David Leonhardt: Thank you! I appreciate that. One of the things that’s fun for me is this is the first interview I’ve done with someone who’s read the site, because all the other ones were before it launched. Let me just tell you: It’s so much more fun. I enjoy talking about it, period, but it’s so much more fun to talk about specific journalism rather than principles of journalism.

Leonhardt: We started with a pretty general idea. What this grew out of was Nate Silver’s departure. Nate left, and I was well known internally as a champion of Nate’s. I was a sort of obvious person to put on a committee to figure out to do after he left.

We decided quite quickly — maybe even in our first meeting — that we didn’t want to go out and replace Nate. Nate has a set of skills that is unusual, in a good way. And not only that, but that 2012 wasn’t going to be repeated. There wasn’t going to be, in all likelihood, another election that went the way that one did. Trying to recapture that lightning in a bottle, when other people out there — including Nate — were going to be out there doing it, seemed like not the right way to go.

On the other hand, we said, you know what? The lessons of FiveThirtyEight are not narrow lessons. They’re consistent with a bunch of whole other lessons we think we’ve heard here. You look all over the paper, in all kinds of different ways, and it’s clear that readers had a demand for this sort of journalism. This funny mix of really substantive on really big, complicated topics, but presented in a really approachable way. Our hugely successful interactives are another example of this. The most visited page in New York Times history is based on an academic study about linguistics, right? That’s amazing.

We realized, when we do this journalism, people like it, and we can do much more than we’re doing. Once we defined it that way, I realized it was a dream job for me, and I got interested in doing it.

Leonhardt: I don’t know that we’ve invented anything totally new. I really do think people want both. I don’t think they conflict. I think that people want information on big, complicated topics that they can grasp, even if they’re not experts. I think that description encompasses what I would define both as explanatory journalism and data journalism.

To me, explanatory journalism is just something that’s written well enough that someone who isn’t an expert really understands it. Understands it so well that they could turn around and explain it to somebody else. That’s my test as a reader. If I get to the end of an article about how X caused Y, and I can’t go then explain to someone else how X caused Y, I think the article has failed.

It’s data journalism because I don’t think it’s possible to write about 99 percent of important topics without using some data. Sometimes it’s a single number, and often it’s very few numbers, but the cliche is, data’s just another word for facts.

Leonhardt: Two different ways. Leo is the better example, but the smaller example is we put two pieces online, one about the methodology and two about the backstory behind the middle class piece.

For a long time, journalists tried to project this image of the infallible authority. Maybe that worked in another time, I don’t know, but it doesn’t work now. People don’t buy it. People are too smart to believe there’s this special class of people called journalists who were infallible.

I think we have more credibility when we’re honest with people about what we know and what we don’t know. “Hey, this clearly seems to be happening, but we don’t know what’s causing it.” Or: “Hey, it looks like this is more likely to happen than not, but it’s not certain.” Or: “These two sides are having a fight and this one side seems to have a bigger claim on the evidence, but we’re not sure about that.”

That is the kind of voice we’re going to write in. I think it’s a voice that readers appreciate. Consistent with that voice is the idea of showing our work. I don’t want to show it most of the time on the first pass, because a lot of our readers don’t want to see all our work. But the beauty of the web is you can publish it and people who want it can go get it.

It is consistent with Jill’s vision. We are going to do a lot of it. One of the things I want to do — I hope we can wait a while before doing it — is say, Here are some of the things we got wrong over the past few months. I love when columnists write that column.

I really want us to be integrated in the newsroom. I really want us to work with other Times reporters — on the national staff, on the political staff, on the science staff — who are interested in doing this kind of journalism. We’re not separate. Our material will run in the newspaper. It will run on the website. Sometimes it will run without even The Upshot label. What I care about is doing good work that gets in front of Times readers.

Leonhardt: We have these outside contributors, like Michael Beschloss and other people. And to some extent we’ll collaborate with academics. I’ve done a lot of that in my career. Today’s story about middle class incomes is effectively a collaboration, even though it’s not a formal collaboration. So, yes.

Leonhardt: I think tone matters. We newspapers adopted this tone that worked for a long time that was based on the idea that people were getting their news from the bundle of newspapers that arrived at their door every morning.

When you are trying to explain the implications of something that people already know about, I think you want to use a voice that’s different. That’s more conversational, that sometimes uses the first and second person. It’s not as if smart people avoid using first and second person when they talk.

Sometimes, it’s hard to be clear about your point when using the 20th-century–form journalism. You can be a lot clearer in using a different form. What we should do is think about the best form for every story.

Leonhardt: It’s all of our jobs. One of the things I want us to do is spend more time on Facebook than many journalists do.

Leonhardt: Specifically Facebook. Journalists really like Twitter. You don’t have to twist most journalists’ arms, particularly the journalists who are doing this kind of work, to spend time on Twitter. It comes naturally to them.

Leonhardt: Me too!

You do have to give them a little nudge to spend time on Facebook. But Facebook’s really important. We’ve created our own Facebook page at The Upshot, which is relatively unusual at The New York Times. We’re also going to encourage people to spend more time on Facebook. Josh Barro is already quite good on Facebook, and he can become a model for the rest of the team.

We want to try some things that the whole paper isn’t necessarily doing. Let’s say some new social media site comes along and we decide we’re going to spend 10 hours a week combined doing stuff on this new social media site — and then after a year we decide, Well, that social media site has fizzled, it’s not worth our time. That’s not that big of a cost. If the entire New York Times had done that, it would have been a big cost.

On the other hand, if we do it and it succeeds, we’ve been this little laboratory that the rest of the Times can then learn from.

Leonhardt: We should be a laboratory! The traditional mistake that longtime, successful organizations make is they say, We can’t do that because we’ve never done it that way, or, We can’t do that, because that will cannibalize us.

When a business is changing, you don’t get to choose whether you’re going to be cannibalized. You only get to choose whether you’re the one who does it, or you let someone else do it. People at the Times — my bosses, Jill and others — very much want us to experiment, and want us to learn from things, so that when we see something that seems to work, the rest of the Times can adopt a version of it. And when we try something and it doesn’t work, we haven’t just tried to turn the huge battleship of The New York Times and then have to try and undo the turn.

Leonhardt: All of the above. One, we want to get audience through social media. Two, The New York Times’ homepage is a massively powerful convener of audiences and we want to benefit from that. Three, we want to do stuff that’s interesting enough that outside sites link to us. And then four — and this isn’t in any order — we want to build an audience for our own page, so that people feel as if, Oh, hey, there’s something interesting there every day, I want to go back and check it out.

To me, the ideal thing is if someone comes to us because they saw a link on another site, or they saw a tweet about us. They see an article of ours in another place, they read it, they think it’s interesting, they decide to check out the site, they see more stuff there, then they come back to the site — and suddenly that person is a New York Times reader.

Leonhardt: I think that one of the things to remember is that Arthur Sulzberger, at least internally, famously liked to use the phrase platform agnostic to talk about paper versus web. A version of that is we should be story-form agnostic when we start thinking about stories.

There’s this fascinating information — should we write a 1,500-word article about that? Should we go send a reporter somewhere to spend three weeks there? Should we do a 10-minute video on it? Should we do a chart on it? Or, wait a second, we should give people 25 questions and let them answer and place them on a map — that’s the right way to do that story.

I think a lot of it starts from that. Today’s a good example: We did do a traditional 1,800-word story on the middle class. We didn’t do a 1,500-word traditional story on our Senate model. Either you or I could write that story. You could take our Senate model — you could write a newspaper story based on it. We didn’t think that was the right form to do that.

Leonhardt: Yes. That, to me, is really important. We don’t just want to be chewing over the same data that everyone’s chewing over. You know, today’s two stories are an example. It’s not as if this was secret data that we had to steal or persuade a whistleblower to give to us — it was out there. We had to spend a lot of time collecting data in the case of the Senate model before we could even begin analyzing it, and we had to persuade a group of academic researchers that we were serious about this project and we were going to spend time to do it seriously.

Going out there and getting new data sets is, to me, crucial. It’s saying: Are you shedding light on something, or are you just saying the same thing that everyone else has said?

Leonhardt: I think it would depend. The Times has a great database reporting team. They’re the ones who did all the stuff with the Medicare doctor data a few weeks ago. I think if you asked me to try to think about what would have happened if that had come out after The Upshot launched, I think the answer is they take the lead, they work with the reporter writing the main story, and we think about: Can we play a supporting role, given the expertise the people on our team have? And can we do something that no one else at the paper is doing? Is there a way to break off other pieces? Is there a way to take some of the knowledge that people at The New York Times already have that ends up on the cutting room floor and put it in front of readers?

This is something we think about with polling. Our polling team here is fabulous. They often are asked by members of our staff to look at polls and analyze them, and they’ll write these emails or memos that make you so much smarter about polling. We should be publishing some of those. In fact, we’re just about to — we’re going to publish something about the polling behind affirmative action. That’s the kind of thing that, in the past, would basically have been an internal document.

Joe Nocera, the op-ed columnist here who used to be a business columnist, sometimes likes to say that he often finds when he talks to a reporter about a story that he or she wrote, that reporter is much smarter about the story than they could have gleaned from what they wrote. One of the jobs of journalism today is to help bridge that gap, help move the story that we publish closer to the knowledge and insight that the journalist has.

Leonhardt: Yep — I agree. I absolutely agree.