Pupil Size Is a Marker of Intelligence | Hacker News Reader

Pupil Size Is a Marker of Intelligence

scientificamerican.com

70 pointsdavesailer25 comments

Doesn't appear to have held up to replication. [1]

>However, unlike Tsukuhara et al., we could establish only small to nonexistent associations between cognitive ability and pupil size (r’s -.01 to .06, r with g specifically = .05)

In fact, a quick scan down the Scholar results shows other replications have found the opposite effect [2]

>The results found that high-Gf adolescents showed smaller tonic pupil size in both tasks

[1] https://www.researchgate.net/profile/Emil-O-W-Kirkegaard/pub... [2] https://onlinelibrary.wiley.com/doi/epdf/10.1002/pchj.397

But their paper argues that the null results were due to A)using tasks that do not have the strongest association (i.e. they did not measure fluid intelligence), and B)setting baseline lighting levels too brightly so that pupils began at the minimum possible size. https://www.sciencedirect.com/science/article/abs/pii/S00100...

So we will see going forward.

Emil Kirkegaard the individual who wants to legalise child porn and reduce age of consent to 13 year olds? https://rationalwiki.org/wiki/Emil_Kirkegaard

If this is a replicable result, they are probably measuring subjects' interest in being in the lab and doing tasks. Some people care, and pay attention and do well. Others are just there for class credit and blow it all off.

If this was a programming/deep-learning thread, we'd all feel like we'd better shut up if we didn't have anything actually insightful/knowledgeable to say.

But when it comes to neuroscience and psychology threads, everyone here thinks they know a lot and wants to comment. But if there is one thing that the majority of HN know next to nothing about its neuroscience and psychology, but boy do they have opinions. Its like a bloody kegger.

Also economics. The database for this message board must contain a million proposals for re-architeching society from scratch.

Don’t small pupils correlate with stress? If a participant did an iq test and was smart they might be more relaxed than a participant who knew they would do badly.

Confidence and skills do not necessarily correlate, see the Dunning–Kruger effect for instance.

Personally, if I enter some kind of contest, I am much more stressed if I know that I can win.

>Confidence and skills do not necessarily correlate, see the Dunning–Kruger effect for instance.

D-K can be summarized roughly as “self-perceived relative skill is directly correlated with actual relative skill, but systematically compressed toward about the 70th percentile”. (For some reason people mistake the finding that those at the bottom of relative skill perceive their relative skill higher than it is, and those at the top perceive theirs lower than it actually is for an inverse relationship, when it involves nothing of the sort.)

So, it is not at all an illustration of non-correlation between confidence and skill.

No, just the opposite. https://www.scientificamerican.com/article/eye-opener-why-do....

Retrospective: microdosing LSD is just cargo-culting this study.

A genuine disservice to science and Scientific American. Some topics tend to attract correlative dross.

"The diameter of the pupil in normal conditions varies within the limits of 2.7-4.8 mm in the dark (22.5-40% of iris diameter), an average of 3.5 mm (approximately 29% of iris diameter). In newborns, pupils are physiologically narrowed (2.32±0.05 mm, that is 19-20% of iris diameter). From the second month, the pupils increase in size gradually an average of 64% up to the sixth month, and more by 8.6 % to the first year of age. This process takes place within the tenth to the eleventh years for a child. Between 10 to 14 years of age, the opposite dynamics are observed, which can be subdivided into three periods. During the first period (15 to 24 years), pupil dimensions are decreased considerably by 17-19%. In the course of the second period (25 to 50 years), narrowing is going less intensively (on average by 6-9% every five years). During the third period (50 to 80 years), the pupils are relatively narrow (2.63±0.06 mm – approximately 22 % of iris diameter) undergoing a smaller decrease in diameter, which is 3-4% every ten years.":

Source: https://pupilanalysis.com/pupillary-dimensions/

If this turns out to have merit, I can imagine all sorts of shady uses for this information. Our faces and biometric info are already out there, and this doesn’t take a fancy ML algorithm to use. However, the other comments seem to indicate it’s bunk, but an idea that’s wrong can still have purchase.

I'm interested to see how this compares to other examples of profiling, such as carding. The government, police, advertisers, etc. are all already doing it to varying degrees.

Pupil size is yet another metric by which to classify individuals in the masses. "Bright eyes to the front of the bus!"

Maybe they are not really testing the IQ, but instead they are just testing peripheral vision.

IQ is a dubious concept anyway. The definition is as unclear as the definition of physical fitness.

Not believing this is true is a marker of intelligent.

I think Phrenology was preferable.

Initially, I misunderstood the title as referring to pupil (as in student) and went, “Weirdly, that makes sense”.

My initial misread was about classroom sizes, and I came to the same conclusion before I thought "huh, that's... not how you reference classroom size".

pupillometry correlates with bloody everything.

Because children have small eyes.

Not sure whether they used pupil size as a ratio of the eye size or other size normalizations. But the article states that pupil sizes shrink with age and only after controlling for age did the size correlate with task performance.

21st century phrenology.

Intelligence is a precisely defined and precisely measurable notion, so of course we can precisely correlate it with physical measurements [/sarcasm]

Reply on news.ycombinator.com