Evan,

Thanks very much for pointing me to your master's thesis. While I only read your blog summary, I am definitely supportive of more scientific studies to support the debate. Since your work, a couple of other works have come out that attempt a similar scientific study, but neither has been significantly conclusive one way or another. One is this one: https://web.cs.ucdavis.edu/~filkov/papers/lang_github.pdf and the other is this one: https://vimeo.com/74354480, the second one being a talk but it contains some large scale data as well. Both point towards a hypothesis that statically typed languages are (very) slightly better at reducing defects but other factors dominate the overall success of the software projects (such as expertise of people involved in the project, and time to delivery). I am sure you will enjoy going through these as well.