This week we talked to interruptor Erin Simpson, CEO of Caerus Associates. She’s a leading strategist and specialist in the application of quantitative social science research methods to conflict environments and complex systems. Phew. She previously worked with NATO ISAF’s field based Counterinsurgency Advisory and Assistance Team, providing analysis and assistance to senior military and civilian officials in Afghanistan, on topics related to campaign metrics and strategic assessments, local defense forces, and anti-corruption. A native of Kansas City, she has a PhD from Harvard University in political science, where she studied intelligence and irregular warfare. And she loves Moneyball and shoes.
FPI: What do you do?
I’m the CEO of Caerus Associates – a strategy and design firm that serves clients working in frontier and conflict environments. Some of those clients include the World Bank and the U.S. Defense Department who need support in data analysis, developing research techniques, and monitoring and evaluation of complex environments often afflicted by violent conflict. We bring together tools from urban design, applied social science, and big data analytics build up a data rich and evidence based picture for these complicated places.
FPI: How did you come to do that?
I joined Caerus after returning from Afghanistan in 2009. I wrote my dissertation on third-party counterinsurgency campaigns and taught Marine officers at the staff college at Quantico after leaving Harvard. When a new program for counter insurgency advisors opened up in Afghanistan, I went over for a year. I worked with various military staffs in Kandahar, Helmand, and Kabul.
I started graduate school on September 12, 2001, three miles from Logan Airport. I had always been interested in protest movements, civil wars, and the nature of political violence. I assumed that I would become a professor. Then we invaded two countries and that didn’t go so well. I ended up studying insurgencies and counter insurgencies. That led me to spending a summer at Rand, work on the counter insurgency field manual, and teaching Marines.
That landed me in Washington in 2007. It all makes sense in retrospect, but it didn’t always seem so straight-forward at the time.
FPI: You work on “big data” – what is that and where did it come from?
“Big data” is a buzzword, but there is a marked difference in the amount of (digital) information available today. People talk about the ubiquity of sensors which (for better or worse) allow us to observe and model more events and behaviors than ever before.
Take one example: When I was an undergraduate at the University of Kansas in the late 1990s, I was a research assistant to two incredible professors, Phil Schrodt and Misty Gerner. I worked with them on something called the Kansas Event Data System or KEDS (you can still find my initials in various dictionary files – something my current research team finds amusing).
KEDS was an idea way ahead of its time, the software worked like this: I would pull stories off Lexus Nexus. This system would take the first line of these wire service stories, match the nouns and verbs with a set of dictionaries –and turn that text into discrete political events. Through these events and bits you could conduct analysis, like forecasting international conflict or mass atrocities. At the time we thought it generate a huge amount of data. Now it would be a charmingly small dataset.
What changed? Well there was an explosion in available source material. Twenty years ago, you had to work pretty hard to get digitized news reporting. Now you just need to check Google. And now there are teams of researchers around the country working on open source tools that we can build on. So now there is vastly more available data and more sophisticated tools.
FPI: How has data evolved?
There is a lot of cool data available – whether it’s the Twitter fire hose or the open.gov info, or GDELT and now there are publicly available tools that weren’t previously available. You don’t have to have a masters in statistics or a PhD in political science to work with this stuff. You can be relatively non-technical and still do some really interesting work. But that means you can also get it wrong. There is a buyer beware element.
Whether you are a professor, or an intelligence analyst, or a data journalist you need to be able to validate that data – understand where it’s coming from, what it’s saying, and then how you should interpret it. In the social sciences it’s called modeling the “data generating process,” but it’s really about interrogating your data so you understand its limits – what’s included and what’s not.
FPI: Data is great. But how can we use it better?
One of the things that may seem parochial is not to let data scientists and computer scientists do all your data work. You need people with substantive backgrounds to get involved. For example, it’s important to have social scientists who are trained in certain kinds of political phenomena and public health people who work with messy incidence data engaged. All data are not equal. People want there to be an engineering solution – that in you pop data and out comes the silver bullet. That’s not how any of this works –and it’s actually dangerous to think about it from that perspective.
There are insights you can derive from flurry of informational sets made available from smart phones or parking meters. New York City has done a lot under the Bloomberg administration to generate more data from the city to improve services. That’s a place that makes a lot of sense –but there are privacy concerns involved.
Doing data well is time intensive. You need actual people with actual expertise who can spend time understanding the data. You can’t just automate the entire process. What Nate Silver did with the election analysis – he knew that data really well. He did one thing carefully. But that sort of thing doesn’t always scale. In journalism there are a lot of issues when a writer has 15 different distractions and is on deadline to say something “interesting” or counter-intuitive every day.
FPI: You’re the CEO of an organization – that’s rare for a woman.
There are lots of reasons why you don’t see women in these C-level positions. My experience has been odd in a lot of ways, so I don’t necessarily see it as a model to be emulated. I definitely have a big personality and strong opinions and that serves me well (except when it doesn’t). Women are supposed to have a different leadership style than men. We’re supposed to be inclusive and patient. But not all of us are! Some of us are “abrupt and mercurial.” People react differently from the behavior of a woman than a man. I don’t know what we do about that exactly. The second piece is there is a latent, low-lying sexism that still pervades DC. I show up to a meeting and people have no idea who I am and why I’m at that meeting. Am I there to bring them coffee? I’m used to it and I try to have fun with it. But a 25-year-old man wouldn’t have to do that; they’d sort of assume he was supposed to be there. That’s the difference.
FPI: Parting thoughts? Advice for fellow interruptors?
Don’t over plan your life. Take opportunities as they arrive. Be thoughtful about what you want to do, but not rigid – I never thought I’d go to Afghanistan, but it was an amazing experience. It’s not about having a path and sticking to it at all costs. It’s knowing your own skills and being open to unique opportunities.
And more importantly, be excellent. Learn how to do something very well. I know everyone wants to follow their passion, but to be successful there is an element of excellence. It’s not about being “the best” – there’s no need to be a social climber and kick people out of the way. But excel at what you’re work at. That’s meaningful and gets recognized.