Women Spot Snakes Faster Before Their Periods

An Irish non-nun spots a snake.
St Patrick's nun

It is nearly St Patrick’s day and so time to revisit the legend of how that holy man drove the snakes from Ireland. Little known is how he did it.

Turns out St Patrick recruited premenstrual nuns from a nearby convent. He directed these religious to walk before him and spot the snakes, which they did with great efficacy. Once spotted, St. Patrick used his staff to persuade the serpents to head back to England.

St Patrick had tried the same stunt a year earlier, but made the mistake of using nuns who were menstruating. Turns out these nuns spotted snakes about 0.2 seconds slower than premenstrual nuns, which gave the snakes the time they needed to avoid the staff.

It has taken modern science about 1,600 years to confirm the experiment of the sainted Pádraig. Or so say researchers Nobuo Masataka and Masahiro Shibasaki of the Primate Research Institute at Kyoto University.

In the peer-reviewed journal Nature: Scientific Reports, the pair report that in women, “snake detection was found to become temporarily enhanced during the luteal phase as compared to early or late follicular phases.”

The luteal phase is that time of the month in which some women suffer PMS. The enhanced snake-finding abilities thus come as no surprise, as science has earlier discovered that many luteal-phase women are adept at spotting horrors of all kinds, especially flaws in their mates, even counterfactual ones.

I want to stress that this is not yours truly saying these things. It is Masataka and Shibasaki who claim, “Mood, cognition, and social behavior may fluctuate in women across the phases of the menstrual cycle according to changes in the blood concentration levels of a variety of hormones.”

Yesterday I had a discussion with a colleague about my opinion that there is too much research and that professors would be better off concentrating on teaching. This is one of the instances of “research” that was better left undone. Anyway, since we started, we might as well examine the particulars of the study.

They took 60 right-handed, period-producing women and broke them into three sections: early follicular, luteal, late follicular phases.

They sat these cycling women in front of computer screens and asked them to touch either a black and white photo of a snake or of a flower. The time it took to hit the snake or flower picture was recorded (response time = RT). The photos

were displayed in a 3-by-3 matrix. Each matrix contained 1 target image from one category and 8 distracter images from the category: a flower matrix would contain a snake target, and a snake matrix would contain a flower target. This yielded two combinations: a snake among flowers, and a flower among snakes.

No word on what kind of snakes or flowers were used. We do know—and this is the kind of detail tossed in to make papers sound scientific—that a “RDT151TU (MITSUBISHI) touch-screen monitor was used to present each image matrix on a 38.1 cm (15-in.) screen.”

The testing was done in two blocks of 24 trials. Preciously, in between each trial “an image of a stuffed animal or a popular character appeared on the screen to keep her attention on the screen.” This is Japan, after all.

And now it becomes really bizarre. “The first and the second testing [blocks] were separated from one another by a 2- to 3-month-period for each participant.” This is important because the women self-reported their phase of the month.

I can’t tell whether each woman was her own control or whether returning women were put into whatever group matched their current phase. This makes a difference in the way the results should be displayed and analyzed. The former would be the correct way, the latter the incorrect way. Given their Figure 1, it looks like they did it the incorrect way. They present results of means of similarly-phase women within groups, and not means of differences between the same women.

It appears that women of any phase are slightly faster at identifying snakes than flowers. Who would have guessed? But it’s not so clear that the menstrual phase had anything to do with picking snakes faster or slower. For example, in “group A” (we’re not sure who these women are, as above), early follicular picked snakes at a mean 1345 ms +/- 97 ms, versus a mean 1149 ms +/- 87 ms in the luteal phase.

That’s 1345 ms and not 1344 ms nor 1345 ms, mind. In other words, the average across women in their early follicular phase was 1.3 seconds versus the average of those women (we suppose) in their luteal phase was 1.1 seconds. As I said, this is the wrong comparison, but even supposing it’s correct, this is a trivial difference, especially considering that these women are self-reporting their cycle (and may have been incorrectly classified after 3 months) and because of the following juicy fact.

What did they do with the really slow or fast values? Why, they threw them out of course. No use cluttering up a pretty theory with messy data! “RTs of incorrect responses as well as extreme RT scores—defined as values more than 2 standard deviations above or below the mean relative to each participant’s mean RT—were excluded from the analyses.”

What makes this study depressing is the extended “analysis” of the “findings.” We hear terms like “biologically relevant threatening stimuli” and that “adult females during the comparable endocrinological phase were likely to be more isolated from other group members”. The amygdala makes an appearance, as expected, as does the “neurotransmitter GABA.” And don’t let’s forget fMRIs, cortisol, and brain-blood barriers, and blah blah blah.

It’s enough to make one despair.


  1. Are you sure you weren’t reading a description of a Japanese game show rather than a (supposed) scientific experiment?

    Of course the model of the monitor was an important fact to include in the description: Think what might happen if someone tried to reproduce the results, but used a monitor with a different scan rate, which could skew the results by several nanoseconds.

  2. Briggs,

    Your analyis is certainly ok, but, overlooks other VERY fundamental factors.

    Specifically, note the subject study is using a proxy–reaction time–for “ability to detect” something. So, before considering anything else, one must first account for the accuracy/precision of the proxy measured for the ‘thing actually being sought.’ (ignoring the finer points between “accuracy” and “precision” for here & now)

    Turns out that menstrual cycle, hormonal balance changes, DOES affect reaction time: http://www.ncbi.nlm.nih.gov/pubmed/21675039 That shouldn’t come as any surprise.

    But, the author’s study under review above that completely ignores this relevant factor.

    It also ignores how & to what extent an emotionally arousing, via fear/disgust, stimulant will affect reaction response vs. a pleasing/relaxing stimulus. I’m not bothering to do any research on this….but….it just seems logical that a stimulus that normally provokes more or less instant disgust & fear will prompt a faster reaction than something that provokes soothing emotions.

    In other words, the study probably didn’t measure “detection” ability at all — more likely it merely demonstrated the trivial & intuitively obvious: that we respond to things that prompt a strong emotional/visceral response faster than when we respond to things that sooth us. No surprise there. Nor is the trivial finding that biochemical balance is a slight factor affecting response time.

    Which is to suggest: This paper may be a very good gussied up example of the infamous Correlation-equals-Causality error.

  3. Of course you could say the nuns were his staff. He just took the credit.

    Then there’s this. I note the pics were placed by Patrick:

    Question: if a snake gives a girl flowers will he get the girl? If so, how long will it be before she correctly identifies the snake?

    Big Mike,
    you never heard of ad placement?

  4. No, no, no. St. Patrick obviously drove PICTURES of snakes displayed on RDT151TU (MITSUBISHI) touch-screen monitors that presented their image matrix on a 38.1 cm (15-in.) screen out of Ireland. Otherwise, this research would be useless.

  5. “Its enough to make one despair.”

    Clearly a bad news/good news situation:

    The bad news is that you are now desperate.

    The good news that you clearly lie about 10 SD past the peak of the sanguinity bell curve or you would have been desperate LONG AGO.

  6. Jim and I might regard the Japanese researchers and their research as being the product of an alien type of human ie another race. Our “racism” is supposedly predicated on fear. Since racism is irrational our fear is supposedly irrational. Beta blockers or alcohol might reduce our levels of irrational fear but would we still be able to raise rational objection? Or would the drug take away our ability to see and hear or interpret what we saw and heard?

    If our levels of fear were reduced would we be more likely to favourably consider the scientific report of Masakata and Shibasaki? Or would we be more likely to enthusiastically ridicule it as being the product of a silly culture?

    There is no racism since there is no race. But there is culturalism since there are diverse cultures and all are to some extent stupid, a great many downright perverse.

Leave a Comment

Your email address will not be published. Required fields are marked *