5 ms·
> Most discovered planets to date are... It's possible to do better than lump all of the discovered planets together and complain that the set is biased. For e
by stromgo 12y ago
> Most discovered planets to date are...
It's possible to do better than lump all of the discovered planets together and complain that the set is biased. For example, Kepler can detect a planet iff there is a chance alignment of its orbit with our line of sight, so for each discovered planet, it's possible to compute the probability that its orbit was correctly aligned, and divide by it to compensate. That way it's possible to make unbiased statements about the distribution of planet sizes and orbits, the fraction of stars with planets, etc.
- grey-area 12y agoWasn't complaining. I'd be interested in a ref for this, most popular science I've seen has skated over the limitations of data collection and made lots of assumptions.
- mturmon 12y agoThis appears to be exactly what they're doing. Here's an exchange on HN a couple of years ago on a similar question: https://news.ycombinator.com/item?id=5668324 https://news.ycombinator.com/item?id=5668324 In that submission, astronomers including Mike Brown of Caltech, Steve Vogt of UCSC, and Sara Seager of MIT all make assertions about typical exoplanet composition. A HN reader asked about the biased sampling issue. You can bet they are not just ignoring the possibility of biased sampling.