An ode to people much smarter than me

There’s great value in the ideas that emerge when two epistemological traditions collide

Throughout this incredible year, many of my fellow John S. Knight Journalism Fellows have profited enormously from academic collaborators, both at Stanford and around the world. Geraldine Moriba has effectively become a member of Maneesh Agrawala’s computer science group on campus, working on applications of artificial intelligence to identify different forms of bias in journalistic work. Others, like Mandy Jenkins and Chris Horne, have found inspiration in ethnographic and psychological research. In classes about nonprofit management at the Graduate School of Business, fellows thought about designing randomized controlled trials — the gold standard of scientific practice — to measure the impact of their publications instead of vaguely gesturing at outcomes.

It’s been thrilling to see their work trampoline off these scientific ideas — and humbling. I’m a science journalist, and because of my background, I was well-positioned to take advantage of the scientific community at Stanford. But I didn’t — not at the beginning, at least. After nine months at Stanford, I’m only now appreciating the opportunities that present themselves when I think of scientists not just as sources, but as potential colleagues and collaborators.

Not that I ever doubted the value of a scientific approach to journalism. In the beginning of my fellowship, I thought a lot about what journalists can learn from the scientific method, inspired by those who seek a shift toward scientific or precision journalism. I believe the transparency and replicability of this approach support two critical goals: establishing public trust in reporting and empowering institutions to take action on open evidence.

But I drew a line between patterning journalism on scientific practices and actively working alongside scientists. I met with academics at the Meta-Research Innovation Center at Stanford (METRICS) early in the year — one of my original motivations for coming to JSK — but didn’t end up working closely with them. I shared a little, they shared a little, and we left it at that.

I shied away from collaboration because, as a newsroom leader, I take my responsibility to the ethical standards of journalism seriously. I’ve reported on issues of open access and reproducibility in science — some of the main topics studied by the METRICS researchers — and I worried that working too closely with those scientists would prevent me from objectively reporting on their work.

In fact, I thought, a collegial relationship with any scientist could jeopardize readers’ trust in my reporting. I can never know what kind of science I’ll find myself covering, which meant all scholarly partnerships were a no-go. (Science journalists don’t often have the luxury of beat specificity; I’ve reported on everything from neural correlates of creativity to the genetics of marijuana to the efficacy of canvassing to change voters’ positions on same-sex marriage.)

But after seeing so many journo-scientific partnerships succeed at JSK, I’m rethinking my position. Scientists will always be my sources, and the reporting that comes out of those arm’s-length relationships is valuable. But there’s also great value in the ideas that emerge when two epistemological traditions collide. If I’m mindful about how I manage and disclose my partnerships, I can work alongside certain scientists as fellow … let’s call us “information professionals.”

In the course of my year, I’ve worked toward building a database of conflicts of interest in science, which are voluntarily disclosed by scientists in the course of publishing their research. I’m excited by the potential to surface biases in the scientific process, and to increase journalists’ awareness of the role business plays in research. I dove into my first real coding project in more than a decade, and thrilled at every small success. But my progress slowed as I realized my coding skills weren’t up to the task. So I started looking outward.

Instead of using the single tool in my box, I broadened my net to other programming languages and found a robust community of scientists that I could use as a springboard for my own work. I am grateful to Stephanie Kovalchik, a statistical researcher conducting sports analysis at Victoria University, for generously updating an R package that I found crucial to some of my work. Vivek Kulkarni, a natural language processing researcher at Stanford, helped me by applying his background summarizing scientific manuscripts in bulk to my simpler problem of collation. I’m in awe of their smarts, and I appreciate their collaboration so much.

Now, the tricky part. I’ve covered sports science as a journalist before, and NLP is definitely in the cards. So if I ever find myself reporting on those areas again? I’ll either avoid talking to Stephanie or Vivek directly (not hard, science is big). Or I’ll simply exercise the same level of transparency I expect from every journalistic organization I trust: I’ll openly disclose my relationship with them.

Which, of course, I should have figured out before — because that’s exactly the point of the scientific conflict of interest statements I’ve been working to collect and organize this year. The mere presence of a conflicted relationship doesn’t necessarily undermine the authority or validity of a piece of research, be it scientific or journalistic. But it’s absolutely necessary to disclose all those relationships, to allow information consumers to weigh the potential biases for themselves.

I’m excited to see what else can come out of collaborations between scientists and journalists. We’ve got plenty of examples already: Any big data journalism project is likely to have consulting social scientists on the roster, who help to design statistically valid methods. As sensor technology becomes more distributed, some journalists are essentially undertaking their own public health research, tracking heat spikes and air pollution in urban communities and delivering results directly to the communities being impacted.

Both science and journalism are powerful ways of collecting and disseminating new information about the world — and each comes with its own strengths and weaknesses. But the impact of both institutions seems primed to increase when we expand our approaches and collaborate. With full disclosure, of course.