Statistical analysis of terrestrial planets formation and observational bias

Abstract

When investigating and researching into extrasolar planets, astronomers must deal with the inherent selection bias that arises because of limitations in technology or resources. Ideally, astronomers should aim to have a database with unbiased samples from across the universe. The ability for astronomers to efficiently use equipment will rely on statistical models employed to discover regions in which astronomical observations for extrasolar planets are lacking. Previous researchers have focused on the observation of Hot-Jupiters, Super Earths etc. This paper will be specific in exploring the selection bias involved with the detection of extrasolar terrestrials. In addition, the paper will investigate the probability of finding such terrestrial planets depending on the characteristics of exoplanet host stars. Ideally, the distribution of terrestrial planets should be random if the Earth is not considered special. The paper will also evaluate this hypothesis and determine whether there is any relationship between detection probability for exoplanets and their distance from Earth. Statistical methods such as logit regression and kernel estimations will be used in Stata to analyse data from the exoplanet encyclopedia. Hopefully, this paper will be able to provide more insight into the distribution of terrestrial planets and bias involved in detecting such planets.