Oh, and by the way, the ‘selection effect’ isn’t just an astronomical concept. It applies to all the sciences, and even to everyday life. Whatever that means.

Astronomers get two really important tools that I’m sure other fields don’t really stress. First, the sheer size of the universe, and how totally insignificant we are to it. Second, what you see from a great distance is only a tiny piece of the puzzle, and it may not even be the piece you need.

Just looking at the raw numbers though, almost 6000 extra-solar planets discovered (with thousands more awaiting confirmation), that large of a sample size alone would seem to statistically confirm the hypothesis of super-Earths and mini-Neptunes being the most common planet types in the galaxy.

But beyond that, the Kepler survey data shows that around 30% of Sun-like stars have super-Earths, but the rate is much higher around M stars (Red dwarfs) – which are the most common stars in the Universe.

This is what astronomers call the “selection effect”; that is, that the nature of some objects makes them easier or harder to detect, and that surveys of these objects tend to mislead us as to their actual distribution and the nature of their populations.

We have to be very careful in our planetary surveys for pretty much the same reason. All the methods we use to detect extrasolar planets work much better on big planets circling very close to small stars. Our techniques also favor planets with highly elliptical orbits that allow them to pass very near their primary star but which spend most of their time at great distances from their primary.

Astronomers are well aware of these pitfalls and have developed data analysis techniques and theoretical assumptions to help correct for these effects, but they also have learned to be skeptical about being too reliant on these maths and theories. Astronomy students are systematically indoctrinated about these issues, Unfortunately, most sicience journalists have not. As in most astronomical research, every improvement in observational tech usually leads to vast revisions of the accepted theories.

Gathering comparative data on many planetary systems is essential to understand the processes that form planets, which may vary substantially between different types of stars. But the data available to our current observational techniques is highly biased due to this selection effect. To this day, we still don’t know if our own solar system is typical, or a special case.

For example, our own Sun is much brighter and massive that the vast majority of other stars, but both mass and luminosity fall pretty much in the mid range
compared to other stars. And yet Sol is usually described as a “fairly typical star”.

Consider the difference between the “mean” and the “median”.