Now in Print: “The Impact of Leadership Removal on Mexican Drug Trafficking Organizations”

My Journal of Quantitative Criminology article “The Impact of Leadership Removal on Mexican Drug Trafficking Organizations” is now in print. For the abstract and other discussions of the research see here, as well as the posts tagged “Mexico,” “drug trafficking,” and “leadership removal“.

Here is a timeline of the research and publication process:

  • Read an article in the Economist about DTO leadership removal, December 2010
  • Preliminary research for a graduate seminar in time series analysis at the University of Houston, Spring 2011
  • Draft of paper incorporating other research on organized crime and political violence for a seminar at Duke University, Fall 2011
  • Revised manuscript rejected from a security studies journal after R&R, Spring 2012
  • Revised manuscript rejected from a political violence journal after R&R, Late summer 2012
  • R&R from JQC, Summer 2013
  • Accepted for publication in JQC, December 2013
  • Published online, March 2014
  • Published in print, December 2014

All in all, a four-year project, with no significant changes to the manuscript in about 18 months previous to the print publication. The paper absolutely improved thanks to feedback from reviewers and quality, but I think you will agree that this is a very long feedback cycle.

Academia to Industry

Last week, Brian Keegan had a great post on moving from doctoral studies to industrial data science. If you have not yet read it, go read the whole thing. In this post I will share a couple of my favorite parts of the post, as well as one area where I strongly disagreed with Brian.

The first key point of the post is to obtain relevant, marketable skills while you are in grad school. There’s just no excuse not to, regardless of your field of study–taking classes and working with scholars in other departments is almost always allowed and frequently encouraged. As Brian puts it:

[I]f you spend 4+ years in graduate school without ever taking classes that demand general programming and/or data analysis skills, I unapologetically believe that your very real illiteracy has held you back from your potential as a scholar and citizen.

Another great nugget in the post is in the context of recruiters, but it is also very descriptive of a prevailing attitude in academia:

This [realizing recruiters’ self-interested motivations] is often hard for academics who have come up through a system that demands deference to others’ agendas under the assumption they have your interests at heart as future advocates.

The final point from the post that I want to discuss may be very attractive and comforting to graduate students doing industry interviews for the first time:

After 4+ years in a PhD program, you’ve earned the privilege to be treated better than the humiliation exercises 20-year old computer science majors are subjected to for software engineering internships.

My response to this is, “no, you haven’t.” This is for exactly the reasons mentioned above–that many graduate students can go through an entire curriculum without being able to code up FizzBuzz. A coding interview is standard for junior and midlevel engineers, even if they have a PhD. Frankly, there are a lot of people trying to pass themselves off as data scientists who can’t code their way out of a paper bag, and a coding interview is a necessary screen. Think of it as a relatively low threshold that greatly enhances the signal-to-noise ratio for the interviewer. If you’re uncomfortable coding in front of another person, spend a few hours pairing with a friend and getting their feedback on your code. Interviewers know that coding on a whiteboard or in a Google Doc is not the most natural environment, and should be able to calibrate for this.

With this one caveat, I heartily recommend the remainder of the original post. This is an interesting topic, and you can expect to hear more about it here in the future.