Hacker News Reader

Comment by natch | Hacker News Reader

One possibility not mentioned is that the expert who eyeballed the old photographic plates to determine that there was a slight dimming did not do so under double blind conditions, and his assessment is tainted.

I don't mean to suggest this supports one hypothesis or another. But I'm not convinced his opinion on the dimming is so valuable if it was done without proper controls. There may very well have been proper controls, but unfortunately that didn't come through in the article.

View on HN

The author used both methods, eye measurements, as well as digitized light curves from scanned plates. The paper has a graph of the latter. It's actually easy to replicate, the data is online.

http://arxiv.org/abs/1601.03256

http://dasch.rc.fas.harvard.edu/lightcurve.php

(You can download this star under the name "J20061546+4427248"; the KIC identifier doesn't seem to work).

The by-eye measurement was done by the author himself, so you have a point.

Still, apparently that measurement was much simpler (using only 131 of the plates; the digital measurement used 1232) and only done to be sure that the dimming was not a byproduct of the digital analysis method.

In other words, the real data for the paper is based on digital analysis, not on the by-eye measurement. More importantly, according to the paper, there were 5 control stars also measured digitally. They didn't show the dimming. The paper's graph (http://arxiv.org/pdf/1601.03256v1.pdf) plots two of them.

Yes, came to say the exact same thing. I am not at all familiar with the photographic plate procedure, but it seems likely there is room for all sorts of tinkering with parameters like gain, lighting conditions, white balance, whatever that would allow any bias to shine through.

The aperture, exposure time, etc would all be recorded with the plate, so any bias would be absolutely minimal. Additionally, photographic plates don't have a single star on them, but rather a field of stars, so any overall bias would be applied equally to the entire field and weighted out.

That could be possible - though we'd like to believe other wise the human eye isn't well calibrated. After all maps of the canals of mars came about from eye balling - and they never existed.

Reply on news.ycombinator.com