Tuesday, November 29, 2016

Men and Women See Things Differently

This will come as a shock to the social construct crowd, but men and women do not see things the same way. In the most literal sense, men and women look at faces differently. It must have something to do with different wiring for different kinds of brains.

Luckily the social construct crowd does not believe in science anyway, so, don’t feel too badly for them.

In any event, I report the research from PsyPost: 

Women and men look at faces and absorb visual information in different ways, which suggests there is a gender difference in understanding visual cues, according to a team of scientists that included psychologists from Queen Mary University of London (QMUL).

The researchers used an eye tracking device on almost 500 participants at the Science Museum over a five-week period to monitor and judge how much eye contact they felt comfortable with while looking at a face on a computer screen.

They found that women looked more at the left-hand side of faces and had a strong left eye bias, but that they also explored the face much more than men. The team observed that it was possible to tell the gender of the participant based on the scanning pattern of how they looked at the face with nearly 80 per cent accuracy. Given the very large sample size the researchers suggest this is not due to chance.

[Addendum: Now, the Daily Mail has reported on the same study. It adds the following Darwinian explanation for the difference between male and female facial recognition:

Many experts believe that the way women see faces goes back to early humans, as females spent their lives as gatherers – they had to recognize close at hand, static objects such as wild berries.

On the other hand, men had the job of hunting prey and keeping predators away, which did not require a detailed view.]


Trigger Warning said...

To begin with, this research was not reported in any journal. It was a paper presented at an annual meeting, and the only description of the data analysis was that they used unspecified "data mining" techniques. Piffle.

Null hypothesis: women = men
Alternative hyp.: women <> men

Reject null at better than chance. Now what? It's possible that women simply followed instructions better.

In fact, I'll bet that if 500 people looked at the wrappings of wrapped Christmas presents, the women's eye tracking data would yield different results than men's. So what?

This is not science. It's just poking around for reportable results. Science is using geological theory and data to find ore deposits. Poking around looking for interesting things is a stroll on the beach with a metal detector.

And it's a perfect example of too much taxpayer $$$ being wasted on bulls**t.

Trigger Warning said...

OT, slightly...

I'm rooting for Nancy Pelosi to retain her leadership. She's done wonders for American politics. Here's a March 2010 picture of her celebrating the upcoming handover of America to the Republican Party:


Stuart Schneiderman said...

The research was reported in the Journal of Vision.

Trigger Warning said...

The Journal of Vision is an open access journal. I searched their site, and found the abstract listed as a conference paper. I also searched via Google. If you can uncover a link to a paper, I'd like to read the whole thing.

And my criticism was not aimed at you or your reporting, but at the general problem of data mining (the metal detector approach) as opposed to doing science.

I presume we are in complete agreement about M/F difference, right down to the neural and cognitive architecture.

Trigger Warning said...

Well, insofar as a difference exists. We may disagree on the details. Dunno.

Trigger Warning said...


Ares Olympus said...

Regarding "The team observed that it was possible to tell the gender of the participant based on the scanning pattern of how they looked at the face with nearly 80 per cent accuracy."

An 80% accuracy on an objective test is pretty impressive, but 80% was also the chance that Hillary was supposed to win the election, so we still need to consider the exceptions.

It would also be interesting to know if mis-identification happen more for men or women. And do the same set of 20% confound the test results every time, or perhaps its 80% accurate on every individual equally, if you observed sequential tests.

And if we found a certain 20% subset consistently were mis-identified, that would be interesting and we could look at what other traits were in common. Like perhaps even things like birth order or sexual orientation or introversion/extroversion, or whatever.

I'm also thinkin I'm not great with names, I sometimes do give extra attention to a person's face whom I want to recognize again. On the other hand, women are also sneaky, and sometimes they'll confuse you and change their hair color, and in general change their hair styles more frequently, and also can make identification more difficult.

Like in crime shows how they'll ask the victim what the suspect looked like, and I admit it seems hard to imagine I'd have sufficient memory to describe, or find the words to describe, or trust my memory of what I think I remember. So unless there's something abnormal, like a mole or something, it seems pretty hard. And if you put a lineup of suspects, I'm sure I'd prefer not to say which one, unless I had a very specific detail that was unique to them. I'd not want to falsely accuse someone.

It's strange how some people stand out as more unique, while others don't. And like at church recently, I feel bad when someone recognizes me, and I can't be sure I recognize them. OTOH, one guy was sure he met me before, but I was not around this summer when he claimed he met me, so I decided it didn't matter to contradict him beyond my open confusion.