Meet Dhrumil Mehta: Political Journalist and Professor

Philosophy major, software engineer, and now political journalist, Dhrumil shares his story.

Seventeen-year-old Dhrumil Mehta had his life planned out. He would triple-major in political science, economics, and international relations. He had every single class planned out. Or so he thought.

By his second semester of college, he had thrown that all out the window to become a philosophy major. Exposed to a mix of eastern and western philosophical concepts through debate in high school and Pathshala at Derasar, he was naturally drawn to it. Yet his exploratory journey only continued as he was exposed to linguistics and eventually natural language processing through his research at Northwestern University. There, he analyzed the Congressional Record, a corpus of every word ever spoken on the congressional floor. This work ultimately motivated him to pursue a Master’s in Computer Science and also led him to become involved with the Knight Lab for media innovation, which marked his introduction to journalism.

Over the course of his academic journey, Dhrumil discovered a rather niche field for himself. After graduating with his Master’s degree, he worked at Amazon as a Software Development Engineer for just under a year. However, he found himself searching for something closer to his interests and eventually landed himself a job as a Database Journalist at FiveThirtyEight, where he covered a range of political topics.

Dhrumil was recently appointed as an Associate Professor of Journalism at Columbia University in New York and Assistant Director of the Tow Center, the university’s research center for digital journalism. We spoke with Dhrumil prior to this news about his work as a data journalist.

Tell us a bit about what you do at FiveThirtyEight.

At FiveThirtyEight, we’re a data-driven news outlet. Everything we produce is somehow based on numbers, statistics, and data to help tell the story of what’s going on in America.

For me, half of my work is coding and the other half is writing. First, I build and maintain the polling database, which is a collection of public opinion polls. For those not familiar, these polls are a tool that social scientists use to measure what people in a population are thinking. For example, pollsters might take a sample of 1,000 people, call them on the phone, and ask them about their opinions, which are broken down by gender, race, political affiliation, etc. They will make sure that it’s representative of the country as a whole and then publish a result that says, you know, 50% of Americans like mint chocolate chip ice cream, or, x% plan to vote for Trump in the election.

That for me is really exciting. There’s never a dull day. While the job description stays the same, the underlying news keeps changing. There’s always something interesting to analyze or someone interesting to talk to.

And I have some algorithms that I’ve written that I set up and watch. There’s a recent one that I wrote that scans CNN, MSNBC, and Fox News for every word that’s spoken on all of these networks. I take note of when they are all talking about the same thing versus when one of them is peeling off and talking about something different. I might say, hey, what’s going on? And it might be that one of the networks has decided to pick up a story that the other two haven’t. I might think about writing a story there.

And then lastly, I think there’s a lot of traditional journalism work. I want to be talking to a lot of people on the ground. Recently, I wrote a piece about South Asian American politicians in Texas. That piece was inspired by listening to a Pakistani Radio in Houston with my parents. They were talking about this candidate that I had not read much about in mainstream media and it ended up being like a really cool big data analysis piece.

Do any particular topics interest you more than others?

When it comes to politics, the public opinion beat is really interesting. Public opinion polls are a mechanism by which you can measure what Americans are thinking, usually within a single-digit margin of error. And then you can observe that change over time. So since the start of Obama’s administration, acceptance of same-sex marriage has gone up quite a bit. It crossed that 50% threshold at some point, and after that, it got legalized in a Supreme Court case. To be able to kind of observe what share of Americans think something with such granularity is really cool. And it’s such a vital part of our democracy, because politicians are paying attention to that stuff, to what we write — it helps inform the laws that they decide to pass or even pursue.

Have you ever found data to be limiting?

We pay attention to the biases in the data. We follow the pollsters through our database, and over time we can look at their election results and note that a given pollster tends to get the result wrong in one direction or another, or they tend to be more conservative or more liberal. So that’s one way of sort of measuring pollster bias.

But, there can be bias even in the way the questions are presented. I read the raw form of the data and try to understand what exactly each question was, how it was presented, and in which order it was asked. That might matter because if the pollsters ask you seven questions on your opinion of plastic killing sea turtles and then ask you a question about the economy, you might be more likely to say, well, I’m more concerned about climate change than the economy right now.

Another issue is depth. A pollster could ask what share of Americans are worried more about climate change than the economy. But I wouldn’t necessarily know why in the same way I might be able to if I was on the ground and talking to people. There’s just a certain depth you get when you talk to a lot of people that you’ll never get through just looking at the numbers. So it’s really important to reach out to the affected people and use their responses to help tell that story. I am trying to do more of that, actually.

So… what happened in 2016?

The 2016 U.S. election undermined a lot of people’s faith in polling, and that is unfortunate for a couple of reasons. You know, whenever I say I work at FiveThirtyEight, I get one of two responses. Either, “oh, man, that must have been embarrassing for you guys like 2016, what a mess, right?” or “y’all were the most right”.

The difference here is a nuanced understanding of statistics, and that’s where I think a lot of news media has failed the public — not helping them understand what these numbers mean.

A day or two before the 2016 election, we published a piece that Trump is within standard polling error of Clinton. When you’re conducting a poll, there’s always some degree of error; you’re taking a statistical sample, you’re not talking to every single American. So there’s always some margin of error. Since 1998, presidential general election polling has had an average error of about 4.3%. In 2019–2020, the average was about 5.0%. And in 2015–2016, it was still about 4.9%. So they weren’t off by an unprecedented amount, but it was a very, very close election, and so, it matters more — you’re at the cusp.

So yes, there are some issues with public opinion polling. But the bigger issue was the way the media portrayed that data to the public. I think a lot of people were not ready to believe that Trump had a chance, even though he did. And polls are not a crystal ball — they don’t tell you exactly what’s gonna happen. Otherwise, why even have elections, right?

How does this class you teach at Harvard fit into your interests and vision?

I designed a course at the Harvard Kennedy School called “Programming and Data for Policymakers” which I teach every January. The course is intended to help students in the public policy school learn programming and data skills. While some of my students do intend to work with code and data as a core function of their job, many of them will never code again. But the goal of the class is to introduce them to technology concepts with a “learning by doing” philosophy. So rather than just talking about, for example, open source technology, students will actually explore an open source codebase, understand the implications of publishing code open source and speak to experts who have done that kind of work in the context of government. I think we saw during the pandemic this year just how important it is for the government to be competent with technology. So many government services these days are delivered over the internet, and the goal of the class is to make sure that anyone who intends to work in government has the knowledge and skills to navigate technical questions that arise in government, both when it comes to making policy and implementing solutions for the public. For any readers who are interested, there is a really neat community of civic technologists out there working in government, nonprofit, and in the private sector who use technology to improve public services. It is a really interesting world to get plugged into and a really great way to have a positive social impact!

Telling your parents you’ve switched to a philosophy major must’ve been… interesting. How was that?

It’s an interesting question — it was definitely a bit of a shock to them, especially after having seen my very meticulous four-year plan.

I don’t know if you’ve ever seen “That ’70s Show,” but there’s a great scene where one of the characters decides to major in philosophy. I think it’s Eric’s sister, and the dad goes, ‘Oh, that’s great. You can go work at the new philosophy factory up in Green Bay!’

I do think it’s important to have a plan for living life, for food and shelter and all the things that are necessary for a living. But that said, I’m not sure that being a philosophy major means you’re going to be struggling to survive. There are plenty of things you can do with a philosophy degree, a lot of transferable skills. You learn how to write, to read, and to think clearly. So it worked out for me, but I also did a master’s in computer science because I really was interested in linguistics and artificial intelligence.

I had a realization when I got to Amazon: I wasn’t really thinking about why or how I got there, or what I was doing, or why I was doing it. I was just kind of going with the flow. And that realization helped me kind of start looking for other jobs that I might be interested in, and journalism came up.

I asked myself, “Can I meet my financial obligations? Can I pay off my loans? Like, can I go home to visit my parents occasionally?” And if I can do all of these things, then maybe it’s fine for me to jump from here to journalism? And, you know, you don’t really go into journalism for the money. But yeah, I guess it turned out okay, for me. I’m much happier, and I think I’m able to contribute a lot more. I’ve had the privilege of being able to sort of pursue my interests.

What advice would you give to someone looking to start their career?

High school students should really have more exposure to a variety of fields. When I was in high school, for example, it didn’t even occur to me to consider a career in journalism, or a major in philosophy. And there were disciplines like anthropology that I barely even knew existed. With linguistics, for example, I imagined it involved just learning a bunch of languages. But it turns out, it’s also about the structure of language, how your brain processes that language, and writing computer programs that do the same sort of thing. And of course, it’s also about learning languages and grammar and syntax and all that, but there’s so much more that I just had no idea about. I discovered linguistics at the end of junior year, and I was like, “Whoa, this is cool”.

So I would advise students to not limit themselves to career paths that are well represented in their own social circles. I think there are some perfectly viable career paths that high school students may not have considered simply because they are underrepresented in our communities. You can make a perfectly good living doing a whole lot of things that you may have not even known existed! If you really want to learn about those disciplines, a good starting point is a course catalogue at a major university. When I was in college, I used to order a physical copy and sit down with a highlighter and read it cover-to-cover, highlighting the courses that I found appealing. It was a great way to learn about all of the different fields of study that I may not have had any exposure to through family and friends or through high school. So I suppose my advice is to explore options broadly with a wide open mind and a sense of curiosity.

Dhrumil Mehta studied Philosophy and Computer Science at Northwestern University and was a Database Journalist at FiveThirtyEight. He is an Adjunct Lecturer at the Harvard Kennedy School of Government, where he teaches a Technology & Policy class and has recently been appointed as an Associate Professor of Journalism at Columbia University in New York.