search

willettk / gzhubble

Reduction and analysis materials for the Galaxy Zoo: Hubble project.
http://hubble.galaxyzoo.org
MIT License
4 stars 5 forks source link

Referee Report: MINOR COMMENT 9 #185

Closed Mel23 closed 8 years ago

Mel23 commented 8 years ago

9) Why 80% of the data (orange bars, Figure 10)? This is an ad hoc choice that doesn't correspond to any typical statistical practice (except perhaps a 10% trimmed mean, but the authors are not computing means here). And how is this information applied in Figure 11, middle panel? This is not clear.

willettk commented 8 years ago

That's a very reasonable point. What about moving to the interquartile range (IQR) and taking the inner 50% instead?

vrooje commented 8 years ago

Or 68% as that's sorta-kinda-not-really-but-we-still-pretend-it-is 1 sigma?

karenlmasters commented 8 years ago

I agree with either Kyle or Brooke's suggestions. Just make this change to something more typically used - I presume it'll be quick.

willettk commented 8 years ago

The code wouldn't be difficult to change, but it is one of the things that ultimately controls parameters in the catalog. So I want to make sure everything else is set before Melanie runs the pipeline again to make the final versions of the data, since that's a moderate amount of work.

On Thu, Aug 11, 2016 at 7:37 AM karenlmasters notifications@github.com wrote:

I agree with either Kyle or Brooke's suggestions. Just make this change to something more typically used - I presume it'll be quick.

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/willettk/gzhubble/issues/185#issuecomment-239137126, or mute the thread https://github.com/notifications/unsubscribe-auth/ACLha5Tk2GkZ3l-AL4WXAL8di2h7lbZ1ks5qewlsgaJpZM4I3qQn .

willettk commented 8 years ago

@Mel23 - has Figure 8 been changed in the master branch to show the 1-sigma limits? If so, can you adjust the caption and then close this issue?