Author Topic: CW/ACW bias study results  (Read 20263 times)

zookeeper_Kate

  • Full Member
  • ***
  • Posts: 140
    • View Profile
Re: CW/ACW bias study results
« Reply #15 on: January 11, 2008, 05:24:43 pm »
And how does NGC4622 fit into this whole CW/ACW discussion?? ;D
Story here:
http://www.msnbc.msn.com/id/22594171/
Indeed!! I should have added the disclaimer that the spiral arms indicate the rotation only about 95% of the time!!!

Rick Nowell

  • Hero Member
  • *****
  • Posts: 1333
  • Bristol, U.K.
    • View Profile
Re: CW/ACW bias study results
« Reply #16 on: January 11, 2008, 05:29:17 pm »
.....with gravity (from all matter) pulling it round...

Does this mean just 'ordinary' matter, baryonic i think is the term, or does it include 'dark' matter?
Are you suggesting that dark matter has a gravitational effect? It has gravity, the same as other
masses?


Alice

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 31782
    • View Profile
Re: CW/ACW bias study results
« Reply #17 on: January 11, 2008, 05:54:53 pm »
If I understand right, gravity - or rather, gravitational effects over huge areas, such as galaxies or clusters of galaxies - seem to be the only sign we can detect of dark matter: that Newtonian physics would mean a galaxy should fly apart judging by the amount of baryonic matter we can see, for example. It doesn't seem to show itself through heat or light, or anything else I can think of . . .

zookeeperChris

  • Admin
  • Hero Member
  • *****
  • Posts: 723
    • View Profile
Re: CW/ACW bias study results
« Reply #18 on: January 11, 2008, 08:36:55 pm »
Yep, we infer the existence of dark matter because we see the effects of its gravity, but don't think it interacts via any other force. As there's lots more dark matter than normal (baryonic) matter, you can even get away with just using dark matter in many models of structure formation. That's good, because then all you have to worry about is gravity, and not all this messy physics and chemistry. Of course, if you want to compare your simulations with observations, then you'll need to convert from dark matter to visible light somehow, and that opens a whole new can of worms...

ballettlover

  • Newbie
  • *
  • Posts: 23
    • View Profile
Re: CW/ACW bias study results
« Reply #19 on: January 13, 2008, 08:41:49 am »
I've copied this post from another topic, as this looks the most appropriate place for it:

Chris, thanks a bunch for your various explanations.

To pick up another point from the first of my two posts in this topic:

- the original study had coherent images ie all the relevant images were CW or ACW   
- the bias study had transformed images, but the original images were still available from the SDSS website
- we know from the forum that in the bias study: a) some people ignored the team's advice and looked at the SDSS images b) others did not realise that the bias experiemnt was taking place and continued to reference the SDSS images

Thus, whereas in the main experiment, people were looking at a coherent set of images, in the bias study, some at least were looking at both CW and ACW images before making a classification of CW or ACW.  Given the above, my gut feel is that if an anomolous result emerges, it is more likely, or at least a possibility, that the anomoly lies with the bias study methodology, rather than the main survey.

A thought: how about a second pilot bias survey (would 1 million be sufficient as a good pointer?) with transformed images but with the SDSS access switched off, so that the incoherent images problem doesn't arise?  If the bias still remains then the original bias study is confirmed.  If a different result arises, then maybe a new full bias study on the new basis could be undertaken.

Floooder

  • Newbie
  • *
  • Posts: 47
    • View Profile
Re: CW/ACW bias study results
« Reply #20 on: January 13, 2008, 08:52:00 am »
I think further biais-study must be made by psychologist, not by astronomers. It's really more complex than at first sight ;D

ballettlover

  • Newbie
  • *
  • Posts: 23
    • View Profile
Re: CW/ACW bias study results
« Reply #21 on: January 13, 2008, 09:45:14 am »
Your comment makes sense to me, Flooder.  GalaxyZoo is not just an experiment about galaxies, but about the process of observation.  And although all experiments have both these characteristics, in this one the people element is a stronger component than for most.  The GalaxyZoo team have initiated a fascinating and innovative research approach in the field of professional astronomy, conducted on an open and inclusive basis.  But, might I suggest that input from a specialist in a field such as Human Factors Psychology or a related field, could prove worthwhile at this stage:

http://en.wikipedia.org/wiki/Human_factors 
« Last Edit: January 13, 2008, 12:58:06 pm by ballettlover »

Bodhi

  • Newbie
  • *
  • Posts: 4
    • View Profile
Re: CW/ACW bias study results
« Reply #22 on: January 13, 2008, 02:53:32 pm »
I agree Battlelover. I am a PhD student in Cognitive Neuroscience, researching selective attention, so this is slightly outside of my field. However, I do believe the bias can be explained by a potential human bias. I have explained roughly (and I do mean ROUGHLY!!) my general argument in this post:

http://www.galaxyzooforum.org/index.php?topic=7956.msg98488#msg98488

There is surprisingly little research on clockwise/anti-clockwise bias, so this could be interesting...

Cheers!

Jim.

Alice

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 31782
    • View Profile
Re: CW/ACW bias study results
« Reply #23 on: January 13, 2008, 05:37:28 pm »
I think it's worth pointing out that for some galaxies, the choice is not clear. Take Anze's Object of the Day on Friday 11th. It's causing so much debate that I've started a poll to see what people think after some conscious arguing, reasoning and interaction with other people. That topic is turning into a good laugh too. Have a look and enjoy (and don't take the fierce arguing seriously)!  :D

ballettlover

  • Newbie
  • *
  • Posts: 23
    • View Profile
Re: CW/ACW bias study results
« Reply #24 on: January 15, 2008, 03:50:09 pm »
A recent example of the confusion that has arisen with the GalaxyZoo and SDSS images in the bias study:

http://www.galaxyzooforum.org/index.php?topic=9619.0

Alice

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 31782
    • View Profile
Re: CW/ACW bias study results
« Reply #25 on: January 15, 2008, 04:12:38 pm »
Thanks for pointing that out  :) but don't worry, I read every post on this forum and make certain that all queries like that get answered. The information saying "this is normal, classify what you see in Galaxy Analysis NOT SDSS etc" has been e-mailed to everybody and is available both on this board and on www.galaxyzoo.org, but we realise people don't always read it. We thought about putting it in Galaxy Analysis too, but were worried that it might change people's classifying behaviour and bias our bias study, as it were.

quarkspin

  • Hero Member
  • *****
  • Posts: 4206
    • View Profile
Re: CW/ACW bias study results
« Reply #26 on: January 16, 2008, 12:52:45 am »
Thanks for pointing that out  :) but don't worry, I read every post on this forum and make certain that all queries like that get answered. The information saying "this is normal, classify what you see in Galaxy Analysis NOT SDSS etc" has been e-mailed to everybody and is available both on this board and on www.galaxyzoo.org, but we realise people don't always read it. We thought about putting it in Galaxy Analysis too, but were worried that it might change people's classifying behaviour and bias our bias study, as it were.
To be honest, do you think the Gz crew bit off a bit more than they can chew?
Never mind we all think you're wonderful.
Answer on a postcard..
As ever

Q

ballettlover

  • Newbie
  • *
  • Posts: 23
    • View Profile
Re: CW/ACW bias study results
« Reply #27 on: January 16, 2008, 04:38:49 pm »
Thanks Alice, but my point was that this example is another indication that in the bias study an unquantified number of people were looking at both CW and ACW images and then perhaps making a classification based on data that was not coherent.  Thus, I believe, and it would seem others here share this view, there is a case to be made that the bias study appears to have been set up on basis that brings its results into question. 

zookeeper_anze

  • Admin
  • Jr. Member
  • *****
  • Posts: 54
    • View Profile
Re: CW/ACW bias study results
« Reply #28 on: January 16, 2008, 05:35:59 pm »
Kate and I got worried about the people classifying the non-flipped images from the SDSS Object explorer. However, we see the votes flipping  for really bright and obvious images: we can use these to put the upper limit on the number of people classifying the images from SDSS Object explorer.

One can cook up really contrived scenarios in which only some people use the SDSS Object explorer for classifying less obvious galaxies, but with some thinking you can convince yourself that you need a lot of fine-tuning to reproduce our results.

You have to realize that the blog is not the whole story, we could have written pages about all the cross-checks, test-runs and other things that we did with the data, but it would boring and too long. Hence we present the somewhat condensed results. The full details will be in the papers and the data will eventually be public so that anybody can convince themselves using their own code.

Half65

  • Hero Member
  • *****
  • Posts: 10288
  • "Eagle Nebula"
    • View Profile
Re: CW/ACW bias study results
« Reply #29 on: January 16, 2008, 07:09:31 pm »
I have lesser time in this period and I can read the result only now.
I wont to thank for the news and the fact that you share the information whit us.
I'm not an astronomer and my opinion could be taken just as supporter and galaxyzoo addicted but for me is brilliant.
Go galaxyzoo.