Bogdan starts by explaining why translators are so important to Gnome.
With the current study the authors wanted to evaluate 1) where people are located that translate Gnome and 2) whether the hypothesis holds that translations by people still living in their native country translate better, since this is known to be true for the translation of literature or poetry.
Location of translators
Bogdan mined mailing lists and looked at headers that contain headers with IP addresses. This way they could see where translators currently live, but also where they lived in the past. Bogdan admits this system is not too reliable, a lot of manual labor had to be done (f.i look at LinkedIn or Twitter) to improve the quality.
It turns out that there is a statistically significant quality difference. In fluency, word choice and impact of mistakes on comprehension, translators in their own country perform better.
This observation leads to some advice for the Gnome community, for instance, review translations of translators that do not live in their own country, or advice translators living abroad to read guidelines. For the current study, they only looked at in-country reviewers, there could also be differences in reviewing quality.