Author Topic: CW/ACW bias study results  (Read 19385 times)

zookeeper_anze

  • Admin
  • Jr. Member
  • *****
  • Posts: 54
    • View Profile
CW/ACW bias study results
« on: January 10, 2008, 04:51:34 pm »
Kate and I spent the last four days hammering the clockwise / anti-clockwise bias testing together.

You can read about the exciting results here.

Infinity

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 11175
    • View Profile
Re: CW/ACW bias study results
« Reply #1 on: January 10, 2008, 07:01:23 pm »
Thanks for the update and the link, great to see the blog in action :)

RBH

  • Newbie
  • *
  • Posts: 7
    • View Profile
Re: CW/ACW bias study results
« Reply #2 on: January 10, 2008, 09:35:29 pm »
I'll repeat my comment on the blog.  I often click through the Galaxy Ref to magnify the image.  For the reversed test images, the Galaxy Ref image is reversed from that on the main classification page.  Hence one doesn't know in which direction to classify the object.  For example, on the classification page this galaxy is reversed from the magnified view.  So which does one use to classify its direction of rotation?

RBH

Edit to add: I have encountered several more such instances since posting the above question, so I've suspended classifying until I know which image is 'canonical.'

In his interview with Phil Plait on Bad Astronomy, Chris Lintott remarked that he might have to become a psychologist to tell what's going on with respect to the design of the interface and human perception/classification of images.  Consulting someone experienced in human perception/cognition might be a real good idea.  (That my Ph.D. is in cognitive science has nothing whatsoever to do with my suggestion.  :))
« Last Edit: January 10, 2008, 09:47:41 pm by RBH »

ElisabethB

  • Hero Member
  • *****
  • Posts: 13200
    • View Profile
Re: CW/ACW bias study results
« Reply #3 on: January 10, 2008, 09:43:36 pm »
Hi RBH,
The one to classify is the image on the analysis (main classification) page.
Happy hunting !  :)

RBH

  • Newbie
  • *
  • Posts: 7
    • View Profile
Re: CW/ACW bias study results
« Reply #4 on: January 10, 2008, 09:49:01 pm »
Hi RBH,
The one to classify is the image on the analysis (main classification) page.
Happy hunting !  :)
Then a couple of my classifications are screwed up.

I now more strongly suggest that you consult someone experienced in the design of stimuli for human perception studies before shooting a whole lot more time and effort.

RBH

  • Newbie
  • *
  • Posts: 7
    • View Profile
Re: CW/ACW bias study results
« Reply #5 on: January 10, 2008, 10:12:23 pm »
I'll add my last comment on the blog post:

But you’d still be ahead to consult with someone who knows something about human perception and the design of interfaces in classification tasks.

Just as a side note, the inability of single-layer neural nets to accurately classify spiral stimuli was one reason for their near-abandonment in research on modeling human perception through the 1970s into the 1980s, before backpropagation was invented for multi-layer neural nets. Interestingly, humans share that inability. Hence the desirability of consulting someone who knows something about human perception at a professional level.

RBH

zookeeperChris

  • Admin
  • Hero Member
  • *****
  • Posts: 723
    • View Profile
Re: CW/ACW bias study results
« Reply #6 on: January 10, 2008, 11:45:34 pm »
I can't find the thread now, but the effect of reversed images is due to our bias study; we have manipulated some of the images, but can't change the sky server images you see when you click through. We thought about removing the link, but that would spoil a lot of the fun people are having. However, as ever, classifications should be made based on what you see in *galaxy zoo*, not on the sky server (and, yes, this makes 'my galaxies' wrong sometimes too.

To deal with the other point, the design we have works perfectly well for doing astronomy; Kate and Anze's hard work lets us show the bias exists and quantify it. Clearly if we get to the stage where we're investigating the bias we'll do things differently (and let people used to doing this sort of study run it - if you're interested in helping, please get in touch).

Alice

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 31782
    • View Profile
Re: CW/ACW bias study results
« Reply #7 on: January 11, 2008, 12:32:21 am »
I like the idea of having edge-on/unclear between the clockwise and anticlockwise buttons - I think I saw that on the blog.

I don't know how I feel about these results. I didn't find out about them long before everybody else did. After months of believing that the Universe appeared to contain more anticlockwise galaxies, and getting used to the incomprehensibility the way I've got used to quantum mechanics and relativity, is it a disappointment or a relief to know that the Universe is actually not as mad as it looked at least in this way?

What I don't yet know is local rotation. I'd be really interested to see if groups of spirals tend to affect each other by their movement, or whether that, too, is random. I bet we'll find out soon.

Rick Nowell

  • Hero Member
  • *****
  • Posts: 1333
  • Bristol, U.K.
    • View Profile
Re: CW/ACW bias study results
« Reply #8 on: January 11, 2008, 09:48:55 am »
Some points that are being discussed on another Forum i frequent include:

1) Should there be a 50/50 split in ACW/CW anyway: the Cosmic Microwave Background Radiation
isn't smooth but rather clumpy. If everything was completely smooth with no anomalies, there might
not be any galaxies at all.

2) Do galaxies precess like gyroscopes?

3) Is there any correlation between CW/ACW rotation and age?


zookeeper_Kate

  • Full Member
  • ***
  • Posts: 140
    • View Profile
Re: CW/ACW bias study results
« Reply #9 on: January 11, 2008, 10:23:13 am »
Quote
1) Should there be a 50/50 split in ACW/CW anyway: the Cosmic Microwave Background Radiation
isn't smooth but rather clumpy. If everything was completely smooth with no anomalies, there might not be any galaxies at all.
Indeed - because of the initial perturbations (the clumpiness) we are lucky enough to have galaxies! But this has nothing to do with the rotation of the galaxies, which is a 'late time' effect (ie. wasn't set up in the early Universe). The rotation of the galaxies should be pretty random (according to our current theories of structure formation), as we think galaxies get their rotation by being 'torqued', or rather tugged, by the gravitational pull of large clumps of mass (like superclusters of galaxies) near them. This effect sets the galaxies off spinning in a pretty random direction (based on their initial orientation with respect to all the large clumps).

Further, because the AC/C ration depends on your location in space, then we also expect it to 50/50 from simple arguments based on the idea that the statistics of the Universe shouldn't depend on an observers location. I hope this helps explain.

Quote
3) Is there any correlation between CW/ACW rotation and age?
Don't know! I imagine there might be a correlation between how fast a galaxy is spinning, and age. But not the direction. (Alas we don't have the info on how fast a galaxy spins in Galaxy Zoo).
K

« Last Edit: January 11, 2008, 10:26:46 am by zookeeper_Kate »

Rick Nowell

  • Hero Member
  • *****
  • Posts: 1333
  • Bristol, U.K.
    • View Profile
Re: CW/ACW bias study results
« Reply #10 on: January 11, 2008, 10:45:02 am »
Which raises the question from me as to what mass does the tugging? You mention nearby superclusters, but, as is
mentioned elsewhere in the Forum, Spirals tend to be more 'on their own'. Does this imply that a mass such as the
unknown 'dark matter' may be setting the Spirals in motion? I guess this is a 'field of continuing research'.
« Last Edit: January 11, 2008, 10:50:51 am by zookeeper_Kate »

zookeeper_Kate

  • Full Member
  • ***
  • Posts: 140
    • View Profile
Re: CW/ACW bias study results
« Reply #11 on: January 11, 2008, 10:50:22 am »
Which raises the question from me as to what mass does the tugging? You mention nearby superclusters, but, as is
mentioned elsewhere in the Forum, Spirals tend to be more 'on their own'. Does this imply that a mass such as the
unknown 'dark matter' may be setting the Spirals in motion? I guess this is a 'field of continuing research'.

Most of the matter in the Universe is dark - so indeed the tugging is always mainly done by dark matter. Even in a supercluster, most of the mass is in dark matter. One of the ways that you can detect the dark matter is by the way it is pulling galaxies around, making them move.


Rick Nowell

  • Hero Member
  • *****
  • Posts: 1333
  • Bristol, U.K.
    • View Profile
Re: CW/ACW bias study results
« Reply #12 on: January 11, 2008, 02:27:05 pm »
Most of the matter in the Universe is dark - so indeed the tugging is always mainly done by dark matter. Even in a supercluster, most of the mass is in dark matter. One of the ways that you can detect the dark matter is by the way it is pulling galaxies around, making them move.

Although this thread isn't about dark matter, perhaps you can bear with me....Does this mean that, in
simple terms, there is more dark matter around spirals than ellipticals? There is more tug or torque which
makes the spirals rotate, whereas there is less around ellipticals. This introduces an age question as well?


zookeeper_Kate

  • Full Member
  • ***
  • Posts: 140
    • View Profile
Re: CW/ACW bias study results
« Reply #13 on: January 11, 2008, 04:49:38 pm »
Although this thread isn't about dark matter, perhaps you can bear with me....Does this mean that, in
simple terms, there is more dark matter around spirals than ellipticals? There is more tug or torque which
makes the spirals rotate, whereas there is less around ellipticals. This introduces an age question as well?
Be careful - Ellipticals rotate too! You just can't see it through only its structure (you'd actually have to look at the velocity of the stars with spectra). Further, we think the arms of a spiral are not because IT is rotating - but more because different bits of it are rotating at different speeds... but the whole community is unsure of exactly how and why ellipticals and spirals form (indeed that is a key goal of GZ!).

I also think that a galaxy is going to rotate more (from this torquing) the more asymmetrical its initial shape is. For example - if you dangle a ball on a string, then gravity just pulls it down and doesn't cause it to rotate. But if you have a bar that is heavier at one end, and you initially hang it out of equilibrium (ie. heavy end up) then it will start to rotate (like a pendulum). So a spiral galaxy is like the bar - with gravity (from all matter) pulling it round (if the spiral is pitched at an angle relative to the gravity force). While ellipticals are a bit more like the ball - less reason to spin.

BUT I have to warn you - I am kind of thinking aloud here - and I have not had any of the above confirmed by another cosmologist!! Further - the spiral arm formation stuff should also be taken with a pinch of salt... I'm getting most of that off wikipedia!
« Last Edit: January 11, 2008, 04:51:20 pm by zookeeper_Kate »

pickanotherid

  • Newbie
  • *
  • Posts: 26
  • ...Major Tom?
    • View Profile
Re: CW/ACW bias study results
« Reply #14 on: January 11, 2008, 04:55:54 pm »
And how does NGC4622 fit into this whole CW/ACW discussion?? ;D

Story here:

http://www.msnbc.msn.com/id/22594171/

Additonal info and pictures at:

http://hubblesite.org/newscenter/archive/releases/2002/03/image/a/



The only dumb question is the one you don't ask.