Like you, I'm first inclined to see her swinging clockwise. But if I concentrate on the rotating foot and tell myself that it's just a 2-d image lacking any depth, some of the time the figure switches directions literally before my eyes.

And yes, this is the weirdest cognitive test I have ever taken.

Counter-clockwise !!! Can't be NO other way (the way I see it :-).
I tried to "fool" my brain so that it looks like clockwise rotation, but I couldn't.

I can only see it counter-clockwise.

I can see both. I don't know what that means.

I once saw it counter-clockwise. Once.

The rest of the time, clockwise. And I can't get counter-clockwise to come back.

I think that it changes according to the computer software underwriting it. I saw it both ways, and then it switched before my eyes. I couldn't see it the other way when it switched either

When I refresh my browser she reverses. I suspect this is more a software test than a cognitive one. When I first went to the site, she started by doing a little stutter step.

It means nothing!! Does that brain analysis really seem scientific to you?? All that matters is which side the foot starts on when you begin looking. Try covering up the whole body with your hand and just look at the foot swinging to one side. Remove your hand. Okay which way is she swinging? Now try moving your hand so you can only see the other foot. Then take your hand away. It changes doesn't it? This is just another M.C. Escher style optical illusion... this is not a "cognitive test!"

By "software test", do people mean that it is somehow a trick of the software? A thought experiment should convince people that the trajectory of any point when seen from the side is the same regardless of the direction of rotation.

Suppose you are facing a person, and they are sticking their right arm out at a 45 degree angle from their body. Seen from the side, in silhouette, there is no way that you can tell if they are facing towards you or facing away from you. (Assume a parallel projection, so there is no foreshortening.)

well, obviously, the real test would be to have two people watch at the same time and see whether they ever see the image spinning in different directions. if so, then it is a very weird thing, if not, it is just a gimmick.

Maybe it's wierd in some sense, but mathematically, it's straightforward. The motion of a point rotating CCW on a circle is x=cos(t), y=sin(t). The motion of a point CW is x=cos(-t), y=sin(-t). Project along the x dimension, and recall that cos(t) equals cos(-t).

j's onto something (maybe?). I see her moving clockwise most of the time. But I process her changing directions and moving counterclockwise by concentrating on her feet. Is it a foot fetish?

Daryl McCullough, a commenter on Kevin's site, gets it right, but it's not nearly as much fun:

"The clockwise/counterclockwise perception is completely determined by whether you see the dancer as standing on her left foot or right foot. Can you see that the same picture can be interpreted both ways? See my modified pictures (h t t p : / / dee-mccullough.com/dancer1.bmp and h t t p : / / dee-mccullough.com/dancer2.bmp) to see how the same picture can be interpreted both ways. (I'm putting extra spaces into the urls so as to fool Kevin's stupid automated moderation program, which holds up any post with urls in it).

If you see her as standing on her left foot, she will appear to rotate clockwise. If you see her as standing on her right foot, she will appear to rotate counter-clockwise."

The left-foot/right-foot thing is another way of thinking about it, but it is a consequence of the math: namely, the projection of any point to a plane does not change when the point is reflected about any other parallel plane.

In other words, reflecting the dancer about a plane parallel to the image plane changes her handedness, but does not change the projected silhouette.

The real test is in determining how much time people put in to developing complex BS theories about why it looks one way or the other and then debunking each others' complex BS theories.

I saw the dancer go both ways simultaneously so I'm either addle-brained or a flippin' genius.

I first saw clockwise, then it reversed -- and I found I could "control" the direction by shifting my focus (yes, between feet). Seems too simple/random to reveal anything much about cognitive style -- I wonder if correlation with other tests has been fully investigated.

Its a software trick. Concentrate on her spinning, and concentrate on the foot on which she spins (it is absolutely unambiguous which foot she spins on at any given time, and this is easiest to determine when she is facing directly away from you-when her butt is square at the computer screen-she is then absolutely unambiguously spinning on either her left or right leg). Periodically, the sofware will switch the leg she is spinning on (and the direction she is spinning). Again, this is unambiguous if you concentrate on her pivot foot, and pay particular attention when she is facing away.

I suspect the optical illusion comes when she is turned 1/2-when her pivot leg and extended leg appear to cross each other (when she is facing either over your left or right shoulders, roughly)-it appears that the direction in which she spins is ambiguous at that point, and thus the direction she is spinning is somehow defined by your mind. but its not.

Sk

More specifically, when she is standing on her right foot, she spins counterclockwise. When she stands on her left foot, she spins clockwise.

Sk

At first I saw it rotating counter-clockwise. After awhile I saw it
rotating clockwise. By focusing on the feet I can switch back and forth
between the two perceptions.

I doubt very much whether this has anything to do with right versus
left hemisphere dominance. You're left visual field is dominated by
the right brain, the right visual field by the left brain. Trying to use
my peripheral vision I can see it rotating either direction using
either the right or left visual field but it's hard to test this
because if I focus at all I'm using both.

I highly doubt the attribution of right or left dominance given on
the website. If there was really evidence for such an implausible
relationship I'm sure they'd cite it.

Wow. Kevin Drum's hint in his UPDATE solved it for me:

Hooray! I finally got her to change direction. For me, the trick is to cover up everything but her legs and stare for a few seconds. Then the direction switches. Do it again, and I can get it to switch back.

I covered everything but her legs and then concentrated on whether the moving leg was passing in *front* of or in *back* of her pivot leg as it passed by.

Telling myself "Now it's passing in front of her pivot" got her spinning clockwise. Telling myself "Now it's passing behind her" got her spinning counter-clockwise.

Very weird.

I can shift at will. The reason is there is no difference between looking at her front or back and one has to make an assumption that she is facing you or has her back turned towards you. Changing that assumption changes the direction of rotation.

I also doubt this has anything to do with right- or left-brain thinking.

The axis of rotation is not aligned with the figure's center of gravity. When I see her rotating clockwise, she seems to be leaning to my left. When I perceive her rotating counter-clockwise, she seems to be leaning into the background. That's the only qualitative difference I can see between the two modes. Are right-brain thinkers more lefty, compared to left-brain thinkers who are more backward? :)

Look again. When she's rotating counterclockwise, she is standing on her right foot. When she is rotating clockwise, she is standing on her left foot. the image is different-she actually is rotating on one or the other foot, and rotating in the clockwise or counterclockwise direction. Its not your brain or your eyes.

Sk

It takes lots of concentration to perceive her movements as counterclockwise.

Both looking at the same computer, my husband saw her going counter-clockwise and I saw her going clockwise. Later I managed to make her switch, but only by looking out of the corner of my right eye (that is, indirectly) - as soon as I look directly, she's going clockwise. At the same time, I doubt that this tests left-brain/right-brain dominance, not least because I have none of the classic "right-brainer" abilities or proclivities. So here's another theory - perhaps it tests which eye is dominant, rather than which side of the brain. Left-eye dominants will see her going clockwise, and vice-versa.

I only see Che Guevara (notice link above)

Poor sk. You broke your brain! Try this: Freeze the animation (right click should give you this option). Let's say you have been thinking that the pivot foot is her right foot, and that right now the dancer is facing you. Is there actually anything in this single image to tell you if it is not in fact her left foot, and that right now the dancer is facing away from you? No? No. There isn't.

I always see her pivoting clockwise on her left foot with her right foot out. I fit the left brain list much better than the right brain list, though.

I saw clockwise but got her to switch pretty quickly.