Galaxy Zoo and human bias

Another blog covering the American Astronomical Society 2008 meeting in Austin is the Galaxy Zoo blog. Jordan is providing some good posts from after hours discussions and his interactions with a recent Nobel Prize winner.

I assume most people looking at this blog are probably aware of Galaxy Zoo but for those that don't it is a huge galaxy sorting project that lets the public join in. In fact over 100,000 of the public joined in. Over the past six months or so a staggering number of classifications have been made of candidate galaxies automatically extracted from the Sloan Digital Sky Survey. The basic aim was to sort galaxies into elliptical galaxies - huge rugby/football shapes of stars - and spiral galaxies - those that look like the water going down your bath plughole. A secondary aim was to further sort the spiral galaxies into clockwise or anti-clockwise (counter-clockwise) spirals.

Back in October The Telegraph mentioned that people seemed to be finding more ant-clockwise galaxies than clockwise ones. That is odd. Very odd. It also seemed to be statistically significant given the number of classifications that were made. Was there something up with the universe? Could there be a preferential direction in the universe that caused one side of the sky to have more anti-clockwise galaxies than the other? Big claims like that require very careful checks for bias and the Galaxy Zoo team got on the case.

Over the past couple of months the Galaxy Zoo website has been showing flipped versions - mirror images - of galaxies to users. The result of that simple test should be to change the observed preference to clockwise galaxies rather than anti-clockwise galaxies if the universe does indeed have a preference. On their blog yesterday, Galaxy Zoo team members Anze and Kate say that when you crank through the stats it would appear that it is a human bias and not a cosmological one. It would appear that the humans that took part in Galaxy Zoo have a slight preference for anti-clockwise spirals over clockwise spirals even when there are equal numbers of both. The blog post does a good job of explaining the results and even has graphs with error bars!

Although Kate is probably a little disappointed - as publishing a paper showing a cosmological bias may have added weight to her 'Axis of Evil' idea - it just shows that you can't always jump to conclusions and you really need to distrust your data until you've checked for bias. Galaxy Zoo appears to have been a great success and I hear mention that Galaxy Zoo 2 is on the way.

If you want to see Chris talking about it, Phil Plait has interviewed him for his blog.

Posted in astro blog by Stuart on Friday 11th Jan 2008 (11:10 GMT) | Permalink
[an error occurred while processing this directive]
[an error occurred while processing this directive]