that this may be because neurotic people have more inflammation or different hormonal profiles due to stress, that neurotic people sleep poorly and their overall health/mental health suffers, or it may be that there's a separate element causing or contributing to PMS and neuroticism. I think jumping right to an assumption that PMS is psychosomatic on the basis of such a link is quite a stretch and is honestly a bit weird.
I found out that my SO's answer on "do you get PMS symptoms that affect your mood?" is 1 and she complains often and has visited doctors often in past about the condition. Looks like lying.
I strongly suspect this correlates to being a Highly Sensitive Person (hsp). This is a scientific term not meant as sarcasm, not saying that's a bad thing at all.
> Now 72% of people with PMS self-describe as neurotic, compared to only 45% without. Aella writes more about this here, and sebjenseb confirms here. I’m less weirded out by this one, because you can imagine that people feel neurotic because of PMS symptoms, but it’s still a surprisingly strong effect.
You can imagine that, though you could imagine a handful of other causations for that particular correlation. If you Google "what causes PMS symptoms," the general consensus that emerges is "we don't know, but it's probably some sort of hormonal imbalance." What causes those hormonal imbalances? We don't know.
Is it possible that they're not feeling neurotic because of PMS symptoms, but rather that they're getting PMS because of their neurosis, or because of the same underlying cause?
What if whether you get mood-related symptoms due to hormonal fluctuations is less about the hormonal fluctuations themselves and more about your ability to regulate your emotions, a releasing-mechanism like interaction effect?
The reason I suspect this is the sky-high prevalence of postpartum depression and psychosis in unmedicated women with bipolar disorder. Almost uniquely among the mental illnesses, bipolar disorder has about equal prevalence in men and women. For this reason we don't really think hormones have anything to do with causing the disorder in the first place. But bipolar women appear to be much more sensitive to their own hormonal fluctuations. They have higher rates of PMS and PMDD (which, technically, probably, they shouldn't get diagnosed with, but they meet the inclusion criteria at higher rates) and, as mentioned, have a curious tendency to land in the hospital not so long after giving birth. (I don't think this can be chalked up to any kind of social performance, internalized or otherwise--post-partum psychosis is not widely known about, even among bipolar women, and having homicidal feelings about your baby because the rapture is coming and you don't want them to suffer through the second coming is very not conforming to expectations. If you don't already believe it, and you don't want to look into the literature supporting it, phenomena like this should increase your credence that hormones can affect mood quite a bit!) Treating bipolar with medication reduces endorsement of PMS symptoms, and, by a large margin, postpartum psychopathology. (For awhile I would take a slightly higher dose of lithium in the week before I menstruated specifically to control the irritability that I struggle with chronically but could really get out of hand then.)
Sebastian Jensen had a charming graphic that proposed an additive effect of hormones on, uh, craziness level, but I'm thinking more of an interaction effect, as mentioned, and I wouldn't be surprised if women with other disorders that have a mood dysregulation or subpar executive function component (ADHD, BPD, substance use/behavioral addiction) also have higher rates of mood issues with PMS.
The extent to which "poor emotional regulation" and "believing in ghosts and witches" might associate I'll leave to others' speculation.
Anyone read Charles Taylor? My theory would be that the correlation of supernatural/PMS is that both are correlated with having a more “porous” as opposed to “buffered” sense of self.
Firstly, I'd submit that the set of people likely to follow Aella are not the same as the mass of the general population, so there is probably a higher chance that for those who do believe in non-conventional views will also believe in energy fields healing or whatever.
Second, what is understood by the supernatural? We covered this already in regards to the survey, but "I believe in God but not ghosts or energy healing" - is that belief in the supernatural or not?
Third, I think I would agree that there is a cultural element to expectations around menstruation, and that may indeed affect how women approach it. For instance, I never heard of the term "cramps" and only encountered it as American slang in books and movies, and famously the name of a rock band, and then later online, and eventually found out what it meant specifically. I was *very* surprised to find out that there seemed to be an entire set of expectations and activities around menstruation if you were American; you got 'cramps' which were not just ordinary stomach or muscle cramps of the kind anyone might get, these were specific period pains; these were supposed to be debilitating to the extent that there was specific for women painkillers (Midol) and conventions such as hot water bottles to the abdomen, taking to your bed, etc.
So far as I knew, I had none of this. Did I get period pains? Well yeah, of course, at times - just like every other woman. Did I get the American version of having periods? Not that I knew. Now, looking back on it, probably I did but I didn't have either the vocabulary or the expectations around it.
Enter PMS when I'm in my 20s or so as a topic coming to public notice. Again, did I consider I got PMS? No.
Did I get more moody/irritable etc. around the time of having my period? Yes. Do I believe in the supernatural (for a limited definition of that?) Also yes.
I think this had much more to do with my particular hormonal levels (I always had very heavy flow) and nothing to do with cultural or religious expectations, though. While I was vaguely aware of the term "the curse", this did not give me expectations that my period would, or should, be painful or debilitating, and believing in God or the fairies had nothing to do with how I experienced each month.
So the relationship from the survey questions is curious, but I don't think strongly connected. A society where the majority of people are religious, or raised in religious families, even weakly so, is going to have a lot of women who believe in 'the supernatural', even if that's the horoscopes/crystals/Wicca/spiritual not religious sort of belief, and that will mean an overlap anyway with "women who have PMS" regardless.
I think the cultural American expectations around menstruation may have more to do with it, but I have no idea how those arose. I'd look into when ads for pain relievers specifically for period pains began, because I think if you have a couple of generations of women who are raised on "this is specifically for women, it's specifically for period pains, that means I should expect periods to be painful" and they see "every month Mom has to take these and go to bed with a heat patch or hot water bottle", there is much more likelihood to be both the expectation of PMS and persuading oneself into "this pain isn't just a pain, it's PMS".
The suggestibility hypothesis is interesting because some, but not all, aspects of PMS seem to be cultural specific.
For instance, in Anglo countries, craving chocolate is associated with PMS. While in East Asian cultures, women with PMS are known to feel abnormally cold. These types of trends led some feminists in the late 20th century to assume that all PMS symptoms were cultural-bound. But beyond some cultural-specific examples, the general PMS profile seems to be pretty consistent: bloating, headache, fatigue, irritability — mostly mild, but the severity varies throughout a woman's life and among different women. Women who experience high rates of PMS also report having more severe post-partum and menopause symptoms (also related to drops in estrogen and progesterone) — it would be surprising if this didn't relate to any inherent mood changes.
So there's something going on physiologically with PMS, but it also seems to rack up these cultural-bound associations. Overdetecting trends, or psychosomatically manifesting them (e.g. "I crave chocolate during this particular time of the month"), seems like a similar phenomenon to the overdetection of trends in horoscopes, etc.
As a woman who experiences PMS (and doesn't believe in the occult):
- It's quite humbling to notice strong feelings, especially negative ones towards others, which I can't explain, and then realize it's been almost two weeks since I ovulated. When this happens, I consciously choose to wait about 48 hours and only engage with the feelings if they don't go away.
- This experience is as a periodic (pun intended) reminder of how strange the concept of "me" is.
- While I think some aspects of PMS might be cultural, based on my experience, other anecdotal evidence and understanding how the hormonal cycle works, I think there is a clear basis for biological causes of PMS.
- I didn’t notice any PMS symptoms while I was on hormonal contraception.
- I think I experience productivity peaks during ovulation, but I wouldn't be terribly surprised if this turned out to be largely psychosomatic.
I suspect that female readers of this blog are less likely to experience strong PMS compared to the general population for three reasons: 1. They may naturally experience fewer hormonal/mood shifts, making them more analytical. 2. They might be too much in their heads to notice it. 3. They are more likely to be on the pill.
I’m not sure which factor is the most significant.
I haven’t yet seen reliable data on women’s mood and productivity relative to their menstrual cycle, specifically data that considers whether they are on any form of hormonal contraception. Even some pretty good studies from period-tracking apps do not collect this information, which makes the data less informative (https://www.lesswrong.com/posts/fnCz8S6PocTxpB2ie/menstrual-cycle-effects-clue-study-summary-and-commentary). Again, self-reporting and self-fulfilling prophecies might be an issue.
From my experience, getting off the pill seemed to increase both my average mood and its variance; I'm happy most of the time, occasionally super happy, but sad and confused during PMS. While on the pill, my mood felt mostly the same without ups or downs.
I think the effects are quite individual, and for some women, the pill might act like a transhumanist-like intervention, improving mood and period symptoms with few drawbacks.
Next time, I'd like to see the data how being on a pill relates to self-reported happiness, anxiety, etc.
I know Scott has written previously about PMS not appearing to be 'internalized social expectations', and I have what I believe to be an interesting anecdote that may support this. I'm still breastfeeding my 13-month-old baby and just had my first menstrual period since before conceiving. There was no way to know exactly when it was coming - I've felt random cramps intermittently for the last few months, and my son is eating slightly less, so I had a hunch it was coming, but with my eldest, I didn't menstruate until I weaned him completely.
However, last week, I was having a moody day, and found myself in the evening thinking "everything is awful, I'm useless, my career is doomed, I'm a joke, I should give up." These are very uncommon thoughts for me - I'm usually quite optimistic and reasonable. It was so uncommon, and such a specific feeling - sudden, crushing hopelessness - that I knew it was PMS and told my partner I'd be getting my first period in over a year two or three days later. And I did!
I suppose there is a chance that there is an unconscious response to the hormonal changes which induces the socially learned behavior of feeling hopeless - in the same way that my son getting hungrier led me to snack more unconsciously to provide more milk without really consciously realizing in the moment either that he was eating more, or I was. But the crushing hopelessness is such a specific and unusual feeling for me that I think it may be directly caused by hormonal fluctuations, and I find this anecdote really rules out a psychosomatic response for me.
I am not sure that an ACX survey is the best way to get solid data on the link between believing in the supernatural and PMS.
First, the ACX readership is very skeptic, 71% replied with an emphatic No (1) when asked about the supernatural, with half of the remainder replying 2, which I would parse as "not really but I try to keep an open mind". Less than 10% replied anything affirmative, i.e. 4-5. Counting all the people who did not deny the supernatural in the harshest possible terms as Believers does not really cut reality at its joints in my opinion.
Then there is the fact that the readership is overwhelmingly male biologically. Only 12% are female. This is uncomfortably close to the Lizardman constant.
I am not an expert on female physiology (understatement of the year, there), but from my understanding not every woman menstruates. Hormonal birth control, being outside a certain age window, pregnancy, breastfeeding and probably a zillion medical conditions seem all likely reasons why some of the cis-female readers might not have their period, and hence also not get PMS. In the public results, I see no question about that, so it is impossible to control for that.
While the trend observed by Aella is clearly visible in the data, this does not mean that we can trust the ACX survey data to be good in the first place.
Furthermore methodically, I think that if you have answers to two questions on a scale then the natural way to compute correlations would be to first plot them into a two-dimensional histogram (with 5 times 5 bins), then do a linear correlation a la Pearson.
If you have answers to a question like the supernatural which are distributed like this:
1: 71.8%
2: 14.2%
3: 5.5%
4: 4.7%
5: 0.39%
(from the overall responses)
then each answer has an average information content of 1.36 bits. If you now lump two to five together, your information content per answer is only 0.86 bits -- less than two thirds.
In a way, mapping {1,2,3,4,5} using identity is just as arbitrary a decision as mapping 1 to zero and everything else to one, but throwing out information does not seem like good practice to me.
(I know that the answer to that is to download the data and analyse it myself. Of course, the only statistics tool I actually know how to use is ROOT, which is terrible even when used for physics. There is probably a scipy package or something though.)
I remember being uncertain how to answer the PMS question on the survey, due to the wording ("do you get PMS symptoms that affect your mood?"). PMS does affect my mood, but in a positive direction. I tend to feel energized and confident in the few days before my period, or more accurately, the few days before the week when I'm supposed to get my period according to the birth control pills. I don't always actually bleed that week, probably because I'm in my late 40's.
Would intensification of feelings/emotions explain this? I would bet there's a correlation between intensification and religious/supernatural beliefs (many people are religious because of some emotional religious experience they had/continue to have). Single data point: my wife has pretty bad PMS and also has very expressive emotions.
From personal experience, I'm going to vote for hormones. My cycle was very unpredictable for a while and I would have a day of being really upset and most, often leading to a right with my husband, and it would be the day before the period came. Since it was irregular, I wasn't expecting to feel bad in anticipation of my regular expected period.
> 19% of women without pre-menstrual symptoms believed in the supernatural, compared to 39% of women with PMS. I can’t do chi-squared tests in my head, but with 1,074 votes this looks significant. Weird!
You got similar results, probably because both of you are drawing disproportionately from rationalists or rationalist-adjacent groups, who are less likely to believe in ghosts and spirits. I don't have a good hypothesis for why this correlation would exist all, though.
First thought was skepticism. People that are especially skeptical of aliens and ghosts are skeptical of PMS. But my hunch is if you ran it back with something like climate change, PMS-skeptics would not be global warming skeptics. So second guess is tribal affiliation. I'd wager PMS skeptics are more likely to believe that Russia colluded with Donald Trump in 2016.
The matter of whether period pains in the abdomen are cramps which resemble muscle cramps is interesting-- I got some pain, but it was more generic than what I experience with muscle cramps. Muscle cramps are clearly localized, and their distinguishing feature is that trying to move in a way the muscle cramp doesn't like is quite painful, and some other movement might release the cramp, or maybe the cramp just needs time to let go.
Those abdominal pains weren't related to movement, at least not for me. I don't know if I would have called them cramps except that cramps was the usual name.
You don't want to dichotomize variables. This loses information and statistical precision. Use a regression model, which also allows you to control for other things. Plots are your friend, SPSS is not. Any LLM will write this kind of basic analysis code for you in R/Python.
I've been suspecting for a while now that Aella's survey style of asking "Do you A? || Do you B?" is flawed. It attracts responders who believe there is a correlation between the two. Second, her surveys are very opt-in and people see the question before they opt in, which I think attracts people who find their answers an interesting fact about themselves.
I'm not surprised by this. People who are more stressed out have more need of religion and spirituality to deal with the stress. My understanding is religiosity is negatively correlated with income. PMS and neuroticism seem like poverty in that they produce more need of the kind of succor people use religion and spirituality for. This would be true whether or not the PMS was psychosomatic.
If PMS is mainly suggestion, that implies people keep really good track of their periods, and so know "this is when I should be upset". I have a tracking app, really dread being caught short at work, and yet I still sometimes forget when it's coming. I can often predict my period by the way I'm feeling (particularly regarding sleep), so this feels like the wrong order of cause and effect for PMS to have a high proportion of suggestibility.
I'd be interested to know if PMS correlates with cramps, and then if chronic pain correlates with belief in the supernatural. From the people I know, people going through things (bereavement, illness, etc) are often more religious. That's one potential causal link I can think of.
Unrelated: there's not much actual research on periods, I'd be very interested in a follow-up survey about periods. I would happily help draft or just participate in some questions.
How does psychosomatic joint mobility work. The joints range is clearly visible isnt. Does the psychosomatic case just bend the joint farther than its supposed to and therefore dislocates it regularly? (Psychosomatically suppressing the pain)
Or is it that the non mobile ppl dont bend the joint as far as it can actually go because they are convinced it cant?
This is anecdotal, but I know a woman who gets mentally wrecked by all big hormonal changes. This has included puberty, pregnancy, all her menstrual periods and now menopause. She takes antidepressants to cope. For women like her, PMS is absolutely real.
One thing that would have been good to do is to reverse the polarity of the questions for half of respondents (agree (0)->disagree(10) for half, disagree(0)->agree(10)), so that you could eliminate individual biases in how people use the scale (which was taught as a real thing back when I was in Cog Sci grad school, though how knows if it survived the replication crisis).
Basically, you have a whole lot of seemingly random measures all correlating with each other on a 0-10 no-to-yes scale. One way that can happen is if some people use teh edges of teh scale more than others, so one person gives more 8s on everything while another gives more 6s, despite their actual states of belief being pretty much identical.
Not saying this is definitely the answer, but when you get weird results on a single instrument it's always good to think about the instrument itself and look for alternate ways to measure the phenomena.
I would say that the common variable for me among experiencing PMS/PMDD, being involved in "spooky stuff" (like ghosts, cryptids, and true crime) and being neurotic is definitely being neurodivergent (Aspergers, ADHD).
It's pretty well known at this point that there is a huge overlap between ADHD/ASD and more severe PMS symptoms. The internal sensitivity to hormonal changes is pretty much exactly the same as sensitivity to external stimuli like temperature and sound (and research shows that hormonal changes during PMS increase sound and smell sensitivity).
I wouldn't say I believe in ghosts as much as I'm "interested" in ghosts (i.e., I'm culturally spooky rather than religiously spooky). I recently went to a ghost hunting event and the vast majority of attendees were neurodivergent women skeptics. Experiencing and then debunking spooky experiences is a great hobby if you have sensory issues and prefer to go to events in the dark, with soft ambient music, and sit in silence rather than talk to people, and then go home and obsessively research mapping/measurement techniques and the history of old buildings.
PMS means pre, not post
Did you try controlling for age? Because of the menopause, I expect you'd see a lot more older women in the non-PMS cohort. If older women are (for whatever reason) inherently less likely to believe in the supernatural, it might look like a correlation between PMS and supernatural beliefs. Ditto for neuroticism.
(I'd probably have expected older generations to have *more* belief in the supernatural, but who knows. It does seem plausible that younger women are more neurotic.)
How do these compare to for men?
(I ask because women generally believe in the supernatural at a higher rate than men, right? So I have to wonder if there could be a common cause between that and PMS...)
Maybe it would be (or have been) useful to ask the same women if they had the same strange beliefs before they had their period. One could imagine that PMS leads to neuroticism (as you suggested) and that neuroticism then leads to strange beliefs.
This seems like confounders galore here, of people who report PMS symptoms; of people who believe in supernatural; and of people who self-report as neurotic.
"This maybe slightly contradicts a suggestibility hypothesis (lots of people say Donald Trump is bad; wouldn’t that convince more suggestible people that Donald Trump is bad?)"
'Suggestibility' here is a tricky concept that is probably *not* best understood as "being suggestible" like "believing things other people tell you". Think something like: how much can you *believe* in something? Can you make tulpas, be hypnotized, have intense reactions to psychoaffective drugs, spot Patterns, daydream for hours, dive deep into fiction? The correlation between these things and "believing what other people tell you" is quite possibly negative. I...have no idea how this should interact with PMS, though, which seems to be suggestible in the other sense in Aella's framework. (In all likelihood, there are multiple axises of "how, phenomenologically, someone is a member of the believing-in-things cluster" that are covered up by studying this in largeish quantitative studies and are, amongst other things, gender-skewed".)
Probably two factors. People who have more easily perterbed emotions in general may see stronger shifts from PMS hormones and have a stronger reaction to their own strange experiences, leading to belief in the occult.
People who don't experience strong mood swings as a general rule may simply be more level-headed and come up with more rational explations for odd experiences.
I'm confused, are people entertaining the possibility that PMS is not real? As in, that there is a significant fraction of (menstruating, cis) women who are miserable for several days before their period? Is it some "growth mindset"-like wishful thinking?
My theory would be that if you have more problems (especially like PMS which makes it clear you are not in total control of your mind) you tend more towards supernatural explanations because they are useful to keep you functioning through the problems, and you also have more neuroticism because bad things do in fact regularly happen to you...
The claim about hormones seems more plausible to me. Most people I've heard about who believe in the supernatural will cite having weird sensations as one cause of believe, and unusual sensations also feature prominently in most supernatural stories. Hormones can fuck with your perceptions and mood quite a lot, so that connection would be rather straightforward, and that PMS is connected to hormones is already the mainstream position anyway.
Though also voting for "likely to be spurious correlation" as well.
Isn't it generally agreed that conditions like PCOS affect mental health, particularly stress responses and anxiety. And hormonal contraception is sometimes prescribed to help with that. So would give precedent for the hormonal explanation. If people feel unexplained anxiety they will subconsciously look for causes and that may influence belief in supernatural
Might be interesting to include question on those kind of conditions in the next survey?
Your number one hypothesis should always be: DEMOGRAPHICS. Is the correlation robust to controlling for age? For country/state? Education level?
9 times out of 10 such things end up being some boring demographic thing like "young women get PMS and believe more in the supernatural".
If it's not some simple demographic thing, perhaps skeptical women are vaguely offended by the existence of PMS and try not to attribute their moods to it. But before jumping to such explanations I want to see demographic checks.
Seems like a proxy for proclivity to see causal patterns. If you only need one or two cases to accept that PMS is (above the threshold you impute the question to present) causing an effect maybe you only need one or 2 cases to conclude prayer of grandma's locket caused something.
I don't claim to know anything about anything but recently I've been thinking about the idea that a lot of stuff people refer to as "body-mind connection" (and a lot of experiential emotion stuff more generally) has to do with like, patterns of muscle tension that people do when having various feelings/experiences. like, I have noticed that when feeling stressed/anxious/afraid I will tend to semi-volitionally tense a lot of muscles, especially abdominal muscles; I think I also do a version of this (maybe differently?) when assigning myself a task I want to not forget? this then has further effects on my felt experience, and if it's a chronic thing then it likely also has effects on very physical body things (e.g. if you are chronically stressed and chronically tense your ab muscles this probably affects how you move and may end up affecting other muscles and stuff in the area). also I think the connection goes both ways somewhat - you can mimic a emotional state by messing with bodily signals.
anyway periods certainly do stuff to muscle tension patterns and it makes some sense to me that this might end up interacting with emotions sometimes and maybe interacting with how people tend to experience emotions in general (though I'm struggling to formulate the exact mechanism of the latter - maybe something like, if someone tends to feel a lot of anxiety with accompanying muscle tension they're likelier to interpret other muscle tension as anxiety or have that sensation trigger anxiety since it's a very familiar state for them? I guess like, sometimes when I've been sufficiently anxious I've given myself actual abdominal pain as a result, and if someone has that experience a lot and then they have period-related abdominal pain that might feel like anxiety or trigger anxiety?)
...but also, if someone does get PMS for say, hormonal reasons, I think one of the emotions it may cause *is anxiety*, in which case it is obviously true that people who get PMS have more anxiety than people who don't
(and also, people who get anxiety more in general will also have more total instances of "anxiety AND near period" and thus more material to infer spurious causality from and are not necessarily *that* careful about checking base rates)
I am a young-ish woman who doesn't take hormonal contraception and I get very strong PMS in the sense that if I know I feeling psychologically particularly unwell I know I will get my period the next few days. I have OCD so there is a very concrete ability to feel my mental stability in realising how strong my obsessions are. I also get my period very unregularly so it's not like I would be able to deduce my period coming from the passed days, so I never doubted PMS's reality. I know many mentally ill women who feel the same. So something I could see from anecdotal evidence is that maybe PMS amplifies preexisting mental conditions and that maybe neurotic people are more likely to believe in ghosts.
Next year ask if people experience bloating before their periods, because that is believed to be a result of hormones as well. Of course bloating could be making people feel 'fat', and that might be effecting their mood independently of any hormonal effects.
Wouldn't it be stranger if a significant shift in hormonal balance *didn't* produce any symptoms?
This makes 100% sense to me if we look at pms less as a fully psychosomatic condition and more as something that will be mediated by sensitivity (physical, cognitive, some neuropsychological in-between) to pain, discomfort and even just change. The best measure for that I can think of is Big5 N -- called neuroticism but in fact related more to how intensely people experience "negative" emotions, not necessarily "being neurotic" in the psychopathology sense.
It seems to me though that self reported "being neurotic" and "proper actual N" will be correlated (maybe highly) because humans as such are oversensitive to threats.
So -- still hypothesising -- regardless of "actual" intensity of PMS (assuming it's even a thing), women with higher N will be more likely to notice it, and report it.
The idea that higher N will also influence beliefs in the supernatural feels (irrationally and intuitively) plausible, if we accept the premise that delusional beliefs (beliefs in things that don't exist NOT SHARED BY OTHERS IN THE COMMUNITY/SOCIETY - that's the take used for example in WHO psychosis protocol), paranoia, anxiety, other fear-related conditions are all part of the same bundle related to higher sensitivity to threat signals, then it also sticks.
Obligatory introspective anecdata: when asked if I experienced PMS when menstruating "naturally", I had trouble even remembering if I had it. I suspect it would have needed to be very very bad for me to notice and register. Then my brain broke post-menopause as a result of one traumatic experience that sensitised me to body-malfunction and now I notice every twinge and ache and any change. Was I still menstruating, I'd probably notice subtle changes I'd not have even registered 10-15 years ago.
This is a terrible question for self-reporting, as it’s going to pick up noise from different women’s theory of the self.
Anecdotally, I’ve encountered women who clearly experience mood/behavioural changes when they’re pre-menstrual but insist they don’t (in one instant, only realising after it was pointed out to them). If they’d answered the survey, they’d have been no||no.
I think the root of this is lots of people (especially men, but plenty of women) have a worldview in which the mind isn’t influenced by the condition of the body. This very easily leads to people thinking they don’t get PMS because it would never occur to them that this was a thing. This could very easily apply to whole cultures, explaining away the culture-bound theory if the idea isn’t floating around in that culture.
The real comparison questions should be “Do you react differently to non-food related situations when you’re hungry?” or “How much are your thoughts affected by chemical processes in your body?”
The second of these correlations isn’t surprising, as if you accept you’re “neurotic” you accept your cognition isn’t purely rational. Belief in ghosts is more of a stretch, but does suggest a massive gulf in worldview which should come up all over the place.
"This is bad practice and for exploratory purposes only"
One thing I've been thinking a lot about is *why* it's bad practice. Obviously there's the p-hacking/multiple comparisons issue, but the alternative to testing lots of correlations is to use solid theory to decide which correlations to test, and that requires a place for the theory to come from. If that place doesn't ultimately ground out in correlations, then where does it ground out?
Over time, I've come up with some answers to this:
* The unit of analysis is a big issue. If you correlate over people, then you shatter any connections that might have been observed at a smaller unit of analysis, e.g. if the periods coincide with some other thing.
* Root causes of phenomena often have high skewness (because skewness tends to diffuse away as you go down through the causal chains, so non-root-causes tend to have lower skewness). If you have a fixed set of quantitative or semi-quantitative questions in a survey, then it's hard to observe information with high skewness, because the highly skewed variables that are relevant vary a lot from person to person. On the other hand, if you looked qualitatively at a single individual, it would be more obvious.
* Correlation adjusts for the variance in both the predictor and the outcome, but it is easier for causality to flow from variables with high variance to variables with low variance than vice versa. Thus one should focus the analysis on variables with especially high variance.
I'm currently working on an alternative survey method that might take these sorts of things better into account, by allowing respondents to define sub-units within their responses, and allowing respondents to rank the relative importance of answers within each sub-unit. Obviously by passing more of the analytic work to respondents, this would make the surveys more dependent on the respondents' reasoning than on my reasoning, but given the weakness of correlation-based research, that might be a compromise we're going to have to accept.
I'm surprised that 38% of (cis-female) ACX readers believe in the supernatural, I would have assumed that number to be way lower.
Is this a case of this community having a higher standard for what counts as non-supernatural than the general population or something?
Edit: Checked the full results in the survey. 28% has 2+ in the believes in the supernatural question 31% is some level of religious or spiritual so I guess this community is just more religiously/spiritually diverse than I was expecting it to be.
More severe PMS coincides with higher adrenaline levels, and is linked to trauma history. This also presents physiological symptoms that people are able to track and observe with common wearable devices. This might be a causal factor towards other correlates.
See the following links and twitter discussion.
https://pubmed.ncbi.nlm.nih.gov/9858034/
https://x.com/mgoldingmd/status/1393508177455452162
https://x.com/captain_mrs/status/1542748142843711488
Here's one testable story which would explain it away via a confound:
rural -> religious -> supernaturalism
rural -> religious -> not on the Pill -> PMS symptoms
I suspect that one person might classify as 'mild', another person might classify as 'severe' and vice versa and perhaps it says as much about as how somebody perceives such symptoms as to the actual symptoms themselves. I wonder if some women pay more attention to the timings of their mood changes than others. Some might see mood differences as more random uncorrelated fluctuations while others might seek an 'explanation' or pattern more.
Regarding the initial Aella poll, here's an online calculator results (after plausible rounding):
"The chi-square statistic is 41.3771. The p-value is < .00001. The result is significant at p < .05."
Maybe ask a gynecologist
I'm a little annoyed that post-menopausal women aren't included. I really mean a little annoyed, I'm not planning to go social justice on your ass. Are there just too few in the survey populations to be worth mentioning?
Anyway, moderate but real symptoms of pain and mood swings when I had periods, nothing debilitating. If I remember correctly, I answered the ACX survey as though I still had periods.
I'm pretty neurotic.
I believe in energy, but not so much in non-physical entities, and not at all in ideas about destiny or in intention controlling reality. I believe in rather naturalistic supernaturalism. Speaking of, I once did an energy manipulation which made PMS drain way and dissolve, but I never got it to work again.
Okay, I’ll bite the bullet that no one else will.
The common denominator is that both phenomena are obviously true.
Of course getting your period affects your mood. Even if it doesn’t have hormonal effects on mood, it would be astonishing if regular bleeding and cramps didn’t exert at least a small pressure towards moodiness (the same way that I would be inclined to be cranky if I was unable to remove a stone from my shoe). Admitting this shows intellectual humility; denying it shows inability or unwillingness to connect obvious evidence out of intellectual pride.
Likewise, there is a spiritual reality to the universe, which exerts such a profound impact on the human soul (giving us beauty, morality and natural law, charity and grace, etc.) that denying its existence says more than anything that your self-knowledge is lacking and you would deny the hand in front of your face in order to assert your intellectual independence.
Of course this is only a theory, the intellectual pride thing; there could be other reasons why there is a group that fails to miss both obvious insights.
I strongly suspect some of this is due to reporting bias. When studies of PMS and PMDD are conducted, it's apparently best practice (I found this out just now, investigating a hypothesis of my own) to use prospective rather than retrospective reporting, because the latter leads to substantial inflation. In one paper (https://doi.org/10.53260/GREM.212016) that looked at this specifically, "The MDQ [Menstrual Distress Questionnaire] total scores in the retrospective trial were significantly greater compared with those recorded in the prospective late-luteal trial (p < 0.001). The average value of the overestimation was 23.7 ± 35.0%." This is among currently menstruating women; among women who have to reach further back in their memory (due to menopause, pregnancy, breastfeeding, contraception, sterilization, or disease) it's probably higher. And this includes less subjective measures than mood. On a scale from 0 - 5 (more on the MEDI-Q here https://doi.org/10.1080/09513590.2023.2227275), mood swings specifically dropped from a mean of 3.00 to 2.15. This study has limitations that would lead you to expect some underreporting in the prospective phase but the problem seems well-established, insofar as lots of people agree that it exists.
Being aware and accurately assessing that your mood is subject to hormonal influence due to your menstrual cycle is not the same thing as believing and reporting that your mood is subject to hormonal influence, and there are probably type II-ish errors as well--women who do get more irritable or sad in before they menstruate, but who haven't really noticed or thought about it much. All of the reasons people might pay more or less attention to, reflect on, track, or put labels on their own emotions are potential confounds.
Fertility counselor here. Most women experience PMS between and before their period, not when their period arrives. This can vary (even in the same woman) from happening right after ovulation, a week after ovulation, or right before their periods arrive - that is a solid 2 week swing that is NOT a woman socially conditioned to "feel bad around when their period comes".
Many women also have irregular menstrual cycles at least part of the time, and do not know when their period will arrive. Yet they still experience mood swings even when they do not know their period is coming. (One time I didn't have my period for many months due to breastfeeding, and one day out of the blue I felt literally insane - jittery, upset, anxious over everything, and I had no idea why - and my period arrived that afternoon. I had no expectation of that happening, I just felt weird and upset for no explainable reason.) Again, this does not support social conditioning because women with irregular cycles are not expecting their periods to come at a particular place in time, but ARE experiencing largely negative mood swings regardless of any expectation.
If PMS arrives unpredictably for most women all the time, and for ALL women at least some of the time, not concurrent with the period arriving but beforehand, and in a way that is not directly linked to a calendar date, then there is simply no way that PMS is psychosomatic.
Doctors have found that many health issues are linked to "neuroticism" including many that are provably true. Some speculate that this may be because neurotic people have more inflammation or different hormonal profiles due to stress, that neurotic people sleep poorly and their overall health/mental health suffers, or it may be that there's a separate element causing or contributing to PMS and neuroticism. I think jumping right to an assumption that PMS is psychosomatic on the basis of such a link is quite a stretch and is honestly a bit weird.
I found out that my SO's answer on "do you get PMS symptoms that affect your mood?" is 1 and she complains often and has visited doctors often in past about the condition. Looks like lying.
I strongly suspect this correlates to being a Highly Sensitive Person (hsp). This is a scientific term not meant as sarcasm, not saying that's a bad thing at all.
> Now 72% of people with PMS self-describe as neurotic, compared to only 45% without. Aella writes more about this here, and sebjenseb confirms here. I’m less weirded out by this one, because you can imagine that people feel neurotic because of PMS symptoms, but it’s still a surprisingly strong effect.
You can imagine that, though you could imagine a handful of other causations for that particular correlation. If you Google "what causes PMS symptoms," the general consensus that emerges is "we don't know, but it's probably some sort of hormonal imbalance." What causes those hormonal imbalances? We don't know.
Is it possible that they're not feeling neurotic because of PMS symptoms, but rather that they're getting PMS because of their neurosis, or because of the same underlying cause?
What if whether you get mood-related symptoms due to hormonal fluctuations is less about the hormonal fluctuations themselves and more about your ability to regulate your emotions, a releasing-mechanism like interaction effect?
The reason I suspect this is the sky-high prevalence of postpartum depression and psychosis in unmedicated women with bipolar disorder. Almost uniquely among the mental illnesses, bipolar disorder has about equal prevalence in men and women. For this reason we don't really think hormones have anything to do with causing the disorder in the first place. But bipolar women appear to be much more sensitive to their own hormonal fluctuations. They have higher rates of PMS and PMDD (which, technically, probably, they shouldn't get diagnosed with, but they meet the inclusion criteria at higher rates) and, as mentioned, have a curious tendency to land in the hospital not so long after giving birth. (I don't think this can be chalked up to any kind of social performance, internalized or otherwise--post-partum psychosis is not widely known about, even among bipolar women, and having homicidal feelings about your baby because the rapture is coming and you don't want them to suffer through the second coming is very not conforming to expectations. If you don't already believe it, and you don't want to look into the literature supporting it, phenomena like this should increase your credence that hormones can affect mood quite a bit!) Treating bipolar with medication reduces endorsement of PMS symptoms, and, by a large margin, postpartum psychopathology. (For awhile I would take a slightly higher dose of lithium in the week before I menstruated specifically to control the irritability that I struggle with chronically but could really get out of hand then.)
Sebastian Jensen had a charming graphic that proposed an additive effect of hormones on, uh, craziness level, but I'm thinking more of an interaction effect, as mentioned, and I wouldn't be surprised if women with other disorders that have a mood dysregulation or subpar executive function component (ADHD, BPD, substance use/behavioral addiction) also have higher rates of mood issues with PMS.
The extent to which "poor emotional regulation" and "believing in ghosts and witches" might associate I'll leave to others' speculation.
Anyone read Charles Taylor? My theory would be that the correlation of supernatural/PMS is that both are correlated with having a more “porous” as opposed to “buffered” sense of self.
Oh, boy. This is one of those vexed questions.
Firstly, I'd submit that the set of people likely to follow Aella are not the same as the mass of the general population, so there is probably a higher chance that for those who do believe in non-conventional views will also believe in energy fields healing or whatever.
Second, what is understood by the supernatural? We covered this already in regards to the survey, but "I believe in God but not ghosts or energy healing" - is that belief in the supernatural or not?
Third, I think I would agree that there is a cultural element to expectations around menstruation, and that may indeed affect how women approach it. For instance, I never heard of the term "cramps" and only encountered it as American slang in books and movies, and famously the name of a rock band, and then later online, and eventually found out what it meant specifically. I was *very* surprised to find out that there seemed to be an entire set of expectations and activities around menstruation if you were American; you got 'cramps' which were not just ordinary stomach or muscle cramps of the kind anyone might get, these were specific period pains; these were supposed to be debilitating to the extent that there was specific for women painkillers (Midol) and conventions such as hot water bottles to the abdomen, taking to your bed, etc.
So far as I knew, I had none of this. Did I get period pains? Well yeah, of course, at times - just like every other woman. Did I get the American version of having periods? Not that I knew. Now, looking back on it, probably I did but I didn't have either the vocabulary or the expectations around it.
Enter PMS when I'm in my 20s or so as a topic coming to public notice. Again, did I consider I got PMS? No.
Did I get more moody/irritable etc. around the time of having my period? Yes. Do I believe in the supernatural (for a limited definition of that?) Also yes.
I think this had much more to do with my particular hormonal levels (I always had very heavy flow) and nothing to do with cultural or religious expectations, though. While I was vaguely aware of the term "the curse", this did not give me expectations that my period would, or should, be painful or debilitating, and believing in God or the fairies had nothing to do with how I experienced each month.
So the relationship from the survey questions is curious, but I don't think strongly connected. A society where the majority of people are religious, or raised in religious families, even weakly so, is going to have a lot of women who believe in 'the supernatural', even if that's the horoscopes/crystals/Wicca/spiritual not religious sort of belief, and that will mean an overlap anyway with "women who have PMS" regardless.
I think the cultural American expectations around menstruation may have more to do with it, but I have no idea how those arose. I'd look into when ads for pain relievers specifically for period pains began, because I think if you have a couple of generations of women who are raised on "this is specifically for women, it's specifically for period pains, that means I should expect periods to be painful" and they see "every month Mom has to take these and go to bed with a heat patch or hot water bottle", there is much more likelihood to be both the expectation of PMS and persuading oneself into "this pain isn't just a pain, it's PMS".
Why is a t-test being used on ordinal data? Am I missing something??
The suggestibility hypothesis is interesting because some, but not all, aspects of PMS seem to be cultural specific.
For instance, in Anglo countries, craving chocolate is associated with PMS. While in East Asian cultures, women with PMS are known to feel abnormally cold. These types of trends led some feminists in the late 20th century to assume that all PMS symptoms were cultural-bound. But beyond some cultural-specific examples, the general PMS profile seems to be pretty consistent: bloating, headache, fatigue, irritability — mostly mild, but the severity varies throughout a woman's life and among different women. Women who experience high rates of PMS also report having more severe post-partum and menopause symptoms (also related to drops in estrogen and progesterone) — it would be surprising if this didn't relate to any inherent mood changes.
So there's something going on physiologically with PMS, but it also seems to rack up these cultural-bound associations. Overdetecting trends, or psychosomatically manifesting them (e.g. "I crave chocolate during this particular time of the month"), seems like a similar phenomenon to the overdetection of trends in horoscopes, etc.
As a woman who experiences PMS (and doesn't believe in the occult):
- It's quite humbling to notice strong feelings, especially negative ones towards others, which I can't explain, and then realize it's been almost two weeks since I ovulated. When this happens, I consciously choose to wait about 48 hours and only engage with the feelings if they don't go away.
- This experience is as a periodic (pun intended) reminder of how strange the concept of "me" is.
- While I think some aspects of PMS might be cultural, based on my experience, other anecdotal evidence and understanding how the hormonal cycle works, I think there is a clear basis for biological causes of PMS.
- I didn’t notice any PMS symptoms while I was on hormonal contraception.
- I think I experience productivity peaks during ovulation, but I wouldn't be terribly surprised if this turned out to be largely psychosomatic.
I suspect that female readers of this blog are less likely to experience strong PMS compared to the general population for three reasons: 1. They may naturally experience fewer hormonal/mood shifts, making them more analytical. 2. They might be too much in their heads to notice it. 3. They are more likely to be on the pill.
I’m not sure which factor is the most significant.
I haven’t yet seen reliable data on women’s mood and productivity relative to their menstrual cycle, specifically data that considers whether they are on any form of hormonal contraception. Even some pretty good studies from period-tracking apps do not collect this information, which makes the data less informative (https://www.lesswrong.com/posts/fnCz8S6PocTxpB2ie/menstrual-cycle-effects-clue-study-summary-and-commentary). Again, self-reporting and self-fulfilling prophecies might be an issue.
From my experience, getting off the pill seemed to increase both my average mood and its variance; I'm happy most of the time, occasionally super happy, but sad and confused during PMS. While on the pill, my mood felt mostly the same without ups or downs.
I think the effects are quite individual, and for some women, the pill might act like a transhumanist-like intervention, improving mood and period symptoms with few drawbacks.
Next time, I'd like to see the data how being on a pill relates to self-reported happiness, anxiety, etc.
I know Scott has written previously about PMS not appearing to be 'internalized social expectations', and I have what I believe to be an interesting anecdote that may support this. I'm still breastfeeding my 13-month-old baby and just had my first menstrual period since before conceiving. There was no way to know exactly when it was coming - I've felt random cramps intermittently for the last few months, and my son is eating slightly less, so I had a hunch it was coming, but with my eldest, I didn't menstruate until I weaned him completely.
However, last week, I was having a moody day, and found myself in the evening thinking "everything is awful, I'm useless, my career is doomed, I'm a joke, I should give up." These are very uncommon thoughts for me - I'm usually quite optimistic and reasonable. It was so uncommon, and such a specific feeling - sudden, crushing hopelessness - that I knew it was PMS and told my partner I'd be getting my first period in over a year two or three days later. And I did!
I suppose there is a chance that there is an unconscious response to the hormonal changes which induces the socially learned behavior of feeling hopeless - in the same way that my son getting hungrier led me to snack more unconsciously to provide more milk without really consciously realizing in the moment either that he was eating more, or I was. But the crushing hopelessness is such a specific and unusual feeling for me that I think it may be directly caused by hormonal fluctuations, and I find this anecdote really rules out a psychosomatic response for me.
As a happily married man I know better than to offer any opinion on a woman’s subjective experience.
I am not sure that an ACX survey is the best way to get solid data on the link between believing in the supernatural and PMS.
First, the ACX readership is very skeptic, 71% replied with an emphatic No (1) when asked about the supernatural, with half of the remainder replying 2, which I would parse as "not really but I try to keep an open mind". Less than 10% replied anything affirmative, i.e. 4-5. Counting all the people who did not deny the supernatural in the harshest possible terms as Believers does not really cut reality at its joints in my opinion.
Then there is the fact that the readership is overwhelmingly male biologically. Only 12% are female. This is uncomfortably close to the Lizardman constant.
I am not an expert on female physiology (understatement of the year, there), but from my understanding not every woman menstruates. Hormonal birth control, being outside a certain age window, pregnancy, breastfeeding and probably a zillion medical conditions seem all likely reasons why some of the cis-female readers might not have their period, and hence also not get PMS. In the public results, I see no question about that, so it is impossible to control for that.
While the trend observed by Aella is clearly visible in the data, this does not mean that we can trust the ACX survey data to be good in the first place.
Furthermore methodically, I think that if you have answers to two questions on a scale then the natural way to compute correlations would be to first plot them into a two-dimensional histogram (with 5 times 5 bins), then do a linear correlation a la Pearson.
If you have answers to a question like the supernatural which are distributed like this:
1: 71.8%
2: 14.2%
3: 5.5%
4: 4.7%
5: 0.39%
(from the overall responses)
then each answer has an average information content of 1.36 bits. If you now lump two to five together, your information content per answer is only 0.86 bits -- less than two thirds.
In a way, mapping {1,2,3,4,5} using identity is just as arbitrary a decision as mapping 1 to zero and everything else to one, but throwing out information does not seem like good practice to me.
(I know that the answer to that is to download the data and analyse it myself. Of course, the only statistics tool I actually know how to use is ROOT, which is terrible even when used for physics. There is probably a scipy package or something though.)
I remember being uncertain how to answer the PMS question on the survey, due to the wording ("do you get PMS symptoms that affect your mood?"). PMS does affect my mood, but in a positive direction. I tend to feel energized and confident in the few days before my period, or more accurately, the few days before the week when I'm supposed to get my period according to the birth control pills. I don't always actually bleed that week, probably because I'm in my late 40's.
Would intensification of feelings/emotions explain this? I would bet there's a correlation between intensification and religious/supernatural beliefs (many people are religious because of some emotional religious experience they had/continue to have). Single data point: my wife has pretty bad PMS and also has very expressive emotions.
From personal experience, I'm going to vote for hormones. My cycle was very unpredictable for a while and I would have a day of being really upset and most, often leading to a right with my husband, and it would be the day before the period came. Since it was irregular, I wasn't expecting to feel bad in anticipation of my regular expected period.
> 19% of women without pre-menstrual symptoms believed in the supernatural, compared to 39% of women with PMS. I can’t do chi-squared tests in my head, but with 1,074 votes this looks significant. Weird!
I know you've pointed out that correlations often survive selection effects, but only 1/3 of respondents believe in the supernatural, which seems wildly out of line with the general population. E.g. https://news.gallup.com/poll/16915/three-four-americans-believe-paranormal.aspx or https://thehill.com/blogs/blog-briefing-room/4400922-americans-ghosts-aliens-devil-survey/
You got similar results, probably because both of you are drawing disproportionately from rationalists or rationalist-adjacent groups, who are less likely to believe in ghosts and spirits. I don't have a good hypothesis for why this correlation would exist all, though.
First thought was skepticism. People that are especially skeptical of aliens and ghosts are skeptical of PMS. But my hunch is if you ran it back with something like climate change, PMS-skeptics would not be global warming skeptics. So second guess is tribal affiliation. I'd wager PMS skeptics are more likely to believe that Russia colluded with Donald Trump in 2016.
The matter of whether period pains in the abdomen are cramps which resemble muscle cramps is interesting-- I got some pain, but it was more generic than what I experience with muscle cramps. Muscle cramps are clearly localized, and their distinguishing feature is that trying to move in a way the muscle cramp doesn't like is quite painful, and some other movement might release the cramp, or maybe the cramp just needs time to let go.
Those abdominal pains weren't related to movement, at least not for me. I don't know if I would have called them cramps except that cramps was the usual name.
You don't want to dichotomize variables. This loses information and statistical precision. Use a regression model, which also allows you to control for other things. Plots are your friend, SPSS is not. Any LLM will write this kind of basic analysis code for you in R/Python.
I've been suspecting for a while now that Aella's survey style of asking "Do you A? || Do you B?" is flawed. It attracts responders who believe there is a correlation between the two. Second, her surveys are very opt-in and people see the question before they opt in, which I think attracts people who find their answers an interesting fact about themselves.
I can't believe that no one's suggested the most parsimonious explanation: PMS is caused by evil spirits.
I'm not surprised by this. People who are more stressed out have more need of religion and spirituality to deal with the stress. My understanding is religiosity is negatively correlated with income. PMS and neuroticism seem like poverty in that they produce more need of the kind of succor people use religion and spirituality for. This would be true whether or not the PMS was psychosomatic.
If PMS is mainly suggestion, that implies people keep really good track of their periods, and so know "this is when I should be upset". I have a tracking app, really dread being caught short at work, and yet I still sometimes forget when it's coming. I can often predict my period by the way I'm feeling (particularly regarding sleep), so this feels like the wrong order of cause and effect for PMS to have a high proportion of suggestibility.
I'd be interested to know if PMS correlates with cramps, and then if chronic pain correlates with belief in the supernatural. From the people I know, people going through things (bereavement, illness, etc) are often more religious. That's one potential causal link I can think of.
Unrelated: there's not much actual research on periods, I'd be very interested in a follow-up survey about periods. I would happily help draft or just participate in some questions.
How does psychosomatic joint mobility work. The joints range is clearly visible isnt. Does the psychosomatic case just bend the joint farther than its supposed to and therefore dislocates it regularly? (Psychosomatically suppressing the pain)
Or is it that the non mobile ppl dont bend the joint as far as it can actually go because they are convinced it cant?
This is anecdotal, but I know a woman who gets mentally wrecked by all big hormonal changes. This has included puberty, pregnancy, all her menstrual periods and now menopause. She takes antidepressants to cope. For women like her, PMS is absolutely real.
She does believe in the supernatural.
One thing that would have been good to do is to reverse the polarity of the questions for half of respondents (agree (0)->disagree(10) for half, disagree(0)->agree(10)), so that you could eliminate individual biases in how people use the scale (which was taught as a real thing back when I was in Cog Sci grad school, though how knows if it survived the replication crisis).
Basically, you have a whole lot of seemingly random measures all correlating with each other on a 0-10 no-to-yes scale. One way that can happen is if some people use teh edges of teh scale more than others, so one person gives more 8s on everything while another gives more 6s, despite their actual states of belief being pretty much identical.
Not saying this is definitely the answer, but when you get weird results on a single instrument it's always good to think about the instrument itself and look for alternate ways to measure the phenomena.
I would say that the common variable for me among experiencing PMS/PMDD, being involved in "spooky stuff" (like ghosts, cryptids, and true crime) and being neurotic is definitely being neurodivergent (Aspergers, ADHD).
It's pretty well known at this point that there is a huge overlap between ADHD/ASD and more severe PMS symptoms. The internal sensitivity to hormonal changes is pretty much exactly the same as sensitivity to external stimuli like temperature and sound (and research shows that hormonal changes during PMS increase sound and smell sensitivity).
I wouldn't say I believe in ghosts as much as I'm "interested" in ghosts (i.e., I'm culturally spooky rather than religiously spooky). I recently went to a ghost hunting event and the vast majority of attendees were neurodivergent women skeptics. Experiencing and then debunking spooky experiences is a great hobby if you have sensory issues and prefer to go to events in the dark, with soft ambient music, and sit in silence rather than talk to people, and then go home and obsessively research mapping/measurement techniques and the history of old buildings.