The Hardest Bias in Astronomy


A nasty sort of bias called Malmquist bias affects almost every astronomical survey, and the only solution is to…keep doing surveys.

Back in the 1970’s and 80’s astronomers first began to identify a structure known as the Great Attractor, which is defined by the common motion of all the galaxies in our nearby vicinity of the cosmos.

But for over a decade, however, many astronomers weren’t exactly convinced that the Great Attractor existed. Their skepticism was well justified because of a common observational effect in astronomy known as Malmquist bias. Named after the Swedish astronomer Gunnar Malmquist, who first elucidated a discussion of this effect in 1922, this bias is a specialized version of a much more common statistical effect known as selection bias.

Most astronomical surveys are limited in brightness. There is a certain floor representing the dimmest possible object that a given telescope with a given exposure will be able to see. But objects in the universe can be dim for two separate reasons: because they are well and truly dim intrinsically, or simply because they are far away. So a typical survey of astronomical objects, like galaxies, will preferentially select for closer and/or brighter ones. In the case of galaxies, the farther out we look from the Milky Way, there more likely we’ll only catch the brightest galaxies at that distance, and miss all of their dimmer siblings.

This bias could potentially distort our understanding of the wider universe, especially if we’re trying to use the velocities of galaxies to map out their bulk motion. In those first surveys in the 1970’s and 80’s, many astronomers argued that we were only seeing the movements of the brightest galaxies, giving the illusion of a general flow towards the Great Attractor, and a more complete census of the local universe would average everything out.

The solution? Even more surveys, with more depth and more completeness, which eventually revealed the reality of the Great Attractor.



Source link